summary refs log tree commit diff
diff options
context:
space:
mode:
-rw-r--r--.buildkite/test_db.dbbin18825216 -> 19279872 bytes
-rw-r--r--.circleci/config.yml6
-rw-r--r--.github/ISSUE_TEMPLATE/BUG_REPORT.md4
-rw-r--r--CHANGES.md448
-rw-r--r--CONTRIBUTING.md6
-rw-r--r--INSTALL.md109
-rw-r--r--README.rst43
-rw-r--r--UPGRADE.rst11
-rw-r--r--changelog.d/67.misc1
-rw-r--r--changelog.d/8275.feature1
-rwxr-xr-xcontrib/cmdclient/console.py18
-rw-r--r--contrib/cmdclient/http.py8
-rw-r--r--contrib/experiments/cursesio.py2
-rw-r--r--contrib/experiments/test_messaging.py4
-rw-r--r--contrib/graph/graph.py2
-rw-r--r--contrib/graph/graph3.py2
-rw-r--r--contrib/jitsimeetbridge/jitsimeetbridge.py2
-rwxr-xr-xcontrib/scripts/kick_users.py8
-rwxr-xr-xdebian/build_virtualenv2
-rw-r--r--debian/changelog67
-rw-r--r--debian/matrix-synapse.default2
-rw-r--r--debian/matrix-synapse.service2
-rw-r--r--debian/synctl.ronn27
-rw-r--r--docker/Dockerfile17
-rw-r--r--docker/README.md29
-rw-r--r--docker/conf/log.config6
-rw-r--r--docs/.sample_config_header.yaml11
-rw-r--r--docs/admin_api/event_reports.rst129
-rw-r--r--docs/admin_api/rooms.md17
-rw-r--r--docs/admin_api/shutdown_room.md29
-rw-r--r--docs/admin_api/user_admin_api.rst54
-rw-r--r--docs/federate.md12
-rw-r--r--docs/metrics-howto.md2
-rw-r--r--docs/password_auth_providers.md1
-rw-r--r--docs/postgres.md14
-rw-r--r--docs/reverse_proxy.md38
-rw-r--r--docs/sample_config.yaml164
-rw-r--r--docs/sample_log_config.yaml46
-rw-r--r--docs/sso_mapping_providers.md14
-rw-r--r--docs/systemd-with-workers/system/matrix-synapse-worker@.service5
-rw-r--r--docs/systemd-with-workers/workers/federation_reader.yaml2
-rw-r--r--docs/user_directory.md2
-rw-r--r--docs/workers.md118
-rw-r--r--mypy.ini63
-rwxr-xr-xscripts-dev/build_debian_packages1
-rwxr-xr-xscripts-dev/check-newsfragment13
-rwxr-xr-xscripts-dev/check_line_terminators.sh34
-rwxr-xr-xscripts-dev/complement.sh22
-rwxr-xr-xscripts-dev/definitions.py4
-rwxr-xr-xscripts-dev/dump_macaroon.py2
-rwxr-xr-xscripts-dev/federation_client.py39
-rw-r--r--scripts-dev/hash_history.py4
-rw-r--r--scripts-dev/mypy_synapse_plugin.py85
-rwxr-xr-xscripts-dev/update_database4
-rwxr-xr-xscripts/move_remote_media_to_new_store.py2
-rwxr-xr-xscripts/register_new_matrix_user2
-rwxr-xr-xscripts/synapse_port_db106
-rwxr-xr-xsetup.py16
-rw-r--r--stubs/frozendict.pyi47
-rw-r--r--synapse/__init__.py14
-rw-r--r--synapse/_scripts/register_new_matrix_user.py2
-rw-r--r--synapse/api/auth.py136
-rw-r--r--synapse/api/auth_blocking.py15
-rw-r--r--synapse/api/constants.py26
-rw-r--r--synapse/api/errors.py70
-rw-r--r--synapse/api/filtering.py19
-rw-r--r--synapse/api/presence.py (renamed from synapse/storage/presence.py)0
-rw-r--r--synapse/api/ratelimiting.py39
-rw-r--r--synapse/api/room_versions.py10
-rw-r--r--synapse/api/urls.py3
-rw-r--r--synapse/app/_base.py29
-rw-r--r--synapse/app/admin_cmd.py23
-rw-r--r--synapse/app/generic_worker.py38
-rw-r--r--synapse/app/homeserver.py48
-rw-r--r--synapse/appservice/__init__.py54
-rw-r--r--synapse/appservice/api.py42
-rw-r--r--synapse/appservice/scheduler.py57
-rw-r--r--synapse/config/_base.py130
-rw-r--r--synapse/config/_base.pyi1
-rw-r--r--synapse/config/_util.py51
-rw-r--r--synapse/config/cache.py2
-rw-r--r--synapse/config/captcha.py3
-rw-r--r--synapse/config/consent_config.py4
-rw-r--r--synapse/config/database.py5
-rw-r--r--synapse/config/emailconfig.py158
-rw-r--r--synapse/config/federation.py27
-rw-r--r--synapse/config/homeserver.py1
-rw-r--r--synapse/config/key.py2
-rw-r--r--synapse/config/logger.py88
-rw-r--r--synapse/config/metrics.py2
-rw-r--r--synapse/config/oidc_config.py14
-rw-r--r--synapse/config/ratelimiting.py25
-rw-r--r--synapse/config/registration.py25
-rw-r--r--synapse/config/room.py2
-rw-r--r--synapse/config/room_directory.py2
-rw-r--r--synapse/config/saml2_config.py88
-rw-r--r--synapse/config/server.py72
-rw-r--r--synapse/config/server_notices_config.py2
-rw-r--r--synapse/config/sso.py37
-rw-r--r--synapse/config/stats.py2
-rw-r--r--synapse/config/tls.py1
-rw-r--r--synapse/config/workers.py37
-rw-r--r--synapse/crypto/context_factory.py22
-rw-r--r--synapse/crypto/keyring.py198
-rw-r--r--synapse/event_auth.py2
-rw-r--r--synapse/events/__init__.py14
-rw-r--r--synapse/events/builder.py92
-rw-r--r--synapse/events/snapshot.py46
-rw-r--r--synapse/events/spamcheck.py37
-rw-r--r--synapse/events/third_party_rules.py39
-rw-r--r--synapse/events/utils.py17
-rw-r--r--synapse/events/validator.py61
-rw-r--r--synapse/federation/federation_base.py2
-rw-r--r--synapse/federation/federation_client.py29
-rw-r--r--synapse/federation/federation_server.py85
-rw-r--r--synapse/federation/persistence.py37
-rw-r--r--synapse/federation/send_queue.py6
-rw-r--r--synapse/federation/sender/__init__.py94
-rw-r--r--synapse/federation/sender/per_destination_queue.py177
-rw-r--r--synapse/federation/sender/transaction_manager.py48
-rw-r--r--synapse/federation/transport/client.py98
-rw-r--r--synapse/federation/transport/server.py27
-rw-r--r--synapse/federation/units.py4
-rw-r--r--synapse/groups/attestations.py29
-rw-r--r--synapse/groups/groups_server.py4
-rw-r--r--synapse/handlers/__init__.py2
-rw-r--r--synapse/handlers/_base.py2
-rw-r--r--synapse/handlers/account_data.py2
-rw-r--r--synapse/handlers/account_validity.py46
-rw-r--r--synapse/handlers/acme.py13
-rw-r--r--synapse/handlers/acme_issuing_service.py4
-rw-r--r--synapse/handlers/admin.py8
-rw-r--r--synapse/handlers/appservice.py12
-rw-r--r--synapse/handlers/auth.py193
-rw-r--r--synapse/handlers/cas_handler.py11
-rw-r--r--synapse/handlers/deactivate_account.py2
-rw-r--r--synapse/handlers/device.py52
-rw-r--r--synapse/handlers/devicemessage.py7
-rw-r--r--synapse/handlers/directory.py8
-rw-r--r--synapse/handlers/e2e_keys.py18
-rw-r--r--synapse/handlers/e2e_room_keys.py2
-rw-r--r--synapse/handlers/events.py65
-rw-r--r--synapse/handlers/federation.py265
-rw-r--r--synapse/handlers/groups_local.py39
-rw-r--r--synapse/handlers/identity.py105
-rw-r--r--synapse/handlers/initial_sync.py116
-rw-r--r--synapse/handlers/message.py264
-rw-r--r--synapse/handlers/oidc_handler.py139
-rw-r--r--synapse/handlers/pagination.py193
-rw-r--r--synapse/handlers/password_policy.py2
-rw-r--r--synapse/handlers/presence.py17
-rw-r--r--synapse/handlers/profile.py83
-rw-r--r--synapse/handlers/read_marker.py2
-rw-r--r--synapse/handlers/receipts.py19
-rw-r--r--synapse/handlers/register.py38
-rw-r--r--synapse/handlers/room.py190
-rw-r--r--synapse/handlers/room_list.py2
-rw-r--r--synapse/handlers/room_member.py216
-rw-r--r--synapse/handlers/room_member_worker.py11
-rw-r--r--synapse/handlers/saml_handler.py192
-rw-r--r--synapse/handlers/search.py12
-rw-r--r--synapse/handlers/set_password.py2
-rw-r--r--synapse/handlers/state_deltas.py2
-rw-r--r--synapse/handlers/stats.py2
-rw-r--r--synapse/handlers/sync.py132
-rw-r--r--synapse/handlers/typing.py23
-rw-r--r--synapse/handlers/ui_auth/checkers.py5
-rw-r--r--synapse/handlers/user_directory.py10
-rw-r--r--synapse/http/__init__.py17
-rw-r--r--synapse/http/client.py297
-rw-r--r--synapse/http/connectproxyclient.py2
-rw-r--r--synapse/http/federation/matrix_federation_agent.py12
-rw-r--r--synapse/http/federation/srv_resolver.py4
-rw-r--r--synapse/http/federation/well_known_resolver.py68
-rw-r--r--synapse/http/matrixfederationclient.py233
-rw-r--r--synapse/http/proxyagent.py16
-rw-r--r--synapse/http/request_metrics.py2
-rw-r--r--synapse/http/server.py127
-rw-r--r--synapse/http/servlet.py7
-rw-r--r--synapse/http/site.py20
-rw-r--r--synapse/logging/_structured.py6
-rw-r--r--synapse/logging/_terse_json.py4
-rw-r--r--synapse/logging/context.py53
-rw-r--r--synapse/logging/formatter.py2
-rw-r--r--synapse/logging/opentracing.py14
-rw-r--r--synapse/logging/scopecontextmanager.py6
-rw-r--r--synapse/logging/utils.py6
-rw-r--r--synapse/metrics/__init__.py131
-rw-r--r--synapse/metrics/_exposition.py60
-rw-r--r--synapse/metrics/background_process_metrics.py40
-rw-r--r--synapse/module_api/__init__.py20
-rw-r--r--synapse/notifier.py223
-rw-r--r--synapse/push/__init__.py2
-rw-r--r--synapse/push/action_generator.py9
-rw-r--r--synapse/push/baserules.py216
-rw-r--r--synapse/push/bulk_push_rule_evaluator.py157
-rw-r--r--synapse/push/emailpusher.py4
-rw-r--r--synapse/push/httppusher.py62
-rw-r--r--synapse/push/mailer.py76
-rw-r--r--synapse/push/presentable_names.py15
-rw-r--r--synapse/push/push_rule_evaluator.py2
-rw-r--r--synapse/push/push_tools.py22
-rw-r--r--synapse/push/pusher.py33
-rw-r--r--synapse/push/pusherpool.py121
-rw-r--r--synapse/python_dependencies.py25
-rw-r--r--synapse/replication/http/_base.py60
-rw-r--r--synapse/replication/http/devices.py4
-rw-r--r--synapse/replication/http/federation.py33
-rw-r--r--synapse/replication/http/login.py4
-rw-r--r--synapse/replication/http/membership.py24
-rw-r--r--synapse/replication/http/presence.py4
-rw-r--r--synapse/replication/http/register.py12
-rw-r--r--synapse/replication/http/send_event.py9
-rw-r--r--synapse/replication/http/streams.py2
-rw-r--r--synapse/replication/slave/storage/_base.py10
-rw-r--r--synapse/replication/slave/storage/_slaved_id_tracker.py14
-rw-r--r--synapse/replication/slave/storage/account_data.py14
-rw-r--r--synapse/replication/slave/storage/appservice.py2
-rw-r--r--synapse/replication/slave/storage/client_ips.py10
-rw-r--r--synapse/replication/slave/storage/deviceinbox.py10
-rw-r--r--synapse/replication/slave/storage/devices.py17
-rw-r--r--synapse/replication/slave/storage/directory.py2
-rw-r--r--synapse/replication/slave/storage/events.py26
-rw-r--r--synapse/replication/slave/storage/filtering.py8
-rw-r--r--synapse/replication/slave/storage/groups.py10
-rw-r--r--synapse/replication/slave/storage/keys.py2
-rw-r--r--synapse/replication/slave/storage/presence.py10
-rw-r--r--synapse/replication/slave/storage/profile.py2
-rw-r--r--synapse/replication/slave/storage/push_rule.py14
-rw-r--r--synapse/replication/slave/storage/pushers.py10
-rw-r--r--synapse/replication/slave/storage/receipts.py10
-rw-r--r--synapse/replication/slave/storage/registration.py2
-rw-r--r--synapse/replication/slave/storage/room.py10
-rw-r--r--synapse/replication/slave/storage/transactions.py2
-rw-r--r--synapse/replication/tcp/client.py20
-rw-r--r--synapse/replication/tcp/commands.py11
-rw-r--r--synapse/replication/tcp/handler.py2
-rw-r--r--synapse/replication/tcp/protocol.py2
-rw-r--r--synapse/replication/tcp/resource.py4
-rw-r--r--synapse/replication/tcp/streams/_base.py12
-rw-r--r--synapse/replication/tcp/streams/events.py8
-rw-r--r--synapse/res/templates/auth_success.html21
-rw-r--r--synapse/res/templates/password_reset_confirmation.html16
-rw-r--r--synapse/res/templates/recaptcha.html38
-rw-r--r--synapse/res/templates/saml_error.html45
-rw-r--r--synapse/res/templates/sso_error.html43
-rw-r--r--synapse/res/templates/terms.html20
-rw-r--r--synapse/rest/__init__.py10
-rw-r--r--synapse/rest/admin/__init__.py11
-rw-r--r--synapse/rest/admin/_base.py4
-rw-r--r--synapse/rest/admin/devices.py17
-rw-r--r--synapse/rest/admin/event_reports.py88
-rw-r--r--synapse/rest/admin/purge_room_servlet.py5
-rw-r--r--synapse/rest/admin/rooms.py16
-rw-r--r--synapse/rest/admin/server_notice_servlet.py9
-rw-r--r--synapse/rest/admin/users.py38
-rw-r--r--synapse/rest/client/transactions.py2
-rw-r--r--synapse/rest/client/v1/directory.py8
-rw-r--r--synapse/rest/client/v1/events.py7
-rw-r--r--synapse/rest/client/v1/initial_sync.py5
-rw-r--r--synapse/rest/client/v1/login.py133
-rw-r--r--synapse/rest/client/v1/logout.py4
-rw-r--r--synapse/rest/client/v1/presence.py2
-rw-r--r--synapse/rest/client/v1/profile.py6
-rw-r--r--synapse/rest/client/v1/push_rule.py37
-rw-r--r--synapse/rest/client/v1/pusher.py6
-rw-r--r--synapse/rest/client/v1/room.py226
-rw-r--r--synapse/rest/client/v1/voip.py2
-rw-r--r--synapse/rest/client/v2_alpha/account.py288
-rw-r--r--synapse/rest/client/v2_alpha/account_data.py4
-rw-r--r--synapse/rest/client/v2_alpha/account_validity.py4
-rw-r--r--synapse/rest/client/v2_alpha/auth.py138
-rw-r--r--synapse/rest/client/v2_alpha/capabilities.py2
-rw-r--r--synapse/rest/client/v2_alpha/devices.py6
-rw-r--r--synapse/rest/client/v2_alpha/filter.py4
-rw-r--r--synapse/rest/client/v2_alpha/groups.py52
-rw-r--r--synapse/rest/client/v2_alpha/keys.py15
-rw-r--r--synapse/rest/client/v2_alpha/notifications.py2
-rw-r--r--synapse/rest/client/v2_alpha/openid.py2
-rw-r--r--synapse/rest/client/v2_alpha/password_policy.py2
-rw-r--r--synapse/rest/client/v2_alpha/read_marker.py2
-rw-r--r--synapse/rest/client/v2_alpha/receipts.py2
-rw-r--r--synapse/rest/client/v2_alpha/register.py178
-rw-r--r--synapse/rest/client/v2_alpha/relations.py26
-rw-r--r--synapse/rest/client/v2_alpha/report_event.py2
-rw-r--r--synapse/rest/client/v2_alpha/room_keys.py6
-rw-r--r--synapse/rest/client/v2_alpha/room_upgrade_rest_servlet.py16
-rw-r--r--synapse/rest/client/v2_alpha/sendtodevice.py2
-rw-r--r--synapse/rest/client/v2_alpha/shared_rooms.py68
-rw-r--r--synapse/rest/client/v2_alpha/sync.py18
-rw-r--r--synapse/rest/client/v2_alpha/tags.py4
-rw-r--r--synapse/rest/client/v2_alpha/thirdparty.py8
-rw-r--r--synapse/rest/client/v2_alpha/tokenrefresh.py2
-rw-r--r--synapse/rest/client/v2_alpha/user_directory.py2
-rw-r--r--synapse/rest/client/versions.py23
-rw-r--r--synapse/rest/consent/consent_resource.py4
-rw-r--r--synapse/rest/health.py31
-rw-r--r--synapse/rest/key/v2/remote_key_resource.py14
-rw-r--r--synapse/rest/media/v1/_base.py12
-rw-r--r--synapse/rest/media/v1/filepath.py21
-rw-r--r--synapse/rest/media/v1/media_repository.py177
-rw-r--r--synapse/rest/media/v1/media_storage.py106
-rw-r--r--synapse/rest/media/v1/preview_url_resource.py35
-rw-r--r--synapse/rest/media/v1/storage_provider.py77
-rw-r--r--synapse/rest/media/v1/thumbnail_resource.py5
-rw-r--r--synapse/rest/media/v1/thumbnailer.py20
-rw-r--r--synapse/rest/media/v1/upload_resource.py4
-rw-r--r--synapse/rest/saml2/response_resource.py16
-rw-r--r--synapse/rest/synapse/__init__.py14
-rw-r--r--synapse/rest/synapse/client/__init__.py14
-rw-r--r--synapse/rest/synapse/client/password_reset.py127
-rw-r--r--synapse/rest/well_known.py6
-rw-r--r--synapse/secrets.py10
-rw-r--r--synapse/server.py470
-rw-r--r--synapse/server.pyi155
-rw-r--r--synapse/server_notices/consent_server_notices.py14
-rw-r--r--synapse/server_notices/resource_limits_server_notices.py45
-rw-r--r--synapse/server_notices/server_notices_manager.py31
-rw-r--r--synapse/server_notices/server_notices_sender.py14
-rw-r--r--synapse/server_notices/worker_server_notices_sender.py19
-rw-r--r--synapse/spam_checker_api/__init__.py21
-rw-r--r--synapse/state/__init__.py469
-rw-r--r--synapse/state/v1.py89
-rw-r--r--synapse/state/v2.py259
-rw-r--r--synapse/storage/__init__.py22
-rw-r--r--synapse/storage/_base.py14
-rw-r--r--synapse/storage/background_updates.py52
-rw-r--r--synapse/storage/data_stores/__init__.py97
-rw-r--r--synapse/storage/data_stores/main/transactions.py269
-rw-r--r--synapse/storage/database.py898
-rw-r--r--synapse/storage/databases/__init__.py119
-rw-r--r--synapse/storage/databases/main/__init__.py (renamed from synapse/storage/data_stores/main/__init__.py)150
-rw-r--r--synapse/storage/databases/main/account_data.py (renamed from synapse/storage/data_stores/main/account_data.py)173
-rw-r--r--synapse/storage/databases/main/appservice.py (renamed from synapse/storage/data_stores/main/appservice.py)76
-rw-r--r--synapse/storage/databases/main/cache.py (renamed from synapse/storage/data_stores/main/cache.py)14
-rw-r--r--synapse/storage/databases/main/censor_events.py (renamed from synapse/storage/data_stores/main/censor_events.py)35
-rw-r--r--synapse/storage/databases/main/client_ips.py (renamed from synapse/storage/data_stores/main/client_ips.py)109
-rw-r--r--synapse/storage/databases/main/deviceinbox.py (renamed from synapse/storage/data_stores/main/deviceinbox.py)139
-rw-r--r--synapse/storage/databases/main/devices.py (renamed from synapse/storage/data_stores/main/devices.py)526
-rw-r--r--synapse/storage/databases/main/directory.py (renamed from synapse/storage/data_stores/main/directory.py)69
-rw-r--r--synapse/storage/databases/main/e2e_room_keys.py (renamed from synapse/storage/data_stores/main/e2e_room_keys.py)109
-rw-r--r--synapse/storage/databases/main/end_to_end_keys.py (renamed from synapse/storage/data_stores/main/end_to_end_keys.py)430
-rw-r--r--synapse/storage/databases/main/event_federation.py (renamed from synapse/storage/data_stores/main/event_federation.py)212
-rw-r--r--synapse/storage/databases/main/event_push_actions.py (renamed from synapse/storage/data_stores/main/event_push_actions.py)396
-rw-r--r--synapse/storage/databases/main/events.py (renamed from synapse/storage/data_stores/main/events.py)148
-rw-r--r--synapse/storage/databases/main/events_bg_updates.py (renamed from synapse/storage/data_stores/main/events_bg_updates.py)90
-rw-r--r--synapse/storage/databases/main/events_worker.py (renamed from synapse/storage/data_stores/main/events_worker.py)481
-rw-r--r--synapse/storage/databases/main/filtering.py (renamed from synapse/storage/data_stores/main/filtering.py)13
-rw-r--r--synapse/storage/databases/main/group_server.py (renamed from synapse/storage/data_stores/main/group_server.py)498
-rw-r--r--synapse/storage/databases/main/keys.py (renamed from synapse/storage/data_stores/main/keys.py)82
-rw-r--r--synapse/storage/databases/main/media_repository.py (renamed from synapse/storage/data_stores/main/media_repository.py)167
-rw-r--r--synapse/storage/databases/main/metrics.py (renamed from synapse/storage/data_stores/main/metrics.py)83
-rw-r--r--synapse/storage/databases/main/monthly_active_users.py (renamed from synapse/storage/data_stores/main/monthly_active_users.py)90
-rw-r--r--synapse/storage/databases/main/openid.py (renamed from synapse/storage/data_stores/main/openid.py)14
-rw-r--r--synapse/storage/databases/main/presence.py (renamed from synapse/storage/data_stores/main/presence.py)45
-rw-r--r--synapse/storage/databases/main/profile.py (renamed from synapse/storage/data_stores/main/profile.py)122
-rw-r--r--synapse/storage/databases/main/purge_events.py (renamed from synapse/storage/data_stores/main/purge_events.py)44
-rw-r--r--synapse/storage/databases/main/push_rule.py (renamed from synapse/storage/data_stores/main/push_rule.py)400
-rw-r--r--synapse/storage/databases/main/pusher.py (renamed from synapse/storage/data_stores/main/pusher.py)118
-rw-r--r--synapse/storage/databases/main/receipts.py (renamed from synapse/storage/data_stores/main/receipts.py)159
-rw-r--r--synapse/storage/databases/main/registration.py (renamed from synapse/storage/data_stores/main/registration.py)749
-rw-r--r--synapse/storage/databases/main/rejections.py (renamed from synapse/storage/data_stores/main/rejections.py)5
-rw-r--r--synapse/storage/databases/main/relations.py (renamed from synapse/storage/data_stores/main/relations.py)124
-rw-r--r--synapse/storage/databases/main/room.py (renamed from synapse/storage/data_stores/main/room.py)387
-rw-r--r--synapse/storage/databases/main/roommember.py (renamed from synapse/storage/data_stores/main/roommember.py)366
-rw-r--r--synapse/storage/databases/main/schema/delta/12/v12.sql (renamed from synapse/storage/data_stores/main/schema/delta/12/v12.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/13/v13.sql (renamed from synapse/storage/data_stores/main/schema/delta/13/v13.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/14/v14.sql (renamed from synapse/storage/data_stores/main/schema/delta/14/v14.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/15/appservice_txns.sql (renamed from synapse/storage/data_stores/main/schema/delta/15/appservice_txns.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/15/presence_indices.sql (renamed from synapse/storage/data_stores/main/schema/delta/15/presence_indices.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/15/v15.sql (renamed from synapse/storage/data_stores/main/schema/delta/15/v15.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/events_order_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/events_order_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/remote_media_cache_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/remote_media_cache_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/remove_duplicates.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/remove_duplicates.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/room_alias_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/room_alias_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/unique_constraints.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/unique_constraints.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/16/users.sql (renamed from synapse/storage/data_stores/main/schema/delta/16/users.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/17/drop_indexes.sql (renamed from synapse/storage/data_stores/main/schema/delta/17/drop_indexes.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/17/server_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/17/server_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/17/user_threepids.sql (renamed from synapse/storage/data_stores/main/schema/delta/17/user_threepids.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/18/server_keys_bigger_ints.sql (renamed from synapse/storage/data_stores/main/schema/delta/18/server_keys_bigger_ints.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/19/event_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/19/event_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/20/dummy.sql (renamed from synapse/storage/data_stores/main/schema/delta/20/dummy.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/20/pushers.py (renamed from synapse/storage/data_stores/main/schema/delta/20/pushers.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/21/end_to_end_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/21/end_to_end_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/21/receipts.sql (renamed from synapse/storage/data_stores/main/schema/delta/21/receipts.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/22/receipts_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/22/receipts_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/22/user_threepids_unique.sql (renamed from synapse/storage/data_stores/main/schema/delta/22/user_threepids_unique.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/24/stats_reporting.sql (renamed from synapse/storage/data_stores/main/schema/delta/24/stats_reporting.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/25/fts.py (renamed from synapse/storage/data_stores/main/schema/delta/25/fts.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/25/guest_access.sql (renamed from synapse/storage/data_stores/main/schema/delta/25/guest_access.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/25/history_visibility.sql (renamed from synapse/storage/data_stores/main/schema/delta/25/history_visibility.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/25/tags.sql (renamed from synapse/storage/data_stores/main/schema/delta/25/tags.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/26/account_data.sql (renamed from synapse/storage/data_stores/main/schema/delta/26/account_data.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/27/account_data.sql (renamed from synapse/storage/data_stores/main/schema/delta/27/account_data.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/27/forgotten_memberships.sql (renamed from synapse/storage/data_stores/main/schema/delta/27/forgotten_memberships.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/27/ts.py (renamed from synapse/storage/data_stores/main/schema/delta/27/ts.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/event_push_actions.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/event_push_actions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/events_room_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/events_room_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/public_roms_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/public_roms_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/receipts_user_id_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/receipts_user_id_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/upgrade_times.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/upgrade_times.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/28/users_is_guest.sql (renamed from synapse/storage/data_stores/main/schema/delta/28/users_is_guest.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/29/push_actions.sql (renamed from synapse/storage/data_stores/main/schema/delta/29/push_actions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/alias_creator.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/alias_creator.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/as_users.py (renamed from synapse/storage/data_stores/main/schema/delta/30/as_users.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/deleted_pushers.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/deleted_pushers.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/presence_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/presence_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/public_rooms.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/public_rooms.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/push_rule_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/push_rule_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/30/threepid_guest_access_tokens.sql (renamed from synapse/storage/data_stores/main/schema/delta/30/threepid_guest_access_tokens.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/31/invites.sql (renamed from synapse/storage/data_stores/main/schema/delta/31/invites.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/31/local_media_repository_url_cache.sql (renamed from synapse/storage/data_stores/main/schema/delta/31/local_media_repository_url_cache.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/31/pushers.py (renamed from synapse/storage/data_stores/main/schema/delta/31/pushers.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/31/pushers_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/31/pushers_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/31/search_update.py (renamed from synapse/storage/data_stores/main/schema/delta/31/search_update.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/32/events.sql (renamed from synapse/storage/data_stores/main/schema/delta/32/events.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/32/openid.sql (renamed from synapse/storage/data_stores/main/schema/delta/32/openid.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/32/pusher_throttle.sql (renamed from synapse/storage/data_stores/main/schema/delta/32/pusher_throttle.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/32/remove_indices.sql (renamed from synapse/storage/data_stores/main/schema/delta/32/remove_indices.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/32/reports.sql (renamed from synapse/storage/data_stores/main/schema/delta/32/reports.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/access_tokens_device_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/33/access_tokens_device_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/devices.sql (renamed from synapse/storage/data_stores/main/schema/delta/33/devices.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql (renamed from synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/event_fields.py (renamed from synapse/storage/data_stores/main/schema/delta/33/event_fields.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/remote_media_ts.py (renamed from synapse/storage/data_stores/main/schema/delta/33/remote_media_ts.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/33/user_ips_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/33/user_ips_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/34/appservice_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/34/appservice_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/34/cache_stream.py (renamed from synapse/storage/data_stores/main/schema/delta/34/cache_stream.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/34/device_inbox.sql (renamed from synapse/storage/data_stores/main/schema/delta/34/device_inbox.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/34/push_display_name_rename.sql (renamed from synapse/storage/data_stores/main/schema/delta/34/push_display_name_rename.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/34/received_txn_purge.py (renamed from synapse/storage/data_stores/main/schema/delta/34/received_txn_purge.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/contains_url.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/contains_url.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/device_outbox.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/device_outbox.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/device_stream_id.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/device_stream_id.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/event_push_actions_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/event_push_actions_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/public_room_list_change_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/public_room_list_change_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/35/stream_order_to_extrem.sql (renamed from synapse/storage/data_stores/main/schema/delta/35/stream_order_to_extrem.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/36/readd_public_rooms.sql (renamed from synapse/storage/data_stores/main/schema/delta/36/readd_public_rooms.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/37/remove_auth_idx.py (renamed from synapse/storage/data_stores/main/schema/delta/37/remove_auth_idx.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/37/user_threepids.sql (renamed from synapse/storage/data_stores/main/schema/delta/37/user_threepids.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/38/postgres_fts_gist.sql (renamed from synapse/storage/data_stores/main/schema/delta/38/postgres_fts_gist.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/39/appservice_room_list.sql (renamed from synapse/storage/data_stores/main/schema/delta/39/appservice_room_list.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/39/device_federation_stream_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/39/device_federation_stream_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/39/event_push_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/39/event_push_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/39/federation_out_position.sql (renamed from synapse/storage/data_stores/main/schema/delta/39/federation_out_position.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/39/membership_profile.sql (renamed from synapse/storage/data_stores/main/schema/delta/39/membership_profile.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/40/current_state_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/40/current_state_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/40/device_inbox.sql (renamed from synapse/storage/data_stores/main/schema/delta/40/device_inbox.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/40/device_list_streams.sql (renamed from synapse/storage/data_stores/main/schema/delta/40/device_list_streams.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/40/event_push_summary.sql (renamed from synapse/storage/data_stores/main/schema/delta/40/event_push_summary.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/40/pushers.sql (renamed from synapse/storage/data_stores/main/schema/delta/40/pushers.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/41/device_list_stream_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/41/device_list_stream_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/41/device_outbound_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/41/device_outbound_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/41/event_search_event_id_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/41/event_search_event_id_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/41/ratelimit.sql (renamed from synapse/storage/data_stores/main/schema/delta/41/ratelimit.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/42/current_state_delta.sql (renamed from synapse/storage/data_stores/main/schema/delta/42/current_state_delta.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/42/device_list_last_id.sql (renamed from synapse/storage/data_stores/main/schema/delta/42/device_list_last_id.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/42/event_auth_state_only.sql (renamed from synapse/storage/data_stores/main/schema/delta/42/event_auth_state_only.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/42/user_dir.py (renamed from synapse/storage/data_stores/main/schema/delta/42/user_dir.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/43/blocked_rooms.sql (renamed from synapse/storage/data_stores/main/schema/delta/43/blocked_rooms.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/43/quarantine_media.sql (renamed from synapse/storage/data_stores/main/schema/delta/43/quarantine_media.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/43/url_cache.sql (renamed from synapse/storage/data_stores/main/schema/delta/43/url_cache.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/43/user_share.sql (renamed from synapse/storage/data_stores/main/schema/delta/43/user_share.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/44/expire_url_cache.sql (renamed from synapse/storage/data_stores/main/schema/delta/44/expire_url_cache.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/45/group_server.sql (renamed from synapse/storage/data_stores/main/schema/delta/45/group_server.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/45/profile_cache.sql (renamed from synapse/storage/data_stores/main/schema/delta/45/profile_cache.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/drop_refresh_tokens.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/drop_refresh_tokens.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/drop_unique_deleted_pushers.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/drop_unique_deleted_pushers.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/group_server.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/group_server.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/local_media_repository_url_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/local_media_repository_url_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/user_dir_null_room_ids.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/user_dir_null_room_ids.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/46/user_dir_typos.sql (renamed from synapse/storage/data_stores/main/schema/delta/46/user_dir_typos.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/47/last_access_media.sql (renamed from synapse/storage/data_stores/main/schema/delta/47/last_access_media.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/47/postgres_fts_gin.sql (renamed from synapse/storage/data_stores/main/schema/delta/47/postgres_fts_gin.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/47/push_actions_staging.sql (renamed from synapse/storage/data_stores/main/schema/delta/47/push_actions_staging.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/add_user_consent.sql (renamed from synapse/storage/data_stores/main/schema/delta/48/add_user_consent.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/add_user_ips_last_seen_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/48/add_user_ips_last_seen_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/deactivated_users.sql (renamed from synapse/storage/data_stores/main/schema/delta/48/deactivated_users.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/group_unique_indexes.py (renamed from synapse/storage/data_stores/main/schema/delta/48/group_unique_indexes.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/groups_joinable.sql (renamed from synapse/storage/data_stores/main/schema/delta/48/groups_joinable.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/48/profiles_batch.sql (renamed from synapse/storage/data_stores/main/schema/delta/48/profiles_batch.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/49/add_user_consent_server_notice_sent.sql (renamed from synapse/storage/data_stores/main/schema/delta/49/add_user_consent_server_notice_sent.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/49/add_user_daily_visits.sql (renamed from synapse/storage/data_stores/main/schema/delta/49/add_user_daily_visits.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/49/add_user_ips_last_seen_only_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/49/add_user_ips_last_seen_only_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/50/add_creation_ts_users_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/50/add_creation_ts_users_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/50/erasure_store.sql (renamed from synapse/storage/data_stores/main/schema/delta/50/erasure_store.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/50/make_event_content_nullable.py (renamed from synapse/storage/data_stores/main/schema/delta/50/make_event_content_nullable.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/50/profiles_deactivated_users.sql (renamed from synapse/storage/data_stores/main/schema/delta/50/profiles_deactivated_users.sql)2
-rw-r--r--synapse/storage/databases/main/schema/delta/51/e2e_room_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/51/e2e_room_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/51/monthly_active_users.sql (renamed from synapse/storage/data_stores/main/schema/delta/51/monthly_active_users.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/52/add_event_to_state_group_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/52/add_event_to_state_group_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/52/device_list_streams_unique_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/52/device_list_streams_unique_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/52/e2e_room_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/52/e2e_room_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/add_user_type_to_users.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/add_user_type_to_users.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/drop_sent_transactions.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/drop_sent_transactions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/event_format_version.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/event_format_version.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/user_dir_populate.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/user_dir_populate.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/user_ips_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/user_ips_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/user_share.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/user_share.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/user_threepid_id.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/user_threepid_id.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/53/users_in_public_rooms.sql (renamed from synapse/storage/data_stores/main/schema/delta/53/users_in_public_rooms.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/account_validity_with_renewal.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/account_validity_with_renewal.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/add_validity_to_server_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/add_validity_to_server_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/delete_forward_extremities.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/delete_forward_extremities.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/drop_legacy_tables.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/drop_legacy_tables.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/drop_presence_list.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/drop_presence_list.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/relations.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/relations.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/stats.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/stats.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/54/stats2.sql (renamed from synapse/storage/data_stores/main/schema/delta/54/stats2.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/55/access_token_expiry.sql (renamed from synapse/storage/data_stores/main/schema/delta/55/access_token_expiry.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/55/profile_replication_status_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/55/profile_replication_status_index.sql)3
-rw-r--r--synapse/storage/databases/main/schema/delta/55/track_threepid_validations.sql (renamed from synapse/storage/data_stores/main/schema/delta/55/track_threepid_validations.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/55/users_alter_deactivated.sql (renamed from synapse/storage/data_stores/main/schema/delta/55/users_alter_deactivated.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/add_spans_to_device_lists.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/add_spans_to_device_lists.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/current_state_events_membership.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/current_state_events_membership_mk2.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership_mk2.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/delete_keys_from_deleted_backups.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/delete_keys_from_deleted_backups.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/destinations_failure_ts.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/destinations_failure_ts.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/destinations_retry_interval_type.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/56/destinations_retry_interval_type.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/device_stream_id_insert.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/device_stream_id_insert.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/devices_last_seen.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/devices_last_seen.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/drop_unused_event_tables.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/drop_unused_event_tables.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/event_expiry.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/event_expiry.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/event_labels.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/event_labels.sql)2
-rw-r--r--synapse/storage/databases/main/schema/delta/56/event_labels_background_update.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/event_labels_background_update.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/fix_room_keys_index.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/fix_room_keys_index.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/hidden_devices.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/hidden_devices.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/hidden_devices_fix.sql.sqlite (renamed from synapse/storage/data_stores/main/schema/delta/56/hidden_devices_fix.sql.sqlite)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/nuke_empty_communities_from_db.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/nuke_empty_communities_from_db.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/public_room_list_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/public_room_list_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/redaction_censor.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/redaction_censor.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/redaction_censor2.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/redaction_censor2.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/redaction_censor4.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/redaction_censor4.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/room_key_etag.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/room_key_etag.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/room_membership_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/room_membership_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/room_retention.sql (renamed from synapse/storage/data_stores/main/schema/delta/55/room_retention.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/signing_keys.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/signing_keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/signing_keys_nonunique_signatures.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/signing_keys_nonunique_signatures.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/stats_separated.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/stats_separated.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/unique_user_filter_index.py (renamed from synapse/storage/data_stores/main/schema/delta/56/unique_user_filter_index.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/user_external_ids.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/user_external_ids.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/56/users_in_public_rooms_idx.sql (renamed from synapse/storage/data_stores/main/schema/delta/56/users_in_public_rooms_idx.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/delete_old_current_state_events.sql (renamed from synapse/storage/data_stores/main/schema/delta/57/delete_old_current_state_events.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/device_list_remote_cache_stale.sql (renamed from synapse/storage/data_stores/main/schema/delta/57/device_list_remote_cache_stale.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/local_current_membership.py (renamed from synapse/storage/data_stores/main/schema/delta/57/local_current_membership.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/remove_sent_outbound_pokes.sql (renamed from synapse/storage/data_stores/main/schema/delta/57/remove_sent_outbound_pokes.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/rooms_version_column.sql (renamed from synapse/storage/data_stores/main/schema/delta/57/rooms_version_column.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.sqlite (renamed from synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.sqlite)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.sqlite (renamed from synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.sqlite)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/02remove_dup_outbound_pokes.sql (renamed from synapse/storage/data_stores/main/schema/delta/58/02remove_dup_outbound_pokes.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/03persist_ui_auth.sql (renamed from synapse/storage/data_stores/main/schema/delta/58/03persist_ui_auth.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/05cache_instance.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/58/05cache_instance.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/06dlols_unique_idx.py (renamed from synapse/storage/data_stores/main/schema/delta/58/06dlols_unique_idx.py)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.postgres33
-rw-r--r--synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.sqlite44
-rw-r--r--synapse/storage/databases/main/schema/delta/58/07persist_ui_auth_ips.sql25
-rw-r--r--synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres (renamed from synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite (renamed from synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/09shadow_ban.sql18
-rw-r--r--synapse/storage/databases/main/schema/delta/58/10_pushrules_enabled_delete_obsolete.sql28
-rw-r--r--synapse/storage/databases/main/schema/delta/58/10drop_local_rejections_stream.sql (renamed from synapse/storage/data_stores/main/schema/delta/58/10drop_local_rejections_stream.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/10federation_pos_instance_name.sql (renamed from synapse/storage/data_stores/main/schema/delta/58/10federation_pos_instance_name.sql)0
-rw-r--r--synapse/storage/databases/main/schema/delta/58/11user_id_seq.py (renamed from synapse/storage/data_stores/main/schema/delta/58/11user_id_seq.py)2
-rw-r--r--synapse/storage/databases/main/schema/delta/58/12room_stats.sql32
-rw-r--r--synapse/storage/databases/main/schema/delta/58/13remove_presence_allow_inbound.sql17
-rw-r--r--synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql16
-rw-r--r--synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql.postgres28
-rw-r--r--synapse/storage/databases/main/schema/delta/58/15_catchup_destination_rooms.sql42
-rw-r--r--synapse/storage/databases/main/schema/delta/58/15unread_count.sql26
-rw-r--r--synapse/storage/databases/main/schema/delta/58/16populate_stats_process_rooms_fix.sql22
-rw-r--r--synapse/storage/databases/main/schema/delta/58/17_catchup_last_successful.sql21
-rw-r--r--synapse/storage/databases/main/schema/delta/58/18stream_positions.sql22
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/application_services.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/application_services.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/event_edges.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/event_edges.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/event_signatures.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/event_signatures.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/im.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/im.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/keys.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/keys.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/media_repository.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/media_repository.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/presence.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/presence.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/profiles.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/profiles.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/push.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/push.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/redactions.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/redactions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/room_aliases.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/room_aliases.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/state.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/state.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/transactions.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/transactions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/16/users.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/16/users.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/54/full.sql.postgres (renamed from synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.postgres)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/54/full.sql.sqlite (renamed from synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.sqlite)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/54/stream_positions.sql (renamed from synapse/storage/data_stores/main/schema/full_schemas/54/stream_positions.sql)0
-rw-r--r--synapse/storage/databases/main/schema/full_schemas/README.md (renamed from synapse/storage/data_stores/main/schema/full_schemas/README.md)0
-rw-r--r--synapse/storage/databases/main/search.py (renamed from synapse/storage/data_stores/main/search.py)120
-rw-r--r--synapse/storage/databases/main/signatures.py (renamed from synapse/storage/data_stores/main/signatures.py)45
-rw-r--r--synapse/storage/databases/main/state.py (renamed from synapse/storage/data_stores/main/state.py)123
-rw-r--r--synapse/storage/databases/main/state_deltas.py (renamed from synapse/storage/data_stores/main/state_deltas.py)25
-rw-r--r--synapse/storage/databases/main/stats.py (renamed from synapse/storage/data_stores/main/stats.py)237
-rw-r--r--synapse/storage/databases/main/stream.py (renamed from synapse/storage/data_stores/main/stream.py)534
-rw-r--r--synapse/storage/databases/main/tags.py (renamed from synapse/storage/data_stores/main/tags.py)122
-rw-r--r--synapse/storage/databases/main/transactions.py476
-rw-r--r--synapse/storage/databases/main/ui_auth.py (renamed from synapse/storage/data_stores/main/ui_auth.py)93
-rw-r--r--synapse/storage/databases/main/user_directory.py (renamed from synapse/storage/data_stores/main/user_directory.py)280
-rw-r--r--synapse/storage/databases/main/user_erasure_store.py (renamed from synapse/storage/data_stores/main/user_erasure_store.py)36
-rw-r--r--synapse/storage/databases/state/__init__.py (renamed from synapse/storage/data_stores/state/__init__.py)2
-rw-r--r--synapse/storage/databases/state/bg_updates.py (renamed from synapse/storage/data_stores/state/bg_updates.py)46
-rw-r--r--synapse/storage/databases/state/schema/delta/23/drop_state_index.sql (renamed from synapse/storage/data_stores/state/schema/delta/23/drop_state_index.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/30/state_stream.sql (renamed from synapse/storage/data_stores/state/schema/delta/30/state_stream.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/32/remove_state_indices.sql (renamed from synapse/storage/data_stores/state/schema/delta/32/remove_state_indices.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/35/add_state_index.sql (renamed from synapse/storage/data_stores/state/schema/delta/35/add_state_index.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/35/state.sql (renamed from synapse/storage/data_stores/state/schema/delta/35/state.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/35/state_dedupe.sql (renamed from synapse/storage/data_stores/state/schema/delta/35/state_dedupe.sql)0
-rw-r--r--synapse/storage/databases/state/schema/delta/47/state_group_seq.py (renamed from synapse/storage/data_stores/state/schema/delta/47/state_group_seq.py)0
-rw-r--r--synapse/storage/databases/state/schema/delta/56/state_group_room_idx.sql (renamed from synapse/storage/data_stores/state/schema/delta/56/state_group_room_idx.sql)0
-rw-r--r--synapse/storage/databases/state/schema/full_schemas/54/full.sql (renamed from synapse/storage/data_stores/state/schema/full_schemas/54/full.sql)0
-rw-r--r--synapse/storage/databases/state/schema/full_schemas/54/sequence.sql.postgres (renamed from synapse/storage/data_stores/state/schema/full_schemas/54/sequence.sql.postgres)0
-rw-r--r--synapse/storage/databases/state/store.py (renamed from synapse/storage/data_stores/state/store.py)104
-rw-r--r--synapse/storage/engines/_base.py17
-rw-r--r--synapse/storage/engines/postgres.py10
-rw-r--r--synapse/storage/engines/sqlite.py10
-rw-r--r--synapse/storage/keys.py2
-rw-r--r--synapse/storage/persist_events.py71
-rw-r--r--synapse/storage/prepare_database.py160
-rw-r--r--synapse/storage/purge_events.py40
-rw-r--r--synapse/storage/relations.py8
-rw-r--r--synapse/storage/roommember.py2
-rw-r--r--synapse/storage/state.py238
-rw-r--r--synapse/storage/util/id_generators.py512
-rw-r--r--synapse/storage/util/sequence.py98
-rw-r--r--synapse/streams/config.py70
-rw-r--r--synapse/streams/events.py26
-rw-r--r--synapse/types.py281
-rw-r--r--synapse/util/__init__.py18
-rw-r--r--synapse/util/async_helpers.py188
-rw-r--r--synapse/util/caches/__init__.py4
-rw-r--r--synapse/util/caches/descriptors.py102
-rw-r--r--synapse/util/caches/dictionary_cache.py4
-rw-r--r--synapse/util/caches/expiringcache.py4
-rw-r--r--synapse/util/caches/lrucache.py4
-rw-r--r--synapse/util/caches/response_cache.py2
-rw-r--r--synapse/util/caches/treecache.py4
-rw-r--r--synapse/util/caches/ttlcache.py4
-rw-r--r--synapse/util/daemonize.py137
-rw-r--r--synapse/util/distributor.py54
-rw-r--r--synapse/util/file_consumer.py2
-rw-r--r--synapse/util/frozenutils.py10
-rw-r--r--synapse/util/jsonobject.py2
-rw-r--r--synapse/util/manhole.py2
-rw-r--r--synapse/util/metrics.py79
-rw-r--r--synapse/util/patch_inline_callbacks.py2
-rw-r--r--synapse/util/ratelimitutils.py4
-rw-r--r--synapse/util/retryutils.py20
-rw-r--r--synapse/util/stringutils.py4
-rw-r--r--synapse/util/wheel_timer.py4
-rw-r--r--synapse/visibility.py30
-rw-r--r--synmark/__init__.py6
-rw-r--r--sytest-blacklist8
-rw-r--r--tests/api/test_auth.py71
-rw-r--r--tests/api/test_filtering.py72
-rw-r--r--tests/api/test_ratelimiting.py73
-rw-r--r--tests/appservice/test_appservice.py89
-rw-r--r--tests/appservice/test_scheduler.py19
-rw-r--r--tests/config/test_base.py82
-rw-r--r--tests/crypto/test_keyring.py175
-rw-r--r--tests/federation/test_complexity.py124
-rw-r--r--tests/federation/test_federation_catch_up.py422
-rw-r--r--tests/federation/test_federation_sender.py12
-rw-r--r--tests/federation/test_federation_server.py33
-rw-r--r--tests/federation/transport/test_server.py2
-rw-r--r--tests/handlers/test_appservice.py7
-rw-r--r--tests/handlers/test_auth.py23
-rw-r--r--tests/handlers/test_device.py11
-rw-r--r--tests/handlers/test_directory.py5
-rw-r--r--tests/handlers/test_e2e_keys.py2
-rw-r--r--tests/handlers/test_e2e_room_keys.py2
-rw-r--r--tests/handlers/test_oidc.py254
-rw-r--r--tests/handlers/test_presence.py2
-rw-r--r--tests/handlers/test_profile.py64
-rw-r--r--tests/handlers/test_register.py81
-rw-r--r--tests/handlers/test_stats.py97
-rw-r--r--tests/handlers/test_typing.py51
-rw-r--r--tests/handlers/test_user_directory.py28
-rw-r--r--tests/http/__init__.py4
-rw-r--r--tests/http/federation/test_matrix_federation_agent.py26
-rw-r--r--tests/http/test_fedclient.py108
-rw-r--r--tests/http/test_servlet.py80
-rw-r--r--tests/http/test_simple_client.py180
-rw-r--r--tests/logging/test_structured.py4
-rw-r--r--tests/module_api/test_api.py2
-rw-r--r--tests/push/test_email.py4
-rw-r--r--tests/replication/_base.py6
-rw-r--r--tests/replication/slave/storage/test_events.py28
-rw-r--r--tests/replication/test_federation_sender_shard.py17
-rw-r--r--tests/rest/admin/test_admin.py4
-rw-r--r--tests/rest/admin/test_device.py2
-rw-r--r--tests/rest/admin/test_event_reports.py382
-rw-r--r--tests/rest/admin/test_room.py63
-rw-r--r--tests/rest/admin/test_user.py119
-rw-r--r--tests/rest/client/test_retention.py98
-rw-r--r--tests/rest/client/test_shadow_banned.py312
-rw-r--r--tests/rest/client/test_third_party_rules.py170
-rw-r--r--tests/rest/client/third_party_rules.py86
-rw-r--r--tests/rest/client/v1/test_login.py150
-rw-r--r--tests/rest/client/v1/test_profile.py4
-rw-r--r--tests/rest/client/v1/test_push_rule_attrs.py448
-rw-r--r--tests/rest/client/v1/test_rooms.py116
-rw-r--r--tests/rest/client/v1/test_typing.py6
-rw-r--r--tests/rest/client/v1/utils.py36
-rw-r--r--tests/rest/client/v2_alpha/test_account.py35
-rw-r--r--tests/rest/client/v2_alpha/test_filter.py8
-rw-r--r--tests/rest/client/v2_alpha/test_register.py6
-rw-r--r--tests/rest/client/v2_alpha/test_shared_rooms.py138
-rw-r--r--tests/rest/client/v2_alpha/test_sync.py157
-rw-r--r--tests/rest/key/v2/test_remote_key_resource.py4
-rw-r--r--tests/rest/media/v1/test_media_storage.py39
-rw-r--r--tests/rest/media/v1/test_url_preview.py6
-rw-r--r--tests/rest/test_health.py34
-rw-r--r--tests/rest/test_well_known.py2
-rw-r--r--tests/server.py23
-rw-r--r--tests/server_notices/test_resource_limits_server_notices.py19
-rw-r--r--tests/state/test_v2.py4
-rw-r--r--tests/storage/test__base.py34
-rw-r--r--tests/storage/test_appservice.py73
-rw-r--r--tests/storage/test_background_update.py20
-rw-r--r--tests/storage/test_base.py81
-rw-r--r--tests/storage/test_cleanup_extrems.py21
-rw-r--r--tests/storage/test_client_ips.py31
-rw-r--r--tests/storage/test_devices.py54
-rw-r--r--tests/storage/test_directory.py38
-rw-r--r--tests/storage/test_end_to_end_keys.py62
-rw-r--r--tests/storage/test_event_federation.py16
-rw-r--r--tests/storage/test_event_metrics.py21
-rw-r--r--tests/storage/test_event_push_actions.py138
-rw-r--r--tests/storage/test_id_generators.py436
-rw-r--r--tests/storage/test_main.py14
-rw-r--r--tests/storage/test_monthly_active_users.py40
-rw-r--r--tests/storage/test_profile.py27
-rw-r--r--tests/storage/test_purge.py56
-rw-r--r--tests/storage/test_redaction.py12
-rw-r--r--tests/storage/test_registration.py95
-rw-r--r--tests/storage/test_room.py50
-rw-r--r--tests/storage/test_roommember.py20
-rw-r--r--tests/storage/test_state.py76
-rw-r--r--tests/storage/test_user_directory.py20
-rw-r--r--tests/test_federation.py72
-rw-r--r--tests/test_server.py44
-rw-r--r--tests/test_state.py106
-rw-r--r--tests/test_utils/__init__.py13
-rw-r--r--tests/test_utils/event_injection.py12
-rw-r--r--tests/test_utils/logging_setup.py3
-rw-r--r--tests/test_visibility.py33
-rw-r--r--tests/unittest.py90
-rw-r--r--tests/util/caches/test_descriptors.py32
-rw-r--r--tests/util/test_file_consumer.py4
-rw-r--r--tests/util/test_retryutils.py46
-rw-r--r--tests/util/test_rwlock.py6
-rw-r--r--tests/util/test_stringutils.py3
-rw-r--r--tests/utils.py26
-rw-r--r--tox.ini62
762 files changed, 21506 insertions, 11692 deletions
diff --git a/.buildkite/test_db.db b/.buildkite/test_db.db

index f20567ba73..361369a581 100644 --- a/.buildkite/test_db.db +++ b/.buildkite/test_db.db
Binary files differdiff --git a/.circleci/config.yml b/.circleci/config.yml
index 98c217dd1d..5bd2ab2b76 100644 --- a/.circleci/config.yml +++ b/.circleci/config.yml
@@ -4,18 +4,16 @@ jobs: machine: true steps: - checkout - - run: docker build -f docker/Dockerfile --label gitsha1=${CIRCLE_SHA1} -t matrixdotorg/synapse:${CIRCLE_TAG} -t matrixdotorg/synapse:${CIRCLE_TAG}-py3 . + - run: docker build -f docker/Dockerfile --label gitsha1=${CIRCLE_SHA1} -t matrixdotorg/synapse:${CIRCLE_TAG} . - run: docker login --username $DOCKER_HUB_USERNAME --password $DOCKER_HUB_PASSWORD - run: docker push matrixdotorg/synapse:${CIRCLE_TAG} - - run: docker push matrixdotorg/synapse:${CIRCLE_TAG}-py3 dockerhubuploadlatest: machine: true steps: - checkout - - run: docker build -f docker/Dockerfile --label gitsha1=${CIRCLE_SHA1} -t matrixdotorg/synapse:latest -t matrixdotorg/synapse:latest-py3 . + - run: docker build -f docker/Dockerfile --label gitsha1=${CIRCLE_SHA1} -t matrixdotorg/synapse:latest . - run: docker login --username $DOCKER_HUB_USERNAME --password $DOCKER_HUB_PASSWORD - run: docker push matrixdotorg/synapse:latest - - run: docker push matrixdotorg/synapse:latest-py3 workflows: version: 2 diff --git a/.github/ISSUE_TEMPLATE/BUG_REPORT.md b/.github/ISSUE_TEMPLATE/BUG_REPORT.md
index 75c9b2c9fe..978b699886 100644 --- a/.github/ISSUE_TEMPLATE/BUG_REPORT.md +++ b/.github/ISSUE_TEMPLATE/BUG_REPORT.md
@@ -4,12 +4,12 @@ about: Create a report to help us improve --- +<!-- + **THIS IS NOT A SUPPORT CHANNEL!** **IF YOU HAVE SUPPORT QUESTIONS ABOUT RUNNING OR CONFIGURING YOUR OWN HOME SERVER**, please ask in **#synapse:matrix.org** (using a matrix.org account if necessary) -<!-- - If you want to report a security issue, please see https://matrix.org/security-disclosure-policy/ This is a bug report template. By following the instructions below and diff --git a/CHANGES.md b/CHANGES.md
index 6c986808eb..38a0814bbf 100644 --- a/CHANGES.md +++ b/CHANGES.md
@@ -1,3 +1,451 @@ +Synapse 1.21.2 (2020-10-15) +=========================== + +Debian packages and Docker images have been rebuilt using the latest versions of dependency libraries, including authlib 0.15.1. Please see bugfixes below. + +Security advisory +----------------- + +* HTML pages served via Synapse were vulnerable to cross-site scripting (XSS) + attacks. All server administrators are encouraged to upgrade. + ([\#8444](https://github.com/matrix-org/synapse/pull/8444)) + ([CVE-2020-26891](https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2020-26891)) + + This fix was originally included in v1.21.0 but was missing a security advisory. + + This was reported by [Denis Kasak](https://github.com/dkasak). + +Bugfixes +-------- + +- Fix rare bug where sending an event would fail due to a racey assertion. ([\#8530](https://github.com/matrix-org/synapse/issues/8530)) +- An updated version of the authlib dependency is included in the Docker and Debian images to fix an issue using OpenID Connect. See [\#8534](https://github.com/matrix-org/synapse/issues/8534) for details. + + +Synapse 1.21.1 (2020-10-13) +=========================== + +This release fixes a regression in v1.21.0 that prevented debian packages from being built. +It is otherwise identical to v1.21.0. + +Synapse 1.21.0 (2020-10-12) +=========================== + +No significant changes since v1.21.0rc3. + +As [noted in +v1.20.0](https://github.com/matrix-org/synapse/blob/release-v1.21.0/CHANGES.md#synapse-1200-2020-09-22), +a future release will drop support for accessing Synapse's +[Admin API](https://github.com/matrix-org/synapse/tree/master/docs/admin_api) under the +`/_matrix/client/*` endpoint prefixes. At that point, the Admin API will only +be accessible under `/_synapse/admin`. + + +Synapse 1.21.0rc3 (2020-10-08) +============================== + +Bugfixes +-------- + +- Fix duplication of events on high traffic servers, caused by PostgreSQL `could not serialize access due to concurrent update` errors. ([\#8456](https://github.com/matrix-org/synapse/issues/8456)) + + +Internal Changes +---------------- + +- Add Groovy Gorilla to the list of distributions we build `.deb`s for. ([\#8475](https://github.com/matrix-org/synapse/issues/8475)) + + +Synapse 1.21.0rc2 (2020-10-02) +============================== + +Features +-------- + +- Convert additional templates from inline HTML to Jinja2 templates. ([\#8444](https://github.com/matrix-org/synapse/issues/8444)) + +Bugfixes +-------- + +- Fix a regression in v1.21.0rc1 which broke thumbnails of remote media. ([\#8438](https://github.com/matrix-org/synapse/issues/8438)) +- Do not expose the experimental `uk.half-shot.msc2778.login.application_service` flow in the login API, which caused a compatibility problem with Element iOS. ([\#8440](https://github.com/matrix-org/synapse/issues/8440)) +- Fix malformed log line in new federation "catch up" logic. ([\#8442](https://github.com/matrix-org/synapse/issues/8442)) +- Fix DB query on startup for negative streams which caused long start up times. Introduced in [\#8374](https://github.com/matrix-org/synapse/issues/8374). ([\#8447](https://github.com/matrix-org/synapse/issues/8447)) + + +Synapse 1.21.0rc1 (2020-10-01) +============================== + +Features +-------- + +- Require the user to confirm that their password should be reset after clicking the email confirmation link. ([\#8004](https://github.com/matrix-org/synapse/issues/8004)) +- Add an admin API `GET /_synapse/admin/v1/event_reports` to read entries of table `event_reports`. Contributed by @dklimpel. ([\#8217](https://github.com/matrix-org/synapse/issues/8217)) +- Consolidate the SSO error template across all configuration. ([\#8248](https://github.com/matrix-org/synapse/issues/8248), [\#8405](https://github.com/matrix-org/synapse/issues/8405)) +- Add a configuration option to specify a whitelist of domains that a user can be redirected to after validating their email or phone number. ([\#8275](https://github.com/matrix-org/synapse/issues/8275), [\#8417](https://github.com/matrix-org/synapse/issues/8417)) +- Add experimental support for sharding event persister. ([\#8294](https://github.com/matrix-org/synapse/issues/8294), [\#8387](https://github.com/matrix-org/synapse/issues/8387), [\#8396](https://github.com/matrix-org/synapse/issues/8396), [\#8419](https://github.com/matrix-org/synapse/issues/8419)) +- Add the room topic and avatar to the room details admin API. ([\#8305](https://github.com/matrix-org/synapse/issues/8305)) +- Add an admin API for querying rooms where a user is a member. Contributed by @dklimpel. ([\#8306](https://github.com/matrix-org/synapse/issues/8306)) +- Add `uk.half-shot.msc2778.login.application_service` login type to allow appservices to login. ([\#8320](https://github.com/matrix-org/synapse/issues/8320)) +- Add a configuration option that allows existing users to log in with OpenID Connect. Contributed by @BBBSnowball and @OmmyZhang. ([\#8345](https://github.com/matrix-org/synapse/issues/8345)) +- Add prometheus metrics for replication requests. ([\#8406](https://github.com/matrix-org/synapse/issues/8406)) +- Support passing additional single sign-on parameters to the client. ([\#8413](https://github.com/matrix-org/synapse/issues/8413)) +- Add experimental reporting of metrics on expensive rooms for state-resolution. ([\#8420](https://github.com/matrix-org/synapse/issues/8420)) +- Add experimental prometheus metric to track numbers of "large" rooms for state resolutiom. ([\#8425](https://github.com/matrix-org/synapse/issues/8425)) +- Add prometheus metrics to track federation delays. ([\#8430](https://github.com/matrix-org/synapse/issues/8430)) + + +Bugfixes +-------- + +- Fix a bug in the media repository where remote thumbnails with the same size but different crop methods would overwrite each other. Contributed by @deepbluev7. ([\#7124](https://github.com/matrix-org/synapse/issues/7124)) +- Fix inconsistent handling of non-existent push rules, and stop tracking the `enabled` state of removed push rules. ([\#7796](https://github.com/matrix-org/synapse/issues/7796)) +- Fix a longstanding bug when storing a media file with an empty `upload_name`. ([\#7905](https://github.com/matrix-org/synapse/issues/7905)) +- Fix messages not being sent over federation until an event is sent into the same room. ([\#8230](https://github.com/matrix-org/synapse/issues/8230), [\#8247](https://github.com/matrix-org/synapse/issues/8247), [\#8258](https://github.com/matrix-org/synapse/issues/8258), [\#8272](https://github.com/matrix-org/synapse/issues/8272), [\#8322](https://github.com/matrix-org/synapse/issues/8322)) +- Fix a longstanding bug where files that could not be thumbnailed would result in an Internal Server Error. ([\#8236](https://github.com/matrix-org/synapse/issues/8236), [\#8435](https://github.com/matrix-org/synapse/issues/8435)) +- Upgrade minimum version of `canonicaljson` to version 1.4.0, to fix an unicode encoding issue. ([\#8262](https://github.com/matrix-org/synapse/issues/8262)) +- Fix longstanding bug which could lead to incomplete database upgrades on SQLite. ([\#8265](https://github.com/matrix-org/synapse/issues/8265)) +- Fix stack overflow when stderr is redirected to the logging system, and the logging system encounters an error. ([\#8268](https://github.com/matrix-org/synapse/issues/8268)) +- Fix a bug which cause the logging system to report errors, if `DEBUG` was enabled and no `context` filter was applied. ([\#8278](https://github.com/matrix-org/synapse/issues/8278)) +- Fix edge case where push could get delayed for a user until a later event was pushed. ([\#8287](https://github.com/matrix-org/synapse/issues/8287)) +- Fix fetching malformed events from remote servers. ([\#8324](https://github.com/matrix-org/synapse/issues/8324)) +- Fix `UnboundLocalError` from occuring when appservices send a malformed register request. ([\#8329](https://github.com/matrix-org/synapse/issues/8329)) +- Don't send push notifications to expired user accounts. ([\#8353](https://github.com/matrix-org/synapse/issues/8353)) +- Fix a regression in v1.19.0 with reactivating users through the admin API. ([\#8362](https://github.com/matrix-org/synapse/issues/8362)) +- Fix a bug where during device registration the length of the device name wasn't limited. ([\#8364](https://github.com/matrix-org/synapse/issues/8364)) +- Include `guest_access` in the fields that are checked for null bytes when updating `room_stats_state`. Broke in v1.7.2. ([\#8373](https://github.com/matrix-org/synapse/issues/8373)) +- Fix theoretical race condition where events are not sent down `/sync` if the synchrotron worker is restarted without restarting other workers. ([\#8374](https://github.com/matrix-org/synapse/issues/8374)) +- Fix a bug which could cause errors in rooms with malformed membership events, on servers using sqlite. ([\#8385](https://github.com/matrix-org/synapse/issues/8385)) +- Fix "Re-starting finished log context" warning when receiving an event we already had over federation. ([\#8398](https://github.com/matrix-org/synapse/issues/8398)) +- Fix incorrect handling of timeouts on outgoing HTTP requests. ([\#8400](https://github.com/matrix-org/synapse/issues/8400)) +- Fix a regression in v1.20.0 in the `synapse_port_db` script regarding the `ui_auth_sessions_ips` table. ([\#8410](https://github.com/matrix-org/synapse/issues/8410)) +- Remove unnecessary 3PID registration check when resetting password via an email address. Bug introduced in v0.34.0rc2. ([\#8414](https://github.com/matrix-org/synapse/issues/8414)) + + +Improved Documentation +---------------------- + +- Add `/_synapse/client` to the reverse proxy documentation. ([\#8227](https://github.com/matrix-org/synapse/issues/8227)) +- Add note to the reverse proxy settings documentation about disabling Apache's mod_security2. Contributed by Julian Fietkau (@jfietkau). ([\#8375](https://github.com/matrix-org/synapse/issues/8375)) +- Improve description of `server_name` config option in `homserver.yaml`. ([\#8415](https://github.com/matrix-org/synapse/issues/8415)) + + +Deprecations and Removals +------------------------- + +- Drop support for `prometheus_client` older than 0.4.0. ([\#8426](https://github.com/matrix-org/synapse/issues/8426)) + + +Internal Changes +---------------- + +- Fix tests on distros which disable TLSv1.0. Contributed by @danc86. ([\#8208](https://github.com/matrix-org/synapse/issues/8208)) +- Simplify the distributor code to avoid unnecessary work. ([\#8216](https://github.com/matrix-org/synapse/issues/8216)) +- Remove the `populate_stats_process_rooms_2` background job and restore functionality to `populate_stats_process_rooms`. ([\#8243](https://github.com/matrix-org/synapse/issues/8243)) +- Clean up type hints for `PaginationConfig`. ([\#8250](https://github.com/matrix-org/synapse/issues/8250), [\#8282](https://github.com/matrix-org/synapse/issues/8282)) +- Track the latest event for every destination and room for catch-up after federation outage. ([\#8256](https://github.com/matrix-org/synapse/issues/8256)) +- Fix non-user visible bug in implementation of `MultiWriterIdGenerator.get_current_token_for_writer`. ([\#8257](https://github.com/matrix-org/synapse/issues/8257)) +- Switch to the JSON implementation from the standard library. ([\#8259](https://github.com/matrix-org/synapse/issues/8259)) +- Add type hints to `synapse.util.async_helpers`. ([\#8260](https://github.com/matrix-org/synapse/issues/8260)) +- Simplify tests that mock asynchronous functions. ([\#8261](https://github.com/matrix-org/synapse/issues/8261)) +- Add type hints to `StreamToken` and `RoomStreamToken` classes. ([\#8279](https://github.com/matrix-org/synapse/issues/8279)) +- Change `StreamToken.room_key` to be a `RoomStreamToken` instance. ([\#8281](https://github.com/matrix-org/synapse/issues/8281)) +- Refactor notifier code to correctly use the max event stream position. ([\#8288](https://github.com/matrix-org/synapse/issues/8288)) +- Use slotted classes where possible. ([\#8296](https://github.com/matrix-org/synapse/issues/8296)) +- Support testing the local Synapse checkout against the [Complement homeserver test suite](https://github.com/matrix-org/complement/). ([\#8317](https://github.com/matrix-org/synapse/issues/8317)) +- Update outdated usages of `metaclass` to python 3 syntax. ([\#8326](https://github.com/matrix-org/synapse/issues/8326)) +- Move lint-related dependencies to package-extra field, update CONTRIBUTING.md to utilise this. ([\#8330](https://github.com/matrix-org/synapse/issues/8330), [\#8377](https://github.com/matrix-org/synapse/issues/8377)) +- Use the `admin_patterns` helper in additional locations. ([\#8331](https://github.com/matrix-org/synapse/issues/8331)) +- Fix test logging to allow braces in log output. ([\#8335](https://github.com/matrix-org/synapse/issues/8335)) +- Remove `__future__` imports related to Python 2 compatibility. ([\#8337](https://github.com/matrix-org/synapse/issues/8337)) +- Simplify `super()` calls to Python 3 syntax. ([\#8344](https://github.com/matrix-org/synapse/issues/8344)) +- Fix bad merge from `release-v1.20.0` branch to `develop`. ([\#8354](https://github.com/matrix-org/synapse/issues/8354)) +- Factor out a `_send_dummy_event_for_room` method. ([\#8370](https://github.com/matrix-org/synapse/issues/8370)) +- Improve logging of state resolution. ([\#8371](https://github.com/matrix-org/synapse/issues/8371)) +- Add type annotations to `SimpleHttpClient`. ([\#8372](https://github.com/matrix-org/synapse/issues/8372)) +- Refactor ID generators to use `async with` syntax. ([\#8383](https://github.com/matrix-org/synapse/issues/8383)) +- Add `EventStreamPosition` type. ([\#8388](https://github.com/matrix-org/synapse/issues/8388)) +- Create a mechanism for marking tests "logcontext clean". ([\#8399](https://github.com/matrix-org/synapse/issues/8399)) +- A pair of tiny cleanups in the federation request code. ([\#8401](https://github.com/matrix-org/synapse/issues/8401)) +- Add checks on startup that PostgreSQL sequences are consistent with their associated tables. ([\#8402](https://github.com/matrix-org/synapse/issues/8402)) +- Do not include appservice users when calculating the total MAU for a server. ([\#8404](https://github.com/matrix-org/synapse/issues/8404)) +- Typing fixes for `synapse.handlers.federation`. ([\#8422](https://github.com/matrix-org/synapse/issues/8422)) +- Various refactors to simplify stream token handling. ([\#8423](https://github.com/matrix-org/synapse/issues/8423)) +- Make stream token serializing/deserializing async. ([\#8427](https://github.com/matrix-org/synapse/issues/8427)) + + +Synapse 1.20.1 (2020-09-24) +=========================== + +Bugfixes +-------- + +- Fix a bug introduced in v1.20.0 which caused the `synapse_port_db` script to fail. ([\#8386](https://github.com/matrix-org/synapse/issues/8386)) +- Fix a bug introduced in v1.20.0 which caused variables to be incorrectly escaped in Jinja2 templates. ([\#8394](https://github.com/matrix-org/synapse/issues/8394)) + + +Synapse 1.20.0 (2020-09-22) +=========================== + +No significant changes since v1.20.0rc5. + +Removal warning +--------------- + +Historically, the [Synapse Admin +API](https://github.com/matrix-org/synapse/tree/master/docs) has been +accessible under the `/_matrix/client/api/v1/admin`, +`/_matrix/client/unstable/admin`, `/_matrix/client/r0/admin` and +`/_synapse/admin` prefixes. In a future release, we will be dropping support +for accessing Synapse's Admin API using the `/_matrix/client/*` prefixes. + +From that point, the Admin API will only be accessible under `/_synapse/admin`. +This makes it easier for homeserver admins to lock down external access to the +Admin API endpoints. + +Synapse 1.20.0rc5 (2020-09-18) +============================== + +In addition to the below, Synapse 1.20.0rc5 also includes the bug fix that was included in 1.19.3. + +Features +-------- + +- Add flags to the `/versions` endpoint for whether new rooms default to using E2EE. ([\#8343](https://github.com/matrix-org/synapse/issues/8343)) + + +Bugfixes +-------- + +- Fix rate limiting of federation `/send` requests. ([\#8342](https://github.com/matrix-org/synapse/issues/8342)) +- Fix a longstanding bug where back pagination over federation could get stuck if it failed to handle a received event. ([\#8349](https://github.com/matrix-org/synapse/issues/8349)) + + +Internal Changes +---------------- + +- Blacklist [MSC2753](https://github.com/matrix-org/matrix-doc/pull/2753) SyTests until it is implemented. ([\#8285](https://github.com/matrix-org/synapse/issues/8285)) + + +Synapse 1.19.3 (2020-09-18) +=========================== + +Bugfixes +-------- + +- Partially mitigate bug where newly joined servers couldn't get past events in a room when there is a malformed event. ([\#8350](https://github.com/matrix-org/synapse/issues/8350)) + + +Synapse 1.20.0rc4 (2020-09-16) +============================== + +Synapse 1.20.0rc4 is identical to 1.20.0rc3, with the addition of the security fix that was included in 1.19.2. + + +Synapse 1.19.2 (2020-09-16) +=========================== + +Due to the issue below server admins are encouraged to upgrade as soon as possible. + +Bugfixes +-------- + +- Fix joining rooms over federation that include malformed events. ([\#8324](https://github.com/matrix-org/synapse/issues/8324)) + + +Synapse 1.20.0rc3 (2020-09-11) +============================== + +Bugfixes +-------- + +- Fix a bug introduced in v1.20.0rc1 where the wrong exception was raised when invalid JSON data is encountered. ([\#8291](https://github.com/matrix-org/synapse/issues/8291)) + + +Synapse 1.20.0rc2 (2020-09-09) +============================== + +Bugfixes +-------- + +- Fix a bug introduced in v1.20.0rc1 causing some features related to notifications to misbehave following the implementation of unread counts. ([\#8280](https://github.com/matrix-org/synapse/issues/8280)) + + +Synapse 1.20.0rc1 (2020-09-08) +============================== + +Removal warning +--------------- + +Some older clients used a [disallowed character](https://matrix.org/docs/spec/client_server/r0.6.1#post-matrix-client-r0-register-email-requesttoken) (`:`) in the `client_secret` parameter of various endpoints. The incorrect behaviour was allowed for backwards compatibility, but is now being removed from Synapse as most users have updated their client. Further context can be found at [\#6766](https://github.com/matrix-org/synapse/issues/6766). + +Features +-------- + +- Add an endpoint to query your shared rooms with another user as an implementation of [MSC2666](https://github.com/matrix-org/matrix-doc/pull/2666). ([\#7785](https://github.com/matrix-org/synapse/issues/7785)) +- Iteratively encode JSON to avoid blocking the reactor. ([\#8013](https://github.com/matrix-org/synapse/issues/8013), [\#8116](https://github.com/matrix-org/synapse/issues/8116)) +- Add support for shadow-banning users (ignoring any message send requests). ([\#8034](https://github.com/matrix-org/synapse/issues/8034), [\#8092](https://github.com/matrix-org/synapse/issues/8092), [\#8095](https://github.com/matrix-org/synapse/issues/8095), [\#8142](https://github.com/matrix-org/synapse/issues/8142), [\#8152](https://github.com/matrix-org/synapse/issues/8152), [\#8157](https://github.com/matrix-org/synapse/issues/8157), [\#8158](https://github.com/matrix-org/synapse/issues/8158), [\#8176](https://github.com/matrix-org/synapse/issues/8176)) +- Use the default template file when its equivalent is not found in a custom template directory. ([\#8037](https://github.com/matrix-org/synapse/issues/8037), [\#8107](https://github.com/matrix-org/synapse/issues/8107), [\#8252](https://github.com/matrix-org/synapse/issues/8252)) +- Add unread messages count to sync responses, as specified in [MSC2654](https://github.com/matrix-org/matrix-doc/pull/2654). ([\#8059](https://github.com/matrix-org/synapse/issues/8059), [\#8254](https://github.com/matrix-org/synapse/issues/8254), [\#8270](https://github.com/matrix-org/synapse/issues/8270), [\#8274](https://github.com/matrix-org/synapse/issues/8274)) +- Optimise `/federation/v1/user/devices/` API by only returning devices with encryption keys. ([\#8198](https://github.com/matrix-org/synapse/issues/8198)) + + +Bugfixes +-------- + +- Fix a memory leak by limiting the length of time that messages will be queued for a remote server that has been unreachable. ([\#7864](https://github.com/matrix-org/synapse/issues/7864)) +- Fix `Re-starting finished log context PUT-nnnn` warning when event persistence failed. ([\#8081](https://github.com/matrix-org/synapse/issues/8081)) +- Synapse now correctly enforces the valid characters in the `client_secret` parameter used in various endpoints. ([\#8101](https://github.com/matrix-org/synapse/issues/8101)) +- Fix a bug introduced in v1.7.2 impacting message retention policies that would allow federated homeservers to dictate a retention period that's lower than the configured minimum allowed duration in the configuration file. ([\#8104](https://github.com/matrix-org/synapse/issues/8104)) +- Fix a long-standing bug where invalid JSON would be accepted by Synapse. ([\#8106](https://github.com/matrix-org/synapse/issues/8106)) +- Fix a bug introduced in Synapse v1.12.0 which could cause `/sync` requests to fail with a 404 if you had a very old outstanding room invite. ([\#8110](https://github.com/matrix-org/synapse/issues/8110)) +- Return a proper error code when the rooms of an invalid group are requested. ([\#8129](https://github.com/matrix-org/synapse/issues/8129)) +- Fix a bug which could cause a leaked postgres connection if synapse was set to daemonize. ([\#8131](https://github.com/matrix-org/synapse/issues/8131)) +- Clarify the error code if a user tries to register with a numeric ID. This bug was introduced in v1.15.0. ([\#8135](https://github.com/matrix-org/synapse/issues/8135)) +- Fix a bug where appservices with ratelimiting disabled would still be ratelimited when joining rooms. This bug was introduced in v1.19.0. ([\#8139](https://github.com/matrix-org/synapse/issues/8139)) +- Fix logging in via OpenID Connect with a provider that uses integer user IDs. ([\#8190](https://github.com/matrix-org/synapse/issues/8190)) +- Fix a longstanding bug where user directory updates could break when unexpected profile data was included in events. ([\#8223](https://github.com/matrix-org/synapse/issues/8223)) +- Fix a longstanding bug where stats updates could break when unexpected profile data was included in events. ([\#8226](https://github.com/matrix-org/synapse/issues/8226)) +- Fix slow start times for large servers by removing a table scan of the `users` table from startup code. ([\#8271](https://github.com/matrix-org/synapse/issues/8271)) + + +Updates to the Docker image +--------------------------- + +- Fix builds of the Docker image on non-x86 platforms. ([\#8144](https://github.com/matrix-org/synapse/issues/8144)) +- Added curl for healthcheck support and readme updates for the change. Contributed by @maquis196. ([\#8147](https://github.com/matrix-org/synapse/issues/8147)) + + +Improved Documentation +---------------------- + +- Link to matrix-synapse-rest-password-provider in the password provider documentation. ([\#8111](https://github.com/matrix-org/synapse/issues/8111)) +- Updated documentation to note that Synapse does not follow `HTTP 308` redirects due to an upstream library not supporting them. Contributed by Ryan Cole. ([\#8120](https://github.com/matrix-org/synapse/issues/8120)) +- Explain better what GDPR-erased means when deactivating a user. ([\#8189](https://github.com/matrix-org/synapse/issues/8189)) + + +Internal Changes +---------------- + +- Add filter `name` to the `/users` admin API, which filters by user ID or displayname. Contributed by Awesome Technologies Innovationslabor GmbH. ([\#7377](https://github.com/matrix-org/synapse/issues/7377), [\#8163](https://github.com/matrix-org/synapse/issues/8163)) +- Reduce run times of some unit tests by advancing the reactor a fewer number of times. ([\#7757](https://github.com/matrix-org/synapse/issues/7757)) +- Don't fail `/submit_token` requests on incorrect session ID if `request_token_inhibit_3pid_errors` is turned on. ([\#7991](https://github.com/matrix-org/synapse/issues/7991)) +- Convert various parts of the codebase to async/await. ([\#8071](https://github.com/matrix-org/synapse/issues/8071), [\#8072](https://github.com/matrix-org/synapse/issues/8072), [\#8074](https://github.com/matrix-org/synapse/issues/8074), [\#8075](https://github.com/matrix-org/synapse/issues/8075), [\#8076](https://github.com/matrix-org/synapse/issues/8076), [\#8087](https://github.com/matrix-org/synapse/issues/8087), [\#8100](https://github.com/matrix-org/synapse/issues/8100), [\#8119](https://github.com/matrix-org/synapse/issues/8119), [\#8121](https://github.com/matrix-org/synapse/issues/8121), [\#8133](https://github.com/matrix-org/synapse/issues/8133), [\#8156](https://github.com/matrix-org/synapse/issues/8156), [\#8162](https://github.com/matrix-org/synapse/issues/8162), [\#8166](https://github.com/matrix-org/synapse/issues/8166), [\#8168](https://github.com/matrix-org/synapse/issues/8168), [\#8173](https://github.com/matrix-org/synapse/issues/8173), [\#8191](https://github.com/matrix-org/synapse/issues/8191), [\#8192](https://github.com/matrix-org/synapse/issues/8192), [\#8193](https://github.com/matrix-org/synapse/issues/8193), [\#8194](https://github.com/matrix-org/synapse/issues/8194), [\#8195](https://github.com/matrix-org/synapse/issues/8195), [\#8197](https://github.com/matrix-org/synapse/issues/8197), [\#8199](https://github.com/matrix-org/synapse/issues/8199), [\#8200](https://github.com/matrix-org/synapse/issues/8200), [\#8201](https://github.com/matrix-org/synapse/issues/8201), [\#8202](https://github.com/matrix-org/synapse/issues/8202), [\#8207](https://github.com/matrix-org/synapse/issues/8207), [\#8213](https://github.com/matrix-org/synapse/issues/8213), [\#8214](https://github.com/matrix-org/synapse/issues/8214)) +- Remove some unused database functions. ([\#8085](https://github.com/matrix-org/synapse/issues/8085)) +- Add type hints to various parts of the codebase. ([\#8090](https://github.com/matrix-org/synapse/issues/8090), [\#8127](https://github.com/matrix-org/synapse/issues/8127), [\#8187](https://github.com/matrix-org/synapse/issues/8187), [\#8241](https://github.com/matrix-org/synapse/issues/8241), [\#8140](https://github.com/matrix-org/synapse/issues/8140), [\#8183](https://github.com/matrix-org/synapse/issues/8183), [\#8232](https://github.com/matrix-org/synapse/issues/8232), [\#8235](https://github.com/matrix-org/synapse/issues/8235), [\#8237](https://github.com/matrix-org/synapse/issues/8237), [\#8244](https://github.com/matrix-org/synapse/issues/8244)) +- Return the previous stream token if a non-member event is a duplicate. ([\#8093](https://github.com/matrix-org/synapse/issues/8093), [\#8112](https://github.com/matrix-org/synapse/issues/8112)) +- Separate `get_current_token` into two since there are two different use cases for it. ([\#8113](https://github.com/matrix-org/synapse/issues/8113)) +- Remove `ChainedIdGenerator`. ([\#8123](https://github.com/matrix-org/synapse/issues/8123)) +- Reduce the amount of whitespace in JSON stored and sent in responses. ([\#8124](https://github.com/matrix-org/synapse/issues/8124)) +- Update the test federation client to handle streaming responses. ([\#8130](https://github.com/matrix-org/synapse/issues/8130)) +- Micro-optimisations to `get_auth_chain_ids`. ([\#8132](https://github.com/matrix-org/synapse/issues/8132)) +- Refactor `StreamIdGenerator` and `MultiWriterIdGenerator` to have the same interface. ([\#8161](https://github.com/matrix-org/synapse/issues/8161)) +- Add functions to `MultiWriterIdGen` used by events stream. ([\#8164](https://github.com/matrix-org/synapse/issues/8164), [\#8179](https://github.com/matrix-org/synapse/issues/8179)) +- Fix tests that were broken due to the merge of 1.19.1. ([\#8167](https://github.com/matrix-org/synapse/issues/8167)) +- Make `SlavedIdTracker.advance` have the same interface as `MultiWriterIDGenerator`. ([\#8171](https://github.com/matrix-org/synapse/issues/8171)) +- Remove unused `is_guest` parameter from, and add safeguard to, `MessageHandler.get_room_data`. ([\#8174](https://github.com/matrix-org/synapse/issues/8174), [\#8181](https://github.com/matrix-org/synapse/issues/8181)) +- Standardize the mypy configuration. ([\#8175](https://github.com/matrix-org/synapse/issues/8175)) +- Refactor some of `LoginRestServlet`'s helper methods, and move them to `AuthHandler` for easier reuse. ([\#8182](https://github.com/matrix-org/synapse/issues/8182)) +- Fix `wait_for_stream_position` to allow multiple waiters on same stream ID. ([\#8196](https://github.com/matrix-org/synapse/issues/8196)) +- Make `MultiWriterIDGenerator` work for streams that use negative values. ([\#8203](https://github.com/matrix-org/synapse/issues/8203)) +- Refactor queries for device keys and cross-signatures. ([\#8204](https://github.com/matrix-org/synapse/issues/8204), [\#8205](https://github.com/matrix-org/synapse/issues/8205), [\#8222](https://github.com/matrix-org/synapse/issues/8222), [\#8224](https://github.com/matrix-org/synapse/issues/8224), [\#8225](https://github.com/matrix-org/synapse/issues/8225), [\#8231](https://github.com/matrix-org/synapse/issues/8231), [\#8233](https://github.com/matrix-org/synapse/issues/8233), [\#8234](https://github.com/matrix-org/synapse/issues/8234)) +- Fix type hints for functions decorated with `@cached`. ([\#8240](https://github.com/matrix-org/synapse/issues/8240)) +- Remove obsolete `order` field from federation send queues. ([\#8245](https://github.com/matrix-org/synapse/issues/8245)) +- Stop sub-classing from object. ([\#8249](https://github.com/matrix-org/synapse/issues/8249)) +- Add more logging to debug slow startup. ([\#8264](https://github.com/matrix-org/synapse/issues/8264)) +- Do not attempt to upgrade database schema on worker processes. ([\#8266](https://github.com/matrix-org/synapse/issues/8266), [\#8276](https://github.com/matrix-org/synapse/issues/8276)) + + +Synapse 1.19.1 (2020-08-27) +=========================== + +No significant changes. + + +Synapse 1.19.1rc1 (2020-08-25) +============================== + +Bugfixes +-------- + +- Fix a bug introduced in v1.19.0 where appservices with ratelimiting disabled would still be ratelimited when joining rooms. ([\#8139](https://github.com/matrix-org/synapse/issues/8139)) +- Fix a bug introduced in v1.19.0 that would cause e.g. profile updates to fail due to incorrect application of rate limits on join requests. ([\#8153](https://github.com/matrix-org/synapse/issues/8153)) + + +Synapse 1.19.0 (2020-08-17) +=========================== + +No significant changes since 1.19.0rc1. + +Removal warning +--------------- + +As outlined in the [previous release](https://github.com/matrix-org/synapse/releases/tag/v1.18.0), we are no longer publishing Docker images with the `-py3` tag suffix. On top of that, we have also removed the `latest-py3` tag. Please see [the announcement in the upgrade notes for 1.18.0](https://github.com/matrix-org/synapse/blob/develop/UPGRADE.rst#upgrading-to-v1180). + + +Synapse 1.19.0rc1 (2020-08-13) +============================== + +Features +-------- + +- Add option to allow server admins to join rooms which fail complexity checks. Contributed by @lugino-emeritus. ([\#7902](https://github.com/matrix-org/synapse/issues/7902)) +- Add an option to purge room or not with delete room admin endpoint (`POST /_synapse/admin/v1/rooms/<room_id>/delete`). Contributed by @dklimpel. ([\#7964](https://github.com/matrix-org/synapse/issues/7964)) +- Add rate limiting to users joining rooms. ([\#8008](https://github.com/matrix-org/synapse/issues/8008)) +- Add a `/health` endpoint to every configured HTTP listener that can be used as a health check endpoint by load balancers. ([\#8048](https://github.com/matrix-org/synapse/issues/8048)) +- Allow login to be blocked based on the values of SAML attributes. ([\#8052](https://github.com/matrix-org/synapse/issues/8052)) +- Allow guest access to the `GET /_matrix/client/r0/rooms/{room_id}/members` endpoint, according to MSC2689. Contributed by Awesome Technologies Innovationslabor GmbH. ([\#7314](https://github.com/matrix-org/synapse/issues/7314)) + + +Bugfixes +-------- + +- Fix a bug introduced in Synapse v1.7.2 which caused inaccurate membership counts in the room directory. ([\#7977](https://github.com/matrix-org/synapse/issues/7977)) +- Fix a long standing bug: 'Duplicate key value violates unique constraint "event_relations_id"' when message retention is configured. ([\#7978](https://github.com/matrix-org/synapse/issues/7978)) +- Fix "no create event in auth events" when trying to reject invitation after inviter leaves. Bug introduced in Synapse v1.10.0. ([\#7980](https://github.com/matrix-org/synapse/issues/7980)) +- Fix various comments and minor discrepencies in server notices code. ([\#7996](https://github.com/matrix-org/synapse/issues/7996)) +- Fix a long standing bug where HTTP HEAD requests resulted in a 400 error. ([\#7999](https://github.com/matrix-org/synapse/issues/7999)) +- Fix a long-standing bug which caused two copies of some log lines to be written when synctl was used along with a MemoryHandler logger. ([\#8011](https://github.com/matrix-org/synapse/issues/8011), [\#8012](https://github.com/matrix-org/synapse/issues/8012)) + + +Updates to the Docker image +--------------------------- + +- We no longer publish Docker images with the `-py3` tag suffix, as [announced in the upgrade notes](https://github.com/matrix-org/synapse/blob/develop/UPGRADE.rst#upgrading-to-v1180). ([\#8056](https://github.com/matrix-org/synapse/issues/8056)) + + +Improved Documentation +---------------------- + +- Document how to set up a client .well-known file and fix several pieces of outdated documentation. ([\#7899](https://github.com/matrix-org/synapse/issues/7899)) +- Improve workers docs. ([\#7990](https://github.com/matrix-org/synapse/issues/7990), [\#8000](https://github.com/matrix-org/synapse/issues/8000)) +- Fix typo in `docs/workers.md`. ([\#7992](https://github.com/matrix-org/synapse/issues/7992)) +- Add documentation for how to undo a room shutdown. ([\#7998](https://github.com/matrix-org/synapse/issues/7998), [\#8010](https://github.com/matrix-org/synapse/issues/8010)) + + +Internal Changes +---------------- + +- Reduce the amount of whitespace in JSON stored and sent in responses. Contributed by David Vo. ([\#7372](https://github.com/matrix-org/synapse/issues/7372)) +- Switch to the JSON implementation from the standard library and bump the minimum version of the canonicaljson library to 1.2.0. ([\#7936](https://github.com/matrix-org/synapse/issues/7936), [\#7979](https://github.com/matrix-org/synapse/issues/7979)) +- Convert various parts of the codebase to async/await. ([\#7947](https://github.com/matrix-org/synapse/issues/7947), [\#7948](https://github.com/matrix-org/synapse/issues/7948), [\#7949](https://github.com/matrix-org/synapse/issues/7949), [\#7951](https://github.com/matrix-org/synapse/issues/7951), [\#7963](https://github.com/matrix-org/synapse/issues/7963), [\#7973](https://github.com/matrix-org/synapse/issues/7973), [\#7975](https://github.com/matrix-org/synapse/issues/7975), [\#7976](https://github.com/matrix-org/synapse/issues/7976), [\#7981](https://github.com/matrix-org/synapse/issues/7981), [\#7987](https://github.com/matrix-org/synapse/issues/7987), [\#7989](https://github.com/matrix-org/synapse/issues/7989), [\#8003](https://github.com/matrix-org/synapse/issues/8003), [\#8014](https://github.com/matrix-org/synapse/issues/8014), [\#8016](https://github.com/matrix-org/synapse/issues/8016), [\#8027](https://github.com/matrix-org/synapse/issues/8027), [\#8031](https://github.com/matrix-org/synapse/issues/8031), [\#8032](https://github.com/matrix-org/synapse/issues/8032), [\#8035](https://github.com/matrix-org/synapse/issues/8035), [\#8042](https://github.com/matrix-org/synapse/issues/8042), [\#8044](https://github.com/matrix-org/synapse/issues/8044), [\#8045](https://github.com/matrix-org/synapse/issues/8045), [\#8061](https://github.com/matrix-org/synapse/issues/8061), [\#8062](https://github.com/matrix-org/synapse/issues/8062), [\#8063](https://github.com/matrix-org/synapse/issues/8063), [\#8066](https://github.com/matrix-org/synapse/issues/8066), [\#8069](https://github.com/matrix-org/synapse/issues/8069), [\#8070](https://github.com/matrix-org/synapse/issues/8070)) +- Move some database-related log lines from the default logger to the database/transaction loggers. ([\#7952](https://github.com/matrix-org/synapse/issues/7952)) +- Add a script to detect source code files using non-unix line terminators. ([\#7965](https://github.com/matrix-org/synapse/issues/7965), [\#7970](https://github.com/matrix-org/synapse/issues/7970)) +- Log the SAML session ID during creation. ([\#7971](https://github.com/matrix-org/synapse/issues/7971)) +- Implement new experimental push rules for some users. ([\#7997](https://github.com/matrix-org/synapse/issues/7997)) +- Remove redundant and unreliable signature check for v1 Identity Service lookup responses. ([\#8001](https://github.com/matrix-org/synapse/issues/8001)) +- Improve the performance of the register endpoint. ([\#8009](https://github.com/matrix-org/synapse/issues/8009)) +- Reduce less useful output in the newsfragment CI step. Add a link to the changelog section of the contributing guide on error. ([\#8024](https://github.com/matrix-org/synapse/issues/8024)) +- Rename storage layer objects to be more sensible. ([\#8033](https://github.com/matrix-org/synapse/issues/8033)) +- Change the default log config to reduce disk I/O and storage for new servers. ([\#8040](https://github.com/matrix-org/synapse/issues/8040)) +- Add an assertion on `prev_events` in `create_new_client_event`. ([\#8041](https://github.com/matrix-org/synapse/issues/8041)) +- Add a comment to `ServerContextFactory` about the use of `SSLv23_METHOD`. ([\#8043](https://github.com/matrix-org/synapse/issues/8043)) +- Log `OPTIONS` requests at `DEBUG` rather than `INFO` level to reduce amount logged at `INFO`. ([\#8049](https://github.com/matrix-org/synapse/issues/8049)) +- Reduce amount of outbound request logging at `INFO` level. ([\#8050](https://github.com/matrix-org/synapse/issues/8050)) +- It is no longer necessary to explicitly define `filters` in the logging configuration. (Continuing to do so is redundant but harmless.) ([\#8051](https://github.com/matrix-org/synapse/issues/8051)) +- Add and improve type hints. ([\#8058](https://github.com/matrix-org/synapse/issues/8058), [\#8064](https://github.com/matrix-org/synapse/issues/8064), [\#8060](https://github.com/matrix-org/synapse/issues/8060), [\#8067](https://github.com/matrix-org/synapse/issues/8067)) + + Synapse 1.18.0 (2020-07-30) =========================== diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index 062413e925..524f82433d 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md
@@ -17,9 +17,9 @@ https://help.github.com/articles/using-pull-requests/) to ask us to pull your changes into our repo. Some other points to follow: - + * Please base your changes on the `develop` branch. - + * Please follow the [code style requirements](#code-style). * Please include a [changelog entry](#changelog) with each PR. @@ -46,7 +46,7 @@ locally. You'll need python 3.6 or later, and to install a number of tools: ``` # Install the dependencies -pip install -U black flake8 flake8-comprehensions isort +pip install -e ".[lint]" # Run the linter script ./scripts-dev/lint.sh diff --git a/INSTALL.md b/INSTALL.md
index b507de7442..22f7b7c029 100644 --- a/INSTALL.md +++ b/INSTALL.md
@@ -1,10 +1,12 @@ - [Choosing your server name](#choosing-your-server-name) +- [Picking a database engine](#picking-a-database-engine) - [Installing Synapse](#installing-synapse) - [Installing from source](#installing-from-source) - [Platform-Specific Instructions](#platform-specific-instructions) - [Prebuilt packages](#prebuilt-packages) - [Setting up Synapse](#setting-up-synapse) - [TLS certificates](#tls-certificates) + - [Client Well-Known URI](#client-well-known-uri) - [Email](#email) - [Registering a user](#registering-a-user) - [Setting up a TURN server](#setting-up-a-turn-server) @@ -27,6 +29,25 @@ that your email address is probably `user@example.com` rather than `user@email.example.com`) - but doing so may require more advanced setup: see [Setting up Federation](docs/federate.md). +# Picking a database engine + +Synapse offers two database engines: + * [PostgreSQL](https://www.postgresql.org) + * [SQLite](https://sqlite.org/) + +Almost all installations should opt to use PostgreSQL. Advantages include: + +* significant performance improvements due to the superior threading and + caching model, smarter query optimiser +* allowing the DB to be run on separate hardware + +For information on how to install and use PostgreSQL, please see +[docs/postgres.md](docs/postgres.md) + +By default Synapse uses SQLite and in doing so trades performance for convenience. +SQLite is only recommended in Synapse for testing purposes or for servers with +light workloads. + # Installing Synapse ## Installing from source @@ -234,9 +255,9 @@ for a number of platforms. There is an offical synapse image available at https://hub.docker.com/r/matrixdotorg/synapse which can be used with -the docker-compose file available at [contrib/docker](contrib/docker). Further information on -this including configuration options is available in the README on -hub.docker.com. +the docker-compose file available at [contrib/docker](contrib/docker). Further +information on this including configuration options is available in the README +on hub.docker.com. Alternatively, Andreas Peters (previously Silvio Fricke) has contributed a Dockerfile to automate a synapse server in a single Docker image, at @@ -244,7 +265,8 @@ https://hub.docker.com/r/avhost/docker-matrix/tags/ Slavi Pantaleev has created an Ansible playbook, which installs the offical Docker image of Matrix Synapse -along with many other Matrix-related services (Postgres database, riot-web, coturn, mxisd, SSL support, etc.). +along with many other Matrix-related services (Postgres database, Element, coturn, +ma1sd, SSL support, etc.). For more details, see https://github.com/spantaleev/matrix-docker-ansible-deploy @@ -277,22 +299,27 @@ The fingerprint of the repository signing key (as shown by `gpg /usr/share/keyrings/matrix-org-archive-keyring.gpg`) is `AAF9AE843A7584B5A3E4CD2BCF45A512DE2DA058`. -#### Downstream Debian/Ubuntu packages +#### Downstream Debian packages -For `buster` and `sid`, Synapse is available in the Debian repositories and -it should be possible to install it with simply: +We do not recommend using the packages from the default Debian `buster` +repository at this time, as they are old and suffer from known security +vulnerabilities. You can install the latest version of Synapse from +[our repository](#matrixorg-packages) or from `buster-backports`. Please +see the [Debian documentation](https://backports.debian.org/Instructions/) +for information on how to use backports. + +If you are using Debian `sid` or testing, Synapse is available in the default +repositories and it should be possible to install it simply with: ``` sudo apt install matrix-synapse ``` -There is also a version of `matrix-synapse` in `stretch-backports`. Please see -the [Debian documentation on -backports](https://backports.debian.org/Instructions/) for information on how -to use them. +#### Downstream Ubuntu packages -We do not recommend using the packages in downstream Ubuntu at this time, as -they are old and suffer from known security vulnerabilities. +We do not recommend using the packages in the default Ubuntu repository +at this time, as they are old and suffer from known security vulnerabilities. +The latest version of Synapse can be installed from [our repository](#matrixorg-packages). ### Fedora @@ -419,6 +446,60 @@ so, you will need to edit `homeserver.yaml`, as follows: For a more detailed guide to configuring your server for federation, see [federate.md](docs/federate.md). +## Client Well-Known URI + +Setting up the client Well-Known URI is optional but if you set it up, it will +allow users to enter their full username (e.g. `@user:<server_name>`) into clients +which support well-known lookup to automatically configure the homeserver and +identity server URLs. This is useful so that users don't have to memorize or think +about the actual homeserver URL you are using. + +The URL `https://<server_name>/.well-known/matrix/client` should return JSON in +the following format. + +``` +{ + "m.homeserver": { + "base_url": "https://<matrix.example.com>" + } +} +``` + +It can optionally contain identity server information as well. + +``` +{ + "m.homeserver": { + "base_url": "https://<matrix.example.com>" + }, + "m.identity_server": { + "base_url": "https://<identity.example.com>" + } +} +``` + +To work in browser based clients, the file must be served with the appropriate +Cross-Origin Resource Sharing (CORS) headers. A recommended value would be +`Access-Control-Allow-Origin: *` which would allow all browser based clients to +view it. + +In nginx this would be something like: +``` +location /.well-known/matrix/client { + return 200 '{"m.homeserver": {"base_url": "https://<matrix.example.com>"}}'; + add_header Content-Type application/json; + add_header Access-Control-Allow-Origin *; +} +``` + +You should also ensure the `public_baseurl` option in `homeserver.yaml` is set +correctly. `public_baseurl` should be set to the URL that clients will use to +connect to your server. This is the same URL you put for the `m.homeserver` +`base_url` above. + +``` +public_baseurl: "https://<matrix.example.com>" +``` ## Email @@ -437,7 +518,7 @@ email will be disabled. ## Registering a user -The easiest way to create a new user is to do so from a client like [Riot](https://riot.im). +The easiest way to create a new user is to do so from a client like [Element](https://element.io/). Alternatively you can do so from the command line if you have installed via pip. diff --git a/README.rst b/README.rst
index f7116b3480..4a189c8bc4 100644 --- a/README.rst +++ b/README.rst
@@ -45,7 +45,7 @@ which handle: - Eventually-consistent cryptographically secure synchronisation of room state across a global open network of federated servers and services - Sending and receiving extensible messages in a room with (optional) - end-to-end encryption[1] + end-to-end encryption - Inviting, joining, leaving, kicking, banning room members - Managing user accounts (registration, login, logout) - Using 3rd Party IDs (3PIDs) such as email addresses, phone numbers, @@ -82,9 +82,6 @@ at the `Matrix spec <https://matrix.org/docs/spec>`_, and experiment with the Thanks for using Matrix! -[1] End-to-end encryption is currently in beta: `blog post <https://matrix.org/blog/2016/11/21/matrixs-olm-end-to-end-encryption-security-assessment-released-and-implemented-cross-platform-on-riot-at-last>`_. - - Support ======= @@ -115,12 +112,11 @@ Unless you are running a test instance of Synapse on your local machine, in general, you will need to enable TLS support before you can successfully connect from a client: see `<INSTALL.md#tls-certificates>`_. -An easy way to get started is to login or register via Riot at -https://riot.im/app/#/login or https://riot.im/app/#/register respectively. +An easy way to get started is to login or register via Element at +https://app.element.io/#/login or https://app.element.io/#/register respectively. You will need to change the server you are logging into from ``matrix.org`` and instead specify a Homeserver URL of ``https://<server_name>:8448`` (or just ``https://<server_name>`` if you are using a reverse proxy). -(Leave the identity server as the default - see `Identity servers`_.) If you prefer to use another client, refer to our `client breakdown <https://matrix.org/docs/projects/clients-matrix>`_. @@ -137,7 +133,7 @@ it, specify ``enable_registration: true`` in ``homeserver.yaml``. (It is then recommended to also set up CAPTCHA - see `<docs/CAPTCHA_SETUP.md>`_.) Once ``enable_registration`` is set to ``true``, it is possible to register a -user via `riot.im <https://riot.im/app/#/register>`_ or other Matrix clients. +user via a Matrix client. Your new user name will be formed partly from the ``server_name``, and partly from a localpart you specify when you create the account. Your name will take @@ -183,30 +179,6 @@ versions of synapse. .. _UPGRADE.rst: UPGRADE.rst - -Using PostgreSQL -================ - -Synapse offers two database engines: - * `PostgreSQL <https://www.postgresql.org>`_ - * `SQLite <https://sqlite.org/>`_ - -Almost all installations should opt to use PostgreSQL. Advantages include: - -* significant performance improvements due to the superior threading and - caching model, smarter query optimiser -* allowing the DB to be run on separate hardware -* allowing basic active/backup high-availability with a "hot spare" synapse - pointing at the same DB master, as well as enabling DB replication in - synapse itself. - -For information on how to install and use PostgreSQL, please see -`docs/postgres.md <docs/postgres.md>`_. - -By default Synapse uses SQLite and in doing so trades performance for convenience. -SQLite is only recommended in Synapse for testing purposes or for servers with -light workloads. - .. _reverse-proxy: Using a reverse proxy with Synapse @@ -255,10 +227,9 @@ email address. Password reset ============== -If a user has registered an email address to their account using an identity -server, they can request a password-reset token via clients such as Riot. - -A manual password reset can be done via direct database access as follows. +Users can reset their password through their client. Alternatively, a server admin +can reset a users password using the `admin API <docs/admin_api/user_admin_api.rst#reset-password>`_ +or by directly editing the database as shown below. First calculate the hash of the new password:: diff --git a/UPGRADE.rst b/UPGRADE.rst
index b2069a0d26..49e86e628f 100644 --- a/UPGRADE.rst +++ b/UPGRADE.rst
@@ -128,17 +128,6 @@ template. These templates are similar, but the parameters are slightly different * A string ``error`` parameter is available that includes a short hint of why a user is seeing the error page. -ThirdPartyEventRules breaking changes -------------------------------------- - -This release introduces a backwards-incompatible change to modules making use of -`ThirdPartyEventRules` in Synapse. - -The `http_client` argument is no longer passed to modules as they are initialised. Instead, -modules are expected to make use of the `http_client` property on the `ModuleApi` class. -Modules are now passed a `module_api` argument during initialisation, which is an instance of -`ModuleApi`. - Upgrading to v1.18.0 ==================== diff --git a/changelog.d/67.misc b/changelog.d/67.misc new file mode 100644
index 0000000000..0a2095e4d4 --- /dev/null +++ b/changelog.d/67.misc
@@ -0,0 +1 @@ +Merge mainline Synapse v1.21.2 into 'dinsic'. \ No newline at end of file diff --git a/changelog.d/8275.feature b/changelog.d/8275.feature deleted file mode 100644
index 17549c3df3..0000000000 --- a/changelog.d/8275.feature +++ /dev/null
@@ -1 +0,0 @@ -Add a config option to specify a whitelist of domains that a user can be redirected to after validating their email or phone number. \ No newline at end of file diff --git a/contrib/cmdclient/console.py b/contrib/cmdclient/console.py
index 77422f5e5d..ab1e1f1f4c 100755 --- a/contrib/cmdclient/console.py +++ b/contrib/cmdclient/console.py
@@ -15,8 +15,6 @@ # limitations under the License. """ Starts a synapse client console. """ -from __future__ import print_function - import argparse import cmd import getpass @@ -609,13 +607,15 @@ class SynapseCmd(cmd.Cmd): @defer.inlineCallbacks def _do_event_stream(self, timeout): - res = yield self.http_client.get_json( - self._url() + "/events", - { - "access_token": self._tok(), - "timeout": str(timeout), - "from": self.event_stream_token, - }, + res = yield defer.ensureDeferred( + self.http_client.get_json( + self._url() + "/events", + { + "access_token": self._tok(), + "timeout": str(timeout), + "from": self.event_stream_token, + }, + ) ) print(json.dumps(res, indent=4)) diff --git a/contrib/cmdclient/http.py b/contrib/cmdclient/http.py
index e2534ee584..345120b612 100644 --- a/contrib/cmdclient/http.py +++ b/contrib/cmdclient/http.py
@@ -13,8 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - import json import urllib from pprint import pformat @@ -24,7 +22,7 @@ from twisted.web.client import Agent, readBody from twisted.web.http_headers import Headers -class HttpClient(object): +class HttpClient: """ Interface for talking json over http """ @@ -169,7 +167,7 @@ class TwistedHttpClient(HttpClient): return d -class _RawProducer(object): +class _RawProducer: def __init__(self, data): self.data = data self.body = data @@ -186,7 +184,7 @@ class _RawProducer(object): pass -class _JsonProducer(object): +class _JsonProducer: """ Used by the twisted http client to create the HTTP body from json """ diff --git a/contrib/experiments/cursesio.py b/contrib/experiments/cursesio.py
index ffefe3bb39..15a22c3a0e 100644 --- a/contrib/experiments/cursesio.py +++ b/contrib/experiments/cursesio.py
@@ -141,7 +141,7 @@ class CursesStdIO: curses.endwin() -class Callback(object): +class Callback: def __init__(self, stdio): self.stdio = stdio diff --git a/contrib/experiments/test_messaging.py b/contrib/experiments/test_messaging.py
index a84ec4ecae..d4c35ff2fc 100644 --- a/contrib/experiments/test_messaging.py +++ b/contrib/experiments/test_messaging.py
@@ -55,7 +55,7 @@ def excpetion_errback(failure): logging.exception(failure) -class InputOutput(object): +class InputOutput: """ This is responsible for basic I/O so that a user can interact with the example app. """ @@ -132,7 +132,7 @@ class IOLoggerHandler(logging.Handler): self.io.print_log(msg) -class Room(object): +class Room: """ Used to store (in memory) the current membership state of a room, and which home servers we should send PDUs associated with the room to. """ diff --git a/contrib/graph/graph.py b/contrib/graph/graph.py
index de33fac1c7..fdbac087bd 100644 --- a/contrib/graph/graph.py +++ b/contrib/graph/graph.py
@@ -1,5 +1,3 @@ -from __future__ import print_function - import argparse import cgi import datetime diff --git a/contrib/graph/graph3.py b/contrib/graph/graph3.py
index 91db98e7ef..dd0c19368b 100644 --- a/contrib/graph/graph3.py +++ b/contrib/graph/graph3.py
@@ -1,5 +1,3 @@ -from __future__ import print_function - import argparse import cgi import datetime diff --git a/contrib/jitsimeetbridge/jitsimeetbridge.py b/contrib/jitsimeetbridge/jitsimeetbridge.py
index 69aa74bd34..b3de468687 100644 --- a/contrib/jitsimeetbridge/jitsimeetbridge.py +++ b/contrib/jitsimeetbridge/jitsimeetbridge.py
@@ -10,8 +10,6 @@ the bridge. Requires: npm install jquery jsdom """ -from __future__ import print_function - import json import subprocess import time diff --git a/contrib/scripts/kick_users.py b/contrib/scripts/kick_users.py
index 372dbd9e4f..f8e0c732fb 100755 --- a/contrib/scripts/kick_users.py +++ b/contrib/scripts/kick_users.py
@@ -1,5 +1,4 @@ #!/usr/bin/env python -from __future__ import print_function import json import sys @@ -8,11 +7,6 @@ from argparse import ArgumentParser import requests -try: - raw_input -except NameError: # Python 3 - raw_input = input - def _mkurl(template, kws): for key in kws: @@ -58,7 +52,7 @@ def main(hs, room_id, access_token, user_id_prefix, why): print("The following user IDs will be kicked from %s" % room_name) for uid in kick_list: print(uid) - doit = raw_input("Continue? [Y]es\n") + doit = input("Continue? [Y]es\n") if len(doit) > 0 and doit.lower() == "y": print("Kicking members...") # encode them all diff --git a/debian/build_virtualenv b/debian/build_virtualenv
index 4c9aabcac3..cbdde93f96 100755 --- a/debian/build_virtualenv +++ b/debian/build_virtualenv
@@ -42,7 +42,7 @@ dh_virtualenv \ --preinstall="mock" \ --extra-pip-arg="--no-cache-dir" \ --extra-pip-arg="--compile" \ - --extras="all,systemd" + --extras="all,systemd,test" PACKAGE_BUILD_DIR="debian/matrix-synapse-py3" VIRTUALENV_DIR="${PACKAGE_BUILD_DIR}${DH_VIRTUALENV_INSTALL_ROOT}/matrix-synapse" diff --git a/debian/changelog b/debian/changelog
index a6d88906ac..8d873a4845 100644 --- a/debian/changelog +++ b/debian/changelog
@@ -1,3 +1,70 @@ +matrix-synapse-py3 (1.21.2) stable; urgency=medium + + [ Synapse Packaging team ] + * New synapse release 1.21.2. + + -- Synapse Packaging team <packages@matrix.org> Thu, 15 Oct 2020 09:23:27 -0400 + +matrix-synapse-py3 (1.21.1) stable; urgency=medium + + [ Synapse Packaging team ] + * New synapse release 1.21.1. + + [ Andrew Morgan ] + * Explicitly install "test" python dependencies. + + -- Synapse Packaging team <packages@matrix.org> Tue, 13 Oct 2020 10:24:13 +0100 + +matrix-synapse-py3 (1.21.0) stable; urgency=medium + + * New synapse release 1.21.0. + + -- Synapse Packaging team <packages@matrix.org> Mon, 12 Oct 2020 15:47:44 +0100 + +matrix-synapse-py3 (1.20.1) stable; urgency=medium + + * New synapse release 1.20.1. + + -- Synapse Packaging team <packages@matrix.org> Thu, 24 Sep 2020 16:25:22 +0100 + +matrix-synapse-py3 (1.20.0) stable; urgency=medium + + [ Synapse Packaging team ] + * New synapse release 1.20.0. + + [ Dexter Chua ] + * Use Type=notify in systemd service + + -- Synapse Packaging team <packages@matrix.org> Tue, 22 Sep 2020 15:19:32 +0100 + +matrix-synapse-py3 (1.19.3) stable; urgency=medium + + * New synapse release 1.19.3. + + -- Synapse Packaging team <packages@matrix.org> Fri, 18 Sep 2020 14:59:30 +0100 + +matrix-synapse-py3 (1.19.2) stable; urgency=medium + + * New synapse release 1.19.2. + + -- Synapse Packaging team <packages@matrix.org> Wed, 16 Sep 2020 12:50:30 +0100 + +matrix-synapse-py3 (1.19.1) stable; urgency=medium + + * New synapse release 1.19.1. + + -- Synapse Packaging team <packages@matrix.org> Thu, 27 Aug 2020 10:50:19 +0100 + +matrix-synapse-py3 (1.19.0) stable; urgency=medium + + [ Synapse Packaging team ] + * New synapse release 1.19.0. + + [ Aaron Raimist ] + * Fix outdated documentation for SYNAPSE_CACHE_FACTOR + + -- Synapse Packaging team <packages@matrix.org> Mon, 17 Aug 2020 14:06:42 +0100 + matrix-synapse-py3 (1.18.0) stable; urgency=medium * New synapse release 1.18.0. diff --git a/debian/matrix-synapse.default b/debian/matrix-synapse.default
index 65dc2f33d8..f402d73bbf 100644 --- a/debian/matrix-synapse.default +++ b/debian/matrix-synapse.default
@@ -1,2 +1,2 @@ # Specify environment variables used when running Synapse -# SYNAPSE_CACHE_FACTOR=1 (default) +# SYNAPSE_CACHE_FACTOR=0.5 (default) diff --git a/debian/matrix-synapse.service b/debian/matrix-synapse.service
index b0a8d72e6d..553babf549 100644 --- a/debian/matrix-synapse.service +++ b/debian/matrix-synapse.service
@@ -2,7 +2,7 @@ Description=Synapse Matrix homeserver [Service] -Type=simple +Type=notify User=matrix-synapse WorkingDirectory=/var/lib/matrix-synapse EnvironmentFile=/etc/default/matrix-synapse diff --git a/debian/synctl.ronn b/debian/synctl.ronn
index a73c832f62..1bad6094f3 100644 --- a/debian/synctl.ronn +++ b/debian/synctl.ronn
@@ -46,19 +46,20 @@ Configuration file may be generated as follows: ## ENVIRONMENT * `SYNAPSE_CACHE_FACTOR`: - Synapse's architecture is quite RAM hungry currently - a lot of - recent room data and metadata is deliberately cached in RAM in - order to speed up common requests. This will be improved in - future, but for now the easiest way to either reduce the RAM usage - (at the risk of slowing things down) is to set the - SYNAPSE_CACHE_FACTOR environment variable. Roughly speaking, a - SYNAPSE_CACHE_FACTOR of 1.0 will max out at around 3-4GB of - resident memory - this is what we currently run the matrix.org - on. The default setting is currently 0.1, which is probably around - a ~700MB footprint. You can dial it down further to 0.02 if - desired, which targets roughly ~512MB. Conversely you can dial it - up if you need performance for lots of users and have a box with a - lot of RAM. + Synapse's architecture is quite RAM hungry currently - we deliberately + cache a lot of recent room data and metadata in RAM in order to speed up + common requests. We'll improve this in the future, but for now the easiest + way to either reduce the RAM usage (at the risk of slowing things down) + is to set the almost-undocumented ``SYNAPSE_CACHE_FACTOR`` environment + variable. The default is 0.5, which can be decreased to reduce RAM usage + in memory constrained enviroments, or increased if performance starts to + degrade. + + However, degraded performance due to a low cache factor, common on + machines with slow disks, often leads to explosions in memory use due + backlogged requests. In this case, reducing the cache factor will make + things worse. Instead, try increasing it drastically. 2.0 is a good + starting value. ## COPYRIGHT diff --git a/docker/Dockerfile b/docker/Dockerfile
index 8b3a4246a5..27512f8600 100644 --- a/docker/Dockerfile +++ b/docker/Dockerfile
@@ -19,11 +19,16 @@ ARG PYTHON_VERSION=3.7 FROM docker.io/python:${PYTHON_VERSION}-slim as builder # install the OS build deps - - RUN apt-get update && apt-get install -y \ build-essential \ + libffi-dev \ + libjpeg-dev \ libpq-dev \ + libssl-dev \ + libwebp-dev \ + libxml++2.6-dev \ + libxslt1-dev \ + zlib1g-dev \ && rm -rf /var/lib/apt/lists/* # Build dependencies that are not available as wheels, to speed up rebuilds @@ -55,9 +60,12 @@ RUN pip install --prefix="/install" --no-warn-script-location \ FROM docker.io/python:${PYTHON_VERSION}-slim RUN apt-get update && apt-get install -y \ + curl \ + gosu \ + libjpeg62-turbo \ libpq5 \ + libwebp6 \ xmlsec1 \ - gosu \ && rm -rf /var/lib/apt/lists/* COPY --from=builder /install /usr/local @@ -69,3 +77,6 @@ VOLUME ["/data"] EXPOSE 8008/tcp 8009/tcp 8448/tcp ENTRYPOINT ["/start.py"] + +HEALTHCHECK --interval=1m --timeout=5s \ + CMD curl -fSs http://localhost:8008/health || exit 1 diff --git a/docker/README.md b/docker/README.md
index 008a9ff708..d0da34778e 100644 --- a/docker/README.md +++ b/docker/README.md
@@ -162,3 +162,32 @@ docker build -t matrixdotorg/synapse -f docker/Dockerfile . You can choose to build a different docker image by changing the value of the `-f` flag to point to another Dockerfile. + +## Disabling the healthcheck + +If you are using a non-standard port or tls inside docker you can disable the healthcheck +whilst running the above `docker run` commands. + +``` + --no-healthcheck +``` +## Setting custom healthcheck on docker run + +If you wish to point the healthcheck at a different port with docker command, add the following + +``` + --health-cmd 'curl -fSs http://localhost:1234/health' +``` + +## Setting the healthcheck in docker-compose file + +You can add the following to set a custom healthcheck in a docker compose file. +You will need version >2.1 for this to work. + +``` +healthcheck: + test: ["CMD", "curl", "-fSs", "http://localhost:8008/health"] + interval: 1m + timeout: 10s + retries: 3 +``` diff --git a/docker/conf/log.config b/docker/conf/log.config
index ed418a57cd..491bbcc87a 100644 --- a/docker/conf/log.config +++ b/docker/conf/log.config
@@ -4,16 +4,10 @@ formatters: precise: format: '%(asctime)s - %(name)s - %(lineno)d - %(levelname)s - %(request)s - %(message)s' -filters: - context: - (): synapse.logging.context.LoggingContextFilter - request: "" - handlers: console: class: logging.StreamHandler formatter: precise - filters: [context] loggers: synapse.storage.SQL: diff --git a/docs/.sample_config_header.yaml b/docs/.sample_config_header.yaml
index 35a591d042..8c9b31acdb 100644 --- a/docs/.sample_config_header.yaml +++ b/docs/.sample_config_header.yaml
@@ -10,5 +10,16 @@ # homeserver.yaml. Instead, if you are starting from scratch, please generate # a fresh config using Synapse by following the instructions in INSTALL.md. +# Configuration options that take a time period can be set using a number +# followed by a letter. Letters have the following meanings: +# s = second +# m = minute +# h = hour +# d = day +# w = week +# y = year +# For example, setting redaction_retention_period: 5m would remove redacted +# messages from the database after 5 minutes, rather than 5 months. + ################################################################################ diff --git a/docs/admin_api/event_reports.rst b/docs/admin_api/event_reports.rst new file mode 100644
index 0000000000..461be01230 --- /dev/null +++ b/docs/admin_api/event_reports.rst
@@ -0,0 +1,129 @@ +Show reported events +==================== + +This API returns information about reported events. + +The api is:: + + GET /_synapse/admin/v1/event_reports?from=0&limit=10 + +To use it, you will need to authenticate by providing an ``access_token`` for a +server admin: see `README.rst <README.rst>`_. + +It returns a JSON body like the following: + +.. code:: jsonc + + { + "event_reports": [ + { + "content": { + "reason": "foo", + "score": -100 + }, + "event_id": "$bNUFCwGzWca1meCGkjp-zwslF-GfVcXukvRLI1_FaVY", + "event_json": { + "auth_events": [ + "$YK4arsKKcc0LRoe700pS8DSjOvUT4NDv0HfInlMFw2M", + "$oggsNXxzPFRE3y53SUNd7nsj69-QzKv03a1RucHu-ws" + ], + "content": { + "body": "matrix.org: This Week in Matrix", + "format": "org.matrix.custom.html", + "formatted_body": "<strong>matrix.org</strong>:<br><a href=\"https://matrix.org/blog/\"><strong>This Week in Matrix</strong></a>", + "msgtype": "m.notice" + }, + "depth": 546, + "hashes": { + "sha256": "xK1//xnmvHJIOvbgXlkI8eEqdvoMmihVDJ9J4SNlsAw" + }, + "origin": "matrix.org", + "origin_server_ts": 1592291711430, + "prev_events": [ + "$YK4arsKKcc0LRoe700pS8DSjOvUT4NDv0HfInlMFw2M" + ], + "prev_state": [], + "room_id": "!ERAgBpSOcCCuTJqQPk:matrix.org", + "sender": "@foobar:matrix.org", + "signatures": { + "matrix.org": { + "ed25519:a_JaEG": "cs+OUKW/iHx5pEidbWxh0UiNNHwe46Ai9LwNz+Ah16aWDNszVIe2gaAcVZfvNsBhakQTew51tlKmL2kspXk/Dg" + } + }, + "type": "m.room.message", + "unsigned": { + "age_ts": 1592291711430, + } + }, + "id": 2, + "reason": "foo", + "received_ts": 1570897107409, + "room_alias": "#alias1:matrix.org", + "room_id": "!ERAgBpSOcCCuTJqQPk:matrix.org", + "sender": "@foobar:matrix.org", + "user_id": "@foo:matrix.org" + }, + { + "content": { + "reason": "bar", + "score": -100 + }, + "event_id": "$3IcdZsDaN_En-S1DF4EMCy3v4gNRKeOJs8W5qTOKj4I", + "event_json": { + // hidden items + // see above + }, + "id": 3, + "reason": "bar", + "received_ts": 1598889612059, + "room_alias": "#alias2:matrix.org", + "room_id": "!eGvUQuTCkHGVwNMOjv:matrix.org", + "sender": "@foobar:matrix.org", + "user_id": "@bar:matrix.org" + } + ], + "next_token": 2, + "total": 4 + } + +To paginate, check for ``next_token`` and if present, call the endpoint again +with ``from`` set to the value of ``next_token``. This will return a new page. + +If the endpoint does not return a ``next_token`` then there are no more +reports to paginate through. + +**URL parameters:** + +- ``limit``: integer - Is optional but is used for pagination, + denoting the maximum number of items to return in this call. Defaults to ``100``. +- ``from``: integer - Is optional but used for pagination, + denoting the offset in the returned results. This should be treated as an opaque value and + not explicitly set to anything other than the return value of ``next_token`` from a previous call. + Defaults to ``0``. +- ``dir``: string - Direction of event report order. Whether to fetch the most recent first (``b``) or the + oldest first (``f``). Defaults to ``b``. +- ``user_id``: string - Is optional and filters to only return users with user IDs that contain this value. + This is the user who reported the event and wrote the reason. +- ``room_id``: string - Is optional and filters to only return rooms with room IDs that contain this value. + +**Response** + +The following fields are returned in the JSON response body: + +- ``id``: integer - ID of event report. +- ``received_ts``: integer - The timestamp (in milliseconds since the unix epoch) when this report was sent. +- ``room_id``: string - The ID of the room in which the event being reported is located. +- ``event_id``: string - The ID of the reported event. +- ``user_id``: string - This is the user who reported the event and wrote the reason. +- ``reason``: string - Comment made by the ``user_id`` in this report. May be blank. +- ``content``: object - Content of reported event. + + - ``reason``: string - Comment made by the ``user_id`` in this report. May be blank. + - ``score``: integer - Content is reported based upon a negative score, where -100 is "most offensive" and 0 is "inoffensive". + +- ``sender``: string - This is the ID of the user who sent the original message/event that was reported. +- ``room_alias``: string - The alias of the room. ``null`` if the room does not have a canonical alias set. +- ``event_json``: object - Details of the original event that was reported. +- ``next_token``: integer - Indication for pagination. See above. +- ``total``: integer - Total number of event reports related to the query (``user_id`` and ``room_id``). + diff --git a/docs/admin_api/rooms.md b/docs/admin_api/rooms.md
index 15b83e9824..fa9b914fa7 100644 --- a/docs/admin_api/rooms.md +++ b/docs/admin_api/rooms.md
@@ -275,6 +275,8 @@ The following fields are possible in the JSON response body: * `room_id` - The ID of the room. * `name` - The name of the room. +* `topic` - The topic of the room. +* `avatar` - The `mxc` URI to the avatar of the room. * `canonical_alias` - The canonical (main) alias address of the room. * `joined_members` - How many users are currently in the room. * `joined_local_members` - How many local users are currently in the room. @@ -304,6 +306,8 @@ Response: { "room_id": "!mscvqgqpHYjBGDxNym:matrix.org", "name": "Music Theory", + "avatar": "mxc://matrix.org/AQDaVFlbkQoErdOgqWRgiGSV", + "topic": "Theory, Composition, Notation, Analysis", "canonical_alias": "#musictheory:matrix.org", "joined_members": 127 "joined_local_members": 2, @@ -369,7 +373,9 @@ to the new room will have power level `-10` by default, and thus be unable to sp If `block` is `True` it prevents new joins to the old room. This API will remove all trace of the old room from your database after removing -all local users. +all local users. If `purge` is `true` (the default), all traces of the old room will +be removed from your database after removing all local users. If you do not want +this to happen, set `purge` to `false`. Depending on the amount of history being purged a call to the API may take several minutes or longer. @@ -388,7 +394,8 @@ with a body of: "new_room_user_id": "@someuser:example.com", "room_name": "Content Violation Notification", "message": "Bad Room has been shutdown due to content violations on this server. Please review our Terms of Service.", - "block": true + "block": true, + "purge": true } ``` @@ -430,8 +437,10 @@ The following JSON body parameters are available: `new_room_user_id` in the new room. Ideally this will clearly convey why the original room was shut down. Defaults to `Sharing illegal content on this server is not permitted and rooms in violation will be blocked.` -* `block` - Optional. If set to `true`, this room will be added to a blocking list, preventing future attempts to - join the room. Defaults to `false`. +* `block` - Optional. If set to `true`, this room will be added to a blocking list, preventing + future attempts to join the room. Defaults to `false`. +* `purge` - Optional. If set to `true`, it will remove all traces of the room from your database. + Defaults to `true`. The JSON body must not be empty. The body must be at least `{}`. diff --git a/docs/admin_api/shutdown_room.md b/docs/admin_api/shutdown_room.md
index 808caeec79..9b1cb1c184 100644 --- a/docs/admin_api/shutdown_room.md +++ b/docs/admin_api/shutdown_room.md
@@ -33,7 +33,7 @@ You will need to authenticate with an access token for an admin user. * `message` - Optional. A string containing the first message that will be sent as `new_room_user_id` in the new room. Ideally this will clearly convey why the original room was shut down. - + If not specified, the default value of `room_name` is "Content Violation Notification". The default value of `message` is "Sharing illegal content on othis server is not permitted and rooms in violation will be blocked." @@ -72,3 +72,30 @@ Response: "new_room_id": "!newroomid:example.com", }, ``` + +## Undoing room shutdowns + +*Note*: This guide may be outdated by the time you read it. By nature of room shutdowns being performed at the database level, +the structure can and does change without notice. + +First, it's important to understand that a room shutdown is very destructive. Undoing a shutdown is not as simple as pretending it +never happened - work has to be done to move forward instead of resetting the past. In fact, in some cases it might not be possible +to recover at all: + +* If the room was invite-only, your users will need to be re-invited. +* If the room no longer has any members at all, it'll be impossible to rejoin. +* The first user to rejoin will have to do so via an alias on a different server. + +With all that being said, if you still want to try and recover the room: + +1. For safety reasons, shut down Synapse. +2. In the database, run `DELETE FROM blocked_rooms WHERE room_id = '!example:example.org';` + * For caution: it's recommended to run this in a transaction: `BEGIN; DELETE ...;`, verify you got 1 result, then `COMMIT;`. + * The room ID is the same one supplied to the shutdown room API, not the Content Violation room. +3. Restart Synapse. + +You will have to manually handle, if you so choose, the following: + +* Aliases that would have been redirected to the Content Violation room. +* Users that would have been booted from the room (and will have been force-joined to the Content Violation room). +* Removal of the Content Violation room if desired. diff --git a/docs/admin_api/user_admin_api.rst b/docs/admin_api/user_admin_api.rst
index be05128b3e..7ca902faba 100644 --- a/docs/admin_api/user_admin_api.rst +++ b/docs/admin_api/user_admin_api.rst
@@ -108,7 +108,7 @@ The api is:: GET /_synapse/admin/v2/users?from=0&limit=10&guests=false -To use it, you will need to authenticate by providing an `access_token` for a +To use it, you will need to authenticate by providing an ``access_token`` for a server admin: see `README.rst <README.rst>`_. The parameter ``from`` is optional but used for pagination, denoting the @@ -119,8 +119,11 @@ from a previous call. The parameter ``limit`` is optional but is used for pagination, denoting the maximum number of items to return in this call. Defaults to ``100``. -The parameter ``user_id`` is optional and filters to only users with user IDs -that contain this value. +The parameter ``user_id`` is optional and filters to only return users with user IDs +that contain this value. This parameter is ignored when using the ``name`` parameter. + +The parameter ``name`` is optional and filters to only return users with user ID localparts +**or** displaynames that contain this value. The parameter ``guests`` is optional and if ``false`` will **exclude** guest users. Defaults to ``true`` to include guest users. @@ -211,9 +214,11 @@ Deactivate Account This API deactivates an account. It removes active access tokens, resets the password, and deletes third-party IDs (to prevent the user requesting a -password reset). It can also mark the user as GDPR-erased (stopping their data -from distributed further, and deleting it entirely if there are no other -references to it). +password reset). + +It can also mark the user as GDPR-erased. This means messages sent by the +user will still be visible by anyone that was in the room when these messages +were sent, but hidden from users joining the room afterwards. The api is:: @@ -299,6 +304,43 @@ To use it, you will need to authenticate by providing an ``access_token`` for a server admin: see `README.rst <README.rst>`_. +List room memberships of an user +================================ +Gets a list of all ``room_id`` that a specific ``user_id`` is member. + +The API is:: + + GET /_synapse/admin/v1/users/<user_id>/joined_rooms + +To use it, you will need to authenticate by providing an ``access_token`` for a +server admin: see `README.rst <README.rst>`_. + +A response body like the following is returned: + +.. code:: json + + { + "joined_rooms": [ + "!DuGcnbhHGaSZQoNQR:matrix.org", + "!ZtSaPCawyWtxfWiIy:matrix.org" + ], + "total": 2 + } + +**Parameters** + +The following parameters should be set in the URL: + +- ``user_id`` - fully qualified: for example, ``@user:server.com``. + +**Response** + +The following fields are returned in the JSON response body: + +- ``joined_rooms`` - An array of ``room_id``. +- ``total`` - Number of rooms. + + User devices ============ diff --git a/docs/federate.md b/docs/federate.md
index a0786b9cf7..b15cd724d1 100644 --- a/docs/federate.md +++ b/docs/federate.md
@@ -47,6 +47,18 @@ you invite them to. This can be caused by an incorrectly-configured reverse proxy: see [reverse_proxy.md](<reverse_proxy.md>) for instructions on how to correctly configure a reverse proxy. +### Known issues + +**HTTP `308 Permanent Redirect` redirects are not followed**: Due to missing features +in the HTTP library used by Synapse, 308 redirects are currently not followed by +federating servers, which can cause `M_UNKNOWN` or `401 Unauthorized` errors. This +may affect users who are redirecting apex-to-www (e.g. `example.com` -> `www.example.com`), +and especially users of the Kubernetes *Nginx Ingress* module, which uses 308 redirect +codes by default. For those Kubernetes users, [this Stackoverflow post](https://stackoverflow.com/a/52617528/5096871) +might be helpful. For other users, switching to a `301 Moved Permanently` code may be +an option. 308 redirect codes will be supported properly in a future +release of Synapse. + ## Running a demo federation of Synapses If you want to get up and running quickly with a trio of homeservers in a diff --git a/docs/metrics-howto.md b/docs/metrics-howto.md
index cf69938a2a..b386ec91c1 100644 --- a/docs/metrics-howto.md +++ b/docs/metrics-howto.md
@@ -27,7 +27,7 @@ different thread to Synapse. This can make it more resilient to heavy load meaning metrics cannot be retrieved, and can be exposed to just internal networks easier. The served metrics are available - over HTTP only, and will be available at `/`. + over HTTP only, and will be available at `/_synapse/metrics`. Add a new listener to homeserver.yaml: diff --git a/docs/password_auth_providers.md b/docs/password_auth_providers.md
index fef1d47e85..7d98d9f255 100644 --- a/docs/password_auth_providers.md +++ b/docs/password_auth_providers.md
@@ -14,6 +14,7 @@ password auth provider module implementations: * [matrix-synapse-ldap3](https://github.com/matrix-org/matrix-synapse-ldap3/) * [matrix-synapse-shared-secret-auth](https://github.com/devture/matrix-synapse-shared-secret-auth) +* [matrix-synapse-rest-password-provider](https://github.com/ma1uta/matrix-synapse-rest-password-provider) ## Required methods diff --git a/docs/postgres.md b/docs/postgres.md
index 70fe29cdcc..c30cc1fd8c 100644 --- a/docs/postgres.md +++ b/docs/postgres.md
@@ -106,6 +106,17 @@ Note that the above may fail with an error about duplicate rows if corruption has already occurred, and such duplicate rows will need to be manually removed. +## Fixing inconsistent sequences error + +Synapse uses Postgres sequences to generate IDs for various tables. A sequence +and associated table can get out of sync if, for example, Synapse has been +downgraded and then upgraded again. + +To fix the issue shut down Synapse (including any and all workers) and run the +SQL command included in the error message. Once done Synapse should start +successfully. + + ## Tuning Postgres The default settings should be fine for most deployments. For larger @@ -188,6 +199,9 @@ to do step 2. It is safe to at any time kill the port script and restart it. +Note that the database may take up significantly more (25% - 100% more) +space on disk after porting to Postgres. + ### Using the port script Firstly, shut down the currently running synapse server and copy its diff --git a/docs/reverse_proxy.md b/docs/reverse_proxy.md
index 7bfb96eff6..46d8f35771 100644 --- a/docs/reverse_proxy.md +++ b/docs/reverse_proxy.md
@@ -11,7 +11,7 @@ privileges. **NOTE**: Your reverse proxy must not `canonicalise` or `normalise` the requested URI in any way (for example, by decoding `%xx` escapes). -Beware that Apache *will* canonicalise URIs unless you specifify +Beware that Apache *will* canonicalise URIs unless you specify `nocanon`. When setting up a reverse proxy, remember that Matrix clients and other @@ -23,6 +23,10 @@ specification](https://matrix.org/docs/spec/server_server/latest#resolving-serve for more details of the algorithm used for federation connections, and [delegate.md](<delegate.md>) for instructions on setting up delegation. +Endpoints that are part of the standardised Matrix specification are +located under `/_matrix`, whereas endpoints specific to Synapse are +located under `/_synapse/client`. + Let's assume that we expect clients to connect to our server at `https://matrix.example.com`, and other servers to connect at `https://example.com:8448`. The following sections detail the configuration of @@ -45,7 +49,7 @@ server { server_name matrix.example.com; - location /_matrix { + location ~* ^(\/_matrix|\/_synapse\/client) { proxy_pass http://localhost:8008; proxy_set_header X-Forwarded-For $remote_addr; # Nginx by default only allows file uploads up to 1M in size @@ -65,6 +69,10 @@ matrix.example.com { proxy /_matrix http://localhost:8008 { transparent } + + proxy /_synapse/client http://localhost:8008 { + transparent + } } example.com:8448 { @@ -79,6 +87,7 @@ example.com:8448 { ``` matrix.example.com { reverse_proxy /_matrix/* http://localhost:8008 + reverse_proxy /_synapse/client/* http://localhost:8008 } example.com:8448 { @@ -96,6 +105,8 @@ example.com:8448 { AllowEncodedSlashes NoDecode ProxyPass /_matrix http://127.0.0.1:8008/_matrix nocanon ProxyPassReverse /_matrix http://127.0.0.1:8008/_matrix + ProxyPass /_synapse/client http://127.0.0.1:8008/_synapse/client nocanon + ProxyPassReverse /_synapse/client http://127.0.0.1:8008/_synapse/client </VirtualHost> <VirtualHost *:8448> @@ -110,6 +121,14 @@ example.com:8448 { **NOTE**: ensure the `nocanon` options are included. +**NOTE 2**: It appears that Synapse is currently incompatible with the ModSecurity module for Apache (`mod_security2`). If you need it enabled for other services on your web server, you can disable it for Synapse's two VirtualHosts by including the following lines before each of the two `</VirtualHost>` above: + +``` +<IfModule security2_module> + SecRuleEngine off +</IfModule> +``` + ### HAProxy ``` @@ -119,6 +138,7 @@ frontend https # Matrix client traffic acl matrix-host hdr(host) -i matrix.example.com acl matrix-path path_beg /_matrix + acl matrix-path path_beg /_synapse/client use_backend matrix if matrix-host matrix-path @@ -139,3 +159,17 @@ client IP addresses are recorded correctly. Having done so, you can then use `https://matrix.example.com` (instead of `https://matrix.example.com:8448`) as the "Custom server" when connecting to Synapse from a client. + + +## Health check endpoint + +Synapse exposes a health check endpoint for use by reverse proxies. +Each configured HTTP listener has a `/health` endpoint which always returns +200 OK (and doesn't get logged). + +## Synapse administration endpoints + +Endpoints for administering your Synapse instance are placed under +`/_synapse/admin`. These require authentication through an access token of an +admin user. However as access to these endpoints grants the caller a lot of power, +we do not recommend exposing them to the public internet without good reason. diff --git a/docs/sample_config.yaml b/docs/sample_config.yaml
index 85932311bd..e4b6fad449 100644 --- a/docs/sample_config.yaml +++ b/docs/sample_config.yaml
@@ -10,6 +10,17 @@ # homeserver.yaml. Instead, if you are starting from scratch, please generate # a fresh config using Synapse by following the instructions in INSTALL.md. +# Configuration options that take a time period can be set using a number +# followed by a letter. Letters have the following meanings: +# s = second +# m = minute +# h = hour +# d = day +# w = week +# y = year +# For example, setting redaction_retention_period: 5m would remove redacted +# messages from the database after 5 minutes, rather than 5 months. + ################################################################################ # Configuration file for Synapse. @@ -22,10 +33,23 @@ ## Server ## -# The domain name of the server, with optional explicit port. -# This is used by remote servers to connect to this server, -# e.g. matrix.org, localhost:8080, etc. -# This is also the last part of your UserID. +# The public-facing domain of the server +# +# The server_name name will appear at the end of usernames and room addresses +# created on this server. For example if the server_name was example.com, +# usernames on this server would be in the format @user:example.com +# +# In most cases you should avoid using a matrix specific subdomain such as +# matrix.example.com or synapse.example.com as the server_name for the same +# reasons you wouldn't use user@email.example.com as your email address. +# See https://github.com/matrix-org/synapse/blob/master/docs/delegate.md +# for information on how to host Synapse on a subdomain while preserving +# a clean server_name. +# +# The server_name cannot be changed later so it is important to +# configure this correctly before you start Synapse. It should be all +# lowercase and may contain an explicit port. +# Examples: matrix.org, localhost:8080 # server_name: "SERVERNAME" @@ -314,6 +338,10 @@ limit_remote_rooms: # #complexity_error: "This room is too complex." + # allow server admins to join complex rooms. Default is false. + # + #admins_can_join: true + # Whether to require a user to be in the room to add an alias to it. # Defaults to 'true'. # @@ -431,11 +459,10 @@ retention: # min_lifetime: 1d # max_lifetime: 1y - # Retention policy limits. If set, a user won't be able to send a - # 'm.room.retention' event which features a 'min_lifetime' or a 'max_lifetime' - # that's not within this range. This is especially useful in closed federations, - # in which server admins can make sure every federating server applies the same - # rules. + # Retention policy limits. If set, and the state of a room contains a + # 'm.room.retention' event in its state which contains a 'min_lifetime' or a + # 'max_lifetime' that's out of these bounds, Synapse will cap the room's policy + # to these limits when running purge jobs. # #allowed_lifetime_min: 1d #allowed_lifetime_max: 1y @@ -461,12 +488,19 @@ retention: # (e.g. every 12h), but not want that purge to be performed by a job that's # iterating over every room it knows, which could be heavy on the server. # + # If any purge job is configured, it is strongly recommended to have at least + # a single job with neither 'shortest_max_lifetime' nor 'longest_max_lifetime' + # set, or one job without 'shortest_max_lifetime' and one job without + # 'longest_max_lifetime' set. Otherwise some rooms might be ignored, even if + # 'allowed_lifetime_min' and 'allowed_lifetime_max' are set, because capping a + # room's policy to these values is done after the policies are retrieved from + # Synapse's database (which is done using the range specified in a purge job's + # configuration). + # #purge_jobs: - # - shortest_max_lifetime: 1d - # longest_max_lifetime: 3d + # - longest_max_lifetime: 3d # interval: 12h # - shortest_max_lifetime: 3d - # longest_max_lifetime: 1y # interval: 1d # Inhibits the /requestToken endpoints from returning an error that might leak @@ -663,6 +697,7 @@ acme: #tls_fingerprints: [{"sha256": "<base64_encoded_sha256_fingerprint>"}] +## Federation ## # Restrict federation to the following whitelist of domains. # N.B. we recommend also firewalling your federation listener to limit @@ -696,6 +731,17 @@ federation_ip_range_blacklist: - 'fe80::/64' - 'fc00::/7' +# Report prometheus metrics on the age of PDUs being sent to and received from +# the following domains. This can be used to give an idea of "delay" on inbound +# and outbound federation, though be aware that any delay can be due to problems +# at either end or with the intermediate network. +# +# By default, no domains are monitored in this way. +# +#federation_metrics_domains: +# - matrix.org +# - example.com + ## Caching ## @@ -819,6 +865,10 @@ log_config: "CONFDIR/SERVERNAME.log.config" # - one for ratelimiting redactions by room admins. If this is not explicitly # set then it uses the same ratelimiting as per rc_message. This is useful # to allow room admins to deal with abuse quickly. +# - two for ratelimiting number of rooms a user can join, "local" for when +# users are joining rooms the server is already in (this is cheap) vs +# "remote" for when users are trying to join rooms not on the server (which +# can be more expensive) # # The defaults are as shown below. # @@ -848,6 +898,14 @@ log_config: "CONFDIR/SERVERNAME.log.config" #rc_admin_redaction: # per_second: 1 # burst_count: 50 +# +#rc_joins: +# local: +# per_second: 0.1 +# burst_count: 3 +# remote: +# per_second: 0.01 +# burst_count: 3 # Ratelimiting settings for incoming federation @@ -1289,24 +1347,6 @@ account_validity: # #default_identity_server: https://matrix.org -# The list of identity servers trusted to verify third party -# identifiers by this server. -# -# Also defines the ID server which will be called when an account is -# deactivated (one will be picked arbitrarily). -# -# Note: This option is deprecated. Since v0.99.4, Synapse has tracked which identity -# server a 3PID has been bound to. For 3PIDs bound before then, Synapse runs a -# background migration script, informing itself that the identity server all of its -# 3PIDs have been bound to is likely one of the below. -# -# As of Synapse v1.4.0, all other functionality of this option has been deprecated, and -# it is now solely used for the purposes of the background migration script, and can be -# removed once it has run. -#trusted_third_party_id_servers: -# - matrix.org -# - vector.im - # If enabled, user IDs, display names and avatar URLs will be replicated # to this server whenever they change. # This is an experimental API currently implemented by sydent to support @@ -1645,11 +1685,14 @@ trusted_key_servers: # At least one of `sp_config` or `config_path` must be set in this section to # enable SAML login. # -# (You will probably also want to set the following options to `false` to +# You will probably also want to set the following options to `false` to # disable the regular login/registration flows: # * enable_registration # * password_config.enabled # +# You will also want to investigate the settings under the "sso" configuration +# section below. +# # Once SAML support is enabled, a metadata file will be exposed at # https://<server>:<port>/_matrix/saml2/metadata.xml, which you may be able to # use to configure your SAML IdP with. Alternatively, you can manually configure @@ -1761,30 +1804,16 @@ saml2_config: # #grandfathered_mxid_source_attribute: upn - # Directory in which Synapse will try to find the template files below. - # If not set, default templates from within the Synapse package will be used. - # - # DO NOT UNCOMMENT THIS SETTING unless you want to customise the templates. - # If you *do* uncomment it, you will need to make sure that all the templates - # below are in the directory. - # - # Synapse will look for the following templates in this directory: - # - # * HTML page to display to users if something goes wrong during the - # authentication process: 'saml_error.html'. - # - # When rendering, this template is given the following variables: - # * code: an HTML error code corresponding to the error that is being - # returned (typically 400 or 500) - # - # * msg: a textual message describing the error. + # It is possible to configure Synapse to only allow logins if SAML attributes + # match particular values. The requirements can be listed under + # `attribute_requirements` as shown below. All of the listed attributes must + # match for the login to be permitted. # - # The variables will automatically be HTML-escaped. - # - # You can see the default templates at: - # https://github.com/matrix-org/synapse/tree/master/synapse/res/templates - # - #template_dir: "res/templates" + #attribute_requirements: + # - attribute: userGroup + # value: "staff" + # - attribute: department + # value: "sales" # OpenID Connect integration. The following settings can be used to make Synapse @@ -1860,6 +1889,11 @@ oidc_config: # #skip_verification: true + # Uncomment to allow a user logging in via OIDC to match a pre-existing account instead + # of failing. This could be used if switching from password logins to OIDC. Defaults to false. + # + #allow_existing_users: true + # An external module can be provided here as a custom solution to mapping # attributes returned from a OIDC provider onto a matrix user. # @@ -1901,6 +1935,14 @@ oidc_config: # #display_name_template: "{{ user.given_name }} {{ user.last_name }}" + # Jinja2 templates for extra attributes to send back to the client during + # login. + # + # Note that these are non-standard and clients will ignore them without modifications. + # + #extra_attributes: + #birthdate: "{{ user.birthdate }}" + # Enable CAS for registration and login. @@ -2175,9 +2217,7 @@ email: # Directory in which Synapse will try to find the template files below. # If not set, default templates from within the Synapse package will be used. # - # DO NOT UNCOMMENT THIS SETTING unless you want to customise the templates. - # If you *do* uncomment it, you will need to make sure that all the templates - # below are in the directory. + # Do not uncomment this setting unless you want to customise the templates. # # Synapse will look for the following templates in this directory: # @@ -2190,9 +2230,13 @@ email: # * The contents of password reset emails sent by the homeserver: # 'password_reset.html' and 'password_reset.txt' # - # * HTML pages for success and failure that a user will see when they follow - # the link in the password reset email: 'password_reset_success.html' and - # 'password_reset_failure.html' + # * An HTML page that a user will see when they follow the link in the password + # reset email. The user will be asked to confirm the action before their + # password is reset: 'password_reset_confirmation.html' + # + # * HTML pages for success and failure that a user will see when they confirm + # the password reset flow using the page above: 'password_reset_success.html' + # and 'password_reset_failure.html' # # * The contents of address verification emails sent during registration: # 'registration.html' and 'registration.txt' diff --git a/docs/sample_log_config.yaml b/docs/sample_log_config.yaml
index 1a2739455e..55a48a9ed6 100644 --- a/docs/sample_log_config.yaml +++ b/docs/sample_log_config.yaml
@@ -11,24 +11,33 @@ formatters: precise: format: '%(asctime)s - %(name)s - %(lineno)d - %(levelname)s - %(request)s - %(message)s' -filters: - context: - (): synapse.logging.context.LoggingContextFilter - request: "" - handlers: file: - class: logging.handlers.RotatingFileHandler + class: logging.handlers.TimedRotatingFileHandler formatter: precise filename: /var/log/matrix-synapse/homeserver.log - maxBytes: 104857600 - backupCount: 10 - filters: [context] + when: midnight + backupCount: 3 # Does not include the current log file. encoding: utf8 + + # Default to buffering writes to log file for efficiency. This means that + # will be a delay for INFO/DEBUG logs to get written, but WARNING/ERROR + # logs will still be flushed immediately. + buffer: + class: logging.handlers.MemoryHandler + target: file + # The capacity is the number of log lines that are buffered before + # being written to disk. Increasing this will lead to better + # performance, at the expensive of it taking longer for log lines to + # be written to disk. + capacity: 10 + flushLevel: 30 # Flush for WARNING logs as well + + # A handler that writes logs to stderr. Unused by default, but can be used + # instead of "buffer" and "file" in the logger handlers. console: class: logging.StreamHandler formatter: precise - filters: [context] loggers: synapse.storage.SQL: @@ -36,8 +45,23 @@ loggers: # information such as access tokens. level: INFO + twisted: + # We send the twisted logging directly to the file handler, + # to work around https://github.com/matrix-org/synapse/issues/3471 + # when using "buffer" logger. Use "console" to log to stderr instead. + handlers: [file] + propagate: false + root: level: INFO - handlers: [file, console] + + # Write logs to the `buffer` handler, which will buffer them together in memory, + # then write them to a file. + # + # Replace "buffer" with "console" to log to stderr instead. (Note that you'll + # also need to update the configuation for the `twisted` logger above, in + # this case.) + # + handlers: [buffer] disable_existing_loggers: false diff --git a/docs/sso_mapping_providers.md b/docs/sso_mapping_providers.md
index abea432343..32b06aa2c5 100644 --- a/docs/sso_mapping_providers.md +++ b/docs/sso_mapping_providers.md
@@ -57,7 +57,7 @@ A custom mapping provider must specify the following methods: - This method must return a string, which is the unique identifier for the user. Commonly the ``sub`` claim of the response. * `map_user_attributes(self, userinfo, token)` - - This method should be async. + - This method must be async. - Arguments: - `userinfo` - A `authlib.oidc.core.claims.UserInfo` object to extract user information from. @@ -66,6 +66,18 @@ A custom mapping provider must specify the following methods: - Returns a dictionary with two keys: - localpart: A required string, used to generate the Matrix ID. - displayname: An optional string, the display name for the user. +* `get_extra_attributes(self, userinfo, token)` + - This method must be async. + - Arguments: + - `userinfo` - A `authlib.oidc.core.claims.UserInfo` object to extract user + information from. + - `token` - A dictionary which includes information necessary to make + further requests to the OpenID provider. + - Returns a dictionary that is suitable to be serialized to JSON. This + will be returned as part of the response during a successful login. + + Note that care should be taken to not overwrite any of the parameters + usually returned as part of the [login response](https://matrix.org/docs/spec/client_server/latest#post-matrix-client-r0-login). ### Default OpenID Mapping Provider diff --git a/docs/systemd-with-workers/system/matrix-synapse-worker@.service b/docs/systemd-with-workers/system/matrix-synapse-worker@.service
index 39bc5e88e8..cb5ac0ac87 100644 --- a/docs/systemd-with-workers/system/matrix-synapse-worker@.service +++ b/docs/systemd-with-workers/system/matrix-synapse-worker@.service
@@ -1,9 +1,14 @@ [Unit] Description=Synapse %i AssertPathExists=/etc/matrix-synapse/workers/%i.yaml + # This service should be restarted when the synapse target is restarted. PartOf=matrix-synapse.target +# if this is started at the same time as the main, let the main process start +# first, to initialise the database schema. +After=matrix-synapse.service + [Service] Type=notify NotifyAccess=main diff --git a/docs/systemd-with-workers/workers/federation_reader.yaml b/docs/systemd-with-workers/workers/federation_reader.yaml
index 5b65c7040d..13e69e62c9 100644 --- a/docs/systemd-with-workers/workers/federation_reader.yaml +++ b/docs/systemd-with-workers/workers/federation_reader.yaml
@@ -1,7 +1,7 @@ worker_app: synapse.app.federation_reader +worker_name: federation_reader1 worker_replication_host: 127.0.0.1 -worker_replication_port: 9092 worker_replication_http_port: 9093 worker_listeners: diff --git a/docs/user_directory.md b/docs/user_directory.md
index 37dc71e751..872fc21979 100644 --- a/docs/user_directory.md +++ b/docs/user_directory.md
@@ -7,6 +7,6 @@ who are present in a publicly viewable room present on the server. The directory info is stored in various tables, which can (typically after DB corruption) get stale or out of sync. If this happens, for now the -solution to fix it is to execute the SQL [here](../synapse/storage/data_stores/main/schema/delta/53/user_dir_populate.sql) +solution to fix it is to execute the SQL [here](../synapse/storage/databases/main/schema/delta/53/user_dir_populate.sql) and then restart synapse. This should then start a background task to flush the current tables and regenerate the directory. diff --git a/docs/workers.md b/docs/workers.md
index 38bd758e57..ad4d8ca9f2 100644 --- a/docs/workers.md +++ b/docs/workers.md
@@ -1,10 +1,10 @@ # Scaling synapse via workers -For small instances it recommended to run Synapse in monolith mode (the -default). For larger instances where performance is a concern it can be helpful -to split out functionality into multiple separate python processes. These -processes are called 'workers', and are (eventually) intended to scale -horizontally independently. +For small instances it recommended to run Synapse in the default monolith mode. +For larger instances where performance is a concern it can be helpful to split +out functionality into multiple separate python processes. These processes are +called 'workers', and are (eventually) intended to scale horizontally +independently. Synapse's worker support is under active development and subject to change as we attempt to rapidly scale ever larger Synapse instances. However we are @@ -23,29 +23,30 @@ The processes communicate with each other via a Synapse-specific protocol called feeds streams of newly written data between processes so they can be kept in sync with the database state. -Additionally, processes may make HTTP requests to each other. Typically this is -used for operations which need to wait for a reply - such as sending an event. +When configured to do so, Synapse uses a +[Redis pub/sub channel](https://redis.io/topics/pubsub) to send the replication +stream between all configured Synapse processes. Additionally, processes may +make HTTP requests to each other, primarily for operations which need to wait +for a reply ─ such as sending an event. -As of Synapse v1.13.0, it is possible to configure Synapse to send replication -via a [Redis pub/sub channel](https://redis.io/topics/pubsub), and is now the -recommended way of configuring replication. This is an alternative to the old -direct TCP connections to the main process: rather than all the workers -connecting to the main process, all the workers and the main process connect to -Redis, which relays replication commands between processes. This can give a -significant cpu saving on the main process and will be a prerequisite for -upcoming performance improvements. +Redis support was added in v1.13.0 with it becoming the recommended method in +v1.18.0. It replaced the old direct TCP connections (which is deprecated as of +v1.18.0) to the main process. With Redis, rather than all the workers connecting +to the main process, all the workers and the main process connect to Redis, +which relays replication commands between processes. This can give a significant +cpu saving on the main process and will be a prerequisite for upcoming +performance improvements. -(See the [Architectural diagram](#architectural-diagram) section at the end for -a visualisation of what this looks like) +See the [Architectural diagram](#architectural-diagram) section at the end for +a visualisation of what this looks like. ## Setting up workers A Redis server is required to manage the communication between the processes. -(The older direct TCP connections are now deprecated.) The Redis server -should be installed following the normal procedure for your distribution (e.g. -`apt install redis-server` on Debian). It is safe to use an existing Redis -deployment if you have one. +The Redis server should be installed following the normal procedure for your +distribution (e.g. `apt install redis-server` on Debian). It is safe to use an +existing Redis deployment if you have one. Once installed, check that Redis is running and accessible from the host running Synapse, for example by executing `echo PING | nc -q1 localhost 6379` and seeing @@ -65,18 +66,31 @@ https://hub.docker.com/r/matrixdotorg/synapse/. To make effective use of the workers, you will need to configure an HTTP reverse-proxy such as nginx or haproxy, which will direct incoming requests to -the correct worker, or to the main synapse instance. See [reverse_proxy.md](reverse_proxy.md) -for information on setting up a reverse proxy. +the correct worker, or to the main synapse instance. See +[reverse_proxy.md](reverse_proxy.md) for information on setting up a reverse +proxy. -To enable workers you should create a configuration file for each worker -process. Each worker configuration file inherits the configuration of the shared -homeserver configuration file. You can then override configuration specific to -that worker, e.g. the HTTP listener that it provides (if any); logging -configuration; etc. You should minimise the number of overrides though to -maintain a usable config. +When using workers, each worker process has its own configuration file which +contains settings specific to that worker, such as the HTTP listener that it +provides (if any), logging configuration, etc. + +Normally, the worker processes are configured to read from a shared +configuration file as well as the worker-specific configuration files. This +makes it easier to keep common configuration settings synchronised across all +the processes. + +The main process is somewhat special in this respect: it does not normally +need its own configuration file and can take all of its configuration from the +shared configuration file. + + +### Shared configuration + +Normally, only a couple of changes are needed to make an existing configuration +file suitable for use with workers. First, you need to enable an "HTTP replication +listener" for the main process; and secondly, you need to enable redis-based +replication. For example: -Next you need to add both a HTTP replication listener and redis config to the -shared Synapse configuration file (`homeserver.yaml`). For example: ```yaml # extend the existing `listeners` section. This defines the ports that the @@ -98,6 +112,9 @@ See the sample config for the full documentation of each option. Under **no circumstances** should the replication listener be exposed to the public internet; it has no authentication and is unencrypted. + +### Worker configuration + In the config file for each worker, you must specify the type of worker application (`worker_app`), and you should specify a unqiue name for the worker (`worker_name`). The currently available worker applications are listed below. @@ -136,6 +153,9 @@ plain HTTP endpoint on port 8083 separately serving various endpoints, e.g. Obviously you should configure your reverse-proxy to route the relevant endpoints to the worker (`localhost:8083` in the above example). + +### Running Synapse with workers + Finally, you need to start your worker processes. This can be done with either `synctl` or your distribution's preferred service manager such as `systemd`. We recommend the use of `systemd` where available: for information on setting up @@ -197,6 +217,7 @@ expressions: ^/_matrix/client/(api/v1|r0|unstable)/joined_groups$ ^/_matrix/client/(api/v1|r0|unstable)/publicised_groups$ ^/_matrix/client/(api/v1|r0|unstable)/publicised_groups/ + ^/_synapse/client/password_reset/email/submit_token$ # Registration/login requests ^/_matrix/client/(api/v1|r0|unstable)/login$ @@ -222,6 +243,22 @@ for the room are in flight: ^/_matrix/client/(api/v1|r0|unstable)/rooms/.*/messages$ +Additionally, the following endpoints should be included if Synapse is configured +to use SSO (you only need to include the ones for whichever SSO provider you're +using): + + # OpenID Connect requests. + ^/_matrix/client/(api/v1|r0|unstable)/login/sso/redirect$ + ^/_synapse/oidc/callback$ + + # SAML requests. + ^/_matrix/client/(api/v1|r0|unstable)/login/sso/redirect$ + ^/_matrix/saml2/authn_response$ + + # CAS requests. + ^/_matrix/client/(api/v1|r0|unstable)/login/(cas|sso)/redirect$ + ^/_matrix/client/(api/v1|r0|unstable)/login/cas/ticket$ + Note that a HTTP listener with `client` and `federation` resources must be configured in the `worker_listeners` option in the worker config. @@ -278,7 +315,7 @@ instance_map: host: localhost port: 8034 -streams_writers: +stream_writers: events: event_persister1 ``` @@ -398,6 +435,23 @@ all these to be folded into the `generic_worker` app and to use config to define which processes handle the various proccessing such as push notifications. +## Migration from old config + +There are two main independent changes that have been made: introducing Redis +support and merging apps into `synapse.app.generic_worker`. Both these changes +are backwards compatible and so no changes to the config are required, however +server admins are encouraged to plan to migrate to Redis as the old style direct +TCP replication config is deprecated. + +To migrate to Redis add the `redis` config as above, and optionally remove the +TCP `replication` listener from master and `worker_replication_port` from worker +config. + +To migrate apps to use `synapse.app.generic_worker` simply update the +`worker_app` option in the worker configs, and where worker are started (e.g. +in systemd service files, but not required for synctl). + + ## Architectural diagram The following shows an example setup using Redis and a reverse proxy: diff --git a/mypy.ini b/mypy.ini
index a61009b197..7986781432 100644 --- a/mypy.ini +++ b/mypy.ini
@@ -1,11 +1,69 @@ [mypy] namespace_packages = True -plugins = mypy_zope:plugin +plugins = mypy_zope:plugin, scripts-dev/mypy_synapse_plugin.py follow_imports = silent check_untyped_defs = True show_error_codes = True show_traceback = True mypy_path = stubs +files = + synapse/api, + synapse/appservice, + synapse/config, + synapse/event_auth.py, + synapse/events/builder.py, + synapse/events/spamcheck.py, + synapse/federation, + synapse/handlers/auth.py, + synapse/handlers/cas_handler.py, + synapse/handlers/directory.py, + synapse/handlers/events.py, + synapse/handlers/federation.py, + synapse/handlers/identity.py, + synapse/handlers/initial_sync.py, + synapse/handlers/message.py, + synapse/handlers/oidc_handler.py, + synapse/handlers/pagination.py, + synapse/handlers/presence.py, + synapse/handlers/room.py, + synapse/handlers/room_member.py, + synapse/handlers/room_member_worker.py, + synapse/handlers/saml_handler.py, + synapse/handlers/sync.py, + synapse/handlers/ui_auth, + synapse/http/federation/well_known_resolver.py, + synapse/http/server.py, + synapse/http/site.py, + synapse/logging, + synapse/metrics, + synapse/module_api, + synapse/notifier.py, + synapse/push/pusherpool.py, + synapse/push/push_rule_evaluator.py, + synapse/replication, + synapse/rest, + synapse/server.py, + synapse/server_notices, + synapse/spam_checker_api, + synapse/state, + synapse/storage/databases/main/events.py, + synapse/storage/databases/main/stream.py, + synapse/storage/databases/main/ui_auth.py, + synapse/storage/database.py, + synapse/storage/engines, + synapse/storage/persist_events.py, + synapse/storage/state.py, + synapse/storage/util, + synapse/streams, + synapse/types.py, + synapse/util/async_helpers.py, + synapse/util/caches/descriptors.py, + synapse/util/caches/stream_change_cache.py, + synapse/util/metrics.py, + tests/replication, + tests/test_utils, + tests/rest/client/v2_alpha/test_auth.py, + tests/util/test_stream_change_cache.py [mypy-pymacaroons.*] ignore_missing_imports = True @@ -81,3 +139,6 @@ ignore_missing_imports = True [mypy-rust_python_jaeger_reporter.*] ignore_missing_imports = True + +[mypy-nacl.*] +ignore_missing_imports = True diff --git a/scripts-dev/build_debian_packages b/scripts-dev/build_debian_packages
index d055cf3287..d0685c8b35 100755 --- a/scripts-dev/build_debian_packages +++ b/scripts-dev/build_debian_packages
@@ -25,6 +25,7 @@ DISTS = ( "ubuntu:xenial", "ubuntu:bionic", "ubuntu:focal", + "ubuntu:groovy", ) DESC = '''\ diff --git a/scripts-dev/check-newsfragment b/scripts-dev/check-newsfragment
index b55368ea6b..d742c522b5 100755 --- a/scripts-dev/check-newsfragment +++ b/scripts-dev/check-newsfragment
@@ -3,6 +3,8 @@ # A script which checks that an appropriate news file has been added on this # branch. +echo -e "+++ \033[32mChecking newsfragment\033[m" + set -e # make sure that origin/dinsic is up to date @@ -16,6 +18,8 @@ pr="$BUILDKITE_PULL_REQUEST" if ! git diff --quiet FETCH_HEAD... -- debian; then if git diff --quiet FETCH_HEAD... -- debian/changelog; then echo "Updates to debian directory, but no update to the changelog." >&2 + echo "!! Please see the contributing guide for help writing your changelog entry:" >&2 + echo "https://github.com/matrix-org/synapse/blob/develop/CONTRIBUTING.md#debian-changelog" >&2 exit 1 fi fi @@ -26,7 +30,12 @@ if ! git diff --name-only FETCH_HEAD... | grep -qv '^debian/'; then exit 0 fi -tox -qe check-newsfragment +# Print a link to the contributing guide if the user makes a mistake +CONTRIBUTING_GUIDE_TEXT="!! Please see the contributing guide for help writing your changelog entry: +https://github.com/matrix-org/synapse/blob/develop/CONTRIBUTING.md#changelog" + +# If check-newsfragment returns a non-zero exit code, print the contributing guide and exit +tox -qe check-newsfragment || (echo -e "$CONTRIBUTING_GUIDE_TEXT" >&2 && exit 1) echo echo "--------------------------" @@ -38,6 +47,7 @@ for f in `git diff --name-only FETCH_HEAD... -- changelog.d`; do lastchar=`tr -d '\n' < $f | tail -c 1` if [ $lastchar != '.' -a $lastchar != '!' ]; then echo -e "\e[31mERROR: newsfragment $f does not end with a '.' or '!'\e[39m" >&2 + echo -e "$CONTRIBUTING_GUIDE_TEXT" >&2 exit 1 fi @@ -47,5 +57,6 @@ done if [[ -n "$pr" && "$matched" -eq 0 ]]; then echo -e "\e[31mERROR: Did not find a news fragment with the right number: expected changelog.d/$pr.*.\e[39m" >&2 + echo -e "$CONTRIBUTING_GUIDE_TEXT" >&2 exit 1 fi diff --git a/scripts-dev/check_line_terminators.sh b/scripts-dev/check_line_terminators.sh new file mode 100755
index 0000000000..c983956231 --- /dev/null +++ b/scripts-dev/check_line_terminators.sh
@@ -0,0 +1,34 @@ +#!/bin/bash +# +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# +# This script checks that line terminators in all repository files (excluding +# those in the .git directory) feature unix line terminators. +# +# Usage: +# +# ./check_line_terminators.sh +# +# The script will emit exit code 1 if any files that do not use unix line +# terminators are found, 0 otherwise. + +# cd to the root of the repository +cd `dirname $0`/.. + +# Find and print files with non-unix line terminators +if find . -path './.git/*' -prune -o -type f -print0 | xargs -0 grep -I -l $'\r$'; then + echo -e '\e[31mERROR: found files with CRLF line endings. See above.\e[39m' + exit 1 +fi diff --git a/scripts-dev/complement.sh b/scripts-dev/complement.sh new file mode 100755
index 0000000000..3cde53f5c0 --- /dev/null +++ b/scripts-dev/complement.sh
@@ -0,0 +1,22 @@ +#! /bin/bash -eu +# This script is designed for developers who want to test their code +# against Complement. +# +# It makes a Synapse image which represents the current checkout, +# then downloads Complement and runs it with that image. + +cd "$(dirname $0)/.." + +# Build the base Synapse image from the local checkout +docker build -t matrixdotorg/synapse:latest -f docker/Dockerfile . + +# Download Complement +wget -N https://github.com/matrix-org/complement/archive/master.tar.gz +tar -xzf master.tar.gz +cd complement-master + +# Build the Synapse image from Complement, based on the above image we just built +docker build -t complement-synapse -f dockerfiles/Synapse.Dockerfile ./dockerfiles + +# Run the tests on the resulting image! +COMPLEMENT_BASE_IMAGE=complement-synapse go test -v -count=1 ./tests diff --git a/scripts-dev/definitions.py b/scripts-dev/definitions.py
index 9eddb6d515..313860df13 100755 --- a/scripts-dev/definitions.py +++ b/scripts-dev/definitions.py
@@ -1,7 +1,5 @@ #! /usr/bin/python -from __future__ import print_function - import argparse import ast import os @@ -13,7 +11,7 @@ import yaml class DefinitionVisitor(ast.NodeVisitor): def __init__(self): - super(DefinitionVisitor, self).__init__() + super().__init__() self.functions = {} self.classes = {} self.names = {} diff --git a/scripts-dev/dump_macaroon.py b/scripts-dev/dump_macaroon.py
index 22b30fa78e..980b5e709f 100755 --- a/scripts-dev/dump_macaroon.py +++ b/scripts-dev/dump_macaroon.py
@@ -1,7 +1,5 @@ #!/usr/bin/env python2 -from __future__ import print_function - import sys import pymacaroons diff --git a/scripts-dev/federation_client.py b/scripts-dev/federation_client.py
index 531010185d..abcec48c4f 100755 --- a/scripts-dev/federation_client.py +++ b/scripts-dev/federation_client.py
@@ -15,16 +15,16 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - import argparse import base64 import json import sys +from typing import Any, Optional from urllib import parse as urlparse import nacl.signing import requests +import signedjson.types import srvlookup import yaml from requests.adapters import HTTPAdapter @@ -69,7 +69,9 @@ def encode_canonical_json(value): ).encode("UTF-8") -def sign_json(json_object, signing_key, signing_name): +def sign_json( + json_object: Any, signing_key: signedjson.types.SigningKey, signing_name: str +) -> Any: signatures = json_object.pop("signatures", {}) unsigned = json_object.pop("unsigned", None) @@ -122,7 +124,14 @@ def read_signing_keys(stream): return keys -def request_json(method, origin_name, origin_key, destination, path, content): +def request( + method: Optional[str], + origin_name: str, + origin_key: signedjson.types.SigningKey, + destination: str, + path: str, + content: Optional[str], +) -> requests.Response: if method is None: if content is None: method = "GET" @@ -159,11 +168,14 @@ def request_json(method, origin_name, origin_key, destination, path, content): if method == "POST": headers["Content-Type"] = "application/json" - result = s.request( - method=method, url=dest, headers=headers, verify=False, data=content + return s.request( + method=method, + url=dest, + headers=headers, + verify=False, + data=content, + stream=True, ) - sys.stderr.write("Status Code: %d\n" % (result.status_code,)) - return result.json() def main(): @@ -222,7 +234,7 @@ def main(): with open(args.signing_key_path) as f: key = read_signing_keys(f)[0] - result = request_json( + result = request( args.method, args.server_name, key, @@ -231,7 +243,12 @@ def main(): content=args.body, ) - json.dump(result, sys.stdout) + sys.stderr.write("Status Code: %d\n" % (result.status_code,)) + + for chunk in result.iter_content(): + # we write raw utf8 to stdout. + sys.stdout.buffer.write(chunk) + print("") @@ -304,7 +321,7 @@ class MatrixConnectionAdapter(HTTPAdapter): url = urlparse.urlunparse( ("https", netloc, parsed.path, parsed.params, parsed.query, parsed.fragment) ) - return super(MatrixConnectionAdapter, self).get_connection(url, proxies) + return super().get_connection(url, proxies) if __name__ == "__main__": diff --git a/scripts-dev/hash_history.py b/scripts-dev/hash_history.py
index bf3862a386..8d6c3d24db 100644 --- a/scripts-dev/hash_history.py +++ b/scripts-dev/hash_history.py
@@ -1,5 +1,3 @@ -from __future__ import print_function - import sqlite3 import sys @@ -15,7 +13,7 @@ from synapse.storage.pdu import PduStore from synapse.storage.signatures import SignatureStore -class Store(object): +class Store: _get_pdu_tuples = PduStore.__dict__["_get_pdu_tuples"] _get_pdu_content_hashes_txn = SignatureStore.__dict__["_get_pdu_content_hashes_txn"] _get_prev_pdu_hashes_txn = SignatureStore.__dict__["_get_prev_pdu_hashes_txn"] diff --git a/scripts-dev/mypy_synapse_plugin.py b/scripts-dev/mypy_synapse_plugin.py new file mode 100644
index 0000000000..a5b88731f1 --- /dev/null +++ b/scripts-dev/mypy_synapse_plugin.py
@@ -0,0 +1,85 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +"""This is a mypy plugin for Synpase to deal with some of the funky typing that +can crop up, e.g the cache descriptors. +""" + +from typing import Callable, Optional + +from mypy.plugin import MethodSigContext, Plugin +from mypy.typeops import bind_self +from mypy.types import CallableType + + +class SynapsePlugin(Plugin): + def get_method_signature_hook( + self, fullname: str + ) -> Optional[Callable[[MethodSigContext], CallableType]]: + if fullname.startswith( + "synapse.util.caches.descriptors._CachedFunction.__call__" + ): + return cached_function_method_signature + return None + + +def cached_function_method_signature(ctx: MethodSigContext) -> CallableType: + """Fixes the `_CachedFunction.__call__` signature to be correct. + + It already has *almost* the correct signature, except: + + 1. the `self` argument needs to be marked as "bound"; and + 2. any `cache_context` argument should be removed. + """ + + # First we mark this as a bound function signature. + signature = bind_self(ctx.default_signature) + + # Secondly, we remove any "cache_context" args. + # + # Note: We should be only doing this if `cache_context=True` is set, but if + # it isn't then the code will raise an exception when its called anyway, so + # its not the end of the world. + context_arg_index = None + for idx, name in enumerate(signature.arg_names): + if name == "cache_context": + context_arg_index = idx + break + + if context_arg_index: + arg_types = list(signature.arg_types) + arg_types.pop(context_arg_index) + + arg_names = list(signature.arg_names) + arg_names.pop(context_arg_index) + + arg_kinds = list(signature.arg_kinds) + arg_kinds.pop(context_arg_index) + + signature = signature.copy_modified( + arg_types=arg_types, arg_names=arg_names, arg_kinds=arg_kinds, + ) + + return signature + + +def plugin(version: str): + # This is the entry point of the plugin, and let's us deal with the fact + # that the mypy plugin interface is *not* stable by looking at the version + # string. + # + # However, since we pin the version of mypy Synapse uses in CI, we don't + # really care. + return SynapsePlugin diff --git a/scripts-dev/update_database b/scripts-dev/update_database
index 94aa8758b4..56365e2b58 100755 --- a/scripts-dev/update_database +++ b/scripts-dev/update_database
@@ -40,7 +40,7 @@ class MockHomeserver(HomeServer): config.server_name, reactor=reactor, config=config, **kwargs ) - self.version_string = "Synapse/"+get_version_string(synapse) + self.version_string = "Synapse/" + get_version_string(synapse) if __name__ == "__main__": @@ -86,7 +86,7 @@ if __name__ == "__main__": store = hs.get_datastore() async def run_background_updates(): - await store.db.updates.run_background_updates(sleep=False) + await store.db_pool.updates.run_background_updates(sleep=False) # Stop the reactor to exit the script once every background update is run. reactor.stop() diff --git a/scripts/move_remote_media_to_new_store.py b/scripts/move_remote_media_to_new_store.py
index b5b63933ab..ab2e763386 100755 --- a/scripts/move_remote_media_to_new_store.py +++ b/scripts/move_remote_media_to_new_store.py
@@ -32,8 +32,6 @@ To use, pipe the above into:: PYTHON_PATH=. ./scripts/move_remote_media_to_new_store.py <source repo> <dest repo> """ -from __future__ import print_function - import argparse import logging import os diff --git a/scripts/register_new_matrix_user b/scripts/register_new_matrix_user
index b450712ab7..8b9d30877d 100755 --- a/scripts/register_new_matrix_user +++ b/scripts/register_new_matrix_user
@@ -14,8 +14,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - from synapse._scripts.register_new_matrix_user import main if __name__ == "__main__": diff --git a/scripts/synapse_port_db b/scripts/synapse_port_db
index b21b8d573d..dd1ebde894 100755 --- a/scripts/synapse_port_db +++ b/scripts/synapse_port_db
@@ -35,32 +35,30 @@ from synapse.logging.context import ( make_deferred_yieldable, run_in_background, ) -from synapse.storage.data_stores.main.client_ips import ClientIpBackgroundUpdateStore -from synapse.storage.data_stores.main.deviceinbox import ( - DeviceInboxBackgroundUpdateStore, -) -from synapse.storage.data_stores.main.devices import DeviceBackgroundUpdateStore -from synapse.storage.data_stores.main.events_bg_updates import ( +from synapse.storage.database import DatabasePool, make_conn +from synapse.storage.databases.main.client_ips import ClientIpBackgroundUpdateStore +from synapse.storage.databases.main.deviceinbox import DeviceInboxBackgroundUpdateStore +from synapse.storage.databases.main.devices import DeviceBackgroundUpdateStore +from synapse.storage.databases.main.events_bg_updates import ( EventsBackgroundUpdatesStore, ) -from synapse.storage.data_stores.main.media_repository import ( +from synapse.storage.databases.main.media_repository import ( MediaRepositoryBackgroundUpdateStore, ) -from synapse.storage.data_stores.main.profile import ProfileStore -from synapse.storage.data_stores.main.registration import ( +from synapse.storage.databases.main.profile import ProfileStore +from synapse.storage.databases.main.registration import ( RegistrationBackgroundUpdateStore, find_max_generated_user_id_localpart, ) -from synapse.storage.data_stores.main.room import RoomBackgroundUpdateStore -from synapse.storage.data_stores.main.roommember import RoomMemberBackgroundUpdateStore -from synapse.storage.data_stores.main.search import SearchBackgroundUpdateStore -from synapse.storage.data_stores.main.state import MainStateBackgroundUpdateStore -from synapse.storage.data_stores.main.stats import StatsStore -from synapse.storage.data_stores.main.user_directory import ( +from synapse.storage.databases.main.room import RoomBackgroundUpdateStore +from synapse.storage.databases.main.roommember import RoomMemberBackgroundUpdateStore +from synapse.storage.databases.main.search import SearchBackgroundUpdateStore +from synapse.storage.databases.main.state import MainStateBackgroundUpdateStore +from synapse.storage.databases.main.stats import StatsStore +from synapse.storage.databases.main.user_directory import ( UserDirectoryBackgroundUpdateStore, ) -from synapse.storage.data_stores.state.bg_updates import StateBackgroundUpdateStore -from synapse.storage.database import Database, make_conn +from synapse.storage.databases.state.bg_updates import StateBackgroundUpdateStore from synapse.storage.engines import create_engine from synapse.storage.prepare_database import prepare_database from synapse.util import Clock @@ -92,6 +90,7 @@ BOOLEAN_COLUMNS = { "redactions": ["have_censored"], "room_stats_state": ["is_federatable"], "local_media_repository": ["safe_from_quarantine"], + "users": ["shadow_banned"], } @@ -147,6 +146,7 @@ IGNORED_TABLES = { # the sessions are transient anyway, so ignore them. "ui_auth_sessions", "ui_auth_sessions_credentials", + "ui_auth_sessions_ips", } @@ -177,14 +177,14 @@ class Store( StatsStore, ): def execute(self, f, *args, **kwargs): - return self.db.runInteraction(f.__name__, f, *args, **kwargs) + return self.db_pool.runInteraction(f.__name__, f, *args, **kwargs) def execute_sql(self, sql, *args): def r(txn): txn.execute(sql, args) return txn.fetchall() - return self.db.runInteraction("execute_sql", r) + return self.db_pool.runInteraction("execute_sql", r) def insert_many_txn(self, txn, table, headers, rows): sql = "INSERT INTO %s (%s) VALUES (%s)" % ( @@ -229,7 +229,7 @@ class Porter(object): async def setup_table(self, table): if table in APPEND_ONLY_TABLES: # It's safe to just carry on inserting. - row = await self.postgres_store.db.simple_select_one( + row = await self.postgres_store.db_pool.simple_select_one( table="port_from_sqlite3", keyvalues={"table_name": table}, retcols=("forward_rowid", "backward_rowid"), @@ -246,7 +246,7 @@ class Porter(object): ) = await self._setup_sent_transactions() backward_chunk = 0 else: - await self.postgres_store.db.simple_insert( + await self.postgres_store.db_pool.simple_insert( table="port_from_sqlite3", values={ "table_name": table, @@ -276,7 +276,7 @@ class Porter(object): await self.postgres_store.execute(delete_all) - await self.postgres_store.db.simple_insert( + await self.postgres_store.db_pool.simple_insert( table="port_from_sqlite3", values={"table_name": table, "forward_rowid": 1, "backward_rowid": 0}, ) @@ -320,7 +320,7 @@ class Porter(object): if table == "user_directory_stream_pos": # We need to make sure there is a single row, `(X, null), as that is # what synapse expects to be there. - await self.postgres_store.db.simple_insert( + await self.postgres_store.db_pool.simple_insert( table=table, values={"stream_id": None} ) self.progress.update(table, table_size) # Mark table as done @@ -361,7 +361,7 @@ class Porter(object): return headers, forward_rows, backward_rows - headers, frows, brows = await self.sqlite_store.db.runInteraction( + headers, frows, brows = await self.sqlite_store.db_pool.runInteraction( "select", r ) @@ -377,7 +377,7 @@ class Porter(object): def insert(txn): self.postgres_store.insert_many_txn(txn, table, headers[1:], rows) - self.postgres_store.db.simple_update_one_txn( + self.postgres_store.db_pool.simple_update_one_txn( txn, table="port_from_sqlite3", keyvalues={"table_name": table}, @@ -415,7 +415,7 @@ class Porter(object): return headers, rows - headers, rows = await self.sqlite_store.db.runInteraction("select", r) + headers, rows = await self.sqlite_store.db_pool.runInteraction("select", r) if rows: forward_chunk = rows[-1][0] + 1 @@ -453,7 +453,7 @@ class Porter(object): ], ) - self.postgres_store.db.simple_update_one_txn( + self.postgres_store.db_pool.simple_update_one_txn( txn, table="port_from_sqlite3", keyvalues={"table_name": "event_search"}, @@ -496,7 +496,7 @@ class Porter(object): db_conn, allow_outdated_version=allow_outdated_version ) prepare_database(db_conn, engine, config=self.hs_config) - store = Store(Database(hs, db_config, engine), db_conn, hs) + store = Store(DatabasePool(hs, db_config, engine), db_conn, hs) db_conn.commit() return store @@ -504,7 +504,7 @@ class Porter(object): async def run_background_updates_on_postgres(self): # Manually apply all background updates on the PostgreSQL database. postgres_ready = ( - await self.postgres_store.db.updates.has_completed_background_updates() + await self.postgres_store.db_pool.updates.has_completed_background_updates() ) if not postgres_ready: @@ -513,9 +513,9 @@ class Porter(object): self.progress.set_state("Running background updates on PostgreSQL") while not postgres_ready: - await self.postgres_store.db.updates.do_next_background_update(100) + await self.postgres_store.db_pool.updates.do_next_background_update(100) postgres_ready = await ( - self.postgres_store.db.updates.has_completed_background_updates() + self.postgres_store.db_pool.updates.has_completed_background_updates() ) async def run(self): @@ -536,7 +536,7 @@ class Porter(object): # Check if all background updates are done, abort if not. updates_complete = ( - await self.sqlite_store.db.updates.has_completed_background_updates() + await self.sqlite_store.db_pool.updates.has_completed_background_updates() ) if not updates_complete: end_error = ( @@ -578,22 +578,24 @@ class Porter(object): ) try: - await self.postgres_store.db.runInteraction("alter_table", alter_table) + await self.postgres_store.db_pool.runInteraction( + "alter_table", alter_table + ) except Exception: # On Error Resume Next pass - await self.postgres_store.db.runInteraction( + await self.postgres_store.db_pool.runInteraction( "create_port_table", create_port_table ) # Step 2. Get tables. self.progress.set_state("Fetching tables") - sqlite_tables = await self.sqlite_store.db.simple_select_onecol( + sqlite_tables = await self.sqlite_store.db_pool.simple_select_onecol( table="sqlite_master", keyvalues={"type": "table"}, retcol="name" ) - postgres_tables = await self.postgres_store.db.simple_select_onecol( + postgres_tables = await self.postgres_store.db_pool.simple_select_onecol( table="information_schema.tables", keyvalues={}, retcol="distinct table_name", @@ -629,6 +631,7 @@ class Porter(object): self.progress.set_state("Setting up sequence generators") await self._setup_state_group_id_seq() await self._setup_user_id_seq() + await self._setup_events_stream_seqs() self.progress.done() except Exception as e: @@ -694,7 +697,7 @@ class Porter(object): return headers, [r for r in rows if r[ts_ind] < yesterday] - headers, rows = await self.sqlite_store.db.runInteraction("select", r) + headers, rows = await self.sqlite_store.db_pool.runInteraction("select", r) rows = self._convert_rows("sent_transactions", headers, rows) @@ -727,7 +730,7 @@ class Porter(object): next_chunk = await self.sqlite_store.execute(get_start_id) next_chunk = max(max_inserted_rowid + 1, next_chunk) - await self.postgres_store.db.simple_insert( + await self.postgres_store.db_pool.simple_insert( table="port_from_sqlite3", values={ "table_name": "sent_transactions", @@ -796,14 +799,37 @@ class Porter(object): next_id = curr_id + 1 txn.execute("ALTER SEQUENCE state_group_id_seq RESTART WITH %s", (next_id,)) - return self.postgres_store.db.runInteraction("setup_state_group_id_seq", r) + return self.postgres_store.db_pool.runInteraction("setup_state_group_id_seq", r) def _setup_user_id_seq(self): def r(txn): next_id = find_max_generated_user_id_localpart(txn) + 1 txn.execute("ALTER SEQUENCE user_id_seq RESTART WITH %s", (next_id,)) - return self.postgres_store.db.runInteraction("setup_user_id_seq", r) + return self.postgres_store.db_pool.runInteraction("setup_user_id_seq", r) + + def _setup_events_stream_seqs(self): + def r(txn): + txn.execute("SELECT MAX(stream_ordering) FROM events") + curr_id = txn.fetchone()[0] + if curr_id: + next_id = curr_id + 1 + txn.execute( + "ALTER SEQUENCE events_stream_seq RESTART WITH %s", (next_id,) + ) + + txn.execute("SELECT -MIN(stream_ordering) FROM events") + curr_id = txn.fetchone()[0] + if curr_id: + next_id = curr_id + 1 + txn.execute( + "ALTER SEQUENCE events_backfill_stream_seq RESTART WITH %s", + (next_id,), + ) + + return self.postgres_store.db_pool.runInteraction( + "_setup_events_stream_seqs", r + ) ############################################## diff --git a/setup.py b/setup.py
index 54ddec8f9f..926b1bc86f 100755 --- a/setup.py +++ b/setup.py
@@ -94,6 +94,22 @@ ALL_OPTIONAL_REQUIREMENTS = dependencies["ALL_OPTIONAL_REQUIREMENTS"] # Make `pip install matrix-synapse[all]` install all the optional dependencies. CONDITIONAL_REQUIREMENTS["all"] = list(ALL_OPTIONAL_REQUIREMENTS) +# Developer dependencies should not get included in "all". +# +# We pin black so that our tests don't start failing on new releases. +CONDITIONAL_REQUIREMENTS["lint"] = [ + "isort==5.0.3", + "black==19.10b0", + "flake8-comprehensions", + "flake8", +] + +# Dependencies which are exclusively required by unit test code. This is +# NOT a list of all modules that are necessary to run the unit tests. +# Tests assume that all optional dependencies are installed. +# +# parameterized_class decorator was introduced in parameterized 0.7.0 +CONDITIONAL_REQUIREMENTS["test"] = ["mock>=2.0", "parameterized>=0.7.0"] setup( name="matrix-synapse", diff --git a/stubs/frozendict.pyi b/stubs/frozendict.pyi new file mode 100644
index 0000000000..3f3af59f26 --- /dev/null +++ b/stubs/frozendict.pyi
@@ -0,0 +1,47 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +# Stub for frozendict. + +from typing import ( + Any, + Hashable, + Iterable, + Iterator, + Mapping, + overload, + Tuple, + TypeVar, +) + +_KT = TypeVar("_KT", bound=Hashable) # Key type. +_VT = TypeVar("_VT") # Value type. + +class frozendict(Mapping[_KT, _VT]): + @overload + def __init__(self, **kwargs: _VT) -> None: ... + @overload + def __init__(self, __map: Mapping[_KT, _VT], **kwargs: _VT) -> None: ... + @overload + def __init__( + self, __iterable: Iterable[Tuple[_KT, _VT]], **kwargs: _VT + ) -> None: ... + def __getitem__(self, key: _KT) -> _VT: ... + def __contains__(self, key: Any) -> bool: ... + def copy(self, **add_or_replace: Any) -> frozendict: ... + def __iter__(self) -> Iterator[_KT]: ... + def __len__(self) -> int: ... + def __repr__(self) -> str: ... + def __hash__(self) -> int: ... diff --git a/synapse/__init__.py b/synapse/__init__.py
index 5155e719a1..83b8e4897f 100644 --- a/synapse/__init__.py +++ b/synapse/__init__.py
@@ -17,6 +17,7 @@ """ This is a reference implementation of a Matrix homeserver. """ +import json import os import sys @@ -25,6 +26,9 @@ if sys.version_info < (3, 5): print("Synapse requires Python 3.5 or above.") sys.exit(1) +# Twisted and canonicaljson will fail to import when this file is executed to +# get the __version__ during a fresh install. That's OK and subsequent calls to +# actually start Synapse will import these libraries fine. try: from twisted.internet import protocol from twisted.internet.protocol import Factory @@ -36,7 +40,15 @@ try: except ImportError: pass -__version__ = "1.18.0" +# Use the standard library json implementation instead of simplejson. +try: + from canonicaljson import set_json_library + + set_json_library(json) +except ImportError: + pass + +__version__ = "1.21.2" if bool(os.environ.get("SYNAPSE_TEST_PATCH_LOG_CONTEXTS", False)): # We import here so that we don't have to install a bunch of deps when diff --git a/synapse/_scripts/register_new_matrix_user.py b/synapse/_scripts/register_new_matrix_user.py
index 55cce2db22..da0996edbc 100644 --- a/synapse/_scripts/register_new_matrix_user.py +++ b/synapse/_scripts/register_new_matrix_user.py
@@ -14,8 +14,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - import argparse import getpass import hashlib diff --git a/synapse/api/auth.py b/synapse/api/auth.py
index 71859c9bc0..c6660fe687 100644 --- a/synapse/api/auth.py +++ b/synapse/api/auth.py
@@ -13,12 +13,11 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging -from typing import Optional +from typing import List, Optional, Tuple import pymacaroons from netaddr import IPAddress -from twisted.internet import defer from twisted.web.server import Request import synapse.types @@ -59,7 +58,7 @@ class _InvalidMacaroonException(Exception): pass -class Auth(object): +class Auth: """ FIXME: This class contains a mix of functions for authenticating users of our client-server API and authenticating events added to room graphs. @@ -80,13 +79,14 @@ class Auth(object): self._track_appservice_user_ips = hs.config.track_appservice_user_ips self._macaroon_secret_key = hs.config.macaroon_secret_key - @defer.inlineCallbacks - def check_from_context(self, room_version: str, event, context, do_sig_check=True): - prev_state_ids = yield context.get_prev_state_ids() - auth_events_ids = yield self.compute_auth_events( + async def check_from_context( + self, room_version: str, event, context, do_sig_check=True + ): + prev_state_ids = await context.get_prev_state_ids() + auth_events_ids = self.compute_auth_events( event, prev_state_ids, for_verification=True ) - auth_events = yield self.store.get_events(auth_events_ids) + auth_events = await self.store.get_events(auth_events_ids) auth_events = {(e.type, e.state_key): e for e in auth_events.values()} room_version_obj = KNOWN_ROOM_VERSIONS[room_version] @@ -94,14 +94,13 @@ class Auth(object): room_version_obj, event, auth_events=auth_events, do_sig_check=do_sig_check ) - @defer.inlineCallbacks - def check_user_in_room( + async def check_user_in_room( self, room_id: str, user_id: str, current_state: Optional[StateMap[EventBase]] = None, allow_departed_users: bool = False, - ): + ) -> EventBase: """Check if the user is in the room, or was at some point. Args: room_id: The room to check. @@ -119,37 +118,35 @@ class Auth(object): Raises: AuthError if the user is/was not in the room. Returns: - Deferred[Optional[EventBase]]: - Membership event for the user if the user was in the - room. This will be the join event if they are currently joined to - the room. This will be the leave event if they have left the room. + Membership event for the user if the user was in the + room. This will be the join event if they are currently joined to + the room. This will be the leave event if they have left the room. """ if current_state: member = current_state.get((EventTypes.Member, user_id), None) else: - member = yield defer.ensureDeferred( - self.state.get_current_state( - room_id=room_id, event_type=EventTypes.Member, state_key=user_id - ) + member = await self.state.get_current_state( + room_id=room_id, event_type=EventTypes.Member, state_key=user_id ) - membership = member.membership if member else None - if membership == Membership.JOIN: - return member + if member: + membership = member.membership - # XXX this looks totally bogus. Why do we not allow users who have been banned, - # or those who were members previously and have been re-invited? - if allow_departed_users and membership == Membership.LEAVE: - forgot = yield self.store.did_forget(user_id, room_id) - if not forgot: + if membership == Membership.JOIN: return member + # XXX this looks totally bogus. Why do we not allow users who have been banned, + # or those who were members previously and have been re-invited? + if allow_departed_users and membership == Membership.LEAVE: + forgot = await self.store.did_forget(user_id, room_id) + if not forgot: + return member + raise AuthError(403, "User %s not in room %s" % (user_id, room_id)) - @defer.inlineCallbacks - def check_host_in_room(self, room_id, host): + async def check_host_in_room(self, room_id, host): with Measure(self.clock, "check_host_in_room"): - latest_event_ids = yield self.store.is_host_joined(room_id, host) + latest_event_ids = await self.store.is_host_joined(room_id, host) return latest_event_ids def can_federate(self, event, auth_events): @@ -160,14 +157,13 @@ class Auth(object): def get_public_keys(self, invite_event): return event_auth.get_public_keys(invite_event) - @defer.inlineCallbacks - def get_user_by_req( + async def get_user_by_req( self, request: Request, allow_guest: bool = False, rights: str = "access", allow_expired: bool = False, - ): + ) -> synapse.types.Requester: """ Get a registered user's ID. Args: @@ -180,7 +176,7 @@ class Auth(object): /login will deliver access tokens regardless of expiration. Returns: - defer.Deferred: resolves to a `synapse.types.Requester` object + Resolves to the requester Raises: InvalidClientCredentialsError if no user by that token exists or the token is invalid. @@ -194,15 +190,14 @@ class Auth(object): access_token = self.get_access_token_from_request(request) - user_id, app_service = yield self._get_appservice_user_id(request) - + user_id, app_service = self._get_appservice_user_id(request) if user_id: request.authenticated_entity = user_id opentracing.set_tag("authenticated_entity", user_id) opentracing.set_tag("appservice_id", app_service.id) if ip_addr and self._track_appservice_user_ips: - yield self.store.insert_client_ip( + await self.store.insert_client_ip( user_id=user_id, access_token=access_token, ip=ip_addr, @@ -212,20 +207,18 @@ class Auth(object): return synapse.types.create_requester(user_id, app_service=app_service) - user_info = yield self.get_user_by_access_token( + user_info = await self.get_user_by_access_token( access_token, rights, allow_expired=allow_expired ) user = user_info["user"] token_id = user_info["token_id"] is_guest = user_info["is_guest"] + shadow_banned = user_info["shadow_banned"] # Deny the request if the user account has expired. if self._account_validity.enabled and not allow_expired: user_id = user.to_string() - expired = yield self.store.is_account_expired( - user_id, self.clock.time_msec() - ) - if expired: + if await self.store.is_account_expired(user_id, self.clock.time_msec()): raise AuthError( 403, "User account has expired", errcode=Codes.EXPIRED_ACCOUNT ) @@ -235,7 +228,7 @@ class Auth(object): device_id = user_info.get("device_id") if user and access_token and ip_addr: - yield self.store.insert_client_ip( + await self.store.insert_client_ip( user_id=user.to_string(), access_token=access_token, ip=ip_addr, @@ -256,7 +249,12 @@ class Auth(object): opentracing.set_tag("device_id", device_id) return synapse.types.create_requester( - user, token_id, is_guest, device_id, app_service=app_service + user, + token_id, + is_guest, + shadow_banned, + device_id, + app_service=app_service, ) except KeyError: raise MissingClientTokenError() @@ -291,10 +289,9 @@ class Auth(object): # ) return user_id, app_service - @defer.inlineCallbacks - def get_user_by_access_token( + async def get_user_by_access_token( self, token: str, rights: str = "access", allow_expired: bool = False, - ): + ) -> dict: """ Validate access token and get user_id from it Args: @@ -304,9 +301,10 @@ class Auth(object): allow_expired: If False, raises an InvalidClientTokenError if the token is expired Returns: - Deferred[dict]: dict that includes: + dict that includes: `user` (UserID) `is_guest` (bool) + `shadow_banned` (bool) `token_id` (int|None): access token id. May be None if guest `device_id` (str|None): device corresponding to access token Raises: @@ -318,7 +316,7 @@ class Auth(object): if rights == "access": # first look in the database - r = yield self._look_up_user_by_access_token(token) + r = await self._look_up_user_by_access_token(token) if r: valid_until_ms = r["valid_until_ms"] if ( @@ -356,7 +354,7 @@ class Auth(object): # It would of course be much easier to store guest access # tokens in the database as well, but that would break existing # guest tokens. - stored_user = yield self.store.get_user_by_id(user_id) + stored_user = await self.store.get_user_by_id(user_id) if not stored_user: raise InvalidClientTokenError("Unknown user_id %s" % user_id) if not stored_user["is_guest"]: @@ -366,6 +364,7 @@ class Auth(object): ret = { "user": user, "is_guest": True, + "shadow_banned": False, "token_id": None, # all guests get the same device id "device_id": GUEST_DEVICE_ID, @@ -375,6 +374,7 @@ class Auth(object): ret = { "user": user, "is_guest": False, + "shadow_banned": False, "token_id": None, "device_id": None, } @@ -486,9 +486,8 @@ class Auth(object): now = self.hs.get_clock().time_msec() return now < expiry - @defer.inlineCallbacks - def _look_up_user_by_access_token(self, token): - ret = yield self.store.get_user_by_access_token(token) + async def _look_up_user_by_access_token(self, token): + ret = await self.store.get_user_by_access_token(token) if not ret: return None @@ -499,6 +498,7 @@ class Auth(object): "user": UserID.from_string(ret.get("name")), "token_id": ret.get("token_id", None), "is_guest": False, + "shadow_banned": ret.get("shadow_banned"), "device_id": ret.get("device_id"), "valid_until_ms": ret.get("valid_until_ms"), } @@ -511,7 +511,7 @@ class Auth(object): logger.warning("Unrecognised appservice access token.") raise InvalidClientTokenError() request.authenticated_entity = service.sender - return defer.succeed(service) + return service async def is_server_admin(self, user: UserID) -> bool: """ Check if the given user is a local server admin. @@ -526,7 +526,7 @@ class Auth(object): def compute_auth_events( self, event, current_state_ids: StateMap[str], for_verification: bool = False, - ): + ) -> List[str]: """Given an event and current state return the list of event IDs used to auth an event. @@ -534,11 +534,11 @@ class Auth(object): should be added to the event's `auth_events`. Returns: - defer.Deferred(list[str]): List of event IDs. + List of event IDs. """ if event.type == EventTypes.Create: - return defer.succeed([]) + return [] # Currently we ignore the `for_verification` flag even though there are # some situations where we can drop particular auth events when adding @@ -557,7 +557,7 @@ class Auth(object): if auth_ev_id: auth_ids.append(auth_ev_id) - return defer.succeed(auth_ids) + return auth_ids async def check_can_change_room_list(self, room_id: str, user: UserID): """Determine whether the user is allowed to edit the room's entry in the @@ -640,10 +640,9 @@ class Auth(object): return query_params[0].decode("ascii") - @defer.inlineCallbacks - def check_user_in_room_or_world_readable( + async def check_user_in_room_or_world_readable( self, room_id: str, user_id: str, allow_departed_users: bool = False - ): + ) -> Tuple[str, Optional[str]]: """Checks that the user is or was in the room or the room is world readable. If it isn't then an exception is raised. @@ -654,10 +653,9 @@ class Auth(object): members but have now departed Returns: - Deferred[tuple[str, str|None]]: Resolves to the current membership of - the user in the room and the membership event ID of the user. If - the user is not in the room and never has been, then - `(Membership.JOIN, None)` is returned. + Resolves to the current membership of the user in the room and the + membership event ID of the user. If the user is not in the room and + never has been, then `(Membership.JOIN, None)` is returned. """ try: @@ -666,15 +664,13 @@ class Auth(object): # * The user is a non-guest user, and was ever in the room # * The user is a guest user, and has joined the room # else it will throw. - member_event = yield self.check_user_in_room( + member_event = await self.check_user_in_room( room_id, user_id, allow_departed_users=allow_departed_users ) return member_event.membership, member_event.event_id except AuthError: - visibility = yield defer.ensureDeferred( - self.state.get_current_state( - room_id, EventTypes.RoomHistoryVisibility, "" - ) + visibility = await self.state.get_current_state( + room_id, EventTypes.RoomHistoryVisibility, "" ) if ( visibility diff --git a/synapse/api/auth_blocking.py b/synapse/api/auth_blocking.py
index 5c499b6b4e..d8fafd7cb8 100644 --- a/synapse/api/auth_blocking.py +++ b/synapse/api/auth_blocking.py
@@ -15,8 +15,6 @@ import logging -from twisted.internet import defer - from synapse.api.constants import LimitBlockingTypes, UserTypes from synapse.api.errors import Codes, ResourceLimitError from synapse.config.server import is_threepid_reserved @@ -24,7 +22,7 @@ from synapse.config.server import is_threepid_reserved logger = logging.getLogger(__name__) -class AuthBlocking(object): +class AuthBlocking: def __init__(self, hs): self.store = hs.get_datastore() @@ -36,8 +34,7 @@ class AuthBlocking(object): self._limit_usage_by_mau = hs.config.limit_usage_by_mau self._mau_limits_reserved_threepids = hs.config.mau_limits_reserved_threepids - @defer.inlineCallbacks - def check_auth_blocking(self, user_id=None, threepid=None, user_type=None): + async def check_auth_blocking(self, user_id=None, threepid=None, user_type=None): """Checks if the user should be rejected for some external reason, such as monthly active user limiting or global disable flag @@ -60,7 +57,7 @@ class AuthBlocking(object): if user_id is not None: if user_id == self._server_notices_mxid: return - if (yield self.store.is_support_user(user_id)): + if await self.store.is_support_user(user_id): return if self._hs_disabled: @@ -76,11 +73,11 @@ class AuthBlocking(object): # If the user is already part of the MAU cohort or a trial user if user_id: - timestamp = yield self.store.user_last_seen_monthly_active(user_id) + timestamp = await self.store.user_last_seen_monthly_active(user_id) if timestamp: return - is_trial = yield self.store.is_trial_user(user_id) + is_trial = await self.store.is_trial_user(user_id) if is_trial: return elif threepid: @@ -93,7 +90,7 @@ class AuthBlocking(object): # allow registration. Support users are excluded from MAU checks. return # Else if there is no room in the MAU bucket, bail - current_mau = yield self.store.get_monthly_active_count() + current_mau = await self.store.get_monthly_active_count() if current_mau >= self._max_mau_value: raise ResourceLimitError( 403, diff --git a/synapse/api/constants.py b/synapse/api/constants.py
index 6a6d32c302..46013cde15 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py
@@ -28,7 +28,7 @@ MAX_ALIAS_LENGTH = 255 MAX_USERID_LENGTH = 255 -class Membership(object): +class Membership: """Represents the membership states of a user in a room.""" @@ -40,7 +40,7 @@ class Membership(object): LIST = (INVITE, JOIN, KNOCK, LEAVE, BAN) -class PresenceState(object): +class PresenceState: """Represents the presence state of a user.""" OFFLINE = "offline" @@ -48,14 +48,14 @@ class PresenceState(object): ONLINE = "online" -class JoinRules(object): +class JoinRules: PUBLIC = "public" KNOCK = "knock" INVITE = "invite" PRIVATE = "private" -class LoginType(object): +class LoginType: PASSWORD = "m.login.password" EMAIL_IDENTITY = "m.login.email.identity" MSISDN = "m.login.msisdn" @@ -65,7 +65,7 @@ class LoginType(object): DUMMY = "m.login.dummy" -class EventTypes(object): +class EventTypes: Member = "m.room.member" Create = "m.room.create" Tombstone = "m.room.tombstone" @@ -96,17 +96,17 @@ class EventTypes(object): Presence = "m.presence" -class RejectedReason(object): +class RejectedReason: AUTH_ERROR = "auth_error" -class RoomCreationPreset(object): +class RoomCreationPreset: PRIVATE_CHAT = "private_chat" PUBLIC_CHAT = "public_chat" TRUSTED_PRIVATE_CHAT = "trusted_private_chat" -class ThirdPartyEntityKind(object): +class ThirdPartyEntityKind: USER = "user" LOCATION = "location" @@ -115,7 +115,7 @@ ServerNoticeMsgType = "m.server_notice" ServerNoticeLimitReached = "m.server_notice.usage_limit_reached" -class UserTypes(object): +class UserTypes: """Allows for user type specific behaviour. With the benefit of hindsight 'admin' and 'guest' users should also be UserTypes. Normal users are type None """ @@ -125,7 +125,7 @@ class UserTypes(object): ALL_USER_TYPES = (SUPPORT, BOT) -class RelationTypes(object): +class RelationTypes: """The types of relations known to this server. """ @@ -134,14 +134,14 @@ class RelationTypes(object): REFERENCE = "m.reference" -class LimitBlockingTypes(object): +class LimitBlockingTypes: """Reasons that a server may be blocked""" MONTHLY_ACTIVE_USER = "monthly_active_user" HS_DISABLED = "hs_disabled" -class EventContentFields(object): +class EventContentFields: """Fields found in events' content, regardless of type.""" # Labels for the event, cf https://github.com/matrix-org/matrix-doc/pull/2326 @@ -152,6 +152,6 @@ class EventContentFields(object): SELF_DESTRUCT_AFTER = "org.matrix.self_destruct_after" -class RoomEncryptionAlgorithms(object): +class RoomEncryptionAlgorithms: MEGOLM_V1_AES_SHA2 = "m.megolm.v1.aes-sha2" DEFAULT = MEGOLM_V1_AES_SHA2 diff --git a/synapse/api/errors.py b/synapse/api/errors.py
index fcfdf8bd2b..90bb01f414 100644 --- a/synapse/api/errors.py +++ b/synapse/api/errors.py
@@ -22,17 +22,17 @@ import typing from http import HTTPStatus from typing import Dict, List, Optional, Union -from canonicaljson import json - from twisted.web import http +from synapse.util import json_decoder + if typing.TYPE_CHECKING: from synapse.types import JsonDict logger = logging.getLogger(__name__) -class Codes(object): +class Codes: UNRECOGNIZED = "M_UNRECOGNIZED" UNAUTHORIZED = "M_UNAUTHORIZED" FORBIDDEN = "M_FORBIDDEN" @@ -88,7 +88,7 @@ class CodeMessageException(RuntimeError): """ def __init__(self, code: Union[int, HTTPStatus], msg: str): - super(CodeMessageException, self).__init__("%d: %s" % (code, msg)) + super().__init__("%d: %s" % (code, msg)) # Some calls to this method pass instances of http.HTTPStatus for `code`. # While HTTPStatus is a subclass of int, it has magic __str__ methods @@ -139,7 +139,7 @@ class SynapseError(CodeMessageException): msg: The human-readable error message. errcode: The matrix error code e.g 'M_FORBIDDEN' """ - super(SynapseError, self).__init__(code, msg) + super().__init__(code, msg) self.errcode = errcode def error_dict(self): @@ -160,7 +160,7 @@ class ProxiedRequestError(SynapseError): errcode: str = Codes.UNKNOWN, additional_fields: Optional[Dict] = None, ): - super(ProxiedRequestError, self).__init__(code, msg, errcode) + super().__init__(code, msg, errcode) if additional_fields is None: self._additional_fields = {} # type: Dict else: @@ -182,7 +182,7 @@ class ConsentNotGivenError(SynapseError): msg: The human-readable error message consent_url: The URL where the user can give their consent """ - super(ConsentNotGivenError, self).__init__( + super().__init__( code=HTTPStatus.FORBIDDEN, msg=msg, errcode=Codes.CONSENT_NOT_GIVEN ) self._consent_uri = consent_uri @@ -202,7 +202,7 @@ class UserDeactivatedError(SynapseError): Args: msg: The human-readable error message """ - super(UserDeactivatedError, self).__init__( + super().__init__( code=HTTPStatus.FORBIDDEN, msg=msg, errcode=Codes.USER_DEACTIVATED ) @@ -226,7 +226,7 @@ class FederationDeniedError(SynapseError): self.destination = destination - super(FederationDeniedError, self).__init__( + super().__init__( code=403, msg="Federation denied with %s." % (self.destination,), errcode=Codes.FORBIDDEN, @@ -239,14 +239,14 @@ class InteractiveAuthIncompleteError(Exception): (This indicates we should return a 401 with 'result' as the body) Attributes: + session_id: The ID of the ongoing interactive auth session. result: the server response to the request, which should be passed back to the client """ - def __init__(self, result: "JsonDict"): - super(InteractiveAuthIncompleteError, self).__init__( - "Interactive auth not yet complete" - ) + def __init__(self, session_id: str, result: "JsonDict"): + super().__init__("Interactive auth not yet complete") + self.session_id = session_id self.result = result @@ -260,14 +260,14 @@ class UnrecognizedRequestError(SynapseError): message = "Unrecognized request" else: message = args[0] - super(UnrecognizedRequestError, self).__init__(400, message, **kwargs) + super().__init__(400, message, **kwargs) class NotFoundError(SynapseError): """An error indicating we can't find the thing you asked for""" def __init__(self, msg: str = "Not found", errcode: str = Codes.NOT_FOUND): - super(NotFoundError, self).__init__(404, msg, errcode=errcode) + super().__init__(404, msg, errcode=errcode) class AuthError(SynapseError): @@ -278,7 +278,7 @@ class AuthError(SynapseError): def __init__(self, *args, **kwargs): if "errcode" not in kwargs: kwargs["errcode"] = Codes.FORBIDDEN - super(AuthError, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) class InvalidClientCredentialsError(SynapseError): @@ -334,7 +334,7 @@ class ResourceLimitError(SynapseError): ): self.admin_contact = admin_contact self.limit_type = limit_type - super(ResourceLimitError, self).__init__(code, msg, errcode=errcode) + super().__init__(code, msg, errcode=errcode) def error_dict(self): return cs_error( @@ -351,7 +351,7 @@ class EventSizeError(SynapseError): def __init__(self, *args, **kwargs): if "errcode" not in kwargs: kwargs["errcode"] = Codes.TOO_LARGE - super(EventSizeError, self).__init__(413, *args, **kwargs) + super().__init__(413, *args, **kwargs) class EventStreamError(SynapseError): @@ -360,7 +360,7 @@ class EventStreamError(SynapseError): def __init__(self, *args, **kwargs): if "errcode" not in kwargs: kwargs["errcode"] = Codes.BAD_PAGINATION - super(EventStreamError, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) class LoginError(SynapseError): @@ -383,7 +383,7 @@ class InvalidCaptchaError(SynapseError): error_url: Optional[str] = None, errcode: str = Codes.CAPTCHA_INVALID, ): - super(InvalidCaptchaError, self).__init__(code, msg, errcode) + super().__init__(code, msg, errcode) self.error_url = error_url def error_dict(self): @@ -401,7 +401,7 @@ class LimitExceededError(SynapseError): retry_after_ms: Optional[int] = None, errcode: str = Codes.LIMIT_EXCEEDED, ): - super(LimitExceededError, self).__init__(code, msg, errcode) + super().__init__(code, msg, errcode) self.retry_after_ms = retry_after_ms def error_dict(self): @@ -417,9 +417,7 @@ class RoomKeysVersionError(SynapseError): Args: current_version: the current version of the store they should have used """ - super(RoomKeysVersionError, self).__init__( - 403, "Wrong room_keys version", Codes.WRONG_ROOM_KEYS_VERSION - ) + super().__init__(403, "Wrong room_keys version", Codes.WRONG_ROOM_KEYS_VERSION) self.current_version = current_version @@ -428,7 +426,7 @@ class UnsupportedRoomVersionError(SynapseError): not support.""" def __init__(self, msg: str = "Homeserver does not support this room version"): - super(UnsupportedRoomVersionError, self).__init__( + super().__init__( code=400, msg=msg, errcode=Codes.UNSUPPORTED_ROOM_VERSION, ) @@ -439,7 +437,7 @@ class ThreepidValidationError(SynapseError): def __init__(self, *args, **kwargs): if "errcode" not in kwargs: kwargs["errcode"] = Codes.FORBIDDEN - super(ThreepidValidationError, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) class IncompatibleRoomVersionError(SynapseError): @@ -450,7 +448,7 @@ class IncompatibleRoomVersionError(SynapseError): """ def __init__(self, room_version: str): - super(IncompatibleRoomVersionError, self).__init__( + super().__init__( code=400, msg="Your homeserver does not support the features required to " "join this room", @@ -472,7 +470,7 @@ class PasswordRefusedError(SynapseError): msg: str = "This password doesn't comply with the server's policy", errcode: str = Codes.WEAK_PASSWORD, ): - super(PasswordRefusedError, self).__init__( + super().__init__( code=400, msg=msg, errcode=errcode, ) @@ -487,7 +485,7 @@ class RequestSendFailed(RuntimeError): """ def __init__(self, inner_exception, can_retry): - super(RequestSendFailed, self).__init__( + super().__init__( "Failed to send request: %s: %s" % (type(inner_exception).__name__, inner_exception) ) @@ -541,7 +539,7 @@ class FederationError(RuntimeError): self.source = source msg = "%s %s: %s" % (level, code, reason) - super(FederationError, self).__init__(msg) + super().__init__(msg) def get_dict(self): return { @@ -569,7 +567,7 @@ class HttpResponseException(CodeMessageException): msg: reason phrase from HTTP response status line response: body of response """ - super(HttpResponseException, self).__init__(code, msg) + super().__init__(code, msg) self.response = response def to_synapse_error(self): @@ -592,7 +590,7 @@ class HttpResponseException(CodeMessageException): # try to parse the body as json, to get better errcode/msg, but # default to M_UNKNOWN with the HTTP status as the error text try: - j = json.loads(self.response.decode("utf-8")) + j = json_decoder.decode(self.response.decode("utf-8")) except ValueError: j = {} @@ -603,3 +601,11 @@ class HttpResponseException(CodeMessageException): errmsg = j.pop("error", self.msg) return ProxiedRequestError(self.code, errmsg, errcode, j) + + +class ShadowBanError(Exception): + """ + Raised when a shadow-banned user attempts to perform an action. + + This should be caught and a proper "fake" success response sent to the user. + """ diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py
index f988f62a1e..5caf336fd0 100644 --- a/synapse/api/filtering.py +++ b/synapse/api/filtering.py
@@ -15,17 +15,15 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. +import json from typing import List import jsonschema -from canonicaljson import json from jsonschema import FormatChecker -from twisted.internet import defer - from synapse.api.constants import EventContentFields from synapse.api.errors import SynapseError -from synapse.storage.presence import UserPresenceState +from synapse.api.presence import UserPresenceState from synapse.types import RoomID, UserID FILTER_SCHEMA = { @@ -132,14 +130,13 @@ def matrix_user_id_validator(user_id_str): return UserID.from_string(user_id_str) -class Filtering(object): +class Filtering: def __init__(self, hs): - super(Filtering, self).__init__() + super().__init__() self.store = hs.get_datastore() - @defer.inlineCallbacks - def get_user_filter(self, user_localpart, filter_id): - result = yield self.store.get_user_filter(user_localpart, filter_id) + async def get_user_filter(self, user_localpart, filter_id): + result = await self.store.get_user_filter(user_localpart, filter_id) return FilterCollection(result) def add_user_filter(self, user_localpart, user_filter): @@ -171,7 +168,7 @@ class Filtering(object): raise SynapseError(400, str(e)) -class FilterCollection(object): +class FilterCollection: def __init__(self, filter_json): self._filter_json = filter_json @@ -252,7 +249,7 @@ class FilterCollection(object): ) -class Filter(object): +class Filter: def __init__(self, filter_json): self.filter_json = filter_json diff --git a/synapse/storage/presence.py b/synapse/api/presence.py
index 18a462f0ee..18a462f0ee 100644 --- a/synapse/storage/presence.py +++ b/synapse/api/presence.py
diff --git a/synapse/api/ratelimiting.py b/synapse/api/ratelimiting.py
index ec6b3a69a2..5d9d5a228f 100644 --- a/synapse/api/ratelimiting.py +++ b/synapse/api/ratelimiting.py
@@ -17,10 +17,11 @@ from collections import OrderedDict from typing import Any, Optional, Tuple from synapse.api.errors import LimitExceededError +from synapse.types import Requester from synapse.util import Clock -class Ratelimiter(object): +class Ratelimiter: """ Ratelimit actions marked by arbitrary keys. @@ -43,6 +44,42 @@ class Ratelimiter(object): # * The rate_hz of this particular entry. This can vary per request self.actions = OrderedDict() # type: OrderedDict[Any, Tuple[float, int, float]] + def can_requester_do_action( + self, + requester: Requester, + rate_hz: Optional[float] = None, + burst_count: Optional[int] = None, + update: bool = True, + _time_now_s: Optional[int] = None, + ) -> Tuple[bool, float]: + """Can the requester perform the action? + + Args: + requester: The requester to key off when rate limiting. The user property + will be used. + rate_hz: The long term number of actions that can be performed in a second. + Overrides the value set during instantiation if set. + burst_count: How many actions that can be performed before being limited. + Overrides the value set during instantiation if set. + update: Whether to count this check as performing the action + _time_now_s: The current time. Optional, defaults to the current time according + to self.clock. Only used by tests. + + Returns: + A tuple containing: + * A bool indicating if they can perform the action now + * The reactor timestamp for when the action can be performed next. + -1 if rate_hz is less than or equal to zero + """ + # Disable rate limiting of users belonging to any AS that is configured + # not to be rate limited in its registration file (rate_limited: true|false). + if requester.app_service and not requester.app_service.is_rate_limited(): + return True, -1.0 + + return self.can_do_action( + requester.user.to_string(), rate_hz, burst_count, update, _time_now_s + ) + def can_do_action( self, key: Any, diff --git a/synapse/api/room_versions.py b/synapse/api/room_versions.py
index d7baf2bc39..f3ecbf36b6 100644 --- a/synapse/api/room_versions.py +++ b/synapse/api/room_versions.py
@@ -18,7 +18,7 @@ from typing import Dict import attr -class EventFormatVersions(object): +class EventFormatVersions: """This is an internal enum for tracking the version of the event format, independently from the room version. """ @@ -35,20 +35,20 @@ KNOWN_EVENT_FORMAT_VERSIONS = { } -class StateResolutionVersions(object): +class StateResolutionVersions: """Enum to identify the state resolution algorithms""" V1 = 1 # room v1 state res V2 = 2 # MSC1442 state res: room v2 and later -class RoomDisposition(object): +class RoomDisposition: STABLE = "stable" UNSTABLE = "unstable" @attr.s(slots=True, frozen=True) -class RoomVersion(object): +class RoomVersion: """An object which describes the unique attributes of a room version.""" identifier = attr.ib() # str; the identifier for this version @@ -69,7 +69,7 @@ class RoomVersion(object): limit_notifications_power_levels = attr.ib(type=bool) -class RoomVersions(object): +class RoomVersions: V1 = RoomVersion( "1", RoomDisposition.STABLE, diff --git a/synapse/api/urls.py b/synapse/api/urls.py
index bd03ebca5a..6379c86dde 100644 --- a/synapse/api/urls.py +++ b/synapse/api/urls.py
@@ -21,6 +21,7 @@ from urllib.parse import urlencode from synapse.config import ConfigError +SYNAPSE_CLIENT_API_PREFIX = "/_synapse/client" CLIENT_API_PREFIX = "/_matrix/client" FEDERATION_PREFIX = "/_matrix/federation" FEDERATION_V1_PREFIX = FEDERATION_PREFIX + "/v1" @@ -33,7 +34,7 @@ MEDIA_PREFIX = "/_matrix/media/r0" LEGACY_MEDIA_PREFIX = "/_matrix/media/v1" -class ConsentURIBuilder(object): +class ConsentURIBuilder: def __init__(self, hs_config): """ Args: diff --git a/synapse/app/_base.py b/synapse/app/_base.py
index 373a80a4a7..fb476ddaf5 100644 --- a/synapse/app/_base.py +++ b/synapse/app/_base.py
@@ -12,7 +12,6 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import gc import logging import os @@ -22,7 +21,6 @@ import sys import traceback from typing import Iterable -from daemonize import Daemonize from typing_extensions import NoReturn from twisted.internet import defer, error, reactor @@ -34,6 +32,7 @@ from synapse.config.server import ListenerConfig from synapse.crypto import context_factory from synapse.logging.context import PreserveLoggingContext from synapse.util.async_helpers import Linearizer +from synapse.util.daemonize import daemonize_process from synapse.util.rlimit import change_resource_limit from synapse.util.versionstring import get_version_string @@ -129,17 +128,8 @@ def start_reactor( if print_pidfile: print(pid_file) - daemon = Daemonize( - app=appname, - pid=pid_file, - action=run, - auto_close_fds=False, - verbose=True, - logger=logger, - ) - daemon.start() - else: - run() + daemonize_process(pid_file, logger) + run() def quit_with_error(error_string: str) -> NoReturn: @@ -278,7 +268,7 @@ def start(hs: "synapse.server.HomeServer", listeners: Iterable[ListenerConfig]): # It is now safe to start your Synapse. hs.start_listening(listeners) - hs.get_datastore().db.start_profiling() + hs.get_datastore().db_pool.start_profiling() hs.get_pusherpool().start() setup_sentry(hs) @@ -344,6 +334,13 @@ def install_dns_limiter(reactor, max_dns_requests_in_flight=100): This is to workaround https://twistedmatrix.com/trac/ticket/9620, where we can run out of file descriptors and infinite loop if we attempt to do too many DNS queries at once + + XXX: I'm confused by this. reactor.nameResolver does not use twisted.names unless + you explicitly install twisted.names as the resolver; rather it uses a GAIResolver + backed by the reactor's default threadpool (which is limited to 10 threads). So + (a) I don't understand why twisted ticket 9620 is relevant, and (b) I don't + understand why we would run out of FDs if we did too many lookups at once. + -- richvdh 2020/08/29 """ new_resolver = _LimitedHostnameResolver( reactor.nameResolver, max_dns_requests_in_flight @@ -352,7 +349,7 @@ def install_dns_limiter(reactor, max_dns_requests_in_flight=100): reactor.installNameResolver(new_resolver) -class _LimitedHostnameResolver(object): +class _LimitedHostnameResolver: """Wraps a IHostnameResolver, limiting the number of in-flight DNS lookups. """ @@ -412,7 +409,7 @@ class _LimitedHostnameResolver(object): yield deferred -class _DeferredResolutionReceiver(object): +class _DeferredResolutionReceiver: """Wraps a IResolutionReceiver and simply resolves the given deferred when resolution is complete """ diff --git a/synapse/app/admin_cmd.py b/synapse/app/admin_cmd.py
index a37818fe9a..7d309b1bb0 100644 --- a/synapse/app/admin_cmd.py +++ b/synapse/app/admin_cmd.py
@@ -14,13 +14,12 @@ # See the License for the specific language governing permissions and # limitations under the License. import argparse +import json import logging import os import sys import tempfile -from canonicaljson import json - from twisted.internet import defer, task import synapse @@ -79,8 +78,7 @@ class AdminCmdServer(HomeServer): pass -@defer.inlineCallbacks -def export_data_command(hs, args): +async def export_data_command(hs, args): """Export data for a user. Args: @@ -91,10 +89,8 @@ def export_data_command(hs, args): user_id = args.user_id directory = args.output_directory - res = yield defer.ensureDeferred( - hs.get_handlers().admin_handler.export_user_data( - user_id, FileExfiltrationWriter(user_id, directory=directory) - ) + res = await hs.get_handlers().admin_handler.export_user_data( + user_id, FileExfiltrationWriter(user_id, directory=directory) ) print(res) @@ -232,14 +228,15 @@ def start(config_options): # We also make sure that `_base.start` gets run before we actually run the # command. - @defer.inlineCallbacks - def run(_reactor): + async def run(): with LoggingContext("command"): - yield _base.start(ss, []) - yield args.func(ss, args) + _base.start(ss, []) + await args.func(ss, args) _base.start_worker_reactor( - "synapse-admin-cmd", config, run_command=lambda: task.react(run) + "synapse-admin-cmd", + config, + run_command=lambda: task.react(lambda _reactor: defer.ensureDeferred(run())), ) diff --git a/synapse/app/generic_worker.py b/synapse/app/generic_worker.py
index c1ab8378e1..15312ef961 100644 --- a/synapse/app/generic_worker.py +++ b/synapse/app/generic_worker.py
@@ -123,17 +123,18 @@ from synapse.rest.client.v2_alpha.account_data import ( from synapse.rest.client.v2_alpha.keys import KeyChangesServlet, KeyQueryServlet from synapse.rest.client.v2_alpha.register import RegisterRestServlet from synapse.rest.client.versions import VersionsRestServlet +from synapse.rest.health import HealthResource from synapse.rest.key.v2 import KeyApiV2Resource -from synapse.server import HomeServer -from synapse.storage.data_stores.main.censor_events import CensorEventsStore -from synapse.storage.data_stores.main.media_repository import MediaRepositoryStore -from synapse.storage.data_stores.main.monthly_active_users import ( +from synapse.server import HomeServer, cache_in_self +from synapse.storage.databases.main.censor_events import CensorEventsStore +from synapse.storage.databases.main.media_repository import MediaRepositoryStore +from synapse.storage.databases.main.monthly_active_users import ( MonthlyActiveUsersWorkerStore, ) -from synapse.storage.data_stores.main.presence import UserPresenceState -from synapse.storage.data_stores.main.search import SearchWorkerStore -from synapse.storage.data_stores.main.ui_auth import UIAuthWorkerStore -from synapse.storage.data_stores.main.user_directory import UserDirectoryStore +from synapse.storage.databases.main.presence import UserPresenceState +from synapse.storage.databases.main.search import SearchWorkerStore +from synapse.storage.databases.main.ui_auth import UIAuthWorkerStore +from synapse.storage.databases.main.user_directory import UserDirectoryStore from synapse.types import ReadReceipt from synapse.util.async_helpers import Linearizer from synapse.util.httpresourcetree import create_resource_tree @@ -151,7 +152,7 @@ class PresenceStatusStubServlet(RestServlet): PATTERNS = client_patterns("/presence/(?P<user_id>[^/]*)/status") def __init__(self, hs): - super(PresenceStatusStubServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() async def on_GET(self, request, user_id): @@ -175,7 +176,7 @@ class KeyUploadServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(KeyUploadServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.http_client = hs.get_simple_http_client() @@ -493,7 +494,10 @@ class GenericWorkerServer(HomeServer): site_tag = listener_config.http_options.tag if site_tag is None: site_tag = port - resources = {} + + # We always include a health resource. + resources = {"/health": HealthResource()} + for res in listener_config.http_options.resources: for name in res.names: if name == "metrics": @@ -628,19 +632,21 @@ class GenericWorkerServer(HomeServer): self.get_tcp_replication().start_replication(self) - def remove_pusher(self, app_id, push_key, user_id): + async def remove_pusher(self, app_id, push_key, user_id): self.get_tcp_replication().send_remove_pusher(app_id, push_key, user_id) - def build_replication_data_handler(self): + @cache_in_self + def get_replication_data_handler(self): return GenericWorkerReplicationHandler(self) - def build_presence_handler(self): + @cache_in_self + def get_presence_handler(self): return GenericWorkerPresence(self) class GenericWorkerReplicationHandler(ReplicationDataHandler): def __init__(self, hs): - super(GenericWorkerReplicationHandler, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.presence_handler = hs.get_presence_handler() # type: GenericWorkerPresence @@ -739,7 +745,7 @@ class GenericWorkerReplicationHandler(ReplicationDataHandler): self.send_handler.wake_destination(server) -class FederationSenderHandler(object): +class FederationSenderHandler: """Processes the fedration replication stream This class is only instantiate on the worker responsible for sending outbound diff --git a/synapse/app/homeserver.py b/synapse/app/homeserver.py
index ec7401f911..dff739e106 100644 --- a/synapse/app/homeserver.py +++ b/synapse/app/homeserver.py
@@ -15,8 +15,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - import gc import logging import math @@ -48,6 +46,7 @@ from synapse.api.urls import ( from synapse.app import _base from synapse.app._base import listen_ssl, listen_tcp, quit_with_error from synapse.config._base import ConfigError +from synapse.config.emailconfig import ThreepidBehaviour from synapse.config.homeserver import HomeServerConfig from synapse.config.server import ListenerConfig from synapse.federation.transport.server import TransportLayerServer @@ -68,6 +67,7 @@ from synapse.replication.http import REPLICATION_PREFIX, ReplicationRestResource from synapse.replication.tcp.resource import ReplicationStreamProtocolFactory from synapse.rest import ClientRestResource from synapse.rest.admin import AdminRestResource +from synapse.rest.health import HealthResource from synapse.rest.key.v2 import KeyApiV2Resource from synapse.rest.well_known import WellKnownResource from synapse.server import HomeServer @@ -98,7 +98,9 @@ class SynapseHomeServer(HomeServer): if site_tag is None: site_tag = port - resources = {} + # We always include a health resource. + resources = {"/health": HealthResource()} + for res in listener_config.http_options.resources: for name in res.names: if name == "openid" and "federation" in res.names: @@ -206,6 +208,15 @@ class SynapseHomeServer(HomeServer): resources["/_matrix/saml2"] = SAML2Resource(self) + if self.get_config().threepid_behaviour_email == ThreepidBehaviour.LOCAL: + from synapse.rest.synapse.client.password_reset import ( + PasswordResetSubmitTokenResource, + ) + + resources[ + "/_synapse/client/password_reset/email/submit_token" + ] = PasswordResetSubmitTokenResource(self) + if name == "consent": from synapse.rest.consent.consent_resource import ConsentResource @@ -380,13 +391,12 @@ def setup(config_options): hs.setup_master() - @defer.inlineCallbacks - def do_acme(): + async def do_acme() -> bool: """ Reprovision an ACME certificate, if it's required. Returns: - Deferred[bool]: Whether the cert has been updated. + Whether the cert has been updated. """ acme = hs.get_acme_handler() @@ -405,30 +415,28 @@ def setup(config_options): provision = True if provision: - yield acme.provision_certificate() + await acme.provision_certificate() return provision - @defer.inlineCallbacks - def reprovision_acme(): + async def reprovision_acme(): """ Provision a certificate from ACME, if required, and reload the TLS certificate if it's renewed. """ - reprovisioned = yield do_acme() + reprovisioned = await do_acme() if reprovisioned: _base.refresh_certificate(hs) - @defer.inlineCallbacks - def start(): + async def start(): try: # Run the ACME provisioning code, if it's enabled. if hs.config.acme_enabled: acme = hs.get_acme_handler() # Start up the webservices which we will respond to ACME # challenges with, and then provision. - yield acme.start_listening() - yield do_acme() + await acme.start_listening() + await do_acme() # Check if it needs to be reprovisioned every day. hs.get_clock().looping_call(reprovision_acme, 24 * 60 * 60 * 1000) @@ -437,12 +445,12 @@ def setup(config_options): if hs.config.oidc_enabled: oidc = hs.get_oidc_handler() # Loading the provider metadata also ensures the provider config is valid. - yield defer.ensureDeferred(oidc.load_metadata()) - yield defer.ensureDeferred(oidc.load_jwks()) + await oidc.load_metadata() + await oidc.load_jwks() _base.start(hs, config.listeners) - hs.get_datastore().db.updates.start_doing_background_updates() + hs.get_datastore().db_pool.updates.start_doing_background_updates() except Exception: # Print the exception and bail out. print("Error during startup:", file=sys.stderr) @@ -454,7 +462,7 @@ def setup(config_options): reactor.stop() sys.exit(1) - reactor.callWhenRunning(start) + reactor.callWhenRunning(lambda: defer.ensureDeferred(start())) return hs @@ -552,8 +560,8 @@ async def phone_stats_home(hs, stats, stats_process=_stats_process): # # This only reports info about the *main* database. - stats["database_engine"] = hs.get_datastore().db.engine.module.__name__ - stats["database_server_version"] = hs.get_datastore().db.engine.server_version + stats["database_engine"] = hs.get_datastore().db_pool.engine.module.__name__ + stats["database_server_version"] = hs.get_datastore().db_pool.engine.server_version logger.info("Reporting stats to %s: %s" % (hs.config.report_stats_endpoint, stats)) try: diff --git a/synapse/appservice/__init__.py b/synapse/appservice/__init__.py
index 0323256472..13ec1f71a6 100644 --- a/synapse/appservice/__init__.py +++ b/synapse/appservice/__init__.py
@@ -14,22 +14,25 @@ # limitations under the License. import logging import re - -from twisted.internet import defer +from typing import TYPE_CHECKING from synapse.api.constants import EventTypes +from synapse.appservice.api import ApplicationServiceApi from synapse.types import GroupID, get_domain_from_id -from synapse.util.caches.descriptors import cachedInlineCallbacks +from synapse.util.caches.descriptors import cached + +if TYPE_CHECKING: + from synapse.storage.databases.main import DataStore logger = logging.getLogger(__name__) -class ApplicationServiceState(object): +class ApplicationServiceState: DOWN = "down" UP = "up" -class AppServiceTransaction(object): +class AppServiceTransaction: """Represents an application service transaction.""" def __init__(self, service, id, events): @@ -37,19 +40,19 @@ class AppServiceTransaction(object): self.id = id self.events = events - def send(self, as_api): + async def send(self, as_api: ApplicationServiceApi) -> bool: """Sends this transaction using the provided AS API interface. Args: - as_api(ApplicationServiceApi): The API to use to send. + as_api: The API to use to send. Returns: - A Deferred which resolves to True if the transaction was sent. + True if the transaction was sent. """ - return as_api.push_bulk( + return await as_api.push_bulk( service=self.service, events=self.events, txn_id=self.id ) - def complete(self, store): + async def complete(self, store: "DataStore") -> None: """Completes this transaction as successful. Marks this transaction ID on the application service and removes the @@ -57,13 +60,11 @@ class AppServiceTransaction(object): Args: store: The database store to operate on. - Returns: - A Deferred which resolves to True if the transaction was completed. """ - return store.complete_appservice_txn(service=self.service, txn_id=self.id) + await store.complete_appservice_txn(service=self.service, txn_id=self.id) -class ApplicationService(object): +class ApplicationService: """Defines an application service. This definition is mostly what is provided to the /register AS API. @@ -172,8 +173,7 @@ class ApplicationService(object): return regex_obj["exclusive"] return False - @defer.inlineCallbacks - def _matches_user(self, event, store): + async def _matches_user(self, event, store): if not event: return False @@ -188,12 +188,12 @@ class ApplicationService(object): if not store: return False - does_match = yield self._matches_user_in_member_list(event.room_id, store) + does_match = await self._matches_user_in_member_list(event.room_id, store) return does_match - @cachedInlineCallbacks(num_args=1, cache_context=True) - def _matches_user_in_member_list(self, room_id, store, cache_context): - member_list = yield store.get_users_in_room( + @cached(num_args=1, cache_context=True) + async def _matches_user_in_member_list(self, room_id, store, cache_context): + member_list = await store.get_users_in_room( room_id, on_invalidate=cache_context.invalidate ) @@ -208,35 +208,33 @@ class ApplicationService(object): return self.is_interested_in_room(event.room_id) return False - @defer.inlineCallbacks - def _matches_aliases(self, event, store): + async def _matches_aliases(self, event, store): if not store or not event: return False - alias_list = yield store.get_aliases_for_room(event.room_id) + alias_list = await store.get_aliases_for_room(event.room_id) for alias in alias_list: if self.is_interested_in_alias(alias): return True return False - @defer.inlineCallbacks - def is_interested(self, event, store=None): + async def is_interested(self, event, store=None) -> bool: """Check if this service is interested in this event. Args: event(Event): The event to check. store(DataStore) Returns: - bool: True if this service would like to know about this event. + True if this service would like to know about this event. """ # Do cheap checks first if self._matches_room_id(event): return True - if (yield self._matches_aliases(event, store)): + if await self._matches_aliases(event, store): return True - if (yield self._matches_user(event, store)): + if await self._matches_user(event, store): return True return False diff --git a/synapse/appservice/api.py b/synapse/appservice/api.py
index 1e0e4d497d..c526c28b93 100644 --- a/synapse/appservice/api.py +++ b/synapse/appservice/api.py
@@ -14,18 +14,20 @@ # limitations under the License. import logging import urllib +from typing import TYPE_CHECKING, Optional from prometheus_client import Counter -from twisted.internet import defer - from synapse.api.constants import EventTypes, ThirdPartyEntityKind from synapse.api.errors import CodeMessageException from synapse.events.utils import serialize_event from synapse.http.client import SimpleHttpClient -from synapse.types import ThirdPartyInstanceID +from synapse.types import JsonDict, ThirdPartyInstanceID from synapse.util.caches.response_cache import ResponseCache +if TYPE_CHECKING: + from synapse.appservice import ApplicationService + logger = logging.getLogger(__name__) sent_transactions_counter = Counter( @@ -86,20 +88,19 @@ class ApplicationServiceApi(SimpleHttpClient): """ def __init__(self, hs): - super(ApplicationServiceApi, self).__init__(hs) + super().__init__(hs) self.clock = hs.get_clock() self.protocol_meta_cache = ResponseCache( hs, "as_protocol_meta", timeout_ms=HOUR_IN_MS ) - @defer.inlineCallbacks - def query_user(self, service, user_id): + async def query_user(self, service, user_id): if service.url is None: return False uri = service.url + ("/users/%s" % urllib.parse.quote(user_id)) try: - response = yield self.get_json(uri, {"access_token": service.hs_token}) + response = await self.get_json(uri, {"access_token": service.hs_token}) if response is not None: # just an empty json object return True except CodeMessageException as e: @@ -110,14 +111,12 @@ class ApplicationServiceApi(SimpleHttpClient): logger.warning("query_user to %s threw exception %s", uri, ex) return False - @defer.inlineCallbacks - def query_alias(self, service, alias): + async def query_alias(self, service, alias): if service.url is None: return False uri = service.url + ("/rooms/%s" % urllib.parse.quote(alias)) - response = None try: - response = yield self.get_json(uri, {"access_token": service.hs_token}) + response = await self.get_json(uri, {"access_token": service.hs_token}) if response is not None: # just an empty json object return True except CodeMessageException as e: @@ -128,8 +127,7 @@ class ApplicationServiceApi(SimpleHttpClient): logger.warning("query_alias to %s threw exception %s", uri, ex) return False - @defer.inlineCallbacks - def query_3pe(self, service, kind, protocol, fields): + async def query_3pe(self, service, kind, protocol, fields): if kind == ThirdPartyEntityKind.USER: required_field = "userid" elif kind == ThirdPartyEntityKind.LOCATION: @@ -146,7 +144,7 @@ class ApplicationServiceApi(SimpleHttpClient): urllib.parse.quote(protocol), ) try: - response = yield self.get_json(uri, fields) + response = await self.get_json(uri, fields) if not isinstance(response, list): logger.warning( "query_3pe to %s returned an invalid response %r", uri, response @@ -167,19 +165,20 @@ class ApplicationServiceApi(SimpleHttpClient): logger.warning("query_3pe to %s threw exception %s", uri, ex) return [] - def get_3pe_protocol(self, service, protocol): + async def get_3pe_protocol( + self, service: "ApplicationService", protocol: str + ) -> Optional[JsonDict]: if service.url is None: return {} - @defer.inlineCallbacks - def _get(): + async def _get() -> Optional[JsonDict]: uri = "%s%s/thirdparty/protocol/%s" % ( service.url, APP_SERVICE_PREFIX, urllib.parse.quote(protocol), ) try: - info = yield self.get_json(uri, {}) + info = await self.get_json(uri) if not _is_valid_3pe_metadata(info): logger.warning( @@ -200,10 +199,9 @@ class ApplicationServiceApi(SimpleHttpClient): return None key = (service.id, protocol) - return self.protocol_meta_cache.wrap(key, _get) + return await self.protocol_meta_cache.wrap(key, _get) - @defer.inlineCallbacks - def push_bulk(self, service, events, txn_id=None): + async def push_bulk(self, service, events, txn_id=None): if service.url is None: return True @@ -218,7 +216,7 @@ class ApplicationServiceApi(SimpleHttpClient): uri = service.url + ("/transactions/%s" % urllib.parse.quote(txn_id)) try: - yield self.put_json( + await self.put_json( uri=uri, json_body={"events": events}, args={"access_token": service.hs_token}, diff --git a/synapse/appservice/scheduler.py b/synapse/appservice/scheduler.py
index 9998f822f1..8eb8c6f51c 100644 --- a/synapse/appservice/scheduler.py +++ b/synapse/appservice/scheduler.py
@@ -50,8 +50,6 @@ components. """ import logging -from twisted.internet import defer - from synapse.appservice import ApplicationServiceState from synapse.logging.context import run_in_background from synapse.metrics.background_process_metrics import run_as_background_process @@ -59,7 +57,7 @@ from synapse.metrics.background_process_metrics import run_as_background_process logger = logging.getLogger(__name__) -class ApplicationServiceScheduler(object): +class ApplicationServiceScheduler: """ Public facing API for this module. Does the required DI to tie the components together. This also serves as the "event_pool", which in this case is a simple array. @@ -73,12 +71,11 @@ class ApplicationServiceScheduler(object): self.txn_ctrl = _TransactionController(self.clock, self.store, self.as_api) self.queuer = _ServiceQueuer(self.txn_ctrl, self.clock) - @defer.inlineCallbacks - def start(self): + async def start(self): logger.info("Starting appservice scheduler") # check for any DOWN ASes and start recoverers for them. - services = yield self.store.get_appservices_by_state( + services = await self.store.get_appservices_by_state( ApplicationServiceState.DOWN ) @@ -89,7 +86,7 @@ class ApplicationServiceScheduler(object): self.queuer.enqueue(service, event) -class _ServiceQueuer(object): +class _ServiceQueuer: """Queue of events waiting to be sent to appservices. Groups events into transactions per-appservice, and sends them on to the @@ -117,8 +114,7 @@ class _ServiceQueuer(object): "as-sender-%s" % (service.id,), self._send_request, service ) - @defer.inlineCallbacks - def _send_request(self, service): + async def _send_request(self, service): # sanity-check: we shouldn't get here if this service already has a sender # running. assert service.id not in self.requests_in_flight @@ -130,14 +126,14 @@ class _ServiceQueuer(object): if not events: return try: - yield self.txn_ctrl.send(service, events) + await self.txn_ctrl.send(service, events) except Exception: logger.exception("AS request failed") finally: self.requests_in_flight.discard(service.id) -class _TransactionController(object): +class _TransactionController: """Transaction manager. Builds AppServiceTransactions and runs their lifecycle. Also starts a Recoverer @@ -162,36 +158,33 @@ class _TransactionController(object): # for UTs self.RECOVERER_CLASS = _Recoverer - @defer.inlineCallbacks - def send(self, service, events): + async def send(self, service, events): try: - txn = yield self.store.create_appservice_txn(service=service, events=events) - service_is_up = yield self._is_service_up(service) + txn = await self.store.create_appservice_txn(service=service, events=events) + service_is_up = await self._is_service_up(service) if service_is_up: - sent = yield txn.send(self.as_api) + sent = await txn.send(self.as_api) if sent: - yield txn.complete(self.store) + await txn.complete(self.store) else: run_in_background(self._on_txn_fail, service) except Exception: logger.exception("Error creating appservice transaction") run_in_background(self._on_txn_fail, service) - @defer.inlineCallbacks - def on_recovered(self, recoverer): + async def on_recovered(self, recoverer): logger.info( "Successfully recovered application service AS ID %s", recoverer.service.id ) self.recoverers.pop(recoverer.service.id) logger.info("Remaining active recoverers: %s", len(self.recoverers)) - yield self.store.set_appservice_state( + await self.store.set_appservice_state( recoverer.service, ApplicationServiceState.UP ) - @defer.inlineCallbacks - def _on_txn_fail(self, service): + async def _on_txn_fail(self, service): try: - yield self.store.set_appservice_state(service, ApplicationServiceState.DOWN) + await self.store.set_appservice_state(service, ApplicationServiceState.DOWN) self.start_recoverer(service) except Exception: logger.exception("Error starting AS recoverer") @@ -211,13 +204,12 @@ class _TransactionController(object): recoverer.recover() logger.info("Now %i active recoverers", len(self.recoverers)) - @defer.inlineCallbacks - def _is_service_up(self, service): - state = yield self.store.get_appservice_state(service) + async def _is_service_up(self, service): + state = await self.store.get_appservice_state(service) return state == ApplicationServiceState.UP or state is None -class _Recoverer(object): +class _Recoverer: """Manages retries and backoff for a DOWN appservice. We have one of these for each appservice which is currently considered DOWN. @@ -254,25 +246,24 @@ class _Recoverer(object): self.backoff_counter += 1 self.recover() - @defer.inlineCallbacks - def retry(self): + async def retry(self): logger.info("Starting retries on %s", self.service.id) try: while True: - txn = yield self.store.get_oldest_unsent_txn(self.service) + txn = await self.store.get_oldest_unsent_txn(self.service) if not txn: # nothing left: we're done! - self.callback(self) + await self.callback(self) return logger.info( "Retrying transaction %s for AS ID %s", txn.id, txn.service.id ) - sent = yield txn.send(self.as_api) + sent = await txn.send(self.as_api) if not sent: break - yield txn.complete(self.store) + await txn.complete(self.store) # reset the backoff counter and then process the next transaction self.backoff_counter = 1 diff --git a/synapse/config/_base.py b/synapse/config/_base.py
index 34a2370e67..33a917a035 100644 --- a/synapse/config/_base.py +++ b/synapse/config/_base.py
@@ -18,13 +18,17 @@ import argparse import errno import os +import time +import urllib.parse from collections import OrderedDict from hashlib import sha256 from io import open as io_open from textwrap import dedent -from typing import Any, List, MutableMapping, Optional +from typing import Any, Callable, List, MutableMapping, Optional import attr +import jinja2 +import pkg_resources import yaml @@ -85,7 +89,7 @@ def path_exists(file_path): return False -class Config(object): +class Config: """ A configuration section, containing configuration keys and values. @@ -101,6 +105,11 @@ class Config(object): def __init__(self, root_config=None): self.root = root_config + # Get the path to the default Synapse template directory + self.default_template_dir = pkg_resources.resource_filename( + "synapse", "res/templates" + ) + def __getattr__(self, item: str) -> Any: """ Try and fetch a configuration option that does not exist on this class. @@ -185,8 +194,102 @@ class Config(object): with io_open(file_path, encoding="utf-8") as file_stream: return file_stream.read() + def read_templates( + self, + filenames: List[str], + custom_template_directory: Optional[str] = None, + autoescape: bool = False, + ) -> List[jinja2.Template]: + """Load a list of template files from disk using the given variables. + + This function will attempt to load the given templates from the default Synapse + template directory. If `custom_template_directory` is supplied, that directory + is tried first. + + Files read are treated as Jinja templates. These templates are not rendered yet. + + Args: + filenames: A list of template filenames to read. + + custom_template_directory: A directory to try to look for the templates + before using the default Synapse template directory instead. + + autoescape: Whether to autoescape variables before inserting them into the + template. + + Raises: + ConfigError: if the file's path is incorrect or otherwise cannot be read. + + Returns: + A list of jinja2 templates. + """ + templates = [] + search_directories = [self.default_template_dir] + + # The loader will first look in the custom template directory (if specified) for the + # given filename. If it doesn't find it, it will use the default template dir instead + if custom_template_directory: + # Check that the given template directory exists + if not self.path_exists(custom_template_directory): + raise ConfigError( + "Configured template directory does not exist: %s" + % (custom_template_directory,) + ) + + # Search the custom template directory as well + search_directories.insert(0, custom_template_directory) + + loader = jinja2.FileSystemLoader(search_directories) + env = jinja2.Environment(loader=loader, autoescape=autoescape) + + # Update the environment with our custom filters + env.filters.update({"format_ts": _format_ts_filter}) + if self.public_baseurl: + env.filters.update( + {"mxc_to_http": _create_mxc_to_http_filter(self.public_baseurl)} + ) + + for filename in filenames: + # Load the template + template = env.get_template(filename) + templates.append(template) -class RootConfig(object): + return templates + + +def _format_ts_filter(value: int, format: str): + return time.strftime(format, time.localtime(value / 1000)) + + +def _create_mxc_to_http_filter(public_baseurl: str) -> Callable: + """Create and return a jinja2 filter that converts MXC urls to HTTP + + Args: + public_baseurl: The public, accessible base URL of the homeserver + """ + + def mxc_to_http_filter(value, width, height, resize_method="crop"): + if value[0:6] != "mxc://": + return "" + + server_and_media_id = value[6:] + fragment = None + if "#" in server_and_media_id: + server_and_media_id, fragment = server_and_media_id.split("#", 1) + fragment = "#" + fragment + + params = {"width": width, "height": height, "method": resize_method} + return "%s_matrix/media/v1/thumbnail/%s?%s%s" % ( + public_baseurl, + server_and_media_id, + urllib.parse.urlencode(params), + fragment or "", + ) + + return mxc_to_http_filter + + +class RootConfig: """ Holder of an application's configuration. @@ -735,11 +838,26 @@ class ShardedWorkerHandlingConfig: def should_handle(self, instance_name: str, key: str) -> bool: """Whether this instance is responsible for handling the given key. """ - - # If multiple instances are not defined we always return true. + # If multiple instances are not defined we always return true if not self.instances or len(self.instances) == 1: return True + return self.get_instance(key) == instance_name + + def get_instance(self, key: str) -> str: + """Get the instance responsible for handling the given key. + + Note: For things like federation sending the config for which instance + is sending is known only to the sender instance if there is only one. + Therefore `should_handle` should be used where possible. + """ + + if not self.instances: + return "master" + + if len(self.instances) == 1: + return self.instances[0] + # We shard by taking the hash, modulo it by the number of instances and # then checking whether this instance matches the instance at that # index. @@ -749,7 +867,7 @@ class ShardedWorkerHandlingConfig: dest_hash = sha256(key.encode("utf8")).digest() dest_int = int.from_bytes(dest_hash, byteorder="little") remainder = dest_int % (len(self.instances)) - return self.instances[remainder] == instance_name + return self.instances[remainder] __all__ = ["Config", "RootConfig", "ShardedWorkerHandlingConfig"] diff --git a/synapse/config/_base.pyi b/synapse/config/_base.pyi
index eb911e8f9f..b8faafa9bd 100644 --- a/synapse/config/_base.pyi +++ b/synapse/config/_base.pyi
@@ -142,3 +142,4 @@ class ShardedWorkerHandlingConfig: instances: List[str] def __init__(self, instances: List[str]) -> None: ... def should_handle(self, instance_name: str, key: str) -> bool: ... + def get_instance(self, key: str) -> str: ... diff --git a/synapse/config/_util.py b/synapse/config/_util.py new file mode 100644
index 0000000000..c74969a977 --- /dev/null +++ b/synapse/config/_util.py
@@ -0,0 +1,51 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from typing import Any, Iterable + +import jsonschema + +from synapse.config._base import ConfigError +from synapse.types import JsonDict + + +def validate_config( + json_schema: JsonDict, config: Any, config_path: Iterable[str] +) -> None: + """Validates a config setting against a JsonSchema definition + + This can be used to validate a section of the config file against a schema + definition. If the validation fails, a ConfigError is raised with a textual + description of the problem. + + Args: + json_schema: the schema to validate against + config: the configuration value to be validated + config_path: the path within the config file. This will be used as a basis + for the error message. + """ + try: + jsonschema.validate(config, json_schema) + except jsonschema.ValidationError as e: + # copy `config_path` before modifying it. + path = list(config_path) + for p in list(e.path): + if isinstance(p, int): + path.append("<item %i>" % p) + else: + path.append(str(p)) + + raise ConfigError( + "Unable to parse configuration: %s at %s" % (e.message, ".".join(path)) + ) diff --git a/synapse/config/cache.py b/synapse/config/cache.py
index aff5b21ab2..8e03f14005 100644 --- a/synapse/config/cache.py +++ b/synapse/config/cache.py
@@ -33,7 +33,7 @@ _DEFAULT_FACTOR_SIZE = 0.5 _DEFAULT_EVENT_CACHE_SIZE = "10K" -class CacheProperties(object): +class CacheProperties: def __init__(self): # The default factor size for all caches self.default_factor_size = float( diff --git a/synapse/config/captcha.py b/synapse/config/captcha.py
index 82f04d7966..cb00958165 100644 --- a/synapse/config/captcha.py +++ b/synapse/config/captcha.py
@@ -28,6 +28,9 @@ class CaptchaConfig(Config): "recaptcha_siteverify_api", "https://www.recaptcha.net/recaptcha/api/siteverify", ) + self.recaptcha_template = self.read_templates( + ["recaptcha.html"], autoescape=True + )[0] def generate_config_section(self, **kwargs): return """\ diff --git a/synapse/config/consent_config.py b/synapse/config/consent_config.py
index aec9c4bbce..6efa59b110 100644 --- a/synapse/config/consent_config.py +++ b/synapse/config/consent_config.py
@@ -77,7 +77,7 @@ class ConsentConfig(Config): section = "consent" def __init__(self, *args): - super(ConsentConfig, self).__init__(*args) + super().__init__(*args) self.user_consent_version = None self.user_consent_template_dir = None @@ -89,6 +89,8 @@ class ConsentConfig(Config): def read_config(self, config, **kwargs): consent_config = config.get("user_consent") + self.terms_template = self.read_templates(["terms.html"], autoescape=True)[0] + if consent_config is None: return self.user_consent_version = str(consent_config["version"]) diff --git a/synapse/config/database.py b/synapse/config/database.py
index 62bccd9ef5..8a18a9ca2a 100644 --- a/synapse/config/database.py +++ b/synapse/config/database.py
@@ -100,7 +100,10 @@ class DatabaseConnectionConfig: self.name = name self.config = db_config - self.data_stores = data_stores + + # The `data_stores` config is actually talking about `databases` (we + # changed the name). + self.databases = data_stores class DatabaseConfig(Config): diff --git a/synapse/config/emailconfig.py b/synapse/config/emailconfig.py
index a63acbdc63..cceffbfee2 100644 --- a/synapse/config/emailconfig.py +++ b/synapse/config/emailconfig.py
@@ -14,7 +14,6 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function # This file can't be called email.py because if it is, we cannot: import email.utils @@ -23,7 +22,6 @@ from enum import Enum from typing import Optional import attr -import pkg_resources from ._base import Config, ConfigError @@ -98,21 +96,18 @@ class EmailConfig(Config): if parsed[1] == "": raise RuntimeError("Invalid notif_from address") + # A user-configurable template directory template_dir = email_config.get("template_dir") - # we need an absolute path, because we change directory after starting (and - # we don't yet know what auxiliary templates like mail.css we will need). - # (Note that loading as package_resources with jinja.PackageLoader doesn't - # work for the same reason.) - if not template_dir: - template_dir = pkg_resources.resource_filename("synapse", "res/templates") - - self.email_template_dir = os.path.abspath(template_dir) + if isinstance(template_dir, str): + # We need an absolute path, because we change directory after starting (and + # we don't yet know what auxiliary templates like mail.css we will need). + template_dir = os.path.abspath(template_dir) + elif template_dir is not None: + # If template_dir is something other than a str or None, warn the user + raise ConfigError("Config option email.template_dir must be type str") self.email_enable_notifs = email_config.get("enable_notifs", False) - account_validity_config = config.get("account_validity") or {} - account_validity_renewal_enabled = account_validity_config.get("renew_at") - self.threepid_behaviour_email = ( # Have Synapse handle the email sending if account_threepid_delegates.email # is not defined @@ -166,19 +161,6 @@ class EmailConfig(Config): email_config.get("validation_token_lifetime", "1h") ) - if ( - self.email_enable_notifs - or account_validity_renewal_enabled - or self.threepid_behaviour_email == ThreepidBehaviour.LOCAL - ): - # make sure we can import the required deps - import bleach - import jinja2 - - # prevent unused warnings - jinja2 - bleach - if self.threepid_behaviour_email == ThreepidBehaviour.LOCAL: missing = [] if not self.email_notif_from: @@ -196,84 +178,90 @@ class EmailConfig(Config): # These email templates have placeholders in them, and thus must be # parsed using a templating engine during a request - self.email_password_reset_template_html = email_config.get( + password_reset_template_html = email_config.get( "password_reset_template_html", "password_reset.html" ) - self.email_password_reset_template_text = email_config.get( + password_reset_template_text = email_config.get( "password_reset_template_text", "password_reset.txt" ) - self.email_registration_template_html = email_config.get( + registration_template_html = email_config.get( "registration_template_html", "registration.html" ) - self.email_registration_template_text = email_config.get( + registration_template_text = email_config.get( "registration_template_text", "registration.txt" ) - self.email_add_threepid_template_html = email_config.get( + add_threepid_template_html = email_config.get( "add_threepid_template_html", "add_threepid.html" ) - self.email_add_threepid_template_text = email_config.get( + add_threepid_template_text = email_config.get( "add_threepid_template_text", "add_threepid.txt" ) - self.email_password_reset_template_failure_html = email_config.get( + password_reset_template_failure_html = email_config.get( "password_reset_template_failure_html", "password_reset_failure.html" ) - self.email_registration_template_failure_html = email_config.get( + registration_template_failure_html = email_config.get( "registration_template_failure_html", "registration_failure.html" ) - self.email_add_threepid_template_failure_html = email_config.get( + add_threepid_template_failure_html = email_config.get( "add_threepid_template_failure_html", "add_threepid_failure.html" ) # These templates do not support any placeholder variables, so we # will read them from disk once during setup - email_password_reset_template_success_html = email_config.get( + password_reset_template_success_html = email_config.get( "password_reset_template_success_html", "password_reset_success.html" ) - email_registration_template_success_html = email_config.get( + registration_template_success_html = email_config.get( "registration_template_success_html", "registration_success.html" ) - email_add_threepid_template_success_html = email_config.get( + add_threepid_template_success_html = email_config.get( "add_threepid_template_success_html", "add_threepid_success.html" ) - # Check templates exist - for f in [ + # Read all templates from disk + ( self.email_password_reset_template_html, self.email_password_reset_template_text, self.email_registration_template_html, self.email_registration_template_text, self.email_add_threepid_template_html, self.email_add_threepid_template_text, + self.email_password_reset_template_confirmation_html, self.email_password_reset_template_failure_html, self.email_registration_template_failure_html, self.email_add_threepid_template_failure_html, - email_password_reset_template_success_html, - email_registration_template_success_html, - email_add_threepid_template_success_html, - ]: - p = os.path.join(self.email_template_dir, f) - if not os.path.isfile(p): - raise ConfigError("Unable to find template file %s" % (p,)) - - # Retrieve content of web templates - filepath = os.path.join( - self.email_template_dir, email_password_reset_template_success_html - ) - self.email_password_reset_template_success_html = self.read_file( - filepath, "email.password_reset_template_success_html" + password_reset_template_success_html_template, + registration_template_success_html_template, + add_threepid_template_success_html_template, + ) = self.read_templates( + [ + password_reset_template_html, + password_reset_template_text, + registration_template_html, + registration_template_text, + add_threepid_template_html, + add_threepid_template_text, + "password_reset_confirmation.html", + password_reset_template_failure_html, + registration_template_failure_html, + add_threepid_template_failure_html, + password_reset_template_success_html, + registration_template_success_html, + add_threepid_template_success_html, + ], + template_dir, ) - filepath = os.path.join( - self.email_template_dir, email_registration_template_success_html - ) - self.email_registration_template_success_html_content = self.read_file( - filepath, "email.registration_template_success_html" + + # Render templates that do not contain any placeholders + self.email_password_reset_template_success_html_content = ( + password_reset_template_success_html_template.render() ) - filepath = os.path.join( - self.email_template_dir, email_add_threepid_template_success_html + self.email_registration_template_success_html_content = ( + registration_template_success_html_template.render() ) - self.email_add_threepid_template_success_html_content = self.read_file( - filepath, "email.add_threepid_template_success_html" + self.email_add_threepid_template_success_html_content = ( + add_threepid_template_success_html_template.render() ) if self.email_enable_notifs: @@ -290,17 +278,19 @@ class EmailConfig(Config): % (", ".join(missing),) ) - self.email_notif_template_html = email_config.get( + notif_template_html = email_config.get( "notif_template_html", "notif_mail.html" ) - self.email_notif_template_text = email_config.get( + notif_template_text = email_config.get( "notif_template_text", "notif_mail.txt" ) - for f in self.email_notif_template_text, self.email_notif_template_html: - p = os.path.join(self.email_template_dir, f) - if not os.path.isfile(p): - raise ConfigError("Unable to find email template file %s" % (p,)) + ( + self.email_notif_template_html, + self.email_notif_template_text, + ) = self.read_templates( + [notif_template_html, notif_template_text], template_dir, + ) self.email_notif_for_new_users = email_config.get( "notif_for_new_users", True @@ -309,18 +299,20 @@ class EmailConfig(Config): "client_base_url", email_config.get("riot_base_url", None) ) - if account_validity_renewal_enabled: - self.email_expiry_template_html = email_config.get( + if self.account_validity.renew_by_email_enabled: + expiry_template_html = email_config.get( "expiry_template_html", "notice_expiry.html" ) - self.email_expiry_template_text = email_config.get( + expiry_template_text = email_config.get( "expiry_template_text", "notice_expiry.txt" ) - for f in self.email_expiry_template_text, self.email_expiry_template_html: - p = os.path.join(self.email_template_dir, f) - if not os.path.isfile(p): - raise ConfigError("Unable to find email template file %s" % (p,)) + ( + self.account_validity_template_html, + self.account_validity_template_text, + ) = self.read_templates( + [expiry_template_html, expiry_template_text], template_dir, + ) subjects_config = email_config.get("subjects", {}) subjects = {} @@ -400,9 +392,7 @@ class EmailConfig(Config): # Directory in which Synapse will try to find the template files below. # If not set, default templates from within the Synapse package will be used. # - # DO NOT UNCOMMENT THIS SETTING unless you want to customise the templates. - # If you *do* uncomment it, you will need to make sure that all the templates - # below are in the directory. + # Do not uncomment this setting unless you want to customise the templates. # # Synapse will look for the following templates in this directory: # @@ -415,9 +405,13 @@ class EmailConfig(Config): # * The contents of password reset emails sent by the homeserver: # 'password_reset.html' and 'password_reset.txt' # - # * HTML pages for success and failure that a user will see when they follow - # the link in the password reset email: 'password_reset_success.html' and - # 'password_reset_failure.html' + # * An HTML page that a user will see when they follow the link in the password + # reset email. The user will be asked to confirm the action before their + # password is reset: 'password_reset_confirmation.html' + # + # * HTML pages for success and failure that a user will see when they confirm + # the password reset flow using the page above: 'password_reset_success.html' + # and 'password_reset_failure.html' # # * The contents of address verification emails sent during registration: # 'registration.html' and 'registration.txt' diff --git a/synapse/config/federation.py b/synapse/config/federation.py
index 2c77d8f85b..ffd8fca54e 100644 --- a/synapse/config/federation.py +++ b/synapse/config/federation.py
@@ -17,7 +17,8 @@ from typing import Optional from netaddr import IPSet -from ._base import Config, ConfigError +from synapse.config._base import Config, ConfigError +from synapse.config._util import validate_config class FederationConfig(Config): @@ -52,8 +53,18 @@ class FederationConfig(Config): "Invalid range(s) provided in federation_ip_range_blacklist: %s" % e ) + federation_metrics_domains = config.get("federation_metrics_domains") or [] + validate_config( + _METRICS_FOR_DOMAINS_SCHEMA, + federation_metrics_domains, + ("federation_metrics_domains",), + ) + self.federation_metrics_domains = set(federation_metrics_domains) + def generate_config_section(self, config_dir_path, server_name, **kwargs): return """\ + ## Federation ## + # Restrict federation to the following whitelist of domains. # N.B. we recommend also firewalling your federation listener to limit # inbound federation traffic as early as possible, rather than relying @@ -85,4 +96,18 @@ class FederationConfig(Config): - '::1/128' - 'fe80::/64' - 'fc00::/7' + + # Report prometheus metrics on the age of PDUs being sent to and received from + # the following domains. This can be used to give an idea of "delay" on inbound + # and outbound federation, though be aware that any delay can be due to problems + # at either end or with the intermediate network. + # + # By default, no domains are monitored in this way. + # + #federation_metrics_domains: + # - matrix.org + # - example.com """ + + +_METRICS_FOR_DOMAINS_SCHEMA = {"type": "array", "items": {"type": "string"}} diff --git a/synapse/config/homeserver.py b/synapse/config/homeserver.py
index 556e291495..be65554524 100644 --- a/synapse/config/homeserver.py +++ b/synapse/config/homeserver.py
@@ -92,5 +92,4 @@ class HomeServerConfig(RootConfig): TracerConfig, WorkerConfig, RedisConfig, - FederationConfig, ] diff --git a/synapse/config/key.py b/synapse/config/key.py
index b529ea5da0..de964dff13 100644 --- a/synapse/config/key.py +++ b/synapse/config/key.py
@@ -82,7 +82,7 @@ logger = logging.getLogger(__name__) @attr.s -class TrustedKeyServer(object): +class TrustedKeyServer: # string: name of the server. server_name = attr.ib() diff --git a/synapse/config/logger.py b/synapse/config/logger.py
index dd775a97e8..13d6f6a3ea 100644 --- a/synapse/config/logger.py +++ b/synapse/config/logger.py
@@ -17,6 +17,7 @@ import logging import logging.config import os import sys +import threading from string import Template import yaml @@ -25,6 +26,7 @@ from twisted.logger import ( ILogObserver, LogBeginner, STDLibLogObserver, + eventAsText, globalLogBeginner, ) @@ -55,24 +57,33 @@ formatters: format: '%(asctime)s - %(name)s - %(lineno)d - %(levelname)s - \ %(request)s - %(message)s' -filters: - context: - (): synapse.logging.context.LoggingContextFilter - request: "" - handlers: file: - class: logging.handlers.RotatingFileHandler + class: logging.handlers.TimedRotatingFileHandler formatter: precise filename: ${log_file} - maxBytes: 104857600 - backupCount: 10 - filters: [context] + when: midnight + backupCount: 3 # Does not include the current log file. encoding: utf8 + + # Default to buffering writes to log file for efficiency. This means that + # will be a delay for INFO/DEBUG logs to get written, but WARNING/ERROR + # logs will still be flushed immediately. + buffer: + class: logging.handlers.MemoryHandler + target: file + # The capacity is the number of log lines that are buffered before + # being written to disk. Increasing this will lead to better + # performance, at the expensive of it taking longer for log lines to + # be written to disk. + capacity: 10 + flushLevel: 30 # Flush for WARNING logs as well + + # A handler that writes logs to stderr. Unused by default, but can be used + # instead of "buffer" and "file" in the logger handlers. console: class: logging.StreamHandler formatter: precise - filters: [context] loggers: synapse.storage.SQL: @@ -80,9 +91,24 @@ loggers: # information such as access tokens. level: INFO + twisted: + # We send the twisted logging directly to the file handler, + # to work around https://github.com/matrix-org/synapse/issues/3471 + # when using "buffer" logger. Use "console" to log to stderr instead. + handlers: [file] + propagate: false + root: level: INFO - handlers: [file, console] + + # Write logs to the `buffer` handler, which will buffer them together in memory, + # then write them to a file. + # + # Replace "buffer" with "console" to log to stderr instead. (Note that you'll + # also need to update the configuation for the `twisted` logger above, in + # this case.) + # + handlers: [buffer] disable_existing_loggers: false """ @@ -168,17 +194,33 @@ def _setup_stdlib_logging(config, log_config, logBeginner: LogBeginner): handler = logging.StreamHandler() handler.setFormatter(formatter) - handler.addFilter(LoggingContextFilter(request="")) logger.addHandler(handler) else: logging.config.dictConfig(log_config) + # We add a log record factory that runs all messages through the + # LoggingContextFilter so that we get the context *at the time we log* + # rather than when we write to a handler. This can be done in config using + # filter options, but care must when using e.g. MemoryHandler to buffer + # writes. + + log_filter = LoggingContextFilter(request="") + old_factory = logging.getLogRecordFactory() + + def factory(*args, **kwargs): + record = old_factory(*args, **kwargs) + log_filter.filter(record) + return record + + logging.setLogRecordFactory(factory) + # Route Twisted's native logging through to the standard library logging # system. observer = STDLibLogObserver() - def _log(event): + threadlocal = threading.local() + def _log(event): if "log_text" in event: if event["log_text"].startswith("DNSDatagramProtocol starting on "): return @@ -189,7 +231,25 @@ def _setup_stdlib_logging(config, log_config, logBeginner: LogBeginner): if event["log_text"].startswith("Timing out client"): return - return observer(event) + # this is a workaround to make sure we don't get stack overflows when the + # logging system raises an error which is written to stderr which is redirected + # to the logging system, etc. + if getattr(threadlocal, "active", False): + # write the text of the event, if any, to the *real* stderr (which may + # be redirected to /dev/null, but there's not much we can do) + try: + event_text = eventAsText(event) + print("logging during logging: %s" % event_text, file=sys.__stderr__) + except Exception: + # gah. + pass + return + + try: + threadlocal.active = True + return observer(event) + finally: + threadlocal.active = False logBeginner.beginLoggingTo([_log], redirectStandardIO=not config.no_redirect_stdio) if not config.no_redirect_stdio: diff --git a/synapse/config/metrics.py b/synapse/config/metrics.py
index 6aad0d37c0..dfd27e1523 100644 --- a/synapse/config/metrics.py +++ b/synapse/config/metrics.py
@@ -22,7 +22,7 @@ from ._base import Config, ConfigError @attr.s -class MetricsFlags(object): +class MetricsFlags: known_servers = attr.ib(default=False, validator=attr.validators.instance_of(bool)) @classmethod diff --git a/synapse/config/oidc_config.py b/synapse/config/oidc_config.py
index e0939bce84..f924116819 100644 --- a/synapse/config/oidc_config.py +++ b/synapse/config/oidc_config.py
@@ -56,6 +56,7 @@ class OIDCConfig(Config): self.oidc_userinfo_endpoint = oidc_config.get("userinfo_endpoint") self.oidc_jwks_uri = oidc_config.get("jwks_uri") self.oidc_skip_verification = oidc_config.get("skip_verification", False) + self.oidc_allow_existing_users = oidc_config.get("allow_existing_users", False) ump_config = oidc_config.get("user_mapping_provider", {}) ump_config.setdefault("module", DEFAULT_USER_MAPPING_PROVIDER) @@ -158,6 +159,11 @@ class OIDCConfig(Config): # #skip_verification: true + # Uncomment to allow a user logging in via OIDC to match a pre-existing account instead + # of failing. This could be used if switching from password logins to OIDC. Defaults to false. + # + #allow_existing_users: true + # An external module can be provided here as a custom solution to mapping # attributes returned from a OIDC provider onto a matrix user. # @@ -198,6 +204,14 @@ class OIDCConfig(Config): # If unset, no displayname will be set. # #display_name_template: "{{{{ user.given_name }}}} {{{{ user.last_name }}}}" + + # Jinja2 templates for extra attributes to send back to the client during + # login. + # + # Note that these are non-standard and clients will ignore them without modifications. + # + #extra_attributes: + #birthdate: "{{{{ user.birthdate }}}}" """.format( mapping_provider=DEFAULT_USER_MAPPING_PROVIDER ) diff --git a/synapse/config/ratelimiting.py b/synapse/config/ratelimiting.py
index b1981d4d15..4fca5b6d96 100644 --- a/synapse/config/ratelimiting.py +++ b/synapse/config/ratelimiting.py
@@ -17,7 +17,7 @@ from typing import Dict from ._base import Config -class RateLimitConfig(object): +class RateLimitConfig: def __init__( self, config: Dict[str, float], @@ -27,7 +27,7 @@ class RateLimitConfig(object): self.burst_count = config.get("burst_count", defaults["burst_count"]) -class FederationRateLimitConfig(object): +class FederationRateLimitConfig: _items_and_default = { "window_size": 1000, "sleep_limit": 10, @@ -96,6 +96,15 @@ class RatelimitConfig(Config): if rc_admin_redaction: self.rc_admin_redaction = RateLimitConfig(rc_admin_redaction) + self.rc_joins_local = RateLimitConfig( + config.get("rc_joins", {}).get("local", {}), + defaults={"per_second": 0.1, "burst_count": 3}, + ) + self.rc_joins_remote = RateLimitConfig( + config.get("rc_joins", {}).get("remote", {}), + defaults={"per_second": 0.01, "burst_count": 3}, + ) + def generate_config_section(self, **kwargs): return """\ ## Ratelimiting ## @@ -123,6 +132,10 @@ class RatelimitConfig(Config): # - one for ratelimiting redactions by room admins. If this is not explicitly # set then it uses the same ratelimiting as per rc_message. This is useful # to allow room admins to deal with abuse quickly. + # - two for ratelimiting number of rooms a user can join, "local" for when + # users are joining rooms the server is already in (this is cheap) vs + # "remote" for when users are trying to join rooms not on the server (which + # can be more expensive) # # The defaults are as shown below. # @@ -152,6 +165,14 @@ class RatelimitConfig(Config): #rc_admin_redaction: # per_second: 1 # burst_count: 50 + # + #rc_joins: + # local: + # per_second: 0.1 + # burst_count: 3 + # remote: + # per_second: 0.01 + # burst_count: 3 # Ratelimiting settings for incoming federation diff --git a/synapse/config/registration.py b/synapse/config/registration.py
index 080b9bc445..36f19d125e 100644 --- a/synapse/config/registration.py +++ b/synapse/config/registration.py
@@ -30,7 +30,7 @@ class AccountValidityConfig(Config): def __init__(self, config, synapse_config): if config is None: return - super(AccountValidityConfig, self).__init__() + super().__init__() self.enabled = config.get("enabled", False) self.renew_by_email_enabled = "renew_at" in config @@ -221,6 +221,11 @@ class RegistrationConfig(Config): session_lifetime = self.parse_duration(session_lifetime) self.session_lifetime = session_lifetime + # The success template used during fallback auth. + self.fallback_success_template = self.read_templates( + ["auth_success.html"], autoescape=True + )[0] + self.bind_new_user_emails_to_sydent = config.get( "bind_new_user_emails_to_sydent" ) @@ -412,24 +417,6 @@ class RegistrationConfig(Config): # #default_identity_server: https://matrix.org - # The list of identity servers trusted to verify third party - # identifiers by this server. - # - # Also defines the ID server which will be called when an account is - # deactivated (one will be picked arbitrarily). - # - # Note: This option is deprecated. Since v0.99.4, Synapse has tracked which identity - # server a 3PID has been bound to. For 3PIDs bound before then, Synapse runs a - # background migration script, informing itself that the identity server all of its - # 3PIDs have been bound to is likely one of the below. - # - # As of Synapse v1.4.0, all other functionality of this option has been deprecated, and - # it is now solely used for the purposes of the background migration script, and can be - # removed once it has run. - #trusted_third_party_id_servers: - # - matrix.org - # - vector.im - # If enabled, user IDs, display names and avatar URLs will be replicated # to this server whenever they change. # This is an experimental API currently implemented by sydent to support diff --git a/synapse/config/room.py b/synapse/config/room.py
index 52cf0b62fc..692d7a1936 100644 --- a/synapse/config/room.py +++ b/synapse/config/room.py
@@ -22,7 +22,7 @@ from ._base import Config, ConfigError logger = logging.Logger(__name__) -class RoomDefaultEncryptionTypes(object): +class RoomDefaultEncryptionTypes: """Possible values for the encryption_enabled_by_default_for_room_type config option""" ALL = "all" diff --git a/synapse/config/room_directory.py b/synapse/config/room_directory.py
index 7ac7699676..6de1f9d103 100644 --- a/synapse/config/room_directory.py +++ b/synapse/config/room_directory.py
@@ -149,7 +149,7 @@ class RoomDirectoryConfig(Config): return False -class _RoomDirectoryRule(object): +class _RoomDirectoryRule: """Helper class to test whether a room directory action is allowed, like creating an alias or publishing a room. """ diff --git a/synapse/config/saml2_config.py b/synapse/config/saml2_config.py
index 293643b2de..99aa8b3bf1 100644 --- a/synapse/config/saml2_config.py +++ b/synapse/config/saml2_config.py
@@ -15,14 +15,15 @@ # limitations under the License. import logging +from typing import Any, List -import jinja2 -import pkg_resources +import attr from synapse.python_dependencies import DependencyException, check_requirements from synapse.util.module_loader import load_module, load_python_module from ._base import Config, ConfigError +from ._util import validate_config logger = logging.getLogger(__name__) @@ -80,6 +81,11 @@ class SAML2Config(Config): self.saml2_enabled = True + attribute_requirements = saml2_config.get("attribute_requirements") or [] + self.attribute_requirements = _parse_attribute_requirements_def( + attribute_requirements + ) + self.saml2_grandfathered_mxid_source_attribute = saml2_config.get( "grandfathered_mxid_source_attribute", "uid" ) @@ -163,16 +169,6 @@ class SAML2Config(Config): saml2_config.get("saml_session_lifetime", "15m") ) - template_dir = saml2_config.get("template_dir") - if not template_dir: - template_dir = pkg_resources.resource_filename("synapse", "res/templates",) - - loader = jinja2.FileSystemLoader(template_dir) - # enable auto-escape here, to having to remember to escape manually in the - # template - env = jinja2.Environment(loader=loader, autoescape=True) - self.saml2_error_html_template = env.get_template("saml_error.html") - def _default_saml_config_dict( self, required_attributes: set, optional_attributes: set ): @@ -225,11 +221,14 @@ class SAML2Config(Config): # At least one of `sp_config` or `config_path` must be set in this section to # enable SAML login. # - # (You will probably also want to set the following options to `false` to + # You will probably also want to set the following options to `false` to # disable the regular login/registration flows: # * enable_registration # * password_config.enabled # + # You will also want to investigate the settings under the "sso" configuration + # section below. + # # Once SAML support is enabled, a metadata file will be exposed at # https://<server>:<port>/_matrix/saml2/metadata.xml, which you may be able to # use to configure your SAML IdP with. Alternatively, you can manually configure @@ -341,30 +340,47 @@ class SAML2Config(Config): # #grandfathered_mxid_source_attribute: upn - # Directory in which Synapse will try to find the template files below. - # If not set, default templates from within the Synapse package will be used. - # - # DO NOT UNCOMMENT THIS SETTING unless you want to customise the templates. - # If you *do* uncomment it, you will need to make sure that all the templates - # below are in the directory. - # - # Synapse will look for the following templates in this directory: - # - # * HTML page to display to users if something goes wrong during the - # authentication process: 'saml_error.html'. - # - # When rendering, this template is given the following variables: - # * code: an HTML error code corresponding to the error that is being - # returned (typically 400 or 500) - # - # * msg: a textual message describing the error. + # It is possible to configure Synapse to only allow logins if SAML attributes + # match particular values. The requirements can be listed under + # `attribute_requirements` as shown below. All of the listed attributes must + # match for the login to be permitted. # - # The variables will automatically be HTML-escaped. - # - # You can see the default templates at: - # https://github.com/matrix-org/synapse/tree/master/synapse/res/templates - # - #template_dir: "res/templates" + #attribute_requirements: + # - attribute: userGroup + # value: "staff" + # - attribute: department + # value: "sales" """ % { "config_dir_path": config_dir_path } + + +@attr.s(frozen=True) +class SamlAttributeRequirement: + """Object describing a single requirement for SAML attributes.""" + + attribute = attr.ib(type=str) + value = attr.ib(type=str) + + JSON_SCHEMA = { + "type": "object", + "properties": {"attribute": {"type": "string"}, "value": {"type": "string"}}, + "required": ["attribute", "value"], + } + + +ATTRIBUTE_REQUIREMENTS_SCHEMA = { + "type": "array", + "items": SamlAttributeRequirement.JSON_SCHEMA, +} + + +def _parse_attribute_requirements_def( + attribute_requirements: Any, +) -> List[SamlAttributeRequirement]: + validate_config( + ATTRIBUTE_REQUIREMENTS_SCHEMA, + attribute_requirements, + config_path=["saml2_config", "attribute_requirements"], + ) + return [SamlAttributeRequirement(**x) for x in attribute_requirements] diff --git a/synapse/config/server.py b/synapse/config/server.py
index 57b0b97164..9a5c7cb0ab 100644 --- a/synapse/config/server.py +++ b/synapse/config/server.py
@@ -26,7 +26,6 @@ import yaml from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.http.endpoint import parse_and_validate_server_name -from synapse.python_dependencies import DependencyException, check_requirements from ._base import Config, ConfigError @@ -431,7 +430,7 @@ class ServerConfig(Config): self.gc_thresholds = read_gc_thresholds(config.get("gc_thresholds", None)) @attr.s - class LimitRemoteRoomsConfig(object): + class LimitRemoteRoomsConfig: enabled = attr.ib( validator=attr.validators.instance_of(bool), default=False ) @@ -445,6 +444,9 @@ class ServerConfig(Config): validator=attr.validators.instance_of(str), default=ROOM_COMPLEXITY_TOO_GREAT, ) + admins_can_join = attr.ib( + validator=attr.validators.instance_of(bool), default=False + ) self.limit_remote_rooms = LimitRemoteRoomsConfig( **(config.get("limit_remote_rooms") or {}) @@ -511,8 +513,6 @@ class ServerConfig(Config): ) ) - _check_resource_config(self.listeners) - self.cleanup_extremities_with_dummy_events = config.get( "cleanup_extremities_with_dummy_events", True ) @@ -647,10 +647,23 @@ class ServerConfig(Config): """\ ## Server ## - # The domain name of the server, with optional explicit port. - # This is used by remote servers to connect to this server, - # e.g. matrix.org, localhost:8080, etc. - # This is also the last part of your UserID. + # The public-facing domain of the server + # + # The server_name name will appear at the end of usernames and room addresses + # created on this server. For example if the server_name was example.com, + # usernames on this server would be in the format @user:example.com + # + # In most cases you should avoid using a matrix specific subdomain such as + # matrix.example.com or synapse.example.com as the server_name for the same + # reasons you wouldn't use user@email.example.com as your email address. + # See https://github.com/matrix-org/synapse/blob/master/docs/delegate.md + # for information on how to host Synapse on a subdomain while preserving + # a clean server_name. + # + # The server_name cannot be changed later so it is important to + # configure this correctly before you start Synapse. It should be all + # lowercase and may contain an explicit port. + # Examples: matrix.org, localhost:8080 # server_name: "%(server_name)s" @@ -927,6 +940,10 @@ class ServerConfig(Config): # #complexity_error: "This room is too complex." + # allow server admins to join complex rooms. Default is false. + # + #admins_can_join: true + # Whether to require a user to be in the room to add an alias to it. # Defaults to 'true'. # @@ -1044,11 +1061,10 @@ class ServerConfig(Config): # min_lifetime: 1d # max_lifetime: 1y - # Retention policy limits. If set, a user won't be able to send a - # 'm.room.retention' event which features a 'min_lifetime' or a 'max_lifetime' - # that's not within this range. This is especially useful in closed federations, - # in which server admins can make sure every federating server applies the same - # rules. + # Retention policy limits. If set, and the state of a room contains a + # 'm.room.retention' event in its state which contains a 'min_lifetime' or a + # 'max_lifetime' that's out of these bounds, Synapse will cap the room's policy + # to these limits when running purge jobs. # #allowed_lifetime_min: 1d #allowed_lifetime_max: 1y @@ -1074,12 +1090,19 @@ class ServerConfig(Config): # (e.g. every 12h), but not want that purge to be performed by a job that's # iterating over every room it knows, which could be heavy on the server. # + # If any purge job is configured, it is strongly recommended to have at least + # a single job with neither 'shortest_max_lifetime' nor 'longest_max_lifetime' + # set, or one job without 'shortest_max_lifetime' and one job without + # 'longest_max_lifetime' set. Otherwise some rooms might be ignored, even if + # 'allowed_lifetime_min' and 'allowed_lifetime_max' are set, because capping a + # room's policy to these values is done after the policies are retrieved from + # Synapse's database (which is done using the range specified in a purge job's + # configuration). + # #purge_jobs: - # - shortest_max_lifetime: 1d - # longest_max_lifetime: 3d + # - longest_max_lifetime: 3d # interval: 12h # - shortest_max_lifetime: 3d - # longest_max_lifetime: 1y # interval: 1d # Inhibits the /requestToken endpoints from returning an error that might leak @@ -1231,20 +1254,3 @@ def _warn_if_webclient_configured(listeners: Iterable[ListenerConfig]) -> None: if name == "webclient": logger.warning(NO_MORE_WEB_CLIENT_WARNING) return - - -def _check_resource_config(listeners: Iterable[ListenerConfig]) -> None: - resource_names = { - res_name - for listener in listeners - if listener.http_options - for res in listener.http_options.resources - for res_name in res.names - } - - for resource in resource_names: - if resource == "consent": - try: - check_requirements("resources.consent") - except DependencyException as e: - raise ConfigError(e.message) diff --git a/synapse/config/server_notices_config.py b/synapse/config/server_notices_config.py
index 6c427b6f92..57f69dc8e2 100644 --- a/synapse/config/server_notices_config.py +++ b/synapse/config/server_notices_config.py
@@ -62,7 +62,7 @@ class ServerNoticesConfig(Config): section = "servernotices" def __init__(self, *args): - super(ServerNoticesConfig, self).__init__(*args) + super().__init__(*args) self.server_notices_mxid = None self.server_notices_mxid_display_name = None self.server_notices_mxid_avatar_url = None diff --git a/synapse/config/sso.py b/synapse/config/sso.py
index 73b7296399..4427676167 100644 --- a/synapse/config/sso.py +++ b/synapse/config/sso.py
@@ -12,11 +12,8 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import os from typing import Any, Dict -import pkg_resources - from ._base import Config @@ -29,22 +26,32 @@ class SSOConfig(Config): def read_config(self, config, **kwargs): sso_config = config.get("sso") or {} # type: Dict[str, Any] - # Pick a template directory in order of: - # * The sso-specific template_dir - # * /path/to/synapse/install/res/templates + # The sso-specific template_dir template_dir = sso_config.get("template_dir") - if not template_dir: - template_dir = pkg_resources.resource_filename("synapse", "res/templates",) - self.sso_template_dir = template_dir - self.sso_account_deactivated_template = self.read_file( - os.path.join(self.sso_template_dir, "sso_account_deactivated.html"), - "sso_account_deactivated_template", + # Read templates from disk + ( + self.sso_redirect_confirm_template, + self.sso_auth_confirm_template, + self.sso_error_template, + sso_account_deactivated_template, + sso_auth_success_template, + ) = self.read_templates( + [ + "sso_redirect_confirm.html", + "sso_auth_confirm.html", + "sso_error.html", + "sso_account_deactivated.html", + "sso_auth_success.html", + ], + template_dir, ) - self.sso_auth_success_template = self.read_file( - os.path.join(self.sso_template_dir, "sso_auth_success.html"), - "sso_auth_success_template", + + # These templates have no placeholders, so render them here + self.sso_account_deactivated_template = ( + sso_account_deactivated_template.render() ) + self.sso_auth_success_template = sso_auth_success_template.render() self.sso_client_whitelist = sso_config.get("client_whitelist") or [] diff --git a/synapse/config/stats.py b/synapse/config/stats.py
index 62485189ea..b559bfa411 100644 --- a/synapse/config/stats.py +++ b/synapse/config/stats.py
@@ -13,8 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import division - import sys from ._base import Config diff --git a/synapse/config/tls.py b/synapse/config/tls.py
index e368ea564d..9ddb8b546b 100644 --- a/synapse/config/tls.py +++ b/synapse/config/tls.py
@@ -471,7 +471,6 @@ class TlsConfig(Config): # or by checking matrix.org/federationtester/api/report?server_name=$host # #tls_fingerprints: [{"sha256": "<base64_encoded_sha256_fingerprint>"}] - """ # Lowercase the string representation of boolean values % { diff --git a/synapse/config/workers.py b/synapse/config/workers.py
index c784a71508..f23e42cdf9 100644 --- a/synapse/config/workers.py +++ b/synapse/config/workers.py
@@ -13,12 +13,24 @@ # See the License for the specific language governing permissions and # limitations under the License. +from typing import List, Union + import attr from ._base import Config, ConfigError, ShardedWorkerHandlingConfig from .server import ListenerConfig, parse_listener_def +def _instance_to_list_converter(obj: Union[str, List[str]]) -> List[str]: + """Helper for allowing parsing a string or list of strings to a config + option expecting a list of strings. + """ + + if isinstance(obj, str): + return [obj] + return obj + + @attr.s class InstanceLocationConfig: """The host and port to talk to an instance via HTTP replication. @@ -33,11 +45,13 @@ class WriterLocations: """Specifies the instances that write various streams. Attributes: - events: The instance that writes to the event and backfill streams. - events: The instance that writes to the typing stream. + events: The instances that write to the event and backfill streams. + typing: The instance that writes to the typing stream. """ - events = attr.ib(default="master", type=str) + events = attr.ib( + default=["master"], type=List[str], converter=_instance_to_list_converter + ) typing = attr.ib(default="master", type=str) @@ -105,15 +119,18 @@ class WorkerConfig(Config): writers = config.get("stream_writers") or {} self.writers = WriterLocations(**writers) - # Check that the configured writer for events and typing also appears in + # Check that the configured writers for events and typing also appears in # `instance_map`. for stream in ("events", "typing"): - instance = getattr(self.writers, stream) - if instance != "master" and instance not in self.instance_map: - raise ConfigError( - "Instance %r is configured to write %s but does not appear in `instance_map` config." - % (instance, stream) - ) + instances = _instance_to_list_converter(getattr(self.writers, stream)) + for instance in instances: + if instance != "master" and instance not in self.instance_map: + raise ConfigError( + "Instance %r is configured to write %s but does not appear in `instance_map` config." + % (instance, stream) + ) + + self.events_shard_config = ShardedWorkerHandlingConfig(self.writers.events) def generate_config_section(self, config_dir_path, server_name, **kwargs): return """\ diff --git a/synapse/crypto/context_factory.py b/synapse/crypto/context_factory.py
index a5a2a7815d..79668a402e 100644 --- a/synapse/crypto/context_factory.py +++ b/synapse/crypto/context_factory.py
@@ -45,9 +45,21 @@ _TLS_VERSION_MAP = { class ServerContextFactory(ContextFactory): """Factory for PyOpenSSL SSL contexts that are used to handle incoming - connections.""" + connections. + + TODO: replace this with an implementation of IOpenSSLServerConnectionCreator, + per https://github.com/matrix-org/synapse/issues/1691 + """ def __init__(self, config): + # TODO: once pyOpenSSL exposes TLS_METHOD and SSL_CTX_set_min_proto_version, + # switch to those (see https://github.com/pyca/cryptography/issues/5379). + # + # note that, despite the confusing name, SSLv23_METHOD does *not* enforce SSLv2 + # or v3, but is a synonym for TLS_METHOD, which allows the client and server + # to negotiate an appropriate version of TLS constrained by the version options + # set with context.set_options. + # self._context = SSL.Context(SSL.SSLv23_METHOD) self.configure_context(self._context, config) @@ -75,7 +87,7 @@ class ServerContextFactory(ContextFactory): @implementer(IPolicyForHTTPS) -class FederationPolicyForHTTPS(object): +class FederationPolicyForHTTPS: """Factory for Twisted SSLClientConnectionCreators that are used to make connections to remote servers for federation. @@ -144,7 +156,7 @@ class FederationPolicyForHTTPS(object): @implementer(IPolicyForHTTPS) -class RegularPolicyForHTTPS(object): +class RegularPolicyForHTTPS: """Factory for Twisted SSLClientConnectionCreators that are used to make connections to remote servers, for other than federation. @@ -181,7 +193,7 @@ def _context_info_cb(ssl_connection, where, ret): @implementer(IOpenSSLClientConnectionCreator) -class SSLClientConnectionCreator(object): +class SSLClientConnectionCreator: """Creates openssl connection objects for client connections. Replaces twisted.internet.ssl.ClientTLSOptions @@ -206,7 +218,7 @@ class SSLClientConnectionCreator(object): return connection -class ConnectionVerifier(object): +class ConnectionVerifier: """Set the SNI, and do cert verification This is a thing which is attached to the TLSMemoryBIOProtocol, and is called by diff --git a/synapse/crypto/keyring.py b/synapse/crypto/keyring.py
index dbfc3e8972..c04ad77cf9 100644 --- a/synapse/crypto/keyring.py +++ b/synapse/crypto/keyring.py
@@ -42,7 +42,6 @@ from synapse.api.errors import ( ) from synapse.logging.context import ( PreserveLoggingContext, - current_context, make_deferred_yieldable, preserve_fn, run_in_background, @@ -57,7 +56,7 @@ logger = logging.getLogger(__name__) @attr.s(slots=True, cmp=False) -class VerifyJsonRequest(object): +class VerifyJsonRequest: """ A request to verify a JSON object. @@ -96,7 +95,7 @@ class KeyLookupError(ValueError): pass -class Keyring(object): +class Keyring: def __init__(self, hs, key_fetchers=None): self.clock = hs.get_clock() @@ -223,8 +222,7 @@ class Keyring(object): return results - @defer.inlineCallbacks - def _start_key_lookups(self, verify_requests): + async def _start_key_lookups(self, verify_requests): """Sets off the key fetches for each verify request Once each fetch completes, verify_request.key_ready will be resolved. @@ -234,8 +232,6 @@ class Keyring(object): """ try: - ctx = current_context() - # map from server name to a set of outstanding request ids server_to_request_ids = {} @@ -245,7 +241,7 @@ class Keyring(object): server_to_request_ids.setdefault(server_name, set()).add(request_id) # Wait for any previous lookups to complete before proceeding. - yield self.wait_for_previous_lookups(server_to_request_ids.keys()) + await self.wait_for_previous_lookups(server_to_request_ids.keys()) # take out a lock on each of the servers by sticking a Deferred in # key_downloads @@ -266,12 +262,8 @@ class Keyring(object): # if there are no more requests for this server, we can drop the lock. if not server_requests: - with PreserveLoggingContext(ctx): - logger.debug("Releasing key lookup lock on %s", server_name) - - # ... but not immediately, as that can cause stack explosions if - # we get a long queue of lookups. - self.clock.call_later(0, drop_server_lock, server_name) + logger.debug("Releasing key lookup lock on %s", server_name) + drop_server_lock(server_name) return res @@ -283,15 +275,14 @@ class Keyring(object): except Exception: logger.exception("Error starting key lookups") - @defer.inlineCallbacks - def wait_for_previous_lookups(self, server_names): + async def wait_for_previous_lookups(self, server_names) -> None: """Waits for any previous key lookups for the given servers to finish. Args: server_names (Iterable[str]): list of servers which we want to look up Returns: - Deferred[None]: resolves once all key lookups for the given servers have + Resolves once all key lookups for the given servers have completed. Follows the synapse rules of logcontext preservation. """ loop_count = 1 @@ -309,7 +300,7 @@ class Keyring(object): loop_count, ) with PreserveLoggingContext(): - yield defer.DeferredList((w[1] for w in wait_on)) + await defer.DeferredList((w[1] for w in wait_on)) loop_count += 1 @@ -326,44 +317,56 @@ class Keyring(object): remaining_requests = {rq for rq in verify_requests if not rq.key_ready.called} - @defer.inlineCallbacks - def do_iterations(): - with Measure(self.clock, "get_server_verify_keys"): - for f in self._key_fetchers: - if not remaining_requests: - return - yield self._attempt_key_fetches_with_fetcher(f, remaining_requests) + async def do_iterations(): + try: + with Measure(self.clock, "get_server_verify_keys"): + for f in self._key_fetchers: + if not remaining_requests: + return + await self._attempt_key_fetches_with_fetcher( + f, remaining_requests + ) - # look for any requests which weren't satisfied - with PreserveLoggingContext(): - for verify_request in remaining_requests: - verify_request.key_ready.errback( + # look for any requests which weren't satisfied + while remaining_requests: + verify_request = remaining_requests.pop() + rq_str = ( + "VerifyJsonRequest(server=%s, key_ids=%s, min_valid=%i)" + % ( + verify_request.server_name, + verify_request.key_ids, + verify_request.minimum_valid_until_ts, + ) + ) + + # If we run the errback immediately, it may cancel our + # loggingcontext while we are still in it, so instead we + # schedule it for the next time round the reactor. + # + # (this also ensures that we don't get a stack overflow if we + # has a massive queue of lookups waiting for this server). + self.clock.call_later( + 0, + verify_request.key_ready.errback, SynapseError( 401, - "No key for %s with ids in %s (min_validity %i)" - % ( - verify_request.server_name, - verify_request.key_ids, - verify_request.minimum_valid_until_ts, - ), + "Failed to find any key to satisfy %s" % (rq_str,), Codes.UNAUTHORIZED, - ) + ), ) + except Exception as err: + # we don't really expect to get here, because any errors should already + # have been caught and logged. But if we do, let's log the error and make + # sure that all of the deferreds are resolved. + logger.error("Unexpected error in _get_server_verify_keys: %s", err) + with PreserveLoggingContext(): + for verify_request in remaining_requests: + if not verify_request.key_ready.called: + verify_request.key_ready.errback(err) - def on_err(err): - # we don't really expect to get here, because any errors should already - # have been caught and logged. But if we do, let's log the error and make - # sure that all of the deferreds are resolved. - logger.error("Unexpected error in _get_server_verify_keys: %s", err) - with PreserveLoggingContext(): - for verify_request in remaining_requests: - if not verify_request.key_ready.called: - verify_request.key_ready.errback(err) - - run_in_background(do_iterations).addErrback(on_err) + run_in_background(do_iterations) - @defer.inlineCallbacks - def _attempt_key_fetches_with_fetcher(self, fetcher, remaining_requests): + async def _attempt_key_fetches_with_fetcher(self, fetcher, remaining_requests): """Use a key fetcher to attempt to satisfy some key requests Args: @@ -390,7 +393,7 @@ class Keyring(object): verify_request.minimum_valid_until_ts, ) - results = yield fetcher.get_keys(missing_keys) + results = await fetcher.get_keys(missing_keys) completed = [] for verify_request in remaining_requests: @@ -412,18 +415,31 @@ class Keyring(object): # key was not valid at this point continue - with PreserveLoggingContext(): - verify_request.key_ready.callback( - (server_name, key_id, fetch_key_result.verify_key) - ) + # we have a valid key for this request. If we run the callback + # immediately, it may cancel our loggingcontext while we are still in + # it, so instead we schedule it for the next time round the reactor. + # + # (this also ensures that we don't get a stack overflow if we had + # a massive queue of lookups waiting for this server). + logger.debug( + "Found key %s:%s for %s", + server_name, + key_id, + verify_request.request_name, + ) + self.clock.call_later( + 0, + verify_request.key_ready.callback, + (server_name, key_id, fetch_key_result.verify_key), + ) completed.append(verify_request) break remaining_requests.difference_update(completed) -class KeyFetcher(object): - def get_keys(self, keys_to_fetch): +class KeyFetcher: + async def get_keys(self, keys_to_fetch): """ Args: keys_to_fetch (dict[str, dict[str, int]]): @@ -442,8 +458,7 @@ class StoreKeyFetcher(KeyFetcher): def __init__(self, hs): self.store = hs.get_datastore() - @defer.inlineCallbacks - def get_keys(self, keys_to_fetch): + async def get_keys(self, keys_to_fetch): """see KeyFetcher.get_keys""" keys_to_fetch = ( @@ -452,20 +467,19 @@ class StoreKeyFetcher(KeyFetcher): for key_id in keys_for_server.keys() ) - res = yield self.store.get_server_verify_keys(keys_to_fetch) + res = await self.store.get_server_verify_keys(keys_to_fetch) keys = {} for (server_name, key_id), key in res.items(): keys.setdefault(server_name, {})[key_id] = key return keys -class BaseV2KeyFetcher(object): +class BaseV2KeyFetcher: def __init__(self, hs): self.store = hs.get_datastore() self.config = hs.get_config() - @defer.inlineCallbacks - def process_v2_response(self, from_server, response_json, time_added_ms): + async def process_v2_response(self, from_server, response_json, time_added_ms): """Parse a 'Server Keys' structure from the result of a /key request This is used to parse either the entirety of the response from @@ -537,7 +551,7 @@ class BaseV2KeyFetcher(object): key_json_bytes = encode_canonical_json(response_json) - yield make_deferred_yieldable( + await make_deferred_yieldable( defer.gatherResults( [ run_in_background( @@ -562,19 +576,17 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): """KeyFetcher impl which fetches keys from the "perspectives" servers""" def __init__(self, hs): - super(PerspectivesKeyFetcher, self).__init__(hs) + super().__init__(hs) self.clock = hs.get_clock() self.client = hs.get_http_client() self.key_servers = self.config.key_servers - @defer.inlineCallbacks - def get_keys(self, keys_to_fetch): + async def get_keys(self, keys_to_fetch): """see KeyFetcher.get_keys""" - @defer.inlineCallbacks - def get_key(key_server): + async def get_key(key_server): try: - result = yield self.get_server_verify_key_v2_indirect( + result = await self.get_server_verify_key_v2_indirect( keys_to_fetch, key_server ) return result @@ -592,7 +604,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): return {} - results = yield make_deferred_yieldable( + results = await make_deferred_yieldable( defer.gatherResults( [run_in_background(get_key, server) for server in self.key_servers], consumeErrors=True, @@ -606,8 +618,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): return union_of_keys - @defer.inlineCallbacks - def get_server_verify_key_v2_indirect(self, keys_to_fetch, key_server): + async def get_server_verify_key_v2_indirect(self, keys_to_fetch, key_server): """ Args: keys_to_fetch (dict[str, dict[str, int]]): @@ -617,7 +628,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): the keys Returns: - Deferred[dict[str, dict[str, synapse.storage.keys.FetchKeyResult]]]: map + dict[str, dict[str, synapse.storage.keys.FetchKeyResult]]: map from server_name -> key_id -> FetchKeyResult Raises: @@ -632,7 +643,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): ) try: - query_response = yield self.client.post_json( + query_response = await self.client.post_json( destination=perspective_name, path="/_matrix/key/v2/query", data={ @@ -668,7 +679,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): try: self._validate_perspectives_response(key_server, response) - processed_response = yield self.process_v2_response( + processed_response = await self.process_v2_response( perspective_name, response, time_added_ms=time_now_ms ) except KeyLookupError as e: @@ -687,7 +698,7 @@ class PerspectivesKeyFetcher(BaseV2KeyFetcher): ) keys.setdefault(server_name, {}).update(processed_response) - yield self.store.store_server_verify_keys( + await self.store.store_server_verify_keys( perspective_name, time_now_ms, added_keys ) @@ -735,28 +746,27 @@ class ServerKeyFetcher(BaseV2KeyFetcher): """KeyFetcher impl which fetches keys from the origin servers""" def __init__(self, hs): - super(ServerKeyFetcher, self).__init__(hs) + super().__init__(hs) self.clock = hs.get_clock() self.client = hs.get_http_client() - def get_keys(self, keys_to_fetch): + async def get_keys(self, keys_to_fetch): """ Args: keys_to_fetch (dict[str, iterable[str]]): the keys to be fetched. server_name -> key_ids Returns: - Deferred[dict[str, dict[str, synapse.storage.keys.FetchKeyResult|None]]]: + dict[str, dict[str, synapse.storage.keys.FetchKeyResult|None]]: map from server_name -> key_id -> FetchKeyResult """ results = {} - @defer.inlineCallbacks - def get_key(key_to_fetch_item): + async def get_key(key_to_fetch_item): server_name, key_ids = key_to_fetch_item try: - keys = yield self.get_server_verify_key_v2_direct(server_name, key_ids) + keys = await self.get_server_verify_key_v2_direct(server_name, key_ids) results[server_name] = keys except KeyLookupError as e: logger.warning( @@ -765,12 +775,10 @@ class ServerKeyFetcher(BaseV2KeyFetcher): except Exception: logger.exception("Error getting keys %s from %s", key_ids, server_name) - return yieldable_gather_results(get_key, keys_to_fetch.items()).addCallback( - lambda _: results - ) + await yieldable_gather_results(get_key, keys_to_fetch.items()) + return results - @defer.inlineCallbacks - def get_server_verify_key_v2_direct(self, server_name, key_ids): + async def get_server_verify_key_v2_direct(self, server_name, key_ids): """ Args: @@ -778,7 +786,7 @@ class ServerKeyFetcher(BaseV2KeyFetcher): key_ids (iterable[str]): Returns: - Deferred[dict[str, FetchKeyResult]]: map from key ID to lookup result + dict[str, FetchKeyResult]: map from key ID to lookup result Raises: KeyLookupError if there was a problem making the lookup @@ -792,7 +800,7 @@ class ServerKeyFetcher(BaseV2KeyFetcher): time_now_ms = self.clock.time_msec() try: - response = yield self.client.get_json( + response = await self.client.get_json( destination=server_name, path="/_matrix/key/v2/server/" + urllib.parse.quote(requested_key_id), @@ -823,12 +831,12 @@ class ServerKeyFetcher(BaseV2KeyFetcher): % (server_name, response["server_name"]) ) - response_keys = yield self.process_v2_response( + response_keys = await self.process_v2_response( from_server=server_name, response_json=response, time_added_ms=time_now_ms, ) - yield self.store.store_server_verify_keys( + await self.store.store_server_verify_keys( server_name, time_now_ms, ((server_name, key_id, key) for key_id, key in response_keys.items()), @@ -838,22 +846,18 @@ class ServerKeyFetcher(BaseV2KeyFetcher): return keys -@defer.inlineCallbacks -def _handle_key_deferred(verify_request): +async def _handle_key_deferred(verify_request) -> None: """Waits for the key to become available, and then performs a verification Args: verify_request (VerifyJsonRequest): - Returns: - Deferred[None] - Raises: SynapseError if there was a problem performing the verification """ server_name = verify_request.server_name with PreserveLoggingContext(): - _, key_id, verify_key = yield verify_request.key_ready + _, key_id, verify_key = await verify_request.key_ready json_object = verify_request.json_object diff --git a/synapse/event_auth.py b/synapse/event_auth.py
index c0981eee62..8c907ad596 100644 --- a/synapse/event_auth.py +++ b/synapse/event_auth.py
@@ -47,7 +47,7 @@ def check( Args: room_version_obj: the version of the room event: the event being checked. - auth_events (dict: event-key -> event): the existing room state. + auth_events: the existing room state. Raises: AuthError if the checks fail diff --git a/synapse/events/__init__.py b/synapse/events/__init__.py
index cc5deca75b..dc49df0812 100644 --- a/synapse/events/__init__.py +++ b/synapse/events/__init__.py
@@ -18,12 +18,12 @@ import abc import os from distutils.util import strtobool -from typing import Dict, Optional, Type +from typing import Dict, Optional, Tuple, Type from unpaddedbase64 import encode_base64 from synapse.api.room_versions import EventFormatVersions, RoomVersion, RoomVersions -from synapse.types import JsonDict +from synapse.types import JsonDict, RoomStreamToken from synapse.util.caches import intern_dict from synapse.util.frozenutils import freeze @@ -96,7 +96,7 @@ class DefaultDictProperty(DictProperty): return instance._dict.get(self.key, self.default) -class _EventInternalMetadata(object): +class _EventInternalMetadata: __slots__ = ["_dict"] def __init__(self, internal_metadata_dict: JsonDict): @@ -118,9 +118,9 @@ class _EventInternalMetadata(object): # XXX: These are set by StreamWorkerStore._set_before_and_after. # I'm pretty sure that these are never persisted to the database, so shouldn't # be here - before = DictProperty("before") # type: str - after = DictProperty("after") # type: str - order = DictProperty("order") # type: int + before = DictProperty("before") # type: RoomStreamToken + after = DictProperty("after") # type: RoomStreamToken + order = DictProperty("order") # type: Tuple[int, int] def get_dict(self) -> JsonDict: return dict(self._dict) @@ -133,6 +133,8 @@ class _EventInternalMetadata(object): rejection. This is needed as those events are marked as outliers, but they still need to be processed as if they're new events (e.g. updating invite state in the database, relaying to clients, etc). + + (Added in synapse 0.99.0, so may be unreliable for events received before that) """ return self._dict.get("out_of_band_membership", False) diff --git a/synapse/events/builder.py b/synapse/events/builder.py
index 0bb216419a..b6c47be646 100644 --- a/synapse/events/builder.py +++ b/synapse/events/builder.py
@@ -12,13 +12,12 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from typing import Optional +from typing import Any, Dict, List, Optional, Tuple, Union import attr from nacl.signing import SigningKey -from twisted.internet import defer - +from synapse.api.auth import Auth from synapse.api.constants import MAX_DEPTH from synapse.api.errors import UnsupportedRoomVersionError from synapse.api.room_versions import ( @@ -29,13 +28,15 @@ from synapse.api.room_versions import ( ) from synapse.crypto.event_signing import add_hashes_and_signatures from synapse.events import EventBase, _EventInternalMetadata, make_event_from_dict +from synapse.state import StateHandler +from synapse.storage.databases.main import DataStore from synapse.types import EventID, JsonDict from synapse.util import Clock from synapse.util.stringutils import random_string @attr.s(slots=True, cmp=False, frozen=True) -class EventBuilder(object): +class EventBuilder: """A format independent event builder used to build up the event content before signing the event. @@ -44,45 +45,46 @@ class EventBuilder(object): Attributes: room_version: Version of the target room - room_id (str) - type (str) - sender (str) - content (dict) - unsigned (dict) - internal_metadata (_EventInternalMetadata) - - _state (StateHandler) - _auth (synapse.api.Auth) - _store (DataStore) - _clock (Clock) - _hostname (str): The hostname of the server creating the event + room_id + type + sender + content + unsigned + internal_metadata + + _state + _auth + _store + _clock + _hostname: The hostname of the server creating the event _signing_key: The signing key to use to sign the event as the server """ - _state = attr.ib() - _auth = attr.ib() - _store = attr.ib() - _clock = attr.ib() - _hostname = attr.ib() - _signing_key = attr.ib() + _state = attr.ib(type=StateHandler) + _auth = attr.ib(type=Auth) + _store = attr.ib(type=DataStore) + _clock = attr.ib(type=Clock) + _hostname = attr.ib(type=str) + _signing_key = attr.ib(type=SigningKey) room_version = attr.ib(type=RoomVersion) - room_id = attr.ib() - type = attr.ib() - sender = attr.ib() + room_id = attr.ib(type=str) + type = attr.ib(type=str) + sender = attr.ib(type=str) - content = attr.ib(default=attr.Factory(dict)) - unsigned = attr.ib(default=attr.Factory(dict)) + content = attr.ib(default=attr.Factory(dict), type=JsonDict) + unsigned = attr.ib(default=attr.Factory(dict), type=JsonDict) # These only exist on a subset of events, so they raise AttributeError if # someone tries to get them when they don't exist. - _state_key = attr.ib(default=None) - _redacts = attr.ib(default=None) - _origin_server_ts = attr.ib(default=None) + _state_key = attr.ib(default=None, type=Optional[str]) + _redacts = attr.ib(default=None, type=Optional[str]) + _origin_server_ts = attr.ib(default=None, type=Optional[int]) internal_metadata = attr.ib( - default=attr.Factory(lambda: _EventInternalMetadata({})) + default=attr.Factory(lambda: _EventInternalMetadata({})), + type=_EventInternalMetadata, ) @property @@ -95,31 +97,35 @@ class EventBuilder(object): def is_state(self): return self._state_key is not None - @defer.inlineCallbacks - def build(self, prev_event_ids): + async def build(self, prev_event_ids: List[str]) -> EventBase: """Transform into a fully signed and hashed event Args: - prev_event_ids (list[str]): The event IDs to use as the prev events + prev_event_ids: The event IDs to use as the prev events Returns: - Deferred[FrozenEvent] + The signed and hashed event. """ - state_ids = yield defer.ensureDeferred( - self._state.get_current_state_ids(self.room_id, prev_event_ids) + state_ids = await self._state.get_current_state_ids( + self.room_id, prev_event_ids ) - auth_ids = yield self._auth.compute_auth_events(self, state_ids) + auth_ids = self._auth.compute_auth_events(self, state_ids) format_version = self.room_version.event_format if format_version == EventFormatVersions.V1: - auth_events = yield self._store.add_event_hashes(auth_ids) - prev_events = yield self._store.add_event_hashes(prev_event_ids) + # The types of auth/prev events changes between event versions. + auth_events = await self._store.add_event_hashes( + auth_ids + ) # type: Union[List[str], List[Tuple[str, Dict[str, str]]]] + prev_events = await self._store.add_event_hashes( + prev_event_ids + ) # type: Union[List[str], List[Tuple[str, Dict[str, str]]]] else: auth_events = auth_ids prev_events = prev_event_ids - old_depth = yield self._store.get_max_depth_of(prev_event_ids) + old_depth = await self._store.get_max_depth_of(prev_event_ids) depth = old_depth + 1 # we cap depth of generated events, to ensure that they are not @@ -137,7 +143,7 @@ class EventBuilder(object): "unsigned": self.unsigned, "depth": depth, "prev_state": [], - } + } # type: Dict[str, Any] if self.is_state(): event_dict["state_key"] = self._state_key @@ -158,7 +164,7 @@ class EventBuilder(object): ) -class EventBuilderFactory(object): +class EventBuilderFactory: def __init__(self, hs): self.clock = hs.get_clock() self.hostname = hs.hostname diff --git a/synapse/events/snapshot.py b/synapse/events/snapshot.py
index f94cdcbaba..afecafe15c 100644 --- a/synapse/events/snapshot.py +++ b/synapse/events/snapshot.py
@@ -12,17 +12,19 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from typing import Optional, Union +from typing import TYPE_CHECKING, Optional, Union import attr from frozendict import frozendict -from twisted.internet import defer - from synapse.appservice import ApplicationService +from synapse.events import EventBase from synapse.logging.context import make_deferred_yieldable, run_in_background from synapse.types import StateMap +if TYPE_CHECKING: + from synapse.storage.databases.main import DataStore + @attr.s(slots=True) class EventContext: @@ -129,8 +131,7 @@ class EventContext: delta_ids=delta_ids, ) - @defer.inlineCallbacks - def serialize(self, event, store): + async def serialize(self, event: EventBase, store: "DataStore") -> dict: """Converts self to a type that can be serialized as JSON, and then deserialized by `deserialize` @@ -146,7 +147,7 @@ class EventContext: # the prev_state_ids, so if we're a state event we include the event # id that we replaced in the state. if event.is_state(): - prev_state_ids = yield self.get_prev_state_ids() + prev_state_ids = await self.get_prev_state_ids() prev_state_id = prev_state_ids.get((event.type, event.state_key)) else: prev_state_id = None @@ -214,8 +215,7 @@ class EventContext: return self._state_group - @defer.inlineCallbacks - def get_current_state_ids(self): + async def get_current_state_ids(self) -> Optional[StateMap[str]]: """ Gets the room state map, including this event - ie, the state in ``state_group`` @@ -224,32 +224,31 @@ class EventContext: ``rejected`` is set. Returns: - Deferred[dict[(str, str), str]|None]: Returns None if state_group - is None, which happens when the associated event is an outlier. + Returns None if state_group is None, which happens when the associated + event is an outlier. - Maps a (type, state_key) to the event ID of the state event matching - this tuple. + Maps a (type, state_key) to the event ID of the state event matching + this tuple. """ if self.rejected: raise RuntimeError("Attempt to access state_ids of rejected event") - yield self._ensure_fetched() + await self._ensure_fetched() return self._current_state_ids - @defer.inlineCallbacks - def get_prev_state_ids(self): + async def get_prev_state_ids(self): """ Gets the room state map, excluding this event. For a non-state event, this will be the same as get_current_state_ids(). Returns: - Deferred[dict[(str, str), str]|None]: Returns None if state_group + dict[(str, str), str]|None: Returns None if state_group is None, which happens when the associated event is an outlier. Maps a (type, state_key) to the event ID of the state event matching this tuple. """ - yield self._ensure_fetched() + await self._ensure_fetched() return self._prev_state_ids def get_cached_current_state_ids(self): @@ -269,8 +268,8 @@ class EventContext: return self._current_state_ids - def _ensure_fetched(self): - return defer.succeed(None) + async def _ensure_fetched(self): + return None @attr.s(slots=True) @@ -303,21 +302,20 @@ class _AsyncEventContextImpl(EventContext): _event_state_key = attr.ib(default=None) _fetching_state_deferred = attr.ib(default=None) - def _ensure_fetched(self): + async def _ensure_fetched(self): if not self._fetching_state_deferred: self._fetching_state_deferred = run_in_background(self._fill_out_state) - return make_deferred_yieldable(self._fetching_state_deferred) + return await make_deferred_yieldable(self._fetching_state_deferred) - @defer.inlineCallbacks - def _fill_out_state(self): + async def _fill_out_state(self): """Called to populate the _current_state_ids and _prev_state_ids attributes by loading from the database. """ if self.state_group is None: return - self._current_state_ids = yield self._storage.state.get_state_ids_for_group( + self._current_state_ids = await self._storage.state.get_state_ids_for_group( self.state_group ) if self._event_state_key is not None: diff --git a/synapse/events/spamcheck.py b/synapse/events/spamcheck.py
index 3c11e317fd..d141519ec4 100644 --- a/synapse/events/spamcheck.py +++ b/synapse/events/spamcheck.py
@@ -15,16 +15,17 @@ # limitations under the License. import inspect -from typing import Any, Dict, List, Optional +from typing import Any, Dict, List, Optional, Tuple -from synapse.spam_checker_api import SpamCheckerApi +from synapse.spam_checker_api import RegistrationBehaviour, SpamCheckerApi +from synapse.types import Collection MYPY = False if MYPY: import synapse.server -class SpamChecker(object): +class SpamChecker: def __init__(self, hs: "synapse.server.HomeServer"): self.spam_checkers = [] # type: List[Any] @@ -219,3 +220,33 @@ class SpamChecker(object): return True return False + + def check_registration_for_spam( + self, + email_threepid: Optional[dict], + username: Optional[str], + request_info: Collection[Tuple[str, str]], + ) -> RegistrationBehaviour: + """Checks if we should allow the given registration request. + + Args: + email_threepid: The email threepid used for registering, if any + username: The request user name, if any + request_info: List of tuples of user agent and IP that + were used during the registration process. + + Returns: + Enum for how the request should be handled + """ + + for spam_checker in self.spam_checkers: + # For backwards compatibility, only run if the method exists on the + # spam checker + checker = getattr(spam_checker, "check_registration_for_spam", None) + if checker: + behaviour = checker(email_threepid, username, request_info) + assert isinstance(behaviour, RegistrationBehaviour) + if behaviour != RegistrationBehaviour.ALLOW: + return behaviour + + return RegistrationBehaviour.ALLOW diff --git a/synapse/events/third_party_rules.py b/synapse/events/third_party_rules.py
index 317a6eb638..1ca77519d5 100644 --- a/synapse/events/third_party_rules.py +++ b/synapse/events/third_party_rules.py
@@ -15,11 +15,12 @@ from typing import Callable from synapse.events import EventBase +from synapse.events.snapshot import EventContext from synapse.module_api import ModuleApi -from synapse.types import StateMap +from synapse.types import Requester, StateMap -class ThirdPartyEventRules(object): +class ThirdPartyEventRules: """Allows server admins to provide a Python module implementing an extra set of rules to apply when processing events. @@ -42,15 +43,17 @@ class ThirdPartyEventRules(object): config=config, module_api=ModuleApi(hs, hs.get_auth_handler()), ) - async def check_event_allowed(self, event, context): + async def check_event_allowed( + self, event: EventBase, context: EventContext + ) -> bool: """Check if a provided event should be allowed in the given context. Args: - event (synapse.events.EventBase): The event to be checked. - context (synapse.events.snapshot.EventContext): The context of the event. + event: The event to be checked. + context: The context of the event. Returns: - defer.Deferred[bool]: True if the event should be allowed, False if not. + True if the event should be allowed, False if not. """ if self.third_party_rules is None: return True @@ -65,17 +68,19 @@ class ThirdPartyEventRules(object): ret = await self.third_party_rules.check_event_allowed(event, state_events) return ret - async def on_create_room(self, requester, config, is_requester_admin): + async def on_create_room( + self, requester: Requester, config: dict, is_requester_admin: bool + ) -> bool: """Intercept requests to create room to allow, deny or update the request config. Args: - requester (Requester) - config (dict): The creation config from the client. - is_requester_admin (bool): If the requester is an admin + requester + config: The creation config from the client. + is_requester_admin: If the requester is an admin Returns: - defer.Deferred[bool]: Whether room creation is allowed or denied. + Whether room creation is allowed or denied. """ if self.third_party_rules is None: @@ -86,16 +91,18 @@ class ThirdPartyEventRules(object): ) return ret - async def check_threepid_can_be_invited(self, medium, address, room_id): + async def check_threepid_can_be_invited( + self, medium: str, address: str, room_id: str + ) -> bool: """Check if a provided 3PID can be invited in the given room. Args: - medium (str): The 3PID's medium. - address (str): The 3PID's address. - room_id (str): The room we want to invite the threepid to. + medium: The 3PID's medium. + address: The 3PID's address. + room_id: The room we want to invite the threepid to. Returns: - defer.Deferred[bool], True if the 3PID can be invited, False if not. + True if the 3PID can be invited, False if not. """ if self.third_party_rules is None: diff --git a/synapse/events/utils.py b/synapse/events/utils.py
index 11f0d34ec8..32c73d3413 100644 --- a/synapse/events/utils.py +++ b/synapse/events/utils.py
@@ -18,8 +18,6 @@ from typing import Any, Mapping, Union from frozendict import frozendict -from twisted.internet import defer - from synapse.api.constants import EventTypes, RelationTypes from synapse.api.errors import Codes, SynapseError from synapse.api.room_versions import RoomVersion @@ -324,7 +322,7 @@ def serialize_event( return d -class EventClientSerializer(object): +class EventClientSerializer: """Serializes events that are to be sent to clients. This is used for bundling extra information with any events to be sent to @@ -337,8 +335,9 @@ class EventClientSerializer(object): hs.config.experimental_msc1849_support_enabled ) - @defer.inlineCallbacks - def serialize_event(self, event, time_now, bundle_aggregations=True, **kwargs): + async def serialize_event( + self, event, time_now, bundle_aggregations=True, **kwargs + ): """Serializes a single event. Args: @@ -348,7 +347,7 @@ class EventClientSerializer(object): **kwargs: Arguments to pass to `serialize_event` Returns: - Deferred[dict]: The serialized event + dict: The serialized event """ # To handle the case of presence events and the like if not isinstance(event, EventBase): @@ -363,8 +362,8 @@ class EventClientSerializer(object): if not event.internal_metadata.is_redacted() and ( self.experimental_msc1849_support_enabled and bundle_aggregations ): - annotations = yield self.store.get_aggregation_groups_for_event(event_id) - references = yield self.store.get_relations_for_event( + annotations = await self.store.get_aggregation_groups_for_event(event_id) + references = await self.store.get_relations_for_event( event_id, RelationTypes.REFERENCE, direction="f" ) @@ -378,7 +377,7 @@ class EventClientSerializer(object): edit = None if event.type == EventTypes.Message: - edit = yield self.store.get_applicable_edit(event_id) + edit = await self.store.get_applicable_edit(event_id) if edit: # If there is an edit replace the content, preserving existing diff --git a/synapse/events/validator.py b/synapse/events/validator.py
index 588d222f36..9df35b54ba 100644 --- a/synapse/events/validator.py +++ b/synapse/events/validator.py
@@ -20,7 +20,7 @@ from synapse.events.utils import validate_canonicaljson from synapse.types import EventID, RoomID, UserID -class EventValidator(object): +class EventValidator: def validate_new(self, event, config): """Validates the event has roughly the right format @@ -74,15 +74,14 @@ class EventValidator(object): ) if event.type == EventTypes.Retention: - self._validate_retention(event, config) + self._validate_retention(event) - def _validate_retention(self, event, config): + def _validate_retention(self, event): """Checks that an event that defines the retention policy for a room respects the - boundaries imposed by the server's administrator. + format enforced by the spec. Args: event (FrozenEvent): The event to validate. - config (Config): The homeserver's configuration. """ min_lifetime = event.content.get("min_lifetime") max_lifetime = event.content.get("max_lifetime") @@ -95,32 +94,6 @@ class EventValidator(object): errcode=Codes.BAD_JSON, ) - if ( - config.retention_allowed_lifetime_min is not None - and min_lifetime < config.retention_allowed_lifetime_min - ): - raise SynapseError( - code=400, - msg=( - "'min_lifetime' can't be lower than the minimum allowed" - " value enforced by the server's administrator" - ), - errcode=Codes.BAD_JSON, - ) - - if ( - config.retention_allowed_lifetime_max is not None - and min_lifetime > config.retention_allowed_lifetime_max - ): - raise SynapseError( - code=400, - msg=( - "'min_lifetime' can't be greater than the maximum allowed" - " value enforced by the server's administrator" - ), - errcode=Codes.BAD_JSON, - ) - if max_lifetime is not None: if not isinstance(max_lifetime, int): raise SynapseError( @@ -129,32 +102,6 @@ class EventValidator(object): errcode=Codes.BAD_JSON, ) - if ( - config.retention_allowed_lifetime_min is not None - and max_lifetime < config.retention_allowed_lifetime_min - ): - raise SynapseError( - code=400, - msg=( - "'max_lifetime' can't be lower than the minimum allowed value" - " enforced by the server's administrator" - ), - errcode=Codes.BAD_JSON, - ) - - if ( - config.retention_allowed_lifetime_max is not None - and max_lifetime > config.retention_allowed_lifetime_max - ): - raise SynapseError( - code=400, - msg=( - "'max_lifetime' can't be greater than the maximum allowed" - " value enforced by the server's administrator" - ), - errcode=Codes.BAD_JSON, - ) - if ( min_lifetime is not None and max_lifetime is not None diff --git a/synapse/federation/federation_base.py b/synapse/federation/federation_base.py
index 420df2385f..38aa47963f 100644 --- a/synapse/federation/federation_base.py +++ b/synapse/federation/federation_base.py
@@ -39,7 +39,7 @@ from synapse.types import JsonDict, get_domain_from_id logger = logging.getLogger(__name__) -class FederationBase(object): +class FederationBase: def __init__(self, hs): self.hs = hs diff --git a/synapse/federation/federation_client.py b/synapse/federation/federation_client.py
index 994e6c8d5a..302b2f69bc 100644 --- a/synapse/federation/federation_client.py +++ b/synapse/federation/federation_client.py
@@ -24,10 +24,12 @@ from typing import ( Dict, Iterable, List, + Mapping, Optional, Sequence, Tuple, TypeVar, + Union, ) from prometheus_client import Counter @@ -54,7 +56,7 @@ from synapse.events import EventBase, builder from synapse.federation.federation_base import FederationBase, event_from_pdu_json from synapse.logging.context import make_deferred_yieldable, preserve_fn from synapse.logging.utils import log_function -from synapse.types import JsonDict +from synapse.types import JsonDict, get_domain_from_id from synapse.util import unwrapFirstError from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.retryutils import NotRetryingDestination @@ -79,7 +81,7 @@ class InvalidResponseError(RuntimeError): class FederationClient(FederationBase): def __init__(self, hs): - super(FederationClient, self).__init__(hs) + super().__init__(hs) self.pdu_destination_tried = {} self._clock.looping_call(self._clear_tried_cache, 60 * 1000) @@ -135,7 +137,7 @@ class FederationClient(FederationBase): and try the request anyway. Returns: - a Deferred which will eventually yield a JSON object from the + a Awaitable which will eventually yield a JSON object from the response """ sent_queries_counter.labels(query_type).inc() @@ -157,7 +159,7 @@ class FederationClient(FederationBase): content (dict): The query content. Returns: - a Deferred which will eventually yield a JSON object from the + an Awaitable which will eventually yield a JSON object from the response """ sent_queries_counter.labels("client_device_keys").inc() @@ -180,7 +182,7 @@ class FederationClient(FederationBase): content (dict): The query content. Returns: - a Deferred which will eventually yield a JSON object from the + an Awaitable which will eventually yield a JSON object from the response """ sent_queries_counter.labels("client_one_time_keys").inc() @@ -217,11 +219,9 @@ class FederationClient(FederationBase): for p in transaction_data["pdus"] ] - # FIXME: We should handle signature failures more gracefully. - pdus[:] = await make_deferred_yieldable( - defer.gatherResults( - self._check_sigs_and_hashes(room_version, pdus), consumeErrors=True, - ).addErrback(unwrapFirstError) + # Check signatures and hash of pdus, removing any from the list that fail checks + pdus[:] = await self._check_sigs_and_hash_and_fetch( + dest, pdus, outlier=True, room_version=room_version ) return pdus @@ -386,10 +386,11 @@ class FederationClient(FederationBase): pdu.event_id, allow_rejected=True, allow_none=True ) - if not res and pdu.origin != origin: + pdu_origin = get_domain_from_id(pdu.sender) + if not res and pdu_origin != origin: try: res = await self.get_pdu( - destinations=[pdu.origin], + destinations=[pdu_origin], event_id=pdu.event_id, room_version=room_version, outlier=outlier, @@ -502,7 +503,7 @@ class FederationClient(FederationBase): user_id: str, membership: str, content: dict, - params: Dict[str, str], + params: Optional[Mapping[str, Union[str, Iterable[str]]]], ) -> Tuple[str, EventBase, RoomVersion]: """ Creates an m.room.member event, with context, without participating in the room. @@ -900,7 +901,7 @@ class FederationClient(FederationBase): party instance Returns: - Deferred[Dict[str, Any]]: The response from the remote server, or None if + Awaitable[Dict[str, Any]]: The response from the remote server, or None if `remote_server` is the same as the local server_name Raises: diff --git a/synapse/federation/federation_server.py b/synapse/federation/federation_server.py
index 11c5d63298..24329dd0e3 100644 --- a/synapse/federation/federation_server.py +++ b/synapse/federation/federation_server.py
@@ -28,8 +28,7 @@ from typing import ( Union, ) -from canonicaljson import json -from prometheus_client import Counter, Histogram +from prometheus_client import Counter, Gauge, Histogram from twisted.internet import defer from twisted.internet.abstract import isIPAddress @@ -63,7 +62,7 @@ from synapse.replication.http.federation import ( ReplicationGetQueryRestServlet, ) from synapse.types import JsonDict, get_domain_from_id -from synapse.util import glob_to_regex, unwrapFirstError +from synapse.util import glob_to_regex, json_decoder, unwrapFirstError from synapse.util.async_helpers import Linearizer, concurrently_execute from synapse.util.caches.response_cache import ResponseCache @@ -89,19 +88,32 @@ pdu_process_time = Histogram( ) +last_pdu_age_metric = Gauge( + "synapse_federation_last_received_pdu_age", + "The age (in seconds) of the last PDU successfully received from the given domain", + labelnames=("server_name",), +) + + class FederationServer(FederationBase): def __init__(self, hs): - super(FederationServer, self).__init__(hs) + super().__init__(hs) self.auth = hs.get_auth() self.handler = hs.get_handlers().federation_handler self.state = hs.get_state_handler() self.device_handler = hs.get_device_handler() + self._federation_ratelimiter = hs.get_federation_ratelimiter() self._server_linearizer = Linearizer("fed_server") self._transaction_linearizer = Linearizer("fed_txn_handler") + # We cache results for transaction with the same ID + self._transaction_resp_cache = ResponseCache( + hs, "fed_txn_handler", timeout_ms=30000 + ) + self.transaction_actions = TransactionActions(self.store) self.registry = hs.get_federation_registry() @@ -113,6 +125,10 @@ class FederationServer(FederationBase): hs, "state_ids_resp", timeout_ms=30000 ) + self._federation_metrics_domains = ( + hs.get_config().federation.federation_metrics_domains + ) + async def on_backfill_request( self, origin: str, room_id: str, versions: List[str], limit: int ) -> Tuple[int, Dict[str, Any]]: @@ -136,22 +152,44 @@ class FederationServer(FederationBase): request_time = self._clock.time_msec() transaction = Transaction(**transaction_data) + transaction_id = transaction.transaction_id # type: ignore - if not transaction.transaction_id: # type: ignore + if not transaction_id: raise Exception("Transaction missing transaction_id") - logger.debug("[%s] Got transaction", transaction.transaction_id) # type: ignore + logger.debug("[%s] Got transaction", transaction_id) - # use a linearizer to ensure that we don't process the same transaction - # multiple times in parallel. - with ( - await self._transaction_linearizer.queue( - (origin, transaction.transaction_id) # type: ignore - ) - ): - result = await self._handle_incoming_transaction( - origin, transaction, request_time - ) + # We wrap in a ResponseCache so that we de-duplicate retried + # transactions. + return await self._transaction_resp_cache.wrap( + (origin, transaction_id), + self._on_incoming_transaction_inner, + origin, + transaction, + request_time, + ) + + async def _on_incoming_transaction_inner( + self, origin: str, transaction: Transaction, request_time: int + ) -> Tuple[int, Dict[str, Any]]: + # Use a linearizer to ensure that transactions from a remote are + # processed in order. + with await self._transaction_linearizer.queue(origin): + # We rate limit here *after* we've queued up the incoming requests, + # so that we don't fill up the ratelimiter with blocked requests. + # + # This is important as the ratelimiter allows N concurrent requests + # at a time, and only starts ratelimiting if there are more requests + # than that being processed at a time. If we queued up requests in + # the linearizer/response cache *after* the ratelimiting then those + # queued up requests would count as part of the allowed limit of N + # concurrent requests. + with self._federation_ratelimiter.ratelimit(origin) as d: + await d + + result = await self._handle_incoming_transaction( + origin, transaction, request_time + ) return result @@ -235,7 +273,11 @@ class FederationServer(FederationBase): pdus_by_room = {} # type: Dict[str, List[EventBase]] + newest_pdu_ts = 0 + for p in transaction.pdus: # type: ignore + # FIXME (richardv): I don't think this works: + # https://github.com/matrix-org/synapse/issues/8429 if "unsigned" in p: unsigned = p["unsigned"] if "age" in unsigned: @@ -273,6 +315,9 @@ class FederationServer(FederationBase): event = event_from_pdu_json(p, room_version) pdus_by_room.setdefault(room_id, []).append(event) + if event.origin_server_ts > newest_pdu_ts: + newest_pdu_ts = event.origin_server_ts + pdu_results = {} # we can process different rooms in parallel (which is useful if they @@ -313,6 +358,10 @@ class FederationServer(FederationBase): process_pdus_for_room, pdus_by_room.keys(), TRANSACTION_CONCURRENCY_LIMIT ) + if newest_pdu_ts and origin in self._federation_metrics_domains: + newest_pdu_age = self._clock.time_msec() - newest_pdu_ts + last_pdu_age_metric.labels(server_name=origin).set(newest_pdu_age / 1000) + return pdu_results async def _handle_edus_in_txn(self, origin: str, transaction: Transaction): @@ -551,7 +600,7 @@ class FederationServer(FederationBase): for device_id, keys in device_keys.items(): for key_id, json_str in keys.items(): json_result.setdefault(user_id, {})[device_id] = { - key_id: json.loads(json_str) + key_id: json_decoder.decode(json_str) } logger.info( @@ -786,7 +835,7 @@ def _acl_entry_matches(server_name: str, acl_entry: str) -> Match: return regex.match(server_name) -class FederationHandlerRegistry(object): +class FederationHandlerRegistry: """Allows classes to register themselves as handlers for a given EDU or query type for incoming federation traffic. """ diff --git a/synapse/federation/persistence.py b/synapse/federation/persistence.py
index d68b4bd670..079e2b2fe0 100644 --- a/synapse/federation/persistence.py +++ b/synapse/federation/persistence.py
@@ -20,13 +20,16 @@ These actions are mostly only used by the :py:mod:`.replication` module. """ import logging +from typing import Optional, Tuple +from synapse.federation.units import Transaction from synapse.logging.utils import log_function +from synapse.types import JsonDict logger = logging.getLogger(__name__) -class TransactionActions(object): +class TransactionActions: """ Defines persistence actions that relate to handling Transactions. """ @@ -34,30 +37,32 @@ class TransactionActions(object): self.store = datastore @log_function - def have_responded(self, origin, transaction): - """ Have we already responded to a transaction with the same id and + async def have_responded( + self, origin: str, transaction: Transaction + ) -> Optional[Tuple[int, JsonDict]]: + """Have we already responded to a transaction with the same id and origin? Returns: - Deferred: Results in `None` if we have not previously responded to - this transaction or a 2-tuple of `(int, dict)` representing the - response code and response body. + `None` if we have not previously responded to this transaction or a + 2-tuple of `(int, dict)` representing the response code and response body. """ - if not transaction.transaction_id: + transaction_id = transaction.transaction_id # type: ignore + if not transaction_id: raise RuntimeError("Cannot persist a transaction with no transaction_id") - return self.store.get_received_txn_response(transaction.transaction_id, origin) + return await self.store.get_received_txn_response(transaction_id, origin) @log_function - def set_response(self, origin, transaction, code, response): - """ Persist how we responded to a transaction. - - Returns: - Deferred + async def set_response( + self, origin: str, transaction: Transaction, code: int, response: JsonDict + ) -> None: + """Persist how we responded to a transaction. """ - if not transaction.transaction_id: + transaction_id = transaction.transaction_id # type: ignore + if not transaction_id: raise RuntimeError("Cannot persist a transaction with no transaction_id") - return self.store.set_received_txn_response( - transaction.transaction_id, origin, code, response + await self.store.set_received_txn_response( + transaction_id, origin, code, response ) diff --git a/synapse/federation/send_queue.py b/synapse/federation/send_queue.py
index 2b0ab2dcbf..8e46957d15 100644 --- a/synapse/federation/send_queue.py +++ b/synapse/federation/send_queue.py
@@ -37,8 +37,8 @@ from sortedcontainers import SortedDict from twisted.internet import defer +from synapse.api.presence import UserPresenceState from synapse.metrics import LaterGauge -from synapse.storage.presence import UserPresenceState from synapse.util.metrics import Measure from .units import Edu @@ -46,7 +46,7 @@ from .units import Edu logger = logging.getLogger(__name__) -class FederationRemoteSendQueue(object): +class FederationRemoteSendQueue: """A drop in replacement for FederationSender""" def __init__(self, hs): @@ -365,7 +365,7 @@ class FederationRemoteSendQueue(object): ) -class BaseFederationRow(object): +class BaseFederationRow: """Base class for rows to be sent in the federation stream. Specifies how to identify, serialize and deserialize the different types. diff --git a/synapse/federation/sender/__init__.py b/synapse/federation/sender/__init__.py
index 6ae6522f87..8bb17b3a05 100644 --- a/synapse/federation/sender/__init__.py +++ b/synapse/federation/sender/__init__.py
@@ -22,6 +22,7 @@ from twisted.internet import defer import synapse import synapse.metrics +from synapse.api.presence import UserPresenceState from synapse.events import EventBase from synapse.federation.sender.per_destination_queue import PerDestinationQueue from synapse.federation.sender.transaction_manager import TransactionManager @@ -39,7 +40,6 @@ from synapse.metrics import ( events_processed_counter, ) from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.storage.presence import UserPresenceState from synapse.types import ReadReceipt from synapse.util.metrics import Measure, measure_func @@ -55,8 +55,17 @@ sent_pdus_destination_dist_total = Counter( "Total number of PDUs queued for sending across all destinations", ) +# Time (in s) after Synapse's startup that we will begin to wake up destinations +# that have catch-up outstanding. +CATCH_UP_STARTUP_DELAY_SEC = 15 -class FederationSender(object): +# Time (in s) to wait in between waking up each destination, i.e. one destination +# will be woken up every <x> seconds after Synapse's startup until we have woken +# every destination has outstanding catch-up. +CATCH_UP_STARTUP_INTERVAL_SEC = 5 + + +class FederationSender: def __init__(self, hs: "synapse.server.HomeServer"): self.hs = hs self.server_name = hs.hostname @@ -108,8 +117,6 @@ class FederationSender(object): ), ) - self._order = 1 - self._is_processing = False self._last_poked_id = -1 @@ -127,6 +134,14 @@ class FederationSender(object): 1000.0 / hs.config.federation_rr_transactions_per_room_per_second ) + # wake up destinations that have outstanding PDUs to be caught up + self._catchup_after_startup_timer = self.clock.call_later( + CATCH_UP_STARTUP_DELAY_SEC, + run_as_background_process, + "wake_destinations_needing_catchup", + self._wake_destinations_needing_catchup, + ) + def _get_per_destination_queue(self, destination: str) -> PerDestinationQueue: """Get or create a PerDestinationQueue for the given destination @@ -211,7 +226,7 @@ class FederationSender(object): logger.debug("Sending %s to %r", event, destinations) if destinations: - self._send_pdu(event, destinations) + await self._send_pdu(event, destinations) now = self.clock.time_msec() ts = await self.store.get_received_ts(event.event_id) @@ -267,14 +282,11 @@ class FederationSender(object): finally: self._is_processing = False - def _send_pdu(self, pdu: EventBase, destinations: Iterable[str]) -> None: + async def _send_pdu(self, pdu: EventBase, destinations: Iterable[str]) -> None: # We loop through all destinations to see whether we already have # a transaction in progress. If we do, stick it in the pending_pdus # table and we'll get back to it later. - order = self._order - self._order += 1 - destinations = set(destinations) destinations.discard(self.server_name) logger.debug("Sending to: %s", str(destinations)) @@ -285,11 +297,17 @@ class FederationSender(object): sent_pdus_destination_dist_total.inc(len(destinations)) sent_pdus_destination_dist_count.inc() + # track the fact that we have a PDU for these destinations, + # to allow us to perform catch-up later on if the remote is unreachable + # for a while. + await self.store.store_destination_rooms_entries( + destinations, pdu.room_id, pdu.internal_metadata.stream_ordering, + ) + for destination in destinations: - self._get_per_destination_queue(destination).send_pdu(pdu, order) + self._get_per_destination_queue(destination).send_pdu(pdu) - @defer.inlineCallbacks - def send_read_receipt(self, receipt: ReadReceipt): + async def send_read_receipt(self, receipt: ReadReceipt) -> None: """Send a RR to any other servers in the room Args: @@ -330,12 +348,10 @@ class FederationSender(object): room_id = receipt.room_id # Work out which remote servers should be poked and poke them. - domains = yield defer.ensureDeferred( - self.state.get_current_hosts_in_room(room_id) - ) + domains_set = await self.state.get_current_hosts_in_room(room_id) domains = [ d - for d in domains + for d in domains_set if d != self.server_name and self._federation_shard_config.should_handle(self._instance_name, d) ] @@ -387,8 +403,7 @@ class FederationSender(object): queue.flush_read_receipts_for_room(room_id) @preserve_fn # the caller should not yield on this - @defer.inlineCallbacks - def send_presence(self, states: List[UserPresenceState]): + async def send_presence(self, states: List[UserPresenceState]): """Send the new presence states to the appropriate destinations. This actually queues up the presence states ready for sending and @@ -423,7 +438,7 @@ class FederationSender(object): if not states_map: break - yield self._process_presence_inner(list(states_map.values())) + await self._process_presence_inner(list(states_map.values())) except Exception: logger.exception("Error sending presence states to servers") finally: @@ -450,14 +465,11 @@ class FederationSender(object): self._get_per_destination_queue(destination).send_presence(states) @measure_func("txnqueue._process_presence") - @defer.inlineCallbacks - def _process_presence_inner(self, states: List[UserPresenceState]): + async def _process_presence_inner(self, states: List[UserPresenceState]): """Given a list of states populate self.pending_presence_by_dest and poke to send a new transaction to each destination """ - hosts_and_states = yield defer.ensureDeferred( - get_interested_remotes(self.store, states, self.state) - ) + hosts_and_states = await get_interested_remotes(self.store, states, self.state) for destinations, states in hosts_and_states: for destination in destinations: @@ -565,3 +577,37 @@ class FederationSender(object): # Dummy implementation for case where federation sender isn't offloaded # to a worker. return [], 0, False + + async def _wake_destinations_needing_catchup(self): + """ + Wakes up destinations that need catch-up and are not currently being + backed off from. + + In order to reduce load spikes, adds a delay between each destination. + """ + + last_processed = None # type: Optional[str] + + while True: + destinations_to_wake = await self.store.get_catch_up_outstanding_destinations( + last_processed + ) + + if not destinations_to_wake: + # finished waking all destinations! + self._catchup_after_startup_timer = None + break + + destinations_to_wake = [ + d + for d in destinations_to_wake + if self._federation_shard_config.should_handle(self._instance_name, d) + ] + + for last_processed in destinations_to_wake: + logger.info( + "Destination %s has outstanding catch-up, waking up.", + last_processed, + ) + self.wake_destination(last_processed) + await self.clock.sleep(CATCH_UP_STARTUP_INTERVAL_SEC) diff --git a/synapse/federation/sender/per_destination_queue.py b/synapse/federation/sender/per_destination_queue.py
index dd150f89a6..bc99af3fdd 100644 --- a/synapse/federation/sender/per_destination_queue.py +++ b/synapse/federation/sender/per_destination_queue.py
@@ -15,7 +15,7 @@ # limitations under the License. import datetime import logging -from typing import TYPE_CHECKING, Dict, Hashable, Iterable, List, Tuple +from typing import TYPE_CHECKING, Dict, Hashable, Iterable, List, Optional, Tuple, cast from prometheus_client import Counter @@ -24,12 +24,12 @@ from synapse.api.errors import ( HttpResponseException, RequestSendFailed, ) +from synapse.api.presence import UserPresenceState from synapse.events import EventBase from synapse.federation.units import Edu from synapse.handlers.presence import format_user_presence_state from synapse.metrics import sent_transactions_counter from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.storage.presence import UserPresenceState from synapse.types import ReadReceipt from synapse.util.retryutils import NotRetryingDestination, get_retry_limiter @@ -53,7 +53,7 @@ sent_edus_by_type = Counter( ) -class PerDestinationQueue(object): +class PerDestinationQueue: """ Manages the per-destination transmission queues. @@ -92,8 +92,23 @@ class PerDestinationQueue(object): self._destination = destination self.transmission_loop_running = False - # a list of tuples of (pending pdu, order) - self._pending_pdus = [] # type: List[Tuple[EventBase, int]] + # True whilst we are sending events that the remote homeserver missed + # because it was unreachable. We start in this state so we can perform + # catch-up at startup. + # New events will only be sent once this is finished, at which point + # _catching_up is flipped to False. + self._catching_up = True # type: bool + + # The stream_ordering of the most recent PDU that was discarded due to + # being in catch-up mode. + self._catchup_last_skipped = 0 # type: int + + # Cache of the last successfully-transmitted stream ordering for this + # destination (we are the only updater so this is safe) + self._last_successful_stream_ordering = None # type: Optional[int] + + # a list of pending PDUs + self._pending_pdus = [] # type: List[EventBase] # XXX this is never actually used: see # https://github.com/matrix-org/synapse/issues/7549 @@ -132,14 +147,19 @@ class PerDestinationQueue(object): + len(self._pending_edus_keyed) ) - def send_pdu(self, pdu: EventBase, order: int) -> None: + def send_pdu(self, pdu: EventBase) -> None: """Add a PDU to the queue, and start the transmission loop if necessary Args: pdu: pdu to send - order """ - self._pending_pdus.append((pdu, order)) + if not self._catching_up or self._last_successful_stream_ordering is None: + # only enqueue the PDU if we are not catching up (False) or do not + # yet know if we have anything to catch up (None) + self._pending_pdus.append(pdu) + else: + self._catchup_last_skipped = pdu.internal_metadata.stream_ordering + self.attempt_new_transaction() def send_presence(self, states: Iterable[UserPresenceState]) -> None: @@ -185,7 +205,7 @@ class PerDestinationQueue(object): returns immediately. Otherwise kicks off the process of sending a transaction in the background. """ - # list of (pending_pdu, deferred, order) + if self.transmission_loop_running: # XXX: this can get stuck on by a never-ending # request at which point pending_pdus just keeps growing. @@ -210,7 +230,7 @@ class PerDestinationQueue(object): ) async def _transaction_transmission_loop(self) -> None: - pending_pdus = [] # type: List[Tuple[EventBase, int]] + pending_pdus = [] # type: List[EventBase] try: self.transmission_loop_running = True @@ -219,6 +239,13 @@ class PerDestinationQueue(object): # hence why we throw the result away. await get_retry_limiter(self._destination, self._clock, self._store) + if self._catching_up: + # we potentially need to catch-up first + await self._catch_up_transmission_loop() + if self._catching_up: + # not caught up yet + return + pending_pdus = [] while True: # We have to keep 2 free slots for presence and rr_edus @@ -326,6 +353,17 @@ class PerDestinationQueue(object): self._last_device_stream_id = device_stream_id self._last_device_list_stream_id = dev_list_id + + if pending_pdus: + # we sent some PDUs and it was successful, so update our + # last_successful_stream_ordering in the destinations table. + final_pdu = pending_pdus[-1] + last_successful_stream_ordering = ( + final_pdu.internal_metadata.stream_ordering + ) + await self._store.set_destination_last_successful_stream_ordering( + self._destination, last_successful_stream_ordering + ) else: break except NotRetryingDestination as e: @@ -337,6 +375,30 @@ class PerDestinationQueue(object): (e.retry_last_ts + e.retry_interval) / 1000.0 ), ) + + if e.retry_interval > 60 * 60 * 1000: + # we won't retry for another hour! + # (this suggests a significant outage) + # We drop pending EDUs because otherwise they will + # rack up indefinitely. + # (Dropping PDUs is already performed by `_start_catching_up`.) + # Note that: + # - the EDUs that are being dropped here are those that we can + # afford to drop (specifically, only typing notifications, + # read receipts and presence updates are being dropped here) + # - Other EDUs such as to_device messages are queued with a + # different mechanism + # - this is all volatile state that would be lost if the + # federation sender restarted anyway + + # dropping read receipts is a bit sad but should be solved + # through another mechanism, because this is all volatile! + self._pending_edus = [] + self._pending_edus_keyed = {} + self._pending_presence = {} + self._pending_rrs = {} + + self._start_catching_up() except FederationDeniedError as e: logger.info(e) except HttpResponseException as e: @@ -346,25 +408,107 @@ class PerDestinationQueue(object): e.code, e, ) + + self._start_catching_up() except RequestSendFailed as e: logger.warning( "TX [%s] Failed to send transaction: %s", self._destination, e ) - for p, _ in pending_pdus: + for p in pending_pdus: logger.info( "Failed to send event %s to %s", p.event_id, self._destination ) + + self._start_catching_up() except Exception: logger.exception("TX [%s] Failed to send transaction", self._destination) - for p, _ in pending_pdus: + for p in pending_pdus: logger.info( "Failed to send event %s to %s", p.event_id, self._destination ) + + self._start_catching_up() finally: # We want to be *very* sure we clear this after we stop processing self.transmission_loop_running = False + async def _catch_up_transmission_loop(self) -> None: + first_catch_up_check = self._last_successful_stream_ordering is None + + if first_catch_up_check: + # first catchup so get last_successful_stream_ordering from database + self._last_successful_stream_ordering = await self._store.get_destination_last_successful_stream_ordering( + self._destination + ) + + if self._last_successful_stream_ordering is None: + # if it's still None, then this means we don't have the information + # in our database ­ we haven't successfully sent a PDU to this server + # (at least since the introduction of the feature tracking + # last_successful_stream_ordering). + # Sadly, this means we can't do anything here as we don't know what + # needs catching up — so catching up is futile; let's stop. + self._catching_up = False + return + + # get at most 50 catchup room/PDUs + while True: + event_ids = await self._store.get_catch_up_room_event_ids( + self._destination, self._last_successful_stream_ordering, + ) + + if not event_ids: + # No more events to catch up on, but we can't ignore the chance + # of a race condition, so we check that no new events have been + # skipped due to us being in catch-up mode + + if self._catchup_last_skipped > self._last_successful_stream_ordering: + # another event has been skipped because we were in catch-up mode + continue + + # we are done catching up! + self._catching_up = False + break + + if first_catch_up_check: + # as this is our check for needing catch-up, we may have PDUs in + # the queue from before we *knew* we had to do catch-up, so + # clear those out now. + self._start_catching_up() + + # fetch the relevant events from the event store + # - redacted behaviour of REDACT is fine, since we only send metadata + # of redacted events to the destination. + # - don't need to worry about rejected events as we do not actively + # forward received events over federation. + catchup_pdus = await self._store.get_events_as_list(event_ids) + if not catchup_pdus: + raise AssertionError( + "No events retrieved when we asked for %r. " + "This should not happen." % event_ids + ) + + if logger.isEnabledFor(logging.INFO): + rooms = [p.room_id for p in catchup_pdus] + logger.info("Catching up rooms to %s: %r", self._destination, rooms) + + success = await self._transaction_manager.send_new_transaction( + self._destination, catchup_pdus, [] + ) + + if not success: + return + + sent_transactions_counter.inc() + final_pdu = catchup_pdus[-1] + self._last_successful_stream_ordering = cast( + int, final_pdu.internal_metadata.stream_ordering + ) + await self._store.set_destination_last_successful_stream_ordering( + self._destination, self._last_successful_stream_ordering + ) + def _get_rr_edus(self, force_flush: bool) -> Iterable[Edu]: if not self._pending_rrs: return @@ -425,3 +569,12 @@ class PerDestinationQueue(object): ] return (edus, stream_id) + + def _start_catching_up(self) -> None: + """ + Marks this destination as being in catch-up mode. + + This throws away the PDU queue. + """ + self._catching_up = True + self._pending_pdus = [] diff --git a/synapse/federation/sender/transaction_manager.py b/synapse/federation/sender/transaction_manager.py
index 8280f8b900..3e07f925e0 100644 --- a/synapse/federation/sender/transaction_manager.py +++ b/synapse/federation/sender/transaction_manager.py
@@ -15,7 +15,7 @@ import logging from typing import TYPE_CHECKING, List -from canonicaljson import json +from prometheus_client import Gauge from synapse.api.errors import HttpResponseException from synapse.events import EventBase @@ -28,6 +28,7 @@ from synapse.logging.opentracing import ( tags, whitelisted_homeserver, ) +from synapse.util import json_decoder from synapse.util.metrics import measure_func if TYPE_CHECKING: @@ -35,8 +36,14 @@ if TYPE_CHECKING: logger = logging.getLogger(__name__) +last_pdu_age_metric = Gauge( + "synapse_federation_last_sent_pdu_age", + "The age (in seconds) of the last PDU successfully sent to the given domain", + labelnames=("server_name",), +) + -class TransactionManager(object): +class TransactionManager: """Helper class which handles building and sending transactions shared between PerDestinationQueue objects @@ -49,13 +56,26 @@ class TransactionManager(object): self._transaction_actions = TransactionActions(self._store) self._transport_layer = hs.get_federation_transport_client() + self._federation_metrics_domains = ( + hs.get_config().federation.federation_metrics_domains + ) + # HACK to get unique tx id self._next_txn_id = int(self.clock.time_msec()) @measure_func("_send_new_transaction") async def send_new_transaction( - self, destination: str, pending_pdus: List[EventBase], pending_edus: List[Edu] - ): + self, destination: str, pdus: List[EventBase], edus: List[Edu], + ) -> bool: + """ + Args: + destination: The destination to send to (e.g. 'example.org') + pdus: In-order list of PDUs to send + edus: List of EDUs to send + + Returns: + True iff the transaction was successful + """ # Make a transaction-sending opentracing span. This span follows on from # all the edus in that transaction. This needs to be done since there is @@ -65,20 +85,14 @@ class TransactionManager(object): span_contexts = [] keep_destination = whitelisted_homeserver(destination) - for edu in pending_edus: + for edu in edus: context = edu.get_context() if context: - span_contexts.append(extract_text_map(json.loads(context))) + span_contexts.append(extract_text_map(json_decoder.decode(context))) if keep_destination: edu.strip_context() with start_active_span_follows_from("send_transaction", span_contexts): - - # Sort based on the order field - pending_pdus.sort(key=lambda t: t[1]) - pdus = [x[0] for x in pending_pdus] - edus = pending_edus - success = True logger.debug("TX [%s] _attempt_new_transaction", destination) @@ -117,6 +131,9 @@ class TransactionManager(object): # FIXME (erikj): This is a bit of a hack to make the Pdu age # keys work + # FIXME (richardv): I also believe it no longer works. We (now?) store + # "age_ts" in "unsigned" rather than at the top level. See + # https://github.com/matrix-org/synapse/issues/8429. def json_data_cb(): data = transaction.get_dict() now = int(self.clock.time_msec()) @@ -165,5 +182,12 @@ class TransactionManager(object): ) success = False + if success and pdus and destination in self._federation_metrics_domains: + last_pdu = pdus[-1] + last_pdu_age = self.clock.time_msec() - last_pdu.origin_server_ts + last_pdu_age_metric.labels(server_name=destination).set( + last_pdu_age / 1000 + ) + set_tag(tags.ERROR, not success) return success diff --git a/synapse/federation/transport/client.py b/synapse/federation/transport/client.py
index e502c12050..4f7996f947 100644 --- a/synapse/federation/transport/client.py +++ b/synapse/federation/transport/client.py
@@ -18,8 +18,6 @@ import logging import urllib from typing import Any, Dict, List, Optional -from twisted.internet import defer - from synapse.api.constants import Membership from synapse.api.errors import Codes, HttpResponseException, SynapseError from synapse.api.urls import ( @@ -32,7 +30,7 @@ from synapse.logging.utils import log_function logger = logging.getLogger(__name__) -class TransportLayerClient(object): +class TransportLayerClient: """Sends federation HTTP requests to other servers""" def __init__(self, hs): @@ -51,7 +49,7 @@ class TransportLayerClient(object): event_id (str): The event we want the context at. Returns: - Deferred: Results in a dict received from the remote homeserver. + Awaitable: Results in a dict received from the remote homeserver. """ logger.debug("get_room_state_ids dest=%s, room=%s", destination, room_id) @@ -75,7 +73,7 @@ class TransportLayerClient(object): giving up. None indicates no timeout. Returns: - Deferred: Results in a dict received from the remote homeserver. + Awaitable: Results in a dict received from the remote homeserver. """ logger.debug("get_pdu dest=%s, event_id=%s", destination, event_id) @@ -96,7 +94,7 @@ class TransportLayerClient(object): limit (int) Returns: - Deferred: Results in a dict received from the remote homeserver. + Awaitable: Results in a dict received from the remote homeserver. """ logger.debug( "backfill dest=%s, room_id=%s, event_tuples=%r, limit=%s", @@ -118,16 +116,15 @@ class TransportLayerClient(object): destination, path=path, args=args, try_trailing_slash_on_400=True ) - @defer.inlineCallbacks @log_function - def send_transaction(self, transaction, json_data_callback=None): + async def send_transaction(self, transaction, json_data_callback=None): """ Sends the given Transaction to its destination Args: transaction (Transaction) Returns: - Deferred: Succeeds when we get a 2xx HTTP response. The result + Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body. Fails with ``HTTPRequestException`` if we get an HTTP response @@ -154,7 +151,7 @@ class TransportLayerClient(object): path = _create_v1_path("/send/%s", transaction.transaction_id) - response = yield self.client.put_json( + response = await self.client.put_json( transaction.destination, path=path, data=json_data, @@ -166,14 +163,13 @@ class TransportLayerClient(object): return response - @defer.inlineCallbacks @log_function - def make_query( + async def make_query( self, destination, query_type, args, retry_on_dns_fail, ignore_backoff=False ): path = _create_v1_path("/query/%s", query_type) - content = yield self.client.get_json( + content = await self.client.get_json( destination=destination, path=path, args=args, @@ -184,9 +180,10 @@ class TransportLayerClient(object): return content - @defer.inlineCallbacks @log_function - def make_membership_event(self, destination, room_id, user_id, membership, params): + async def make_membership_event( + self, destination, room_id, user_id, membership, params + ): """Asks a remote server to build and sign us a membership event Note that this does not append any events to any graphs. @@ -200,7 +197,7 @@ class TransportLayerClient(object): request. Returns: - Deferred: Succeeds when we get a 2xx HTTP response. The result + Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body (ie, the new event). Fails with ``HTTPRequestException`` if we get an HTTP response @@ -231,7 +228,7 @@ class TransportLayerClient(object): ignore_backoff = True retry_on_dns_fail = True - content = yield self.client.get_json( + content = await self.client.get_json( destination=destination, path=path, args=params, @@ -242,34 +239,31 @@ class TransportLayerClient(object): return content - @defer.inlineCallbacks @log_function - def send_join_v1(self, destination, room_id, event_id, content): + async def send_join_v1(self, destination, room_id, event_id, content): path = _create_v1_path("/send_join/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content ) return response - @defer.inlineCallbacks @log_function - def send_join_v2(self, destination, room_id, event_id, content): + async def send_join_v2(self, destination, room_id, event_id, content): path = _create_v2_path("/send_join/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content ) return response - @defer.inlineCallbacks @log_function - def send_leave_v1(self, destination, room_id, event_id, content): + async def send_leave_v1(self, destination, room_id, event_id, content): path = _create_v1_path("/send_leave/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content, @@ -282,12 +276,11 @@ class TransportLayerClient(object): return response - @defer.inlineCallbacks @log_function - def send_leave_v2(self, destination, room_id, event_id, content): + async def send_leave_v2(self, destination, room_id, event_id, content): path = _create_v2_path("/send_leave/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content, @@ -300,31 +293,28 @@ class TransportLayerClient(object): return response - @defer.inlineCallbacks @log_function - def send_invite_v1(self, destination, room_id, event_id, content): + async def send_invite_v1(self, destination, room_id, event_id, content): path = _create_v1_path("/invite/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content, ignore_backoff=True ) return response - @defer.inlineCallbacks @log_function - def send_invite_v2(self, destination, room_id, event_id, content): + async def send_invite_v2(self, destination, room_id, event_id, content): path = _create_v2_path("/invite/%s/%s", room_id, event_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=content, ignore_backoff=True ) return response - @defer.inlineCallbacks @log_function - def get_public_rooms( + async def get_public_rooms( self, remote_server: str, limit: Optional[int] = None, @@ -355,7 +345,7 @@ class TransportLayerClient(object): data["filter"] = search_filter try: - response = yield self.client.post_json( + response = await self.client.post_json( destination=remote_server, path=path, data=data, ignore_backoff=True ) except HttpResponseException as e: @@ -381,7 +371,7 @@ class TransportLayerClient(object): args["since"] = [since_token] try: - response = yield self.client.get_json( + response = await self.client.get_json( destination=remote_server, path=path, args=args, ignore_backoff=True ) except HttpResponseException as e: @@ -396,29 +386,26 @@ class TransportLayerClient(object): return response - @defer.inlineCallbacks @log_function - def exchange_third_party_invite(self, destination, room_id, event_dict): + async def exchange_third_party_invite(self, destination, room_id, event_dict): path = _create_v1_path("/exchange_third_party_invite/%s", room_id) - response = yield self.client.put_json( + response = await self.client.put_json( destination=destination, path=path, data=event_dict ) return response - @defer.inlineCallbacks @log_function - def get_event_auth(self, destination, room_id, event_id): + async def get_event_auth(self, destination, room_id, event_id): path = _create_v1_path("/event_auth/%s/%s", room_id, event_id) - content = yield self.client.get_json(destination=destination, path=path) + content = await self.client.get_json(destination=destination, path=path) return content - @defer.inlineCallbacks @log_function - def query_client_keys(self, destination, query_content, timeout): + async def query_client_keys(self, destination, query_content, timeout): """Query the device keys for a list of user ids hosted on a remote server. @@ -453,14 +440,13 @@ class TransportLayerClient(object): """ path = _create_v1_path("/user/keys/query") - content = yield self.client.post_json( + content = await self.client.post_json( destination=destination, path=path, data=query_content, timeout=timeout ) return content - @defer.inlineCallbacks @log_function - def query_user_devices(self, destination, user_id, timeout): + async def query_user_devices(self, destination, user_id, timeout): """Query the devices for a user id hosted on a remote server. Response: @@ -493,14 +479,13 @@ class TransportLayerClient(object): """ path = _create_v1_path("/user/devices/%s", user_id) - content = yield self.client.get_json( + content = await self.client.get_json( destination=destination, path=path, timeout=timeout ) return content - @defer.inlineCallbacks @log_function - def claim_client_keys(self, destination, query_content, timeout): + async def claim_client_keys(self, destination, query_content, timeout): """Claim one-time keys for a list of devices hosted on a remote server. Request: @@ -532,14 +517,13 @@ class TransportLayerClient(object): path = _create_v1_path("/user/keys/claim") - content = yield self.client.post_json( + content = await self.client.post_json( destination=destination, path=path, data=query_content, timeout=timeout ) return content - @defer.inlineCallbacks @log_function - def get_missing_events( + async def get_missing_events( self, destination, room_id, @@ -551,7 +535,7 @@ class TransportLayerClient(object): ): path = _create_v1_path("/get_missing_events/%s", room_id) - content = yield self.client.post_json( + content = await self.client.post_json( destination=destination, path=path, data={ diff --git a/synapse/federation/transport/server.py b/synapse/federation/transport/server.py
index b518dace8a..7b4baddbf8 100644 --- a/synapse/federation/transport/server.py +++ b/synapse/federation/transport/server.py
@@ -46,7 +46,6 @@ from synapse.logging.opentracing import ( ) from synapse.server import HomeServer from synapse.types import ThirdPartyInstanceID, get_domain_from_id -from synapse.util.ratelimitutils import FederationRateLimiter from synapse.util.versionstring import get_version_string logger = logging.getLogger(__name__) @@ -70,12 +69,10 @@ class TransportLayerServer(JsonResource): self.clock = hs.get_clock() self.servlet_groups = servlet_groups - super(TransportLayerServer, self).__init__(hs, canonical_json=False) + super().__init__(hs, canonical_json=False) self.authenticator = Authenticator(hs) - self.ratelimiter = FederationRateLimiter( - self.clock, config=hs.config.rc_federation - ) + self.ratelimiter = hs.get_federation_ratelimiter() self.register_servlets() @@ -101,7 +98,7 @@ class NoAuthenticationError(AuthenticationError): pass -class Authenticator(object): +class Authenticator: def __init__(self, hs: HomeServer): self._clock = hs.get_clock() self.keyring = hs.get_keyring() @@ -229,7 +226,7 @@ def _parse_auth_header(header_bytes): ) -class BaseFederationServlet(object): +class BaseFederationServlet: """Abstract base class for federation servlet classes. The servlet object should have a PATH attribute which takes the form of a regexp to @@ -273,6 +270,8 @@ class BaseFederationServlet(object): PREFIX = FEDERATION_V1_PREFIX # Allows specifying the API version + RATELIMIT = True # Whether to rate limit requests or not + def __init__(self, handler, authenticator, ratelimiter, server_name): self.handler = handler self.authenticator = authenticator @@ -336,7 +335,7 @@ class BaseFederationServlet(object): ) with scope: - if origin: + if origin and self.RATELIMIT: with ratelimiter.ratelimit(origin) as d: await d if request._disconnected: @@ -373,10 +372,12 @@ class BaseFederationServlet(object): class FederationSendServlet(BaseFederationServlet): PATH = "/send/(?P<transaction_id>[^/]*)/?" + # We ratelimit manually in the handler as we queue up the requests and we + # don't want to fill up the ratelimiter with blocked requests. + RATELIMIT = False + def __init__(self, handler, server_name, **kwargs): - super(FederationSendServlet, self).__init__( - handler, server_name=server_name, **kwargs - ) + super().__init__(handler, server_name=server_name, **kwargs) self.server_name = server_name # This is when someone is trying to send us a bunch of data. @@ -771,9 +772,7 @@ class PublicRoomList(BaseFederationServlet): PATH = "/publicRooms" def __init__(self, handler, authenticator, ratelimiter, server_name, allow_access): - super(PublicRoomList, self).__init__( - handler, authenticator, ratelimiter, server_name - ) + super().__init__(handler, authenticator, ratelimiter, server_name) self.allow_access = allow_access async def on_GET(self, origin, content, query): diff --git a/synapse/federation/units.py b/synapse/federation/units.py
index 6b32e0dcbf..64d98fc8f6 100644 --- a/synapse/federation/units.py +++ b/synapse/federation/units.py
@@ -107,9 +107,7 @@ class Transaction(JsonEncodedObject): if "edus" in kwargs and not kwargs["edus"]: del kwargs["edus"] - super(Transaction, self).__init__( - transaction_id=transaction_id, pdus=pdus, **kwargs - ) + super().__init__(transaction_id=transaction_id, pdus=pdus, **kwargs) @staticmethod def create_new(pdus, **kwargs): diff --git a/synapse/groups/attestations.py b/synapse/groups/attestations.py
index dab13c243f..a86b3debc5 100644 --- a/synapse/groups/attestations.py +++ b/synapse/groups/attestations.py
@@ -41,8 +41,6 @@ from typing import Tuple from signedjson.sign import sign_json -from twisted.internet import defer - from synapse.api.errors import HttpResponseException, RequestSendFailed, SynapseError from synapse.metrics.background_process_metrics import run_as_background_process from synapse.types import get_domain_from_id @@ -62,7 +60,7 @@ DEFAULT_ATTESTATION_JITTER = (0.9, 1.3) UPDATE_ATTESTATION_TIME_MS = 1 * 24 * 60 * 60 * 1000 -class GroupAttestationSigning(object): +class GroupAttestationSigning: """Creates and verifies group attestations. """ @@ -72,8 +70,9 @@ class GroupAttestationSigning(object): self.server_name = hs.hostname self.signing_key = hs.signing_key - @defer.inlineCallbacks - def verify_attestation(self, attestation, group_id, user_id, server_name=None): + async def verify_attestation( + self, attestation, group_id, user_id, server_name=None + ): """Verifies that the given attestation matches the given parameters. An optional server_name can be supplied to explicitly set which server's @@ -102,7 +101,7 @@ class GroupAttestationSigning(object): if valid_until_ms < now: raise SynapseError(400, "Attestation expired") - yield self.keyring.verify_json_for_server( + await self.keyring.verify_json_for_server( server_name, attestation, now, "Group attestation" ) @@ -125,7 +124,7 @@ class GroupAttestationSigning(object): ) -class GroupAttestionRenewer(object): +class GroupAttestionRenewer: """Responsible for sending and receiving attestation updates. """ @@ -142,8 +141,7 @@ class GroupAttestionRenewer(object): self._start_renew_attestations, 30 * 60 * 1000 ) - @defer.inlineCallbacks - def on_renew_attestation(self, group_id, user_id, content): + async def on_renew_attestation(self, group_id, user_id, content): """When a remote updates an attestation """ attestation = content["attestation"] @@ -151,11 +149,11 @@ class GroupAttestionRenewer(object): if not self.is_mine_id(group_id) and not self.is_mine_id(user_id): raise SynapseError(400, "Neither user not group are on this server") - yield self.attestations.verify_attestation( + await self.attestations.verify_attestation( attestation, user_id=user_id, group_id=group_id ) - yield self.store.update_remote_attestion(group_id, user_id, attestation) + await self.store.update_remote_attestion(group_id, user_id, attestation) return {} @@ -172,8 +170,7 @@ class GroupAttestionRenewer(object): now + UPDATE_ATTESTATION_TIME_MS ) - @defer.inlineCallbacks - def _renew_attestation(group_user: Tuple[str, str]): + async def _renew_attestation(group_user: Tuple[str, str]): group_id, user_id = group_user try: if not self.is_mine_id(group_id): @@ -186,16 +183,16 @@ class GroupAttestionRenewer(object): user_id, group_id, ) - yield self.store.remove_attestation_renewal(group_id, user_id) + await self.store.remove_attestation_renewal(group_id, user_id) return attestation = self.attestations.create_attestation(group_id, user_id) - yield self.transport_client.renew_group_attestation( + await self.transport_client.renew_group_attestation( destination, group_id, user_id, content={"attestation": attestation} ) - yield self.store.update_attestation_renewal( + await self.store.update_attestation_renewal( group_id, user_id, attestation ) except (RequestSendFailed, HttpResponseException) as e: diff --git a/synapse/groups/groups_server.py b/synapse/groups/groups_server.py
index 8cb922ddc7..e5f85b472d 100644 --- a/synapse/groups/groups_server.py +++ b/synapse/groups/groups_server.py
@@ -32,7 +32,7 @@ logger = logging.getLogger(__name__) # TODO: Flairs -class GroupsServerWorkerHandler(object): +class GroupsServerWorkerHandler: def __init__(self, hs): self.hs = hs self.store = hs.get_datastore() @@ -336,7 +336,7 @@ class GroupsServerWorkerHandler(object): class GroupsServerHandler(GroupsServerWorkerHandler): def __init__(self, hs): - super(GroupsServerHandler, self).__init__(hs) + super().__init__(hs) # Ensure attestations get renewed hs.get_groups_attestation_renewer() diff --git a/synapse/handlers/__init__.py b/synapse/handlers/__init__.py
index 2dd183018a..286f0054be 100644 --- a/synapse/handlers/__init__.py +++ b/synapse/handlers/__init__.py
@@ -20,7 +20,7 @@ from .identity import IdentityHandler from .search import SearchHandler -class Handlers(object): +class Handlers: """ Deprecated. A collection of handlers. diff --git a/synapse/handlers/_base.py b/synapse/handlers/_base.py
index ba2bf99800..0206320e96 100644 --- a/synapse/handlers/_base.py +++ b/synapse/handlers/_base.py
@@ -25,7 +25,7 @@ from synapse.types import UserID logger = logging.getLogger(__name__) -class BaseHandler(object): +class BaseHandler: """ Common base class for the event handlers. """ diff --git a/synapse/handlers/account_data.py b/synapse/handlers/account_data.py
index a8d3fbc6de..9112a0ab86 100644 --- a/synapse/handlers/account_data.py +++ b/synapse/handlers/account_data.py
@@ -14,7 +14,7 @@ # limitations under the License. -class AccountDataEventSource(object): +class AccountDataEventSource: def __init__(self, hs): self.store = hs.get_datastore() diff --git a/synapse/handlers/account_validity.py b/synapse/handlers/account_validity.py
index 0c2bcda4d0..5162cc9a58 100644 --- a/synapse/handlers/account_validity.py +++ b/synapse/handlers/account_validity.py
@@ -20,23 +20,16 @@ from email.mime.multipart import MIMEMultipart from email.mime.text import MIMEText from typing import List -from twisted.internet import defer - from synapse.api.errors import StoreError from synapse.logging.context import make_deferred_yieldable from synapse.metrics.background_process_metrics import run_as_background_process from synapse.types import UserID from synapse.util import stringutils -try: - from synapse.push.mailer import load_jinja2_templates -except ImportError: - load_jinja2_templates = None - logger = logging.getLogger(__name__) -class AccountValidityHandler(object): +class AccountValidityHandler: def __init__(self, hs): self.hs = hs self.config = hs.config @@ -51,9 +44,11 @@ class AccountValidityHandler(object): if ( self._account_validity.enabled and self._account_validity.renew_by_email_enabled - and load_jinja2_templates ): # Don't do email-specific configuration if renewal by email is disabled. + self._template_html = self.config.account_validity_template_html + self._template_text = self.config.account_validity_template_text + try: app_name = self.hs.config.email_app_name @@ -69,17 +64,6 @@ class AccountValidityHandler(object): self._raw_from = email.utils.parseaddr(self._from_string)[1] - self._template_html, self._template_text = load_jinja2_templates( - self.config.email_template_dir, - [ - self.config.email_expiry_template_html, - self.config.email_expiry_template_text, - ], - apply_format_ts_filter=True, - apply_mxc_to_http_filter=True, - public_baseurl=self.config.public_baseurl, - ) - # Check the renewal emails to send and send them every 30min. def send_emails(): # run as a background process to make sure that the database transactions @@ -90,12 +74,17 @@ class AccountValidityHandler(object): self.clock.looping_call(send_emails, 30 * 60 * 1000) - # If account_validity is enabled,check every hour to remove expired users from - # the user directory + # Mark users as inactive when they expired. Check once every hour if self._account_validity.enabled: - self.clock.looping_call( - self._mark_expired_users_as_inactive, 60 * 60 * 1000 - ) + + def mark_expired_users_as_inactive(): + # run as a background process to allow async functions to work + return run_as_background_process( + "_mark_expired_users_as_inactive", + self._mark_expired_users_as_inactive, + ) + + self.clock.looping_call(mark_expired_users_as_inactive, 60 * 60 * 1000) async def _send_renewal_emails(self): """Gets the list of users whose account is expiring in the amount of time @@ -286,8 +275,7 @@ class AccountValidityHandler(object): return expiration_ts - @defer.inlineCallbacks - def _mark_expired_users_as_inactive(self): + async def _mark_expired_users_as_inactive(self): """Iterate over active, expired users. Mark them as inactive in order to hide them from the user directory. @@ -295,7 +283,7 @@ class AccountValidityHandler(object): Deferred """ # Get active, expired users - active_expired_users = yield self.store.get_expired_users() + active_expired_users = await self.store.get_expired_users() # Mark each as non-active - yield self.profile_handler.set_active(active_expired_users, False, True) + await self.profile_handler.set_active(active_expired_users, False, True) diff --git a/synapse/handlers/acme.py b/synapse/handlers/acme.py
index a2d7959abe..8476256a59 100644 --- a/synapse/handlers/acme.py +++ b/synapse/handlers/acme.py
@@ -17,7 +17,6 @@ import logging import twisted import twisted.internet.error -from twisted.internet import defer from twisted.web import server, static from twisted.web.resource import Resource @@ -35,14 +34,13 @@ solutions, please read https://github.com/matrix-org/synapse/blob/master/docs/AC --------------------------------------------------------------------------------""" -class AcmeHandler(object): +class AcmeHandler: def __init__(self, hs): self.hs = hs self.reactor = hs.get_reactor() self._acme_domain = hs.config.acme_domain - @defer.inlineCallbacks - def start_listening(self): + async def start_listening(self): from synapse.handlers import acme_issuing_service # Configure logging for txacme, if you need to debug @@ -82,18 +80,17 @@ class AcmeHandler(object): self._issuer._registered = False try: - yield self._issuer._ensure_registered() + await self._issuer._ensure_registered() except Exception: logger.error(ACME_REGISTER_FAIL_ERROR) raise - @defer.inlineCallbacks - def provision_certificate(self): + async def provision_certificate(self): logger.warning("Reprovisioning %s", self._acme_domain) try: - yield self._issuer.issue_cert(self._acme_domain) + await self._issuer.issue_cert(self._acme_domain) except Exception: logger.exception("Fail!") raise diff --git a/synapse/handlers/acme_issuing_service.py b/synapse/handlers/acme_issuing_service.py
index e1d4224e74..7294649d71 100644 --- a/synapse/handlers/acme_issuing_service.py +++ b/synapse/handlers/acme_issuing_service.py
@@ -76,9 +76,9 @@ def create_issuing_service(reactor, acme_url, account_key_file, well_known_resou ) -@attr.s +@attr.s(slots=True) @implementer(ICertificateStore) -class ErsatzStore(object): +class ErsatzStore: """ A store that only stores in memory. """ diff --git a/synapse/handlers/admin.py b/synapse/handlers/admin.py
index 506bb2b275..1ce2091b46 100644 --- a/synapse/handlers/admin.py +++ b/synapse/handlers/admin.py
@@ -28,7 +28,7 @@ logger = logging.getLogger(__name__) class AdminHandler(BaseHandler): def __init__(self, hs): - super(AdminHandler, self).__init__(hs) + super().__init__(hs) self.storage = hs.get_storage() self.state_store = self.storage.state @@ -125,8 +125,8 @@ class AdminHandler(BaseHandler): else: stream_ordering = room.stream_ordering - from_key = str(RoomStreamToken(0, 0)) - to_key = str(RoomStreamToken(None, stream_ordering)) + from_key = RoomStreamToken(0, 0) + to_key = RoomStreamToken(None, stream_ordering) written_events = set() # Events that we've processed in this room @@ -197,7 +197,7 @@ class AdminHandler(BaseHandler): return writer.finished() -class ExfiltrationWriter(object): +class ExfiltrationWriter: """Interface used to specify how to write exported data. """ diff --git a/synapse/handlers/appservice.py b/synapse/handlers/appservice.py
index 92d4c6e16c..9d4e87dad6 100644 --- a/synapse/handlers/appservice.py +++ b/synapse/handlers/appservice.py
@@ -27,7 +27,6 @@ from synapse.metrics import ( event_processing_loop_room_count, ) from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.util import log_failure from synapse.util.metrics import Measure logger = logging.getLogger(__name__) @@ -35,7 +34,7 @@ logger = logging.getLogger(__name__) events_processed_counter = Counter("synapse_handlers_appservice_events_processed", "") -class ApplicationServicesHandler(object): +class ApplicationServicesHandler: def __init__(self, hs): self.store = hs.get_datastore() self.is_mine_id = hs.is_mine_id @@ -100,10 +99,11 @@ class ApplicationServicesHandler(object): if not self.started_scheduler: - def start_scheduler(): - return self.scheduler.start().addErrback( - log_failure, "Application Services Failure" - ) + async def start_scheduler(): + try: + return await self.scheduler.start() + except Exception: + logger.error("Application Services Failure") run_as_background_process("as_scheduler", start_scheduler) self.started_scheduler = True diff --git a/synapse/handlers/auth.py b/synapse/handlers/auth.py
index c7d921c21a..00eae92052 100644 --- a/synapse/handlers/auth.py +++ b/synapse/handlers/auth.py
@@ -42,9 +42,9 @@ from synapse.http.site import SynapseRequest from synapse.logging.context import defer_to_thread from synapse.metrics.background_process_metrics import run_as_background_process from synapse.module_api import ModuleApi -from synapse.push.mailer import load_jinja2_templates -from synapse.types import Requester, UserID +from synapse.types import JsonDict, Requester, UserID from synapse.util import stringutils as stringutils +from synapse.util.msisdn import phone_number_to_msisdn from synapse.util.threepids import canonicalise_email from ._base import BaseHandler @@ -52,6 +52,100 @@ from ._base import BaseHandler logger = logging.getLogger(__name__) +def convert_client_dict_legacy_fields_to_identifier( + submission: JsonDict, +) -> Dict[str, str]: + """ + Convert a legacy-formatted login submission to an identifier dict. + + Legacy login submissions (used in both login and user-interactive authentication) + provide user-identifying information at the top-level instead. + + These are now deprecated and replaced with identifiers: + https://matrix.org/docs/spec/client_server/r0.6.1#identifier-types + + Args: + submission: The client dict to convert + + Returns: + The matching identifier dict + + Raises: + SynapseError: If the format of the client dict is invalid + """ + identifier = submission.get("identifier", {}) + + # Generate an m.id.user identifier if "user" parameter is present + user = submission.get("user") + if user: + identifier = {"type": "m.id.user", "user": user} + + # Generate an m.id.thirdparty identifier if "medium" and "address" parameters are present + medium = submission.get("medium") + address = submission.get("address") + if medium and address: + identifier = { + "type": "m.id.thirdparty", + "medium": medium, + "address": address, + } + + # We've converted valid, legacy login submissions to an identifier. If the + # submission still doesn't have an identifier, it's invalid + if not identifier: + raise SynapseError(400, "Invalid login submission", Codes.INVALID_PARAM) + + # Ensure the identifier has a type + if "type" not in identifier: + raise SynapseError( + 400, "'identifier' dict has no key 'type'", errcode=Codes.MISSING_PARAM, + ) + + return identifier + + +def login_id_phone_to_thirdparty(identifier: JsonDict) -> Dict[str, str]: + """ + Convert a phone login identifier type to a generic threepid identifier. + + Args: + identifier: Login identifier dict of type 'm.id.phone' + + Returns: + An equivalent m.id.thirdparty identifier dict + """ + if "country" not in identifier or ( + # The specification requires a "phone" field, while Synapse used to require a "number" + # field. Accept both for backwards compatibility. + "phone" not in identifier + and "number" not in identifier + ): + raise SynapseError( + 400, "Invalid phone-type identifier", errcode=Codes.INVALID_PARAM + ) + + # Accept both "phone" and "number" as valid keys in m.id.phone + phone_number = identifier.get("phone", identifier["number"]) + + # Convert user-provided phone number to a consistent representation + msisdn = phone_number_to_msisdn(identifier["country"], phone_number) + + return { + "type": "m.id.thirdparty", + "medium": "msisdn", + "address": msisdn, + } + + +@attr.s(slots=True) +class SsoLoginExtraAttributes: + """Data we track about SAML2 sessions""" + + # time the session was created, in milliseconds + creation_time = attr.ib(type=int) + extra_attributes = attr.ib(type=JsonDict) + + class AuthHandler(BaseHandler): SESSION_EXPIRE_MS = 48 * 60 * 60 * 1000 @@ -60,7 +154,7 @@ class AuthHandler(BaseHandler): Args: hs (synapse.server.HomeServer): """ - super(AuthHandler, self).__init__(hs) + super().__init__(hs) self.checkers = {} # type: Dict[str, UserInteractiveAuthChecker] for auth_checker_class in INTERACTIVE_AUTH_CHECKERS: @@ -132,18 +226,17 @@ class AuthHandler(BaseHandler): # after the SSO completes and before redirecting them back to their client. # It notifies the user they are about to give access to their matrix account # to the client. - self._sso_redirect_confirm_template = load_jinja2_templates( - hs.config.sso_template_dir, ["sso_redirect_confirm.html"], - )[0] + self._sso_redirect_confirm_template = hs.config.sso_redirect_confirm_template + # The following template is shown during user interactive authentication # in the fallback auth scenario. It notifies the user that they are # authenticating for an operation to occur on their account. - self._sso_auth_confirm_template = load_jinja2_templates( - hs.config.sso_template_dir, ["sso_auth_confirm.html"], - )[0] + self._sso_auth_confirm_template = hs.config.sso_auth_confirm_template + # The following template is shown after a successful user interactive # authentication session. It tells the user they can close the window. self._sso_auth_success_template = hs.config.sso_auth_success_template + # The following template is shown during the SSO authentication process if # the account is deactivated. self._sso_account_deactivated_template = ( @@ -155,6 +248,10 @@ class AuthHandler(BaseHandler): # cast to tuple for use with str.startswith self._whitelisted_sso_clients = tuple(hs.config.sso_client_whitelist) + # A mapping of user ID to extra attributes to include in the login + # response. + self._extra_attributes = {} # type: Dict[str, SsoLoginExtraAttributes] + async def validate_user_via_ui_auth( self, requester: Requester, @@ -162,7 +259,7 @@ class AuthHandler(BaseHandler): request_body: Dict[str, Any], clientip: str, description: str, - ) -> dict: + ) -> Tuple[dict, str]: """ Checks that the user is who they claim to be, via a UI auth. @@ -183,9 +280,14 @@ class AuthHandler(BaseHandler): describes the operation happening on their account. Returns: - The parameters for this request (which may + A tuple of (params, session_id). + + 'params' contains the parameters for this request (which may have been given only in a previous call). + 'session_id' is the ID of this session, either passed in by the + client or assigned by this call + Raises: InteractiveAuthIncompleteError if the client has not yet completed any of the permitted login flows @@ -207,7 +309,7 @@ class AuthHandler(BaseHandler): flows = [[login_type] for login_type in self._supported_ui_auth_types] try: - result, params, _ = await self.check_auth( + result, params, session_id = await self.check_ui_auth( flows, request, request_body, clientip, description ) except LoginError: @@ -230,7 +332,7 @@ class AuthHandler(BaseHandler): if user_id != requester.user.to_string(): raise AuthError(403, "Invalid auth") - return params + return params, session_id def get_enabled_auth_types(self): """Return the enabled user-interactive authentication types @@ -240,7 +342,7 @@ class AuthHandler(BaseHandler): """ return self.checkers.keys() - async def check_auth( + async def check_ui_auth( self, flows: List[List[str]], request: SynapseRequest, @@ -361,9 +463,17 @@ class AuthHandler(BaseHandler): # authentication flow. await self.store.set_ui_auth_clientdict(sid, clientdict) + user_agent = request.requestHeaders.getRawHeaders(b"User-Agent", default=[b""])[ + 0 + ].decode("ascii", "surrogateescape") + + await self.store.add_user_agent_ip_to_ui_auth_session( + session.session_id, user_agent, clientip + ) + if not authdict: raise InteractiveAuthIncompleteError( - self._auth_dict_for_flows(flows, session.session_id) + session.session_id, self._auth_dict_for_flows(flows, session.session_id) ) # check auth type currently being presented @@ -410,7 +520,7 @@ class AuthHandler(BaseHandler): ret = self._auth_dict_for_flows(flows, session.session_id) ret["completed"] = list(creds) ret.update(errordict) - raise InteractiveAuthIncompleteError(ret) + raise InteractiveAuthIncompleteError(session.session_id, ret) async def add_oob_auth( self, stagetype: str, authdict: Dict[str, Any], clientip: str @@ -1068,6 +1178,7 @@ class AuthHandler(BaseHandler): registered_user_id: str, request: SynapseRequest, client_redirect_url: str, + extra_attributes: Optional[JsonDict] = None, ): """Having figured out a mxid for this user, complete the HTTP request @@ -1076,6 +1187,8 @@ class AuthHandler(BaseHandler): request: The request to complete. client_redirect_url: The URL to which to redirect the user at the end of the process. + extra_attributes: Extra attributes which will be passed to the client + during successful login. Must be JSON serializable. """ # If the account has been deactivated, do not proceed with the login # flow. @@ -1084,19 +1197,30 @@ class AuthHandler(BaseHandler): respond_with_html(request, 403, self._sso_account_deactivated_template) return - self._complete_sso_login(registered_user_id, request, client_redirect_url) + self._complete_sso_login( + registered_user_id, request, client_redirect_url, extra_attributes + ) def _complete_sso_login( self, registered_user_id: str, request: SynapseRequest, client_redirect_url: str, + extra_attributes: Optional[JsonDict] = None, ): """ The synchronous portion of complete_sso_login. This exists purely for backwards compatibility of synapse.module_api.ModuleApi. """ + # Store any extra attributes which will be passed in the login response. + # Note that this is per-user so it may overwrite a previous value, this + # is considered OK since the newest SSO attributes should be most valid. + if extra_attributes: + self._extra_attributes[registered_user_id] = SsoLoginExtraAttributes( + self._clock.time_msec(), extra_attributes, + ) + # Create a login token login_token = self.macaroon_gen.generate_short_term_login_token( registered_user_id @@ -1129,6 +1253,37 @@ class AuthHandler(BaseHandler): ) respond_with_html(request, 200, html) + async def _sso_login_callback(self, login_result: JsonDict) -> None: + """ + A login callback which might add additional attributes to the login response. + + Args: + login_result: The data to be sent to the client. Includes the user + ID and access token. + """ + # Expire attributes before processing. Note that there shouldn't be any + # valid logins that still have extra attributes. + self._expire_sso_extra_attributes() + + extra_attributes = self._extra_attributes.get(login_result["user_id"]) + if extra_attributes: + login_result.update(extra_attributes.extra_attributes) + + def _expire_sso_extra_attributes(self) -> None: + """ + Iterate through the mapping of user IDs to extra attributes and remove any that are no longer valid. + """ + # TODO This should match the amount of time the macaroon is valid for. + LOGIN_TOKEN_EXPIRATION_TIME = 2 * 60 * 1000 + expire_before = self._clock.time_msec() - LOGIN_TOKEN_EXPIRATION_TIME + to_expire = set() + for user_id, data in self._extra_attributes.items(): + if data.creation_time < expire_before: + to_expire.add(user_id) + for user_id in to_expire: + logger.debug("Expiring extra attributes for user %s", user_id) + del self._extra_attributes[user_id] + @staticmethod def add_query_param_to_url(url: str, param_name: str, param: Any): url_parts = list(urllib.parse.urlparse(url)) @@ -1138,8 +1293,8 @@ class AuthHandler(BaseHandler): return urllib.parse.urlunparse(url_parts) -@attr.s -class MacaroonGenerator(object): +@attr.s(slots=True) +class MacaroonGenerator: hs = attr.ib() diff --git a/synapse/handlers/cas_handler.py b/synapse/handlers/cas_handler.py
index 786e608fa2..a4cc4b9a5a 100644 --- a/synapse/handlers/cas_handler.py +++ b/synapse/handlers/cas_handler.py
@@ -35,6 +35,7 @@ class CasHandler: """ def __init__(self, hs): + self.hs = hs self._hostname = hs.hostname self._auth_handler = hs.get_auth_handler() self._registration_handler = hs.get_registration_handler() @@ -210,8 +211,16 @@ class CasHandler: else: if not registered_user_id: + # Pull out the user-agent and IP from the request. + user_agent = request.requestHeaders.getRawHeaders( + b"User-Agent", default=[b""] + )[0].decode("ascii", "surrogateescape") + ip_address = self.hs.get_ip_from_request(request) + registered_user_id = await self._registration_handler.register_user( - localpart=localpart, default_display_name=user_display_name + localpart=localpart, + default_display_name=user_display_name, + user_agent_ips=(user_agent, ip_address), ) await self._auth_handler.complete_sso_login( diff --git a/synapse/handlers/deactivate_account.py b/synapse/handlers/deactivate_account.py
index 0e26a32750..af50f4f167 100644 --- a/synapse/handlers/deactivate_account.py +++ b/synapse/handlers/deactivate_account.py
@@ -29,7 +29,7 @@ class DeactivateAccountHandler(BaseHandler): """Handler which deals with deactivating user accounts.""" def __init__(self, hs): - super(DeactivateAccountHandler, self).__init__(hs) + super().__init__(hs) self.hs = hs self._auth_handler = hs.get_auth_handler() self._device_handler = hs.get_device_handler() diff --git a/synapse/handlers/device.py b/synapse/handlers/device.py
index db417d60de..b9d9098104 100644 --- a/synapse/handlers/device.py +++ b/synapse/handlers/device.py
@@ -20,6 +20,7 @@ from typing import Any, Dict, List, Optional from synapse.api import errors from synapse.api.constants import EventTypes from synapse.api.errors import ( + Codes, FederationDeniedError, HttpResponseException, RequestSendFailed, @@ -28,7 +29,7 @@ from synapse.api.errors import ( from synapse.logging.opentracing import log_kv, set_tag, trace from synapse.metrics.background_process_metrics import run_as_background_process from synapse.types import ( - RoomStreamToken, + StreamToken, get_domain_from_id, get_verify_key_from_cross_signing_key, ) @@ -47,7 +48,7 @@ MAX_DEVICE_DISPLAY_NAME_LEN = 100 class DeviceWorkerHandler(BaseHandler): def __init__(self, hs): - super(DeviceWorkerHandler, self).__init__(hs) + super().__init__(hs) self.hs = hs self.state = hs.get_state_handler() @@ -104,18 +105,14 @@ class DeviceWorkerHandler(BaseHandler): @trace @measure_func("device.get_user_ids_changed") - async def get_user_ids_changed(self, user_id, from_token): + async def get_user_ids_changed(self, user_id: str, from_token: StreamToken): """Get list of users that have had the devices updated, or have newly joined a room, that `user_id` may be interested in. - - Args: - user_id (str) - from_token (StreamToken) """ set_tag("user_id", user_id) set_tag("from_token", from_token) - now_room_key = await self.store.get_room_events_max_id() + now_room_key = self.store.get_room_max_token() room_ids = await self.store.get_rooms_for_user(user_id) @@ -142,7 +139,7 @@ class DeviceWorkerHandler(BaseHandler): ) rooms_changed.update(event.room_id for event in member_events) - stream_ordering = RoomStreamToken.parse_stream_token(from_token.room_key).stream + stream_ordering = from_token.room_key.stream possibly_changed = set(changed) possibly_left = set() @@ -234,7 +231,9 @@ class DeviceWorkerHandler(BaseHandler): return result async def on_federation_query_user_devices(self, user_id): - stream_id, devices = await self.store.get_devices_with_keys_by_user(user_id) + stream_id, devices = await self.store.get_e2e_device_keys_for_federation_query( + user_id + ) master_key = await self.store.get_e2e_cross_signing_key(user_id, "master") self_signing_key = await self.store.get_e2e_cross_signing_key( user_id, "self_signing" @@ -251,7 +250,7 @@ class DeviceWorkerHandler(BaseHandler): class DeviceHandler(DeviceWorkerHandler): def __init__(self, hs): - super(DeviceHandler, self).__init__(hs) + super().__init__(hs) self.federation_sender = hs.get_federation_sender() @@ -265,6 +264,24 @@ class DeviceHandler(DeviceWorkerHandler): hs.get_distributor().observe("user_left_room", self.user_left_room) + def _check_device_name_length(self, name: str): + """ + Checks whether a device name is longer than the maximum allowed length. + + Args: + name: The name of the device. + + Raises: + SynapseError: if the device name is too long. + """ + if name and len(name) > MAX_DEVICE_DISPLAY_NAME_LEN: + raise SynapseError( + 400, + "Device display name is too long (max %i)" + % (MAX_DEVICE_DISPLAY_NAME_LEN,), + errcode=Codes.TOO_LARGE, + ) + async def check_device_registered( self, user_id, device_id, initial_device_display_name=None ): @@ -282,6 +299,9 @@ class DeviceHandler(DeviceWorkerHandler): Returns: str: device id (generated if none was supplied) """ + + self._check_device_name_length(initial_device_display_name) + if device_id is not None: new_device = await self.store.store_device( user_id=user_id, @@ -397,12 +417,8 @@ class DeviceHandler(DeviceWorkerHandler): # Reject a new displayname which is too long. new_display_name = content.get("display_name") - if new_display_name and len(new_display_name) > MAX_DEVICE_DISPLAY_NAME_LEN: - raise SynapseError( - 400, - "Device display name is too long (max %i)" - % (MAX_DEVICE_DISPLAY_NAME_LEN,), - ) + + self._check_device_name_length(new_display_name) try: await self.store.update_device( @@ -495,7 +511,7 @@ def _update_device_from_client_ips(device, client_ips): device.update({"last_seen_ts": ip.get("last_seen"), "last_seen_ip": ip.get("ip")}) -class DeviceListUpdater(object): +class DeviceListUpdater: "Handles incoming device list updates from federation and updates the DB" def __init__(self, hs, device_handler): diff --git a/synapse/handlers/devicemessage.py b/synapse/handlers/devicemessage.py
index 610b08d00b..64ef7f63ab 100644 --- a/synapse/handlers/devicemessage.py +++ b/synapse/handlers/devicemessage.py
@@ -16,8 +16,6 @@ import logging from typing import Any, Dict -from canonicaljson import json - from synapse.api.errors import SynapseError from synapse.logging.context import run_in_background from synapse.logging.opentracing import ( @@ -27,12 +25,13 @@ from synapse.logging.opentracing import ( start_active_span, ) from synapse.types import UserID, get_domain_from_id +from synapse.util import json_encoder from synapse.util.stringutils import random_string logger = logging.getLogger(__name__) -class DeviceMessageHandler(object): +class DeviceMessageHandler: def __init__(self, hs): """ Args: @@ -174,7 +173,7 @@ class DeviceMessageHandler(object): "sender": sender_user_id, "type": message_type, "message_id": message_id, - "org.matrix.opentracing_context": json.dumps(context), + "org.matrix.opentracing_context": json_encoder.encode(context), } log_kv({"local_messages": local_messages}) diff --git a/synapse/handlers/directory.py b/synapse/handlers/directory.py
index af9936f7e2..654e76f2c9 100644 --- a/synapse/handlers/directory.py +++ b/synapse/handlers/directory.py
@@ -23,6 +23,7 @@ from synapse.api.errors import ( CodeMessageException, Codes, NotFoundError, + ShadowBanError, StoreError, SynapseError, ) @@ -36,7 +37,7 @@ logger = logging.getLogger(__name__) class DirectoryHandler(BaseHandler): def __init__(self, hs): - super(DirectoryHandler, self).__init__(hs) + super().__init__(hs) self.state = hs.get_state_handler() self.appservice_handler = hs.get_application_service_handler() @@ -200,6 +201,8 @@ class DirectoryHandler(BaseHandler): try: await self._update_canonical_alias(requester, user_id, room_id, room_alias) + except ShadowBanError as e: + logger.info("Failed to update alias events due to shadow-ban: %s", e) except AuthError as e: logger.info("Failed to update alias events: %s", e) @@ -293,6 +296,9 @@ class DirectoryHandler(BaseHandler): """ Send an updated canonical alias event if the removed alias was set as the canonical alias or listed in the alt_aliases field. + + Raises: + ShadowBanError if the requester has been shadow-banned. """ alias_event = await self.state.get_current_state( room_id, EventTypes.CanonicalAlias, "" diff --git a/synapse/handlers/e2e_keys.py b/synapse/handlers/e2e_keys.py
index 84169c1022..dd40fd1299 100644 --- a/synapse/handlers/e2e_keys.py +++ b/synapse/handlers/e2e_keys.py
@@ -19,7 +19,7 @@ import logging from typing import Dict, List, Optional, Tuple import attr -from canonicaljson import encode_canonical_json, json +from canonicaljson import encode_canonical_json from signedjson.key import VerifyKey, decode_verify_key_bytes from signedjson.sign import SignatureVerifyException, verify_signed_json from unpaddedbase64 import decode_base64 @@ -35,7 +35,7 @@ from synapse.types import ( get_domain_from_id, get_verify_key_from_cross_signing_key, ) -from synapse.util import unwrapFirstError +from synapse.util import json_decoder, unwrapFirstError from synapse.util.async_helpers import Linearizer from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.retryutils import NotRetryingDestination @@ -43,7 +43,7 @@ from synapse.util.retryutils import NotRetryingDestination logger = logging.getLogger(__name__) -class E2eKeysHandler(object): +class E2eKeysHandler: def __init__(self, hs): self.store = hs.get_datastore() self.federation = hs.get_federation_client() @@ -353,7 +353,7 @@ class E2eKeysHandler(object): # make sure that each queried user appears in the result dict result_dict[user_id] = {} - results = await self.store.get_e2e_device_keys(local_query) + results = await self.store.get_e2e_device_keys_for_cs_api(local_query) # Build the result structure for user_id, device_keys in results.items(): @@ -404,7 +404,7 @@ class E2eKeysHandler(object): for device_id, keys in device_keys.items(): for key_id, json_bytes in keys.items(): json_result.setdefault(user_id, {})[device_id] = { - key_id: json.loads(json_bytes) + key_id: json_decoder.decode(json_bytes) } @trace @@ -734,7 +734,7 @@ class E2eKeysHandler(object): # fetch our stored devices. This is used to 1. verify # signatures on the master key, and 2. to compare with what # was sent if the device was signed - devices = await self.store.get_e2e_device_keys([(user_id, None)]) + devices = await self.store.get_e2e_device_keys_for_cs_api([(user_id, None)]) if user_id not in devices: raise NotFoundError("No device keys found") @@ -1186,7 +1186,7 @@ def _exception_to_failure(e): def _one_time_keys_match(old_key_json, new_key): - old_key = json.loads(old_key_json) + old_key = json_decoder.decode(old_key_json) # if either is a string rather than an object, they must match exactly if not isinstance(old_key, dict) or not isinstance(new_key, dict): @@ -1201,7 +1201,7 @@ def _one_time_keys_match(old_key_json, new_key): return old_key == new_key_copy -@attr.s +@attr.s(slots=True) class SignatureListItem: """An item in the signature list as used by upload_signatures_for_device_keys. """ @@ -1212,7 +1212,7 @@ class SignatureListItem: signature = attr.ib() -class SigningKeyEduUpdater(object): +class SigningKeyEduUpdater: """Handles incoming signing key updates from federation and updates the DB""" def __init__(self, hs, e2e_keys_handler): diff --git a/synapse/handlers/e2e_room_keys.py b/synapse/handlers/e2e_room_keys.py
index 0bb983dc28..f01b090772 100644 --- a/synapse/handlers/e2e_room_keys.py +++ b/synapse/handlers/e2e_room_keys.py
@@ -29,7 +29,7 @@ from synapse.util.async_helpers import Linearizer logger = logging.getLogger(__name__) -class E2eRoomKeysHandler(object): +class E2eRoomKeysHandler: """ Implements an optional realtime backup mechanism for encrypted E2E megolm room keys. This gives a way for users to store and recover their megolm keys if they lose all diff --git a/synapse/handlers/events.py b/synapse/handlers/events.py
index 71a89f09c7..539b4fc32e 100644 --- a/synapse/handlers/events.py +++ b/synapse/handlers/events.py
@@ -15,32 +15,29 @@ import logging import random +from typing import TYPE_CHECKING, Iterable, List, Optional from synapse.api.constants import EventTypes, Membership from synapse.api.errors import AuthError, SynapseError from synapse.events import EventBase from synapse.handlers.presence import format_user_presence_state from synapse.logging.utils import log_function -from synapse.types import UserID +from synapse.streams.config import PaginationConfig +from synapse.types import JsonDict, UserID from synapse.visibility import filter_events_for_client from ._base import BaseHandler -logger = logging.getLogger(__name__) +if TYPE_CHECKING: + from synapse.server import HomeServer -class EventStreamHandler(BaseHandler): - def __init__(self, hs): - super(EventStreamHandler, self).__init__(hs) +logger = logging.getLogger(__name__) - # Count of active streams per user - self._streams_per_user = {} - # Grace timers per user to delay the "stopped" signal - self._stop_timer_per_user = {} - self.distributor = hs.get_distributor() - self.distributor.declare("started_user_eventstream") - self.distributor.declare("stopped_user_eventstream") +class EventStreamHandler(BaseHandler): + def __init__(self, hs: "HomeServer"): + super().__init__(hs) self.clock = hs.get_clock() @@ -52,18 +49,15 @@ class EventStreamHandler(BaseHandler): @log_function async def get_stream( self, - auth_user_id, - pagin_config, - timeout=0, - as_client_event=True, - affect_presence=True, - only_keys=None, - room_id=None, - is_guest=False, - ): + auth_user_id: str, + pagin_config: PaginationConfig, + timeout: int = 0, + as_client_event: bool = True, + affect_presence: bool = True, + room_id: Optional[str] = None, + is_guest: bool = False, + ) -> JsonDict: """Fetches the events stream for a given user. - - If `only_keys` is not None, events from keys will be sent down. """ if room_id: @@ -93,7 +87,6 @@ class EventStreamHandler(BaseHandler): auth_user, pagin_config, timeout, - only_keys=only_keys, is_guest=is_guest, explicit_room_id=room_id, ) @@ -102,7 +95,7 @@ class EventStreamHandler(BaseHandler): # When the user joins a new room, or another user joins a currently # joined room, we need to send down presence for those users. - to_add = [] + to_add = [] # type: List[JsonDict] for event in events: if not isinstance(event, EventBase): continue @@ -114,7 +107,7 @@ class EventStreamHandler(BaseHandler): # Send down presence for everyone in the room. users = await self.state.get_current_users_in_room( event.room_id - ) + ) # type: Iterable[str] else: users = [event.state_key] @@ -140,28 +133,30 @@ class EventStreamHandler(BaseHandler): chunk = { "chunk": chunks, - "start": tokens[0].to_string(), - "end": tokens[1].to_string(), + "start": await tokens[0].to_string(self.store), + "end": await tokens[1].to_string(self.store), } return chunk class EventHandler(BaseHandler): - def __init__(self, hs): - super(EventHandler, self).__init__(hs) + def __init__(self, hs: "HomeServer"): + super().__init__(hs) self.storage = hs.get_storage() - async def get_event(self, user, room_id, event_id): + async def get_event( + self, user: UserID, room_id: Optional[str], event_id: str + ) -> Optional[EventBase]: """Retrieve a single specified event. Args: - user (synapse.types.UserID): The user requesting the event - room_id (str|None): The expected room id. We'll return None if the + user: The user requesting the event + room_id: The expected room id. We'll return None if the event's room does not match. - event_id (str): The event ID to obtain. + event_id: The event ID to obtain. Returns: - dict: An event, or None if there is no event matching this ID. + An event, or None if there is no event matching this ID. Raises: SynapseError if there was a problem retrieving this event, or AuthError if the user does not have the rights to inspect this diff --git a/synapse/handlers/federation.py b/synapse/handlers/federation.py
index 8a7ceb0df2..29ec18e25e 100644 --- a/synapse/handlers/federation.py +++ b/synapse/handlers/federation.py
@@ -21,7 +21,7 @@ import itertools import logging from collections.abc import Container from http import HTTPStatus -from typing import Dict, Iterable, List, Optional, Sequence, Tuple, Union +from typing import TYPE_CHECKING, Dict, Iterable, List, Optional, Sequence, Tuple, Union import attr from signedjson.key import decode_verify_key_bytes @@ -69,20 +69,29 @@ from synapse.replication.http.federation import ( ReplicationFederationSendEventsRestServlet, ReplicationStoreRoomOnInviteRestServlet, ) -from synapse.replication.http.membership import ReplicationUserJoinedLeftRoomRestServlet -from synapse.state import StateResolutionStore, resolve_events_with_store -from synapse.storage.data_stores.main.events_worker import EventRedactBehaviour -from synapse.types import JsonDict, StateMap, UserID, get_domain_from_id +from synapse.state import StateResolutionStore +from synapse.storage.databases.main.events_worker import EventRedactBehaviour +from synapse.types import ( + JsonDict, + MutableStateMap, + PersistedEventPosition, + RoomStreamToken, + StateMap, + UserID, + get_domain_from_id, +) from synapse.util.async_helpers import Linearizer, concurrently_execute -from synapse.util.distributor import user_joined_room from synapse.util.retryutils import NotRetryingDestination from synapse.util.stringutils import shortstr from synapse.visibility import filter_events_for_server +if TYPE_CHECKING: + from synapse.server import HomeServer + logger = logging.getLogger(__name__) -@attr.s +@attr.s(slots=True) class _NewEventInfo: """Holds information about a received event, ready for passing to _handle_new_events @@ -96,7 +105,7 @@ class _NewEventInfo: event = attr.ib(type=EventBase) state = attr.ib(type=Optional[Sequence[EventBase]], default=None) - auth_events = attr.ib(type=Optional[StateMap[EventBase]], default=None) + auth_events = attr.ib(type=Optional[MutableStateMap[EventBase]], default=None) class FederationHandler(BaseHandler): @@ -110,8 +119,8 @@ class FederationHandler(BaseHandler): rooms. """ - def __init__(self, hs): - super(FederationHandler, self).__init__(hs) + def __init__(self, hs: "HomeServer"): + super().__init__(hs) self.hs = hs @@ -120,11 +129,11 @@ class FederationHandler(BaseHandler): self.state_store = self.storage.state self.federation_client = hs.get_federation_client() self.state_handler = hs.get_state_handler() + self._state_resolution_handler = hs.get_state_resolution_handler() self.server_name = hs.hostname self.keyring = hs.get_keyring() self.action_generator = hs.get_action_generator() self.is_mine_id = hs.is_mine_id - self.pusher_pool = hs.get_pusherpool() self.spam_checker = hs.get_spam_checker() self.event_creation_handler = hs.get_event_creation_handler() self._message_handler = hs.get_message_handler() @@ -135,9 +144,6 @@ class FederationHandler(BaseHandler): self._replication = hs.get_replication_data_handler() self._send_events = ReplicationFederationSendEventsRestServlet.make_client(hs) - self._notify_user_membership_change = ReplicationUserJoinedLeftRoomRestServlet.make_client( - hs - ) self._clean_room_for_join_client = ReplicationCleanRoomRestServlet.make_client( hs ) @@ -153,8 +159,9 @@ class FederationHandler(BaseHandler): self._device_list_updater = hs.get_device_handler().device_list_updater self._maybe_store_room_on_invite = self.store.maybe_store_room_on_invite - # When joining a room we need to queue any events for that room up - self.room_queues = {} + # When joining a room we need to queue any events for that room up. + # For each room, a list of (pdu, origin) tuples. + self.room_queues = {} # type: Dict[str, List[Tuple[EventBase, str]]] self._room_pdu_linearizer = Linearizer("fed_room_pdu") self.third_party_event_rules = hs.get_third_party_event_rules() @@ -279,7 +286,7 @@ class FederationHandler(BaseHandler): raise Exception( "Error fetching missing prev_events for %s: %s" % (event_id, e) - ) + ) from e # Update the set of things we've seen after trying to # fetch the missing stuff @@ -383,8 +390,7 @@ class FederationHandler(BaseHandler): event_map[x.event_id] = x room_version = await self.store.get_room_version_id(room_id) - state_map = await resolve_events_with_store( - self.clock, + state_map = await self._state_resolution_handler.resolve_events_with_store( room_id, room_version, state_maps, @@ -437,11 +443,11 @@ class FederationHandler(BaseHandler): if not prevs - seen: return - latest = await self.store.get_latest_event_ids_in_room(room_id) + latest_list = await self.store.get_latest_event_ids_in_room(room_id) # We add the prev events that we have seen to the latest # list to ensure the remote server doesn't give them to us - latest = set(latest) + latest = set(latest_list) latest |= seen logger.info( @@ -701,31 +707,10 @@ class FederationHandler(BaseHandler): logger.debug("[%s %s] Processing event: %s", room_id, event_id, event) try: - context = await self._handle_new_event(origin, event, state=state) + await self._handle_new_event(origin, event, state=state) except AuthError as e: raise FederationError("ERROR", e.code, e.msg, affected=event.event_id) - if event.type == EventTypes.Member: - if event.membership == Membership.JOIN: - # Only fire user_joined_room if the user has acutally - # joined the room. Don't bother if the user is just - # changing their profile info. - newly_joined = True - - prev_state_ids = await context.get_prev_state_ids() - - prev_state_id = prev_state_ids.get((event.type, event.state_key)) - if prev_state_id: - prev_state = await self.store.get_event( - prev_state_id, allow_none=True - ) - if prev_state and prev_state.membership == Membership.JOIN: - newly_joined = False - - if newly_joined: - user = UserID.from_string(event.state_key) - await self.user_joined_room(user, room_id) - # For encrypted messages we check that we know about the sending device, # if we don't then we mark the device cache for that user as stale. if event.type == EventTypes.Encrypted: @@ -778,7 +763,7 @@ class FederationHandler(BaseHandler): # keys across all devices. current_keys = [ key - for device in cached_devices + for device in cached_devices.values() for key in device.get("keys", {}).get("keys", {}).values() ] @@ -836,6 +821,9 @@ class FederationHandler(BaseHandler): dest, room_id, limit=limit, extremities=extremities ) + if not events: + return [] + # ideally we'd sanity check the events here for excess prev_events etc, # but it's hard to reject events at this point without completely # breaking backfill in the same way that it is currently broken by @@ -920,7 +908,8 @@ class FederationHandler(BaseHandler): ) ) - await self._handle_new_events(dest, ev_infos, backfilled=True) + if ev_infos: + await self._handle_new_events(dest, room_id, ev_infos, backfilled=True) # Step 2: Persist the rest of the events in the chunk one by one events.sort(key=lambda e: e.depth) @@ -940,15 +929,26 @@ class FederationHandler(BaseHandler): return events - async def maybe_backfill(self, room_id, current_depth): + async def maybe_backfill( + self, room_id: str, current_depth: int, limit: int + ) -> bool: """Checks the database to see if we should backfill before paginating, and if so do. + + Args: + room_id + current_depth: The depth from which we're paginating from. This is + used to decide if we should backfill and what extremities to + use. + limit: The number of events that the pagination request will + return. This is used as part of the heuristic to decide if we + should back paginate. """ extremities = await self.store.get_oldest_events_with_depth_in_room(room_id) if not extremities: logger.debug("Not backfilling as no extremeties found.") - return + return False # We only want to paginate if we can actually see the events we'll get, # as otherwise we'll just spend a lot of resources to get redacted @@ -1001,16 +1001,54 @@ class FederationHandler(BaseHandler): sorted_extremeties_tuple = sorted(extremities.items(), key=lambda e: -int(e[1])) max_depth = sorted_extremeties_tuple[0][1] + # If we're approaching an extremity we trigger a backfill, otherwise we + # no-op. + # + # We chose twice the limit here as then clients paginating backwards + # will send pagination requests that trigger backfill at least twice + # using the most recent extremity before it gets removed (see below). We + # chose more than one times the limit in case of failure, but choosing a + # much larger factor will result in triggering a backfill request much + # earlier than necessary. + if current_depth - 2 * limit > max_depth: + logger.debug( + "Not backfilling as we don't need to. %d < %d - 2 * %d", + max_depth, + current_depth, + limit, + ) + return False + + logger.debug( + "room_id: %s, backfill: current_depth: %s, max_depth: %s, extrems: %s", + room_id, + current_depth, + max_depth, + sorted_extremeties_tuple, + ) + + # We ignore extremities that have a greater depth than our current depth + # as: + # 1. we don't really care about getting events that have happened + # before our current position; and + # 2. we have likely previously tried and failed to backfill from that + # extremity, so to avoid getting "stuck" requesting the same + # backfill repeatedly we drop those extremities. + filtered_sorted_extremeties_tuple = [ + t for t in sorted_extremeties_tuple if int(t[1]) <= current_depth + ] + + # However, we need to check that the filtered extremities are non-empty. + # If they are empty then either we can a) bail or b) still attempt to + # backill. We opt to try backfilling anyway just in case we do get + # relevant events. + if filtered_sorted_extremeties_tuple: + sorted_extremeties_tuple = filtered_sorted_extremeties_tuple + # We don't want to specify too many extremities as it causes the backfill # request URI to be too long. extremities = dict(sorted_extremeties_tuple[:5]) - if current_depth > max_depth: - logger.debug( - "Not backfilling as we don't need to. %d < %d", max_depth, current_depth - ) - return - # Now we need to decide which hosts to hit first. # First we try hosts that are already in the room @@ -1213,7 +1251,7 @@ class FederationHandler(BaseHandler): event_infos.append(_NewEventInfo(event, None, auth)) await self._handle_new_events( - destination, event_infos, + destination, room_id, event_infos, ) def _sanity_check_event(self, ev): @@ -1360,15 +1398,15 @@ class FederationHandler(BaseHandler): ) max_stream_id = await self._persist_auth_tree( - origin, auth_chain, state, event, room_version_obj + origin, room_id, auth_chain, state, event, room_version_obj ) # We wait here until this instance has seen the events come down # replication (if we're using replication) as the below uses caches. - # - # TODO: Currently the events stream is written to from master await self._replication.wait_for_stream_position( - self.config.worker.writers.events, "events", max_stream_id + self.config.worker.events_shard_config.get_instance(room_id), + "events", + max_stream_id, ) # Check whether this room is the result of an upgrade of a room we already know @@ -1547,11 +1585,6 @@ class FederationHandler(BaseHandler): event.signatures, ) - if event.type == EventTypes.Member: - if event.content["membership"] == Membership.JOIN: - user = UserID.from_string(event.state_key) - await self.user_joined_room(user, event.room_id) - prev_state_ids = await context.get_prev_state_ids() state_ids = list(prev_state_ids.values()) @@ -1629,7 +1662,7 @@ class FederationHandler(BaseHandler): ) context = await self.state_handler.compute_event_context(event) - await self.persist_events_and_notify([(event, context)]) + await self.persist_events_and_notify(event.room_id, [(event, context)]) return event @@ -1656,7 +1689,9 @@ class FederationHandler(BaseHandler): await self.federation_client.send_leave(host_list, event) context = await self.state_handler.compute_event_context(event) - stream_id = await self.persist_events_and_notify([(event, context)]) + stream_id = await self.persist_events_and_notify( + event.room_id, [(event, context)] + ) return event, stream_id @@ -1787,9 +1822,7 @@ class FederationHandler(BaseHandler): """Returns the state at the event. i.e. not including said event. """ - event = await self.store.get_event( - event_id, allow_none=False, check_room_id=room_id - ) + event = await self.store.get_event(event_id, check_room_id=room_id) state_groups = await self.state_store.get_state_groups(room_id, [event_id]) @@ -1815,9 +1848,7 @@ class FederationHandler(BaseHandler): async def get_state_ids_for_pdu(self, room_id: str, event_id: str) -> List[str]: """Returns the state at the event. i.e. not including said event. """ - event = await self.store.get_event( - event_id, allow_none=False, check_room_id=room_id - ) + event = await self.store.get_event(event_id, check_room_id=room_id) state_groups = await self.state_store.get_state_groups_ids(room_id, [event_id]) @@ -1887,8 +1918,8 @@ class FederationHandler(BaseHandler): else: return None - def get_min_depth_for_context(self, context): - return self.store.get_min_depth(context) + async def get_min_depth_for_context(self, context): + return await self.store.get_min_depth(context) async def _handle_new_event( self, origin, event, state=None, auth_events=None, backfilled=False @@ -1908,7 +1939,7 @@ class FederationHandler(BaseHandler): ) await self.persist_events_and_notify( - [(event, context)], backfilled=backfilled + event.room_id, [(event, context)], backfilled=backfilled ) except Exception: run_in_background( @@ -1921,6 +1952,7 @@ class FederationHandler(BaseHandler): async def _handle_new_events( self, origin: str, + room_id: str, event_infos: Iterable[_NewEventInfo], backfilled: bool = False, ) -> None: @@ -1952,6 +1984,7 @@ class FederationHandler(BaseHandler): ) await self.persist_events_and_notify( + room_id, [ (ev_info.event, context) for ev_info, context in zip(event_infos, contexts) @@ -1962,6 +1995,7 @@ class FederationHandler(BaseHandler): async def _persist_auth_tree( self, origin: str, + room_id: str, auth_events: List[EventBase], state: List[EventBase], event: EventBase, @@ -1976,6 +2010,7 @@ class FederationHandler(BaseHandler): Args: origin: Where the events came from + room_id, auth_events state event @@ -2050,31 +2085,34 @@ class FederationHandler(BaseHandler): events_to_context[e.event_id].rejected = RejectedReason.AUTH_ERROR await self.persist_events_and_notify( + room_id, [ (e, events_to_context[e.event_id]) for e in itertools.chain(auth_events, state) - ] + ], ) new_event_context = await self.state_handler.compute_event_context( event, old_state=state ) - return await self.persist_events_and_notify([(event, new_event_context)]) + return await self.persist_events_and_notify( + room_id, [(event, new_event_context)] + ) async def _prep_event( self, origin: str, event: EventBase, state: Optional[Iterable[EventBase]], - auth_events: Optional[StateMap[EventBase]], + auth_events: Optional[MutableStateMap[EventBase]], backfilled: bool, ) -> EventContext: context = await self.state_handler.compute_event_context(event, old_state=state) if not auth_events: prev_state_ids = await context.get_prev_state_ids() - auth_events_ids = await self.auth.compute_auth_events( + auth_events_ids = self.auth.compute_auth_events( event, prev_state_ids, for_verification=True ) auth_events_x = await self.store.get_events(auth_events_ids) @@ -2117,8 +2155,8 @@ class FederationHandler(BaseHandler): if backfilled or event.internal_metadata.is_outlier(): return - extrem_ids = await self.store.get_latest_event_ids_in_room(event.room_id) - extrem_ids = set(extrem_ids) + extrem_ids_list = await self.store.get_latest_event_ids_in_room(event.room_id) + extrem_ids = set(extrem_ids_list) prev_event_ids = set(event.prev_event_ids()) if extrem_ids == prev_event_ids: @@ -2143,15 +2181,17 @@ class FederationHandler(BaseHandler): # given state at the event. This should correctly handle cases # like bans, especially with state res v2. - state_sets = await self.state_store.get_state_groups( + state_sets_d = await self.state_store.get_state_groups( event.room_id, extrem_ids ) - state_sets = list(state_sets.values()) + state_sets = list(state_sets_d.values()) # type: List[Iterable[EventBase]] state_sets.append(state) - current_state_ids = await self.state_handler.resolve_events( + current_states = await self.state_handler.resolve_events( room_version, state_sets, event ) - current_state_ids = {k: e.event_id for k, e in current_state_ids.items()} + current_state_ids = { + k: e.event_id for k, e in current_states.items() + } # type: StateMap[str] else: current_state_ids = await self.state_handler.get_current_state_ids( event.room_id, latest_event_ids=extrem_ids @@ -2163,11 +2203,13 @@ class FederationHandler(BaseHandler): # Now check if event pass auth against said current state auth_types = auth_types_for_event(event) - current_state_ids = [e for k, e in current_state_ids.items() if k in auth_types] + current_state_ids_list = [ + e for k, e in current_state_ids.items() if k in auth_types + ] - current_auth_events = await self.store.get_events(current_state_ids) + auth_events_map = await self.store.get_events(current_state_ids_list) current_auth_events = { - (e.type, e.state_key): e for e in current_auth_events.values() + (e.type, e.state_key): e for e in auth_events_map.values() } try: @@ -2183,9 +2225,7 @@ class FederationHandler(BaseHandler): if not in_room: raise AuthError(403, "Host not in room.") - event = await self.store.get_event( - event_id, allow_none=False, check_room_id=room_id - ) + event = await self.store.get_event(event_id, check_room_id=room_id) # Just go through and process each event in `remote_auth_chain`. We # don't want to fall into the trap of `missing` being wrong. @@ -2237,7 +2277,7 @@ class FederationHandler(BaseHandler): origin: str, event: EventBase, context: EventContext, - auth_events: StateMap[EventBase], + auth_events: MutableStateMap[EventBase], ) -> EventContext: """ @@ -2288,7 +2328,7 @@ class FederationHandler(BaseHandler): origin: str, event: EventBase, context: EventContext, - auth_events: StateMap[EventBase], + auth_events: MutableStateMap[EventBase], ) -> EventContext: """Helper for do_auth. See there for docs. @@ -2480,7 +2520,7 @@ class FederationHandler(BaseHandler): } current_state_ids = await context.get_current_state_ids() - current_state_ids = dict(current_state_ids) + current_state_ids = dict(current_state_ids) # type: ignore current_state_ids.update(state_updates) @@ -2909,6 +2949,7 @@ class FederationHandler(BaseHandler): async def persist_events_and_notify( self, + room_id: str, event_and_contexts: Sequence[Tuple[EventBase, EventContext]], backfilled: bool = False, ) -> int: @@ -2916,20 +2957,26 @@ class FederationHandler(BaseHandler): necessary. Args: - event_and_contexts: + room_id: The room ID of events being persisted. + event_and_contexts: Sequence of events with their associated + context that should be persisted. All events must belong to + the same room. backfilled: Whether these events are a result of backfilling or not """ - if self.config.worker.writers.events != self._instance_name: + instance = self.config.worker.events_shard_config.get_instance(room_id) + if instance != self._instance_name: result = await self._send_events( - instance_name=self.config.worker.writers.events, + instance_name=instance, store=self.store, + room_id=room_id, event_and_contexts=event_and_contexts, backfilled=backfilled, ) return result["max_stream_id"] else: - max_stream_id = await self.storage.persistence.persist_events( + assert self.storage.persistence + max_stream_token = await self.storage.persistence.persist_events( event_and_contexts, backfilled=backfilled ) @@ -2940,12 +2987,12 @@ class FederationHandler(BaseHandler): if not backfilled: # Never notify for backfilled events for event, _ in event_and_contexts: - await self._notify_persisted_event(event, max_stream_id) + await self._notify_persisted_event(event, max_stream_token) - return max_stream_id + return max_stream_token.stream async def _notify_persisted_event( - self, event: EventBase, max_stream_id: int + self, event: EventBase, max_stream_token: RoomStreamToken ) -> None: """Checks to see if notifier/pushers should be notified about the event or not. @@ -2971,13 +3018,13 @@ class FederationHandler(BaseHandler): elif event.internal_metadata.is_outlier(): return - event_stream_id = event.internal_metadata.stream_ordering + event_pos = PersistedEventPosition( + self._instance_name, event.internal_metadata.stream_ordering + ) self.notifier.on_new_room_event( - event, event_stream_id, max_stream_id, extra_users=extra_users + event, event_pos, max_stream_token, extra_users=extra_users ) - await self.pusher_pool.on_new_notifications(event_stream_id, max_stream_id) - async def _clean_room_for_join(self, room_id: str) -> None: """Called to clean up any data in DB for a given room, ready for the server to join the room. @@ -2990,16 +3037,6 @@ class FederationHandler(BaseHandler): else: await self.store.clean_room_for_join(room_id) - async def user_joined_room(self, user: UserID, room_id: str) -> None: - """Called when a new user has joined the room - """ - if self.config.worker_app: - await self._notify_user_membership_change( - room_id=room_id, user_id=user.to_string(), change="joined" - ) - else: - user_joined_room(self.distributor, user, room_id) - async def get_room_complexity( self, remote_room_hosts: List[str], room_id: str ) -> Optional[dict]: diff --git a/synapse/handlers/groups_local.py b/synapse/handlers/groups_local.py
index ecdb12a7bf..9684e60fc8 100644 --- a/synapse/handlers/groups_local.py +++ b/synapse/handlers/groups_local.py
@@ -23,43 +23,36 @@ logger = logging.getLogger(__name__) def _create_rerouter(func_name): - """Returns a function that looks at the group id and calls the function + """Returns an async function that looks at the group id and calls the function on federation or the local group server if the group is local """ - def f(self, group_id, *args, **kwargs): + async def f(self, group_id, *args, **kwargs): if self.is_mine_id(group_id): - return getattr(self.groups_server_handler, func_name)( + return await getattr(self.groups_server_handler, func_name)( group_id, *args, **kwargs ) else: destination = get_domain_from_id(group_id) - d = getattr(self.transport_client, func_name)( - destination, group_id, *args, **kwargs - ) - # Capture errors returned by the remote homeserver and - # re-throw specific errors as SynapseErrors. This is so - # when the remote end responds with things like 403 Not - # In Group, we can communicate that to the client instead - # of a 500. - def http_response_errback(failure): - failure.trap(HttpResponseException) - e = failure.value + try: + return await getattr(self.transport_client, func_name)( + destination, group_id, *args, **kwargs + ) + except HttpResponseException as e: + # Capture errors returned by the remote homeserver and + # re-throw specific errors as SynapseErrors. This is so + # when the remote end responds with things like 403 Not + # In Group, we can communicate that to the client instead + # of a 500. raise e.to_synapse_error() - - def request_failed_errback(failure): - failure.trap(RequestSendFailed) + except RequestSendFailed: raise SynapseError(502, "Failed to contact group server") - d.addErrback(http_response_errback) - d.addErrback(request_failed_errback) - return d - return f -class GroupsLocalWorkerHandler(object): +class GroupsLocalWorkerHandler: def __init__(self, hs): self.hs = hs self.store = hs.get_datastore() @@ -247,7 +240,7 @@ class GroupsLocalWorkerHandler(object): class GroupsLocalHandler(GroupsLocalWorkerHandler): def __init__(self, hs): - super(GroupsLocalHandler, self).__init__(hs) + super().__init__(hs) # Ensure attestations get renewed hs.get_groups_attestation_renewer() diff --git a/synapse/handlers/identity.py b/synapse/handlers/identity.py
index 3cf76e954e..c5ff072d5a 100644 --- a/synapse/handlers/identity.py +++ b/synapse/handlers/identity.py
@@ -21,14 +21,6 @@ import logging import urllib.parse from typing import Awaitable, Callable, Dict, List, Optional, Tuple -from canonicaljson import json -from signedjson.key import decode_verify_key_bytes -from signedjson.sign import verify_signed_json -from unpaddedbase64 import decode_base64 - -from twisted.internet import defer -from twisted.internet.error import TimeoutError - from synapse.api.errors import ( AuthError, CodeMessageException, @@ -38,8 +30,10 @@ from synapse.api.errors import ( SynapseError, ) from synapse.config.emailconfig import ThreepidBehaviour +from synapse.http import RequestTimedOutError from synapse.http.client import SimpleHttpClient from synapse.types import JsonDict, Requester +from synapse.util import json_decoder from synapse.util.hash import sha256_and_url_safe_base64 from synapse.util.stringutils import assert_valid_client_secret, random_string @@ -50,7 +44,7 @@ logger = logging.getLogger(__name__) class IdentityHandler(BaseHandler): def __init__(self, hs): - super(IdentityHandler, self).__init__(hs) + super().__init__(hs) self.hs = hs self.http_client = hs.get_simple_http_client() @@ -112,7 +106,7 @@ class IdentityHandler(BaseHandler): try: data = await self.http_client.get_json(url, query_params) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except HttpResponseException as e: logger.info( @@ -197,10 +191,10 @@ class IdentityHandler(BaseHandler): if e.code != 404 or not use_v2: logger.error("3PID bind failed with Matrix error: %r", e) raise e.to_synapse_error() - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except CodeMessageException as e: - data = json.loads(e.msg) # XXX WAT? + data = json_decoder.decode(e.msg) # XXX WAT? return data logger.info("Got 404 when POSTing JSON %s, falling back to v1 URL", bind_url) @@ -304,7 +298,7 @@ class IdentityHandler(BaseHandler): else: logger.error("Failed to unbind threepid on identity server: %s", e) raise SynapseError(500, "Failed to contact identity server") - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") await self.store.remove_user_bound_threepid( @@ -475,7 +469,7 @@ class IdentityHandler(BaseHandler): except HttpResponseException as e: logger.info("Proxied requestToken failed: %r", e) raise e.to_synapse_error() - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") async def requestMsisdnToken( @@ -531,7 +525,7 @@ class IdentityHandler(BaseHandler): except HttpResponseException as e: logger.info("Proxied requestToken failed: %r", e) raise e.to_synapse_error() - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") assert self.hs.config.public_baseurl @@ -613,7 +607,7 @@ class IdentityHandler(BaseHandler): id_server + "/_matrix/identity/api/v1/validate/msisdn/submitToken", body, ) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except HttpResponseException as e: logger.warning("Error contacting msisdn account_threepid_delegate: %s", e) @@ -623,18 +617,19 @@ class IdentityHandler(BaseHandler): # the CS API. They should be consolidated with those in RoomMemberHandler # https://github.com/matrix-org/synapse-dinsic/issues/25 - @defer.inlineCallbacks - def proxy_lookup_3pid(self, id_server, medium, address): + async def proxy_lookup_3pid( + self, id_server: str, medium: str, address: str + ) -> JsonDict: """Looks up a 3pid in the passed identity server. Args: - id_server (str): The server name (including port, if required) + id_server: The server name (including port, if required) of the identity server to use. - medium (str): The type of the third party identifier (e.g. "email"). - address (str): The third party identifier (e.g. "foo@example.com"). + medium: The type of the third party identifier (e.g. "email"). + address: The third party identifier (e.g. "foo@example.com"). Returns: - Deferred[dict]: The result of the lookup. See + The result of the lookup. See https://matrix.org/docs/spec/identity_service/r0.1.0.html#association-lookup for details """ @@ -646,16 +641,11 @@ class IdentityHandler(BaseHandler): id_server_url = self.rewrite_id_server_url(id_server, add_https=True) try: - data = yield self.http_client.get_json( + data = await self.http_client.get_json( "%s/_matrix/identity/api/v1/lookup" % (id_server_url,), {"medium": medium, "address": address}, ) - if "mxid" in data: - if "signatures" not in data: - raise AuthError(401, "No signatures on 3pid binding") - yield self._verify_any_signature(data, id_server) - except HttpResponseException as e: logger.info("Proxied lookup failed: %r", e) raise e.to_synapse_error() @@ -665,18 +655,19 @@ class IdentityHandler(BaseHandler): return data - @defer.inlineCallbacks - def proxy_bulk_lookup_3pid(self, id_server, threepids): + async def proxy_bulk_lookup_3pid( + self, id_server: str, threepids: List[List[str]] + ) -> JsonDict: """Looks up given 3pids in the passed identity server. Args: - id_server (str): The server name (including port, if required) + id_server: The server name (including port, if required) of the identity server to use. - threepids ([[str, str]]): The third party identifiers to lookup, as + threepids: The third party identifiers to lookup, as a list of 2-string sized lists ([medium, address]). Returns: - Deferred[dict]: The result of the lookup. See + The result of the lookup. See https://matrix.org/docs/spec/identity_service/r0.1.0.html#association-lookup for details """ @@ -688,7 +679,7 @@ class IdentityHandler(BaseHandler): id_server_url = self.rewrite_id_server_url(id_server, add_https=True) try: - data = yield self.http_client.post_json_get_json( + data = await self.http_client.post_json_get_json( "%s/_matrix/identity/api/v1/bulk_lookup" % (id_server_url,), {"threepids": threepids}, ) @@ -700,7 +691,7 @@ class IdentityHandler(BaseHandler): logger.info("Failed to contact %s: %s", id_server, e) raise ProxiedRequestError(503, "Failed to contact identity server") - defer.returnValue(data) + return data async def lookup_3pid( self, @@ -770,11 +761,11 @@ class IdentityHandler(BaseHandler): ) if "mxid" in data: - if "signatures" not in data: - raise AuthError(401, "No signatures on 3pid binding") - await self._verify_any_signature(data, id_server) + # note: we used to verify the identity server's signature here, but no longer + # require or validate it. See the following for context: + # https://github.com/matrix-org/synapse/issues/5253#issuecomment-666246950 return data["mxid"] - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except IOError as e: logger.warning("Error from v1 identity server lookup: %s" % (e,)) @@ -801,7 +792,7 @@ class IdentityHandler(BaseHandler): "%s/_matrix/identity/v2/hash_details" % (id_server_url,), {"access_token": id_access_token}, ) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") if not isinstance(hash_details, dict): @@ -872,7 +863,7 @@ class IdentityHandler(BaseHandler): }, headers=headers, ) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except Exception as e: logger.warning("Error when performing a v2 3pid lookup: %s", e) @@ -891,31 +882,6 @@ class IdentityHandler(BaseHandler): mxid = lookup_results["mappings"].get(lookup_value) return mxid - async def _verify_any_signature(self, data, id_server): - if id_server not in data["signatures"]: - raise AuthError(401, "No signature from server %s" % (id_server,)) - - for key_name, signature in data["signatures"][id_server].items(): - id_server_url = self.rewrite_id_server_url(id_server, add_https=True) - - key_data = await self.http_client.get_json( - "%s/_matrix/identity/api/v1/pubkey/%s" % (id_server_url, key_name) - ) - if "public_key" not in key_data: - raise AuthError( - 401, "No public key named %s from %s" % (key_name, id_server) - ) - verify_signed_json( - data, - id_server, - decode_verify_key_bytes( - key_name, decode_base64(key_data["public_key"]) - ), - ) - return - - raise AuthError(401, "No signature from server %s" % (id_server,)) - async def ask_id_server_for_third_party_invite( self, requester: Requester, @@ -995,7 +961,7 @@ class IdentityHandler(BaseHandler): invite_config, {"Authorization": create_id_access_token_header(id_access_token)}, ) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except HttpResponseException as e: if e.code != 404: @@ -1012,7 +978,7 @@ class IdentityHandler(BaseHandler): data = await self.blacklisting_http_client.post_json_get_json( url, invite_config ) - except TimeoutError: + except RequestTimedOutError: raise SynapseError(500, "Timed out contacting identity server") except HttpResponseException as e: logger.warning( @@ -1068,6 +1034,9 @@ class IdentityHandler(BaseHandler): """ # Extract the domain name from the IS URL as we store IS domains instead of URLs id_server = urllib.parse.urlparse(id_server_url).hostname + if not id_server: + # We were unable to determine the hostname, bail out + return # id_server_url is assumed to have no trailing slashes url = id_server_url + "/_matrix/identity/internal/bind" diff --git a/synapse/handlers/initial_sync.py b/synapse/handlers/initial_sync.py
index f88bad5f25..39a85801c1 100644 --- a/synapse/handlers/initial_sync.py +++ b/synapse/handlers/initial_sync.py
@@ -14,6 +14,7 @@ # limitations under the License. import logging +from typing import TYPE_CHECKING from twisted.internet import defer @@ -22,8 +23,9 @@ from synapse.api.errors import SynapseError from synapse.events.validator import EventValidator from synapse.handlers.presence import format_user_presence_state from synapse.logging.context import make_deferred_yieldable, run_in_background +from synapse.storage.roommember import RoomsForUser from synapse.streams.config import PaginationConfig -from synapse.types import StreamToken, UserID +from synapse.types import JsonDict, Requester, RoomStreamToken, StreamToken, UserID from synapse.util import unwrapFirstError from synapse.util.async_helpers import concurrently_execute from synapse.util.caches.response_cache import ResponseCache @@ -31,12 +33,16 @@ from synapse.visibility import filter_events_for_client from ._base import BaseHandler +if TYPE_CHECKING: + from synapse.server import HomeServer + + logger = logging.getLogger(__name__) class InitialSyncHandler(BaseHandler): - def __init__(self, hs): - super(InitialSyncHandler, self).__init__(hs) + def __init__(self, hs: "HomeServer"): + super().__init__(hs) self.hs = hs self.state = hs.get_state_handler() self.clock = hs.get_clock() @@ -48,27 +54,25 @@ class InitialSyncHandler(BaseHandler): def snapshot_all_rooms( self, - user_id=None, - pagin_config=None, - as_client_event=True, - include_archived=False, - ): + user_id: str, + pagin_config: PaginationConfig, + as_client_event: bool = True, + include_archived: bool = False, + ) -> JsonDict: """Retrieve a snapshot of all rooms the user is invited or has joined. This snapshot may include messages for all rooms where the user is joined, depending on the pagination config. Args: - user_id (str): The ID of the user making the request. - pagin_config (synapse.api.streams.PaginationConfig): The pagination - config used to determine how many messages *PER ROOM* to return. - as_client_event (bool): True to get events in client-server format. - include_archived (bool): True to get rooms that the user has left + user_id: The ID of the user making the request. + pagin_config: The pagination config used to determine how many + messages *PER ROOM* to return. + as_client_event: True to get events in client-server format. + include_archived: True to get rooms that the user has left Returns: - A list of dicts with "room_id" and "membership" keys for all rooms - the user is currently invited or joined in on. Rooms where the user - is joined on, may return a "messages" key with messages, depending - on the specified PaginationConfig. + A JsonDict with the same format as the response to `/intialSync` + API """ key = ( user_id, @@ -91,11 +95,11 @@ class InitialSyncHandler(BaseHandler): async def _snapshot_all_rooms( self, - user_id=None, - pagin_config=None, - as_client_event=True, - include_archived=False, - ): + user_id: str, + pagin_config: PaginationConfig, + as_client_event: bool = True, + include_archived: bool = False, + ) -> JsonDict: memberships = [Membership.INVITE, Membership.JOIN] if include_archived: @@ -109,17 +113,16 @@ class InitialSyncHandler(BaseHandler): rooms_ret = [] - now_token = await self.hs.get_event_sources().get_current_token() + now_token = self.hs.get_event_sources().get_current_token() presence_stream = self.hs.get_event_sources().sources["presence"] - pagination_config = PaginationConfig(from_token=now_token) - presence, _ = await presence_stream.get_pagination_rows( - user, pagination_config.get_source_config("presence"), None + presence, _ = await presence_stream.get_new_events( + user, from_key=None, include_offline=False ) - receipt_stream = self.hs.get_event_sources().sources["receipt"] - receipt, _ = await receipt_stream.get_pagination_rows( - user, pagination_config.get_source_config("receipt"), None + joined_rooms = [r.room_id for r in room_list if r.membership == Membership.JOIN] + receipt = await self.store.get_linearized_receipts_for_rooms( + joined_rooms, to_key=int(now_token.receipt_key), ) tags_by_room = await self.store.get_tags_for_user(user_id) @@ -134,7 +137,7 @@ class InitialSyncHandler(BaseHandler): if limit is None: limit = 10 - async def handle_room(event): + async def handle_room(event: RoomsForUser): d = { "room_id": event.room_id, "membership": event.membership, @@ -164,7 +167,7 @@ class InitialSyncHandler(BaseHandler): self.state_handler.get_current_state, event.room_id ) elif event.membership == Membership.LEAVE: - room_end_token = "s%d" % (event.stream_ordering,) + room_end_token = RoomStreamToken(None, event.stream_ordering,) deferred_room_state = run_in_background( self.state_store.get_state_for_events, [event.event_id] ) @@ -200,8 +203,8 @@ class InitialSyncHandler(BaseHandler): messages, time_now=time_now, as_client_event=as_client_event ) ), - "start": start_token.to_string(), - "end": end_token.to_string(), + "start": await start_token.to_string(self.store), + "end": await end_token.to_string(self.store), } d["state"] = await self._event_serializer.serialize_events( @@ -246,22 +249,23 @@ class InitialSyncHandler(BaseHandler): ], "account_data": account_data_events, "receipts": receipt, - "end": now_token.to_string(), + "end": await now_token.to_string(self.store), } return ret - async def room_initial_sync(self, requester, room_id, pagin_config=None): + async def room_initial_sync( + self, requester: Requester, room_id: str, pagin_config: PaginationConfig + ) -> JsonDict: """Capture the a snapshot of a room. If user is currently a member of the room this will be what is currently in the room. If the user left the room this will be what was in the room when they left. Args: - requester(Requester): The user to get a snapshot for. - room_id(str): The room to get a snapshot of. - pagin_config(synapse.streams.config.PaginationConfig): - The pagination config used to determine how many messages to - return. + requester: The user to get a snapshot for. + room_id: The room to get a snapshot of. + pagin_config: The pagination config used to determine how many + messages to return. Raises: AuthError if the user wasn't in the room. Returns: @@ -305,8 +309,14 @@ class InitialSyncHandler(BaseHandler): return result async def _room_initial_sync_parted( - self, user_id, room_id, pagin_config, membership, member_event_id, is_peeking - ): + self, + user_id: str, + room_id: str, + pagin_config: PaginationConfig, + membership: Membership, + member_event_id: str, + is_peeking: bool, + ) -> JsonDict: room_state = await self.state_store.get_state_for_events([member_event_id]) room_state = room_state[member_event_id] @@ -315,7 +325,8 @@ class InitialSyncHandler(BaseHandler): if limit is None: limit = 10 - stream_token = await self.store.get_stream_token_for_event(member_event_id) + leave_position = await self.store.get_position_for_event(member_event_id) + stream_token = leave_position.to_room_stream_token() messages, token = await self.store.get_recent_events_for_room( room_id, limit=limit, end_token=stream_token @@ -337,8 +348,8 @@ class InitialSyncHandler(BaseHandler): "chunk": ( await self._event_serializer.serialize_events(messages, time_now) ), - "start": start_token.to_string(), - "end": end_token.to_string(), + "start": await start_token.to_string(self.store), + "end": await end_token.to_string(self.store), }, "state": ( await self._event_serializer.serialize_events( @@ -350,8 +361,13 @@ class InitialSyncHandler(BaseHandler): } async def _room_initial_sync_joined( - self, user_id, room_id, pagin_config, membership, is_peeking - ): + self, + user_id: str, + room_id: str, + pagin_config: PaginationConfig, + membership: Membership, + is_peeking: bool, + ) -> JsonDict: current_state = await self.state.get_current_state(room_id=room_id) # TODO: These concurrently @@ -360,7 +376,7 @@ class InitialSyncHandler(BaseHandler): current_state.values(), time_now ) - now_token = await self.hs.get_event_sources().get_current_token() + now_token = self.hs.get_event_sources().get_current_token() limit = pagin_config.limit if pagin_config else None if limit is None: @@ -431,8 +447,8 @@ class InitialSyncHandler(BaseHandler): "chunk": ( await self._event_serializer.serialize_events(messages, time_now) ), - "start": start_token.to_string(), - "end": end_token.to_string(), + "start": await start_token.to_string(self.store), + "end": await end_token.to_string(self.store), }, "state": state, "presence": presence, diff --git a/synapse/handlers/message.py b/synapse/handlers/message.py
index a5e8dc1323..8949343801 100644 --- a/synapse/handlers/message.py +++ b/synapse/handlers/message.py
@@ -15,9 +15,10 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging -from typing import TYPE_CHECKING, List, Optional, Tuple +import random +from typing import TYPE_CHECKING, Dict, List, Optional, Tuple -from canonicaljson import encode_canonical_json, json +from canonicaljson import encode_canonical_json from twisted.internet.interfaces import IDelayedCall @@ -34,6 +35,7 @@ from synapse.api.errors import ( Codes, ConsentNotGivenError, NotFoundError, + ShadowBanError, SynapseError, ) from synapse.api.room_versions import KNOWN_ROOM_VERSIONS, RoomVersions @@ -45,16 +47,10 @@ from synapse.events.validator import EventValidator from synapse.logging.context import run_in_background from synapse.metrics.background_process_metrics import run_as_background_process from synapse.replication.http.send_event import ReplicationSendEventRestServlet -from synapse.storage.data_stores.main.events_worker import EventRedactBehaviour +from synapse.storage.databases.main.events_worker import EventRedactBehaviour from synapse.storage.state import StateFilter -from synapse.types import ( - Collection, - Requester, - RoomAlias, - StreamToken, - UserID, - create_requester, -) +from synapse.types import Requester, RoomAlias, StreamToken, UserID, create_requester +from synapse.util import json_decoder from synapse.util.async_helpers import Linearizer from synapse.util.frozenutils import frozendict_json_encoder from synapse.util.metrics import measure_func @@ -69,7 +65,7 @@ if TYPE_CHECKING: logger = logging.getLogger(__name__) -class MessageHandler(object): +class MessageHandler: """Contains some read only APIs to get state about a room """ @@ -93,12 +89,7 @@ class MessageHandler(object): ) async def get_room_data( - self, - user_id: str = None, - room_id: str = None, - event_type: Optional[str] = None, - state_key: str = "", - is_guest: bool = False, + self, user_id: str, room_id: str, event_type: str, state_key: str, ) -> dict: """ Get data from a room. @@ -107,11 +98,10 @@ class MessageHandler(object): room_id event_type state_key - is_guest Returns: The path data content. Raises: - SynapseError if something went wrong. + SynapseError or AuthError if the user is not in the room """ ( membership, @@ -128,6 +118,16 @@ class MessageHandler(object): [membership_event_id], StateFilter.from_types([key]) ) data = room_state[membership_event_id].get(key) + else: + # check_user_in_room_or_world_readable, if it doesn't raise an AuthError, should + # only ever return a Membership.JOIN/LEAVE object + # + # Safeguard in case it returned something else + logger.error( + "Attempted to retrieve data from a room for a user that has never been in it. " + "This should not have happened." + ) + raise SynapseError(403, "User not in room", errcode=Codes.FORBIDDEN) return data @@ -362,7 +362,7 @@ class MessageHandler(object): _DUMMY_EVENT_ROOM_EXCLUSION_EXPIRY = 7 * 24 * 60 * 60 * 1000 -class EventCreationHandler(object): +class EventCreationHandler: def __init__(self, hs: "HomeServer"): self.hs = hs self.auth = hs.get_auth() @@ -377,9 +377,8 @@ class EventCreationHandler(object): self.notifier = hs.get_notifier() self.config = hs.config self.require_membership_for_aliases = hs.config.require_membership_for_aliases - self._is_event_writer = ( - self.config.worker.writers.events == hs.get_instance_name() - ) + self._events_shard_config = self.config.worker.events_shard_config + self._instance_name = hs.get_instance_name() self.room_invite_state_types = self.hs.config.room_invite_state_types @@ -388,8 +387,6 @@ class EventCreationHandler(object): # This is only used to get at ratelimit function, and maybe_kick_guest_users self.base_handler = BaseHandler(hs) - self.pusher_pool = hs.get_pusherpool() - # We arbitrarily limit concurrent event creation for a room to 5. # This is to stop us from diverging history *too* much. self.limiter = Linearizer(max_count=5, name="room_event_creation_limit") @@ -410,7 +407,7 @@ class EventCreationHandler(object): # # map from room id to time-of-last-attempt. # - self._rooms_to_exclude_from_dummy_event_insertion = {} # type: dict[str, int] + self._rooms_to_exclude_from_dummy_event_insertion = {} # type: Dict[str, int] # we need to construct a ConsentURIBuilder here, as it checks that the necessary # config options, but *only* if we have a configuration for which we are @@ -442,7 +439,7 @@ class EventCreationHandler(object): event_dict: dict, token_id: Optional[str] = None, txn_id: Optional[str] = None, - prev_event_ids: Optional[Collection[str]] = None, + prev_event_ids: Optional[List[str]] = None, require_consent: bool = True, ) -> Tuple[EventBase, EventContext]: """ @@ -647,37 +644,48 @@ class EventCreationHandler(object): event: EventBase, context: EventContext, ratelimit: bool = True, + ignore_shadow_ban: bool = False, ) -> int: """ Persists and notifies local clients and federation of an event. Args: - requester - event the event to send. - context: the context of the event. + requester: The requester sending the event. + event: The event to send. + context: The context of the event. ratelimit: Whether to rate limit this send. + ignore_shadow_ban: True if shadow-banned users should be allowed to + send this event. Return: The stream_id of the persisted event. + + Raises: + ShadowBanError if the requester has been shadow-banned. """ if event.type == EventTypes.Member: raise SynapseError( 500, "Tried to send member event through non-member codepath" ) + if not ignore_shadow_ban and requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() + user = UserID.from_string(event.sender) assert self.hs.is_mine(user), "User must be our own: %s" % (user,) if event.is_state(): - prev_state = await self.deduplicate_state_event(event, context) - if prev_state is not None: + prev_event = await self.deduplicate_state_event(event, context) + if prev_event is not None: logger.info( "Not bothering to persist state event %s duplicated by %s", event.event_id, - prev_state.event_id, + prev_event.event_id, ) - return prev_state + return await self.store.get_stream_id_for_event(prev_event.event_id) return await self.handle_new_client_event( requester=requester, event=event, context=context, ratelimit=ratelimit @@ -685,40 +693,61 @@ class EventCreationHandler(object): async def deduplicate_state_event( self, event: EventBase, context: EventContext - ) -> None: + ) -> Optional[EventBase]: """ Checks whether event is in the latest resolved state in context. - If so, returns the version of the event in context. - Otherwise, returns None. + Args: + event: The event to check for duplication. + context: The event context. + + Returns: + The previous verion of the event is returned, if it is found in the + event context. Otherwise, None is returned. """ prev_state_ids = await context.get_prev_state_ids() prev_event_id = prev_state_ids.get((event.type, event.state_key)) if not prev_event_id: - return + return None prev_event = await self.store.get_event(prev_event_id, allow_none=True) if not prev_event: - return + return None if prev_event and event.user_id == prev_event.user_id: prev_content = encode_canonical_json(prev_event.content) next_content = encode_canonical_json(event.content) if prev_content == next_content: return prev_event - return + return None async def create_and_send_nonmember_event( self, requester: Requester, - event_dict: EventBase, + event_dict: dict, ratelimit: bool = True, txn_id: Optional[str] = None, + ignore_shadow_ban: bool = False, ) -> Tuple[EventBase, int]: """ Creates an event, then sends it. See self.create_event and self.send_nonmember_event. + + Args: + requester: The requester sending the event. + event_dict: An entire event. + ratelimit: Whether to rate limit this send. + txn_id: The transaction ID. + ignore_shadow_ban: True if shadow-banned users should be allowed to + send this event. + + Raises: + ShadowBanError if the requester has been shadow-banned. """ + if not ignore_shadow_ban and requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() # We limit the number of concurrent event sends in a room so that we # don't fork the DAG too much. If we don't limit then we can end up in @@ -737,7 +766,11 @@ class EventCreationHandler(object): raise SynapseError(403, spam_error, Codes.FORBIDDEN) stream_id = await self.send_nonmember_event( - requester, event, context, ratelimit=ratelimit + requester, + event, + context, + ratelimit=ratelimit, + ignore_shadow_ban=ignore_shadow_ban, ) return event, stream_id @@ -746,7 +779,7 @@ class EventCreationHandler(object): self, builder: EventBuilder, requester: Optional[Requester] = None, - prev_event_ids: Optional[Collection[str]] = None, + prev_event_ids: Optional[List[str]] = None, ) -> Tuple[EventBase, EventContext]: """Create a new event for a local client @@ -771,6 +804,15 @@ class EventCreationHandler(object): else: prev_event_ids = await self.store.get_prev_events_for_room(builder.room_id) + # we now ought to have some prev_events (unless it's a create event). + # + # do a quick sanity check here, rather than waiting until we've created the + # event and then try to auth it (which fails with a somewhat confusing "No + # create event in auth events") + assert ( + builder.type == EventTypes.Create or len(prev_event_ids) > 0 + ), "Attempting to create an event with no prev_events" + event = await builder.build(prev_event_ids=prev_event_ids) context = await self.state.compute_event_context(event) if requester: @@ -853,7 +895,7 @@ class EventCreationHandler(object): # Ensure that we can round trip before trying to persist in db try: dump = frozendict_json_encoder.encode(event.content) - json.loads(dump) + json_decoder.decode(dump) except Exception: logger.exception("Failed to encode content: %r", event.content) raise @@ -862,9 +904,10 @@ class EventCreationHandler(object): try: # If we're a worker we need to hit out to the master. - if not self._is_event_writer: + writer_instance = self._events_shard_config.get_instance(event.room_id) + if writer_instance != self._instance_name: result = await self.send_event( - instance_name=self.config.worker.writers.events, + instance_name=writer_instance, event_id=event.event_id, store=self.store, requester=requester, @@ -885,9 +928,7 @@ class EventCreationHandler(object): except Exception: # Ensure that we actually remove the entries in the push actions # staging area, if we calculated them. - run_in_background( - self.store.remove_push_actions_from_staging, event.event_id - ) + await self.store.remove_push_actions_from_staging(event.event_id) raise async def _validate_canonical_alias( @@ -934,7 +975,10 @@ class EventCreationHandler(object): This should only be run on the instance in charge of persisting events. """ - assert self._is_event_writer + assert self.storage.persistence is not None + assert self._events_shard_config.should_handle( + self._instance_name, event.room_id + ) if ratelimit: # We check if this is a room admin redacting an event so that we @@ -951,7 +995,7 @@ class EventCreationHandler(object): allow_none=True, ) - is_admin_redaction = ( + is_admin_redaction = bool( original_event and event.sender != original_event.sender ) @@ -965,7 +1009,7 @@ class EventCreationHandler(object): # Validate a newly added alias or newly added alt_aliases. original_alias = None - original_alt_aliases = set() + original_alt_aliases = [] # type: List[str] original_event_id = event.unsigned.get("replaces_state") if original_event_id: @@ -1013,6 +1057,10 @@ class EventCreationHandler(object): current_state_ids = await context.get_current_state_ids() + # We know this event is not an outlier, so this must be + # non-None. + assert current_state_ids is not None + state_to_include_ids = [ e_id for k, e_id in current_state_ids.items() @@ -1064,11 +1112,11 @@ class EventCreationHandler(object): raise SynapseError(400, "Cannot redact event from a different room") prev_state_ids = await context.get_prev_state_ids() - auth_events_ids = await self.auth.compute_auth_events( + auth_events_ids = self.auth.compute_auth_events( event, prev_state_ids, for_verification=True ) - auth_events = await self.store.get_events(auth_events_ids) - auth_events = {(e.type, e.state_key): e for e in auth_events.values()} + auth_events_map = await self.store.get_events(auth_events_ids) + auth_events = {(e.type, e.state_key): e for e in auth_events_map.values()} room_version = await self.store.get_room_version_id(event.room_id) room_version_obj = KNOWN_ROOM_VERSIONS[room_version] @@ -1093,7 +1141,7 @@ class EventCreationHandler(object): if prev_state_ids: raise AuthError(403, "Changing the room create event is forbidden") - event_stream_id, max_stream_id = await self.storage.persistence.persist_event( + event_pos, max_stream_token = await self.storage.persistence.persist_event( event, context=context ) @@ -1101,12 +1149,10 @@ class EventCreationHandler(object): # If there's an expiry timestamp on the event, schedule its expiry. self._message_handler.maybe_schedule_expiry(event) - await self.pusher_pool.on_new_notifications(event_stream_id, max_stream_id) - def _notify(): try: self.notifier.on_new_room_event( - event, event_stream_id, max_stream_id, extra_users=extra_users + event, event_pos, max_stream_token, extra_users=extra_users ) except Exception: logger.exception("Error notifying about new room event") @@ -1118,7 +1164,7 @@ class EventCreationHandler(object): # matters as sometimes presence code can take a while. run_in_background(self._bump_active_time, requester.user) - return event_stream_id + return event_pos.stream async def _bump_active_time(self, user: UserID) -> None: try: @@ -1139,48 +1185,7 @@ class EventCreationHandler(object): ) for room_id in room_ids: - # For each room we need to find a joined member we can use to send - # the dummy event with. - - latest_event_ids = await self.store.get_prev_events_for_room(room_id) - - members = await self.state.get_current_users_in_room( - room_id, latest_event_ids=latest_event_ids - ) - dummy_event_sent = False - for user_id in members: - if not self.hs.is_mine_id(user_id): - continue - requester = create_requester(user_id) - try: - event, context = await self.create_event( - requester, - { - "type": "org.matrix.dummy_event", - "content": {}, - "room_id": room_id, - "sender": user_id, - }, - prev_event_ids=latest_event_ids, - ) - - event.internal_metadata.proactively_send = False - - await self.send_nonmember_event( - requester, event, context, ratelimit=False - ) - dummy_event_sent = True - break - except ConsentNotGivenError: - logger.info( - "Failed to send dummy event into room %s for user %s due to " - "lack of consent. Will try another user" % (room_id, user_id) - ) - except AuthError: - logger.info( - "Failed to send dummy event into room %s for user %s due to " - "lack of power. Will try another user" % (room_id, user_id) - ) + dummy_event_sent = await self._send_dummy_event_for_room(room_id) if not dummy_event_sent: # Did not find a valid user in the room, so remove from future attempts @@ -1193,6 +1198,59 @@ class EventCreationHandler(object): now = self.clock.time_msec() self._rooms_to_exclude_from_dummy_event_insertion[room_id] = now + async def _send_dummy_event_for_room(self, room_id: str) -> bool: + """Attempt to send a dummy event for the given room. + + Args: + room_id: room to try to send an event from + + Returns: + True if a dummy event was successfully sent. False if no user was able + to send an event. + """ + + # For each room we need to find a joined member we can use to send + # the dummy event with. + latest_event_ids = await self.store.get_prev_events_for_room(room_id) + members = await self.state.get_current_users_in_room( + room_id, latest_event_ids=latest_event_ids + ) + for user_id in members: + if not self.hs.is_mine_id(user_id): + continue + requester = create_requester(user_id) + try: + event, context = await self.create_event( + requester, + { + "type": "org.matrix.dummy_event", + "content": {}, + "room_id": room_id, + "sender": user_id, + }, + prev_event_ids=latest_event_ids, + ) + + event.internal_metadata.proactively_send = False + + # Since this is a dummy-event it is OK if it is sent by a + # shadow-banned user. + await self.send_nonmember_event( + requester, event, context, ratelimit=False, ignore_shadow_ban=True, + ) + return True + except ConsentNotGivenError: + logger.info( + "Failed to send dummy event into room %s for user %s due to " + "lack of consent. Will try another user" % (room_id, user_id) + ) + except AuthError: + logger.info( + "Failed to send dummy event into room %s for user %s due to " + "lack of power. Will try another user" % (room_id, user_id) + ) + return False + def _expire_rooms_to_exclude_from_dummy_event_insertion(self): expire_before = self.clock.time_msec() - _DUMMY_EVENT_ROOM_EXCLUSION_EXPIRY to_expire = set() diff --git a/synapse/handlers/oidc_handler.py b/synapse/handlers/oidc_handler.py
index 87f0c5e197..19cd652675 100644 --- a/synapse/handlers/oidc_handler.py +++ b/synapse/handlers/oidc_handler.py
@@ -12,9 +12,8 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import json import logging -from typing import Dict, Generic, List, Optional, Tuple, TypeVar +from typing import TYPE_CHECKING, Dict, Generic, List, Optional, Tuple, TypeVar from urllib.parse import urlencode import attr @@ -38,9 +37,11 @@ from synapse.config import ConfigError from synapse.http.server import respond_with_html from synapse.http.site import SynapseRequest from synapse.logging.context import make_deferred_yieldable -from synapse.push.mailer import load_jinja2_templates -from synapse.server import HomeServer -from synapse.types import UserID, map_username_to_mxid_localpart +from synapse.types import JsonDict, UserID, map_username_to_mxid_localpart +from synapse.util import json_decoder + +if TYPE_CHECKING: + from synapse.server import HomeServer logger = logging.getLogger(__name__) @@ -91,7 +92,8 @@ class OidcHandler: """Handles requests related to the OpenID Connect login flow. """ - def __init__(self, hs: HomeServer): + def __init__(self, hs: "HomeServer"): + self.hs = hs self._callback_url = hs.config.oidc_callback_url # type: str self._scopes = hs.config.oidc_scopes # type: List[str] self._client_auth = ClientAuth( @@ -112,6 +114,7 @@ class OidcHandler: hs.config.oidc_user_mapping_provider_config ) # type: OidcMappingProvider self._skip_verification = hs.config.oidc_skip_verification # type: bool + self._allow_existing_users = hs.config.oidc_allow_existing_users # type: bool self._http_client = hs.get_proxied_http_client() self._auth_handler = hs.get_auth_handler() @@ -121,9 +124,7 @@ class OidcHandler: self._hostname = hs.hostname # type: str self._server_name = hs.config.server_name # type: str self._macaroon_secret_key = hs.config.macaroon_secret_key - self._error_template = load_jinja2_templates( - hs.config.sso_template_dir, ["sso_error.html"] - )[0] + self._error_template = hs.config.sso_error_template # identifier for the external_ids table self._auth_provider_id = "oidc" @@ -131,10 +132,10 @@ class OidcHandler: def _render_error( self, request, error: str, error_description: Optional[str] = None ) -> None: - """Renders the error template and respond with it. + """Render the error template and respond to the request with it. This is used to show errors to the user. The template of this page can - be found under ``synapse/res/templates/sso_error.html``. + be found under `synapse/res/templates/sso_error.html`. Args: request: The incoming request from the browser. @@ -368,7 +369,7 @@ class OidcHandler: # and check for an error field. If not, we respond with a generic # error message. try: - resp = json.loads(resp_body.decode("utf-8")) + resp = json_decoder.decode(resp_body.decode("utf-8")) error = resp["error"] description = resp.get("error_description", error) except (ValueError, KeyError): @@ -385,7 +386,7 @@ class OidcHandler: # Since it is a not a 5xx code, body should be a valid JSON. It will # raise if not. - resp = json.loads(resp_body.decode("utf-8")) + resp = json_decoder.decode(resp_body.decode("utf-8")) if "error" in resp: error = resp["error"] @@ -690,14 +691,31 @@ class OidcHandler: self._render_error(request, "invalid_token", str(e)) return + # Pull out the user-agent and IP from the request. + user_agent = request.requestHeaders.getRawHeaders(b"User-Agent", default=[b""])[ + 0 + ].decode("ascii", "surrogateescape") + ip_address = self.hs.get_ip_from_request(request) + # Call the mapper to register/login the user try: - user_id = await self._map_userinfo_to_user(userinfo, token) + user_id = await self._map_userinfo_to_user( + userinfo, token, user_agent, ip_address + ) except MappingException as e: logger.exception("Could not map user") self._render_error(request, "mapping_error", str(e)) return + # Mapping providers might not have get_extra_attributes: only call this + # method if it exists. + extra_attributes = None + get_extra_attributes = getattr( + self._user_mapping_provider, "get_extra_attributes", None + ) + if get_extra_attributes: + extra_attributes = await get_extra_attributes(userinfo, token) + # and finally complete the login if ui_auth_session_id: await self._auth_handler.complete_sso_ui_auth( @@ -705,7 +723,7 @@ class OidcHandler: ) else: await self._auth_handler.complete_sso_login( - user_id, request, client_redirect_url + user_id, request, client_redirect_url, extra_attributes ) def _generate_oidc_session_token( @@ -829,7 +847,9 @@ class OidcHandler: now = self._clock.time_msec() return now < expiry - async def _map_userinfo_to_user(self, userinfo: UserInfo, token: Token) -> str: + async def _map_userinfo_to_user( + self, userinfo: UserInfo, token: Token, user_agent: str, ip_address: str + ) -> str: """Maps a UserInfo object to a mxid. UserInfo should have a claim that uniquely identifies users. This claim @@ -839,11 +859,14 @@ class OidcHandler: If we don't find the user that way, we should register the user, mapping the localpart and the display name from the UserInfo. - If a user already exists with the mxid we've mapped, raise an exception. + If a user already exists with the mxid we've mapped and allow_existing_users + is disabled, raise an exception. Args: userinfo: an object representing the user token: a dict with the tokens obtained from the provider + user_agent: The user agent of the client making the request. + ip_address: The IP address of the client making the request. Raises: MappingException: if there was an error while mapping some properties @@ -857,6 +880,9 @@ class OidcHandler: raise MappingException( "Failed to extract subject from OIDC response: %s" % (e,) ) + # Some OIDC providers use integer IDs, but Synapse expects external IDs + # to be strings. + remote_user_id = str(remote_user_id) logger.info( "Looking for existing mapping for user %s:%s", @@ -890,19 +916,31 @@ class OidcHandler: localpart = map_username_to_mxid_localpart(attributes["localpart"]) - user_id = UserID(localpart, self._hostname) - if await self._datastore.get_users_by_id_case_insensitive(user_id.to_string()): - # This mxid is taken - raise MappingException( - "mxid '{}' is already taken".format(user_id.to_string()) + user_id = UserID(localpart, self._hostname).to_string() + users = await self._datastore.get_users_by_id_case_insensitive(user_id) + if users: + if self._allow_existing_users: + if len(users) == 1: + registered_user_id = next(iter(users)) + elif user_id in users: + registered_user_id = user_id + else: + raise MappingException( + "Attempted to login as '{}' but it matches more than one user inexactly: {}".format( + user_id, list(users.keys()) + ) + ) + else: + # This mxid is taken + raise MappingException("mxid '{}' is already taken".format(user_id)) + else: + # It's the first time this user is logging in and the mapped mxid was + # not taken, register the user + registered_user_id = await self._registration_handler.register_user( + localpart=localpart, + default_display_name=attributes["display_name"], + user_agent_ips=(user_agent, ip_address), ) - - # It's the first time this user is logging in and the mapped mxid was - # not taken, register the user - registered_user_id = await self._registration_handler.register_user( - localpart=localpart, default_display_name=attributes["display_name"], - ) - await self._datastore.record_user_external_id( self._auth_provider_id, remote_user_id, registered_user_id, ) @@ -955,7 +993,7 @@ class OidcMappingProvider(Generic[C]): async def map_user_attributes( self, userinfo: UserInfo, token: Token ) -> UserAttribute: - """Map a ``UserInfo`` objects into user attributes. + """Map a `UserInfo` object into user attributes. Args: userinfo: An object representing the user given by the OIDC provider @@ -966,6 +1004,18 @@ class OidcMappingProvider(Generic[C]): """ raise NotImplementedError() + async def get_extra_attributes(self, userinfo: UserInfo, token: Token) -> JsonDict: + """Map a `UserInfo` object into additional attributes passed to the client during login. + + Args: + userinfo: An object representing the user given by the OIDC provider + token: A dict with the tokens returned by the provider + + Returns: + A dict containing additional attributes. Must be JSON serializable. + """ + return {} + # Used to clear out "None" values in templates def jinja_finalize(thing): @@ -980,6 +1030,7 @@ class JinjaOidcMappingConfig: subject_claim = attr.ib() # type: str localpart_template = attr.ib() # type: Template display_name_template = attr.ib() # type: Optional[Template] + extra_attributes = attr.ib() # type: Dict[str, Template] class JinjaOidcMappingProvider(OidcMappingProvider[JinjaOidcMappingConfig]): @@ -1018,10 +1069,28 @@ class JinjaOidcMappingProvider(OidcMappingProvider[JinjaOidcMappingConfig]): % (e,) ) + extra_attributes = {} # type Dict[str, Template] + if "extra_attributes" in config: + extra_attributes_config = config.get("extra_attributes") or {} + if not isinstance(extra_attributes_config, dict): + raise ConfigError( + "oidc_config.user_mapping_provider.config.extra_attributes must be a dict" + ) + + for key, value in extra_attributes_config.items(): + try: + extra_attributes[key] = env.from_string(value) + except Exception as e: + raise ConfigError( + "invalid jinja template for oidc_config.user_mapping_provider.config.extra_attributes.%s: %r" + % (key, e) + ) + return JinjaOidcMappingConfig( subject_claim=subject_claim, localpart_template=localpart_template, display_name_template=display_name_template, + extra_attributes=extra_attributes, ) def get_remote_user_id(self, userinfo: UserInfo) -> str: @@ -1042,3 +1111,13 @@ class JinjaOidcMappingProvider(OidcMappingProvider[JinjaOidcMappingConfig]): display_name = None return UserAttribute(localpart=localpart, display_name=display_name) + + async def get_extra_attributes(self, userinfo: UserInfo, token: Token) -> JsonDict: + extras = {} # type: Dict[str, str] + for key, template in self._config.extra_attributes.items(): + try: + extras[key] = template.render(user=userinfo).strip() + except Exception as e: + # Log an error and skip this value (don't break login for this). + logger.error("Failed to render OIDC extra attribute %s: %s" % (key, e)) + return extras diff --git a/synapse/handlers/pagination.py b/synapse/handlers/pagination.py
index da06582d4b..2c2a633938 100644 --- a/synapse/handlers/pagination.py +++ b/synapse/handlers/pagination.py
@@ -14,23 +14,30 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from typing import TYPE_CHECKING, Any, Dict, Optional, Set from twisted.python.failure import Failure from synapse.api.constants import EventTypes, Membership from synapse.api.errors import SynapseError +from synapse.api.filtering import Filter from synapse.logging.context import run_in_background from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage.state import StateFilter -from synapse.types import RoomStreamToken +from synapse.streams.config import PaginationConfig +from synapse.types import Requester from synapse.util.async_helpers import ReadWriteLock from synapse.util.stringutils import random_string from synapse.visibility import filter_events_for_client +if TYPE_CHECKING: + from synapse.server import HomeServer + + logger = logging.getLogger(__name__) -class PurgeStatus(object): +class PurgeStatus: """Object tracking the status of a purge request This class contains information on the progress of a purge request, for @@ -58,14 +65,14 @@ class PurgeStatus(object): return {"status": PurgeStatus.STATUS_TEXT[self.status]} -class PaginationHandler(object): +class PaginationHandler: """Handles pagination and purge history requests. These are in the same handler due to the fact we need to block clients paginating during a purge. """ - def __init__(self, hs): + def __init__(self, hs: "HomeServer"): self.hs = hs self.auth = hs.get_auth() self.store = hs.get_datastore() @@ -75,13 +82,16 @@ class PaginationHandler(object): self._server_name = hs.hostname self.pagination_lock = ReadWriteLock() - self._purges_in_progress_by_room = set() + self._purges_in_progress_by_room = set() # type: Set[str] # map from purge id to PurgeStatus - self._purges_by_id = {} + self._purges_by_id = {} # type: Dict[str, PurgeStatus] self._event_serializer = hs.get_event_client_serializer() self._retention_default_max_lifetime = hs.config.retention_default_max_lifetime + self._retention_allowed_lifetime_min = hs.config.retention_allowed_lifetime_min + self._retention_allowed_lifetime_max = hs.config.retention_allowed_lifetime_max + if hs.config.retention_enabled: # Run the purge jobs described in the configuration file. for job in hs.config.retention_purge_jobs: @@ -96,7 +106,9 @@ class PaginationHandler(object): job["longest_max_lifetime"], ) - async def purge_history_for_rooms_in_range(self, min_ms, max_ms): + async def purge_history_for_rooms_in_range( + self, min_ms: Optional[int], max_ms: Optional[int] + ): """Purge outdated events from rooms within the given retention range. If a default retention policy is defined in the server's configuration and its @@ -104,14 +116,14 @@ class PaginationHandler(object): retention policy. Args: - min_ms (int|None): Duration in milliseconds that define the lower limit of + min_ms: Duration in milliseconds that define the lower limit of the range to handle (exclusive). If None, it means that the range has no lower limit. - max_ms (int|None): Duration in milliseconds that define the upper limit of + max_ms: Duration in milliseconds that define the upper limit of the range to handle (inclusive). If None, it means that the range has no upper limit. """ - # We want the storage layer to to include rooms with no retention policy in its + # We want the storage layer to include rooms with no retention policy in its # return value only if a default retention policy is defined in the server's # configuration and that policy's 'max_lifetime' is either lower (or equal) than # max_ms or higher than min_ms (or both). @@ -152,13 +164,32 @@ class PaginationHandler(object): ) continue - max_lifetime = retention_policy["max_lifetime"] + # If max_lifetime is None, it means that the room has no retention policy. + # Given we only retrieve such rooms when there's a default retention policy + # defined in the server's configuration, we can safely assume that's the + # case and use it for this room. + max_lifetime = ( + retention_policy["max_lifetime"] or self._retention_default_max_lifetime + ) - if max_lifetime is None: - # If max_lifetime is None, it means that include_null equals True, - # therefore we can safely assume that there is a default policy defined - # in the server's configuration. - max_lifetime = self._retention_default_max_lifetime + # Cap the effective max_lifetime to be within the range allowed in the + # config. + # We do this in two steps: + # 1. Make sure it's higher or equal to the minimum allowed value, and if + # it's not replace it with that value. This is because the server + # operator can be required to not delete information before a given + # time, e.g. to comply with freedom of information laws. + # 2. Make sure the resulting value is lower or equal to the maximum allowed + # value, and if it's not replace it with that value. This is because the + # server operator can be required to delete any data after a specific + # amount of time. + if self._retention_allowed_lifetime_min is not None: + max_lifetime = max(self._retention_allowed_lifetime_min, max_lifetime) + + if self._retention_allowed_lifetime_max is not None: + max_lifetime = min(max_lifetime, self._retention_allowed_lifetime_max) + + logger.debug("[purge] max_lifetime for room %s: %s", room_id, max_lifetime) # Figure out what token we should start purging at. ts = self.clock.time_msec() - max_lifetime @@ -195,18 +226,19 @@ class PaginationHandler(object): "_purge_history", self._purge_history, purge_id, room_id, token, True, ) - def start_purge_history(self, room_id, token, delete_local_events=False): + def start_purge_history( + self, room_id: str, token: str, delete_local_events: bool = False + ) -> str: """Start off a history purge on a room. Args: - room_id (str): The room to purge from - - token (str): topological token to delete events before - delete_local_events (bool): True to delete local events as well as + room_id: The room to purge from + token: topological token to delete events before + delete_local_events: True to delete local events as well as remote ones Returns: - str: unique ID for this purge transaction. + unique ID for this purge transaction. """ if room_id in self._purges_in_progress_by_room: raise SynapseError( @@ -225,15 +257,16 @@ class PaginationHandler(object): ) return purge_id - async def _purge_history(self, purge_id, room_id, token, delete_local_events): + async def _purge_history( + self, purge_id: str, room_id: str, token: str, delete_local_events: bool + ) -> None: """Carry out a history purge on a room. Args: - purge_id (str): The id for this purge - room_id (str): The room to purge from - token (str): topological token to delete events before - delete_local_events (bool): True to delete local events as well as - remote ones + purge_id: The id for this purge + room_id: The room to purge from + token: topological token to delete events before + delete_local_events: True to delete local events as well as remote ones """ self._purges_in_progress_by_room.add(room_id) try: @@ -258,20 +291,17 @@ class PaginationHandler(object): self.hs.get_reactor().callLater(24 * 3600, clear_purge) - def get_purge_status(self, purge_id): + def get_purge_status(self, purge_id: str) -> Optional[PurgeStatus]: """Get the current status of an active purge Args: - purge_id (str): purge_id returned by start_purge_history - - Returns: - PurgeStatus|None + purge_id: purge_id returned by start_purge_history """ return self._purges_by_id.get(purge_id) - async def purge_room(self, room_id): + async def purge_room(self, room_id: str) -> None: """Purge the given room from the database""" - with (await self.pagination_lock.write(room_id)): + with await self.pagination_lock.write(room_id): # check we know about the room await self.store.get_room_version_id(room_id) @@ -285,43 +315,38 @@ class PaginationHandler(object): async def get_messages( self, - requester, - room_id=None, - pagin_config=None, - as_client_event=True, - event_filter=None, - ): + requester: Requester, + room_id: str, + pagin_config: PaginationConfig, + as_client_event: bool = True, + event_filter: Optional[Filter] = None, + ) -> Dict[str, Any]: """Get messages in a room. Args: - requester (Requester): The user requesting messages. - room_id (str): The room they want messages from. - pagin_config (synapse.api.streams.PaginationConfig): The pagination - config rules to apply, if any. - as_client_event (bool): True to get events in client-server format. - event_filter (Filter): Filter to apply to results or None + requester: The user requesting messages. + room_id: The room they want messages from. + pagin_config: The pagination config rules to apply, if any. + as_client_event: True to get events in client-server format. + event_filter: Filter to apply to results or None Returns: - dict: Pagination API results + Pagination API results """ user_id = requester.user.to_string() if pagin_config.from_token: - room_token = pagin_config.from_token.room_key + from_token = pagin_config.from_token else: - pagin_config.from_token = ( - await self.hs.get_event_sources().get_current_token_for_pagination() - ) - room_token = pagin_config.from_token.room_key + from_token = self.hs.get_event_sources().get_current_token_for_pagination() - room_token = RoomStreamToken.parse(room_token) - - pagin_config.from_token = pagin_config.from_token.copy_and_replace( - "room_key", str(room_token) - ) + if pagin_config.limit is None: + # This shouldn't happen as we've set a default limit before this + # gets called. + raise Exception("limit not set") - source_config = pagin_config.get_source_config("room") + room_token = from_token.room_key - with (await self.pagination_lock.read(room_id)): + with await self.pagination_lock.read(room_id): ( membership, member_event_id, @@ -329,13 +354,13 @@ class PaginationHandler(object): room_id, user_id, allow_departed_users=True ) - if source_config.direction == "b": + if pagin_config.direction == "b": # if we're going backwards, we might need to backfill. This # requires that we have a topo token. if room_token.topological: - max_topo = room_token.topological + curr_topo = room_token.topological else: - max_topo = await self.store.get_max_topological_token( + curr_topo = await self.store.get_current_topological_token( room_id, room_token.stream ) @@ -343,27 +368,39 @@ class PaginationHandler(object): # If they have left the room then clamp the token to be before # they left the room, to save the effort of loading from the # database. + + # This is only None if the room is world_readable, in which + # case "JOIN" would have been returned. + assert member_event_id + leave_token = await self.store.get_topological_token_for_event( member_event_id ) - leave_token = RoomStreamToken.parse(leave_token) - if leave_token.topological < max_topo: - source_config.from_key = str(leave_token) + assert leave_token.topological is not None + + if leave_token.topological < curr_topo: + from_token = from_token.copy_and_replace( + "room_key", leave_token + ) await self.hs.get_handlers().federation_handler.maybe_backfill( - room_id, max_topo + room_id, curr_topo, limit=pagin_config.limit, ) + to_room_key = None + if pagin_config.to_token: + to_room_key = pagin_config.to_token.room_key + events, next_key = await self.store.paginate_room_events( room_id=room_id, - from_key=source_config.from_key, - to_key=source_config.to_key, - direction=source_config.direction, - limit=source_config.limit, + from_key=from_token.room_key, + to_key=to_room_key, + direction=pagin_config.direction, + limit=pagin_config.limit, event_filter=event_filter, ) - next_token = pagin_config.from_token.copy_and_replace("room_key", next_key) + next_token = from_token.copy_and_replace("room_key", next_key) if events: if event_filter: @@ -376,8 +413,8 @@ class PaginationHandler(object): if not events: return { "chunk": [], - "start": pagin_config.from_token.to_string(), - "end": next_token.to_string(), + "start": await from_token.to_string(self.store), + "end": await next_token.to_string(self.store), } state = None @@ -394,8 +431,8 @@ class PaginationHandler(object): ) if state_ids: - state = await self.store.get_events(list(state_ids.values())) - state = state.values() + state_dict = await self.store.get_events(list(state_ids.values())) + state = state_dict.values() time_now = self.clock.time_msec() @@ -405,8 +442,8 @@ class PaginationHandler(object): events, time_now, as_client_event=as_client_event ) ), - "start": pagin_config.from_token.to_string(), - "end": next_token.to_string(), + "start": await from_token.to_string(self.store), + "end": await next_token.to_string(self.store), } if state: diff --git a/synapse/handlers/password_policy.py b/synapse/handlers/password_policy.py
index d06b110269..88e2f87200 100644 --- a/synapse/handlers/password_policy.py +++ b/synapse/handlers/password_policy.py
@@ -22,7 +22,7 @@ from synapse.api.errors import Codes, PasswordRefusedError logger = logging.getLogger(__name__) -class PasswordPolicyHandler(object): +class PasswordPolicyHandler: def __init__(self, hs): self.policy = hs.config.password_policy self.enabled = hs.config.password_policy_enabled diff --git a/synapse/handlers/presence.py b/synapse/handlers/presence.py
index b3a3bb8c3f..1000ac95ff 100644 --- a/synapse/handlers/presence.py +++ b/synapse/handlers/presence.py
@@ -33,14 +33,14 @@ from typing_extensions import ContextManager import synapse.metrics from synapse.api.constants import EventTypes, Membership, PresenceState from synapse.api.errors import SynapseError +from synapse.api.presence import UserPresenceState from synapse.logging.context import run_in_background from synapse.logging.utils import log_function from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process from synapse.state import StateHandler -from synapse.storage.data_stores.main import DataStore -from synapse.storage.presence import UserPresenceState -from synapse.types import JsonDict, UserID, get_domain_from_id +from synapse.storage.databases.main import DataStore +from synapse.types import Collection, JsonDict, UserID, get_domain_from_id from synapse.util.async_helpers import Linearizer from synapse.util.caches.descriptors import cached from synapse.util.metrics import Measure @@ -319,7 +319,7 @@ class PresenceHandler(BasePresenceHandler): is some spurious presence changes that will self-correct. """ # If the DB pool has already terminated, don't try updating - if not self.store.db.is_running(): + if not self.store.db_pool.is_running(): return logger.info( @@ -1010,7 +1010,7 @@ def format_user_presence_state(state, now, include_user_id=True): return content -class PresenceEventSource(object): +class PresenceEventSource: def __init__(self, hs): # We can't call get_presence_handler here because there's a cycle: # @@ -1108,9 +1108,6 @@ class PresenceEventSource(object): def get_current_key(self): return self.store.get_current_presence_token() - async def get_pagination_rows(self, user, pagination_config, key): - return await self.get_new_events(user, from_key=None, include_offline=False) - @cached(num_args=2, cache_context=True) async def _get_interested_in(self, user, explicit_room_id, cache_context): """Returns the set of users that the given user should see presence @@ -1318,7 +1315,7 @@ async def get_interested_parties( async def get_interested_remotes( store: DataStore, states: List[UserPresenceState], state_handler: StateHandler -) -> List[Tuple[List[str], List[UserPresenceState]]]: +) -> List[Tuple[Collection[str], List[UserPresenceState]]]: """Given a list of presence states figure out which remote servers should be sent which. @@ -1334,7 +1331,7 @@ async def get_interested_remotes( each tuple the list of UserPresenceState should be sent to each destination """ - hosts_and_states = [] + hosts_and_states = [] # type: List[Tuple[Collection[str], List[UserPresenceState]]] # First we look up the rooms each user is in (as well as any explicit # subscriptions), then for each distinct room we look up the remote diff --git a/synapse/handlers/profile.py b/synapse/handlers/profile.py
index b05aa89455..c301ac0631 100644 --- a/synapse/handlers/profile.py +++ b/synapse/handlers/profile.py
@@ -15,11 +15,12 @@ # limitations under the License. import logging +import random from typing import List from signedjson.sign import sign_json -from twisted.internet import defer, reactor +from twisted.internet import reactor from synapse.api.errors import ( AuthError, @@ -52,7 +53,7 @@ class BaseProfileHandler(BaseHandler): PROFILE_REPLICATE_INTERVAL = 2 * 60 * 1000 def __init__(self, hs): - super(BaseProfileHandler, self).__init__(hs) + super().__init__(hs) self.federation = hs.get_federation_client() hs.get_federation_registry().register_query_handler( @@ -73,36 +74,45 @@ class BaseProfileHandler(BaseHandler): ) if len(self.hs.config.replicate_user_profiles_to) > 0: - reactor.callWhenRunning(self._assign_profile_replication_batches) - reactor.callWhenRunning(self._replicate_profiles) + reactor.callWhenRunning(self._do_assign_profile_replication_batches) + reactor.callWhenRunning(self._start_replicate_profiles) # Add a looping call to replicate_profiles: this handles retries # if the replication is unsuccessful when the user updated their # profile. self.clock.looping_call( - self._replicate_profiles, self.PROFILE_REPLICATE_INTERVAL + self._start_replicate_profiles, self.PROFILE_REPLICATE_INTERVAL ) - @defer.inlineCallbacks - def _assign_profile_replication_batches(self): + def _do_assign_profile_replication_batches(self): + return run_as_background_process( + "_assign_profile_replication_batches", + self._assign_profile_replication_batches, + ) + + def _start_replicate_profiles(self): + return run_as_background_process( + "_replicate_profiles", self._replicate_profiles + ) + + async def _assign_profile_replication_batches(self): """If no profile replication has been done yet, allocate replication batch numbers to each profile to start the replication process. """ logger.info("Assigning profile batch numbers...") total = 0 while True: - assigned = yield self.store.assign_profile_batch() + assigned = await self.store.assign_profile_batch() total += assigned if assigned == 0: break logger.info("Assigned %d profile batch numbers", total) - @defer.inlineCallbacks - def _replicate_profiles(self): + async def _replicate_profiles(self): """If any profile data has been updated and not pushed to the replication targets, replicate it. """ - host_batches = yield self.store.get_replication_hosts() - latest_batch = yield self.store.get_latest_profile_replication_batch_number() + host_batches = await self.store.get_replication_hosts() + latest_batch = await self.store.get_latest_profile_replication_batch_number() if latest_batch is None: latest_batch = -1 for repl_host in self.hs.config.replicate_user_profiles_to: @@ -110,16 +120,15 @@ class BaseProfileHandler(BaseHandler): host_batches[repl_host] = -1 try: for i in range(host_batches[repl_host] + 1, latest_batch + 1): - yield self._replicate_host_profile_batch(repl_host, i) + await self._replicate_host_profile_batch(repl_host, i) except Exception: logger.exception( "Exception while replicating to %s: aborting for now", repl_host ) - @defer.inlineCallbacks - def _replicate_host_profile_batch(self, host, batchnum): + async def _replicate_host_profile_batch(self, host, batchnum): logger.info("Replicating profile batch %d to %s", batchnum, host) - batch_rows = yield self.store.get_profile_batch(batchnum) + batch_rows = await self.store.get_profile_batch(batchnum) batch = { UserID(r["user_id"], self.hs.hostname).to_string(): ( {"display_name": r["displayname"], "avatar_url": r["avatar_url"]} @@ -133,9 +142,11 @@ class BaseProfileHandler(BaseHandler): body = {"batchnum": batchnum, "batch": batch, "origin_server": self.hs.hostname} signed_body = sign_json(body, self.hs.hostname, self.hs.config.signing_key[0]) try: - yield self.http_client.post_json_get_json(url, signed_body) - yield self.store.update_replication_batch_for_host(host, batchnum) - logger.info("Sucessfully replicated profile batch %d to %s", batchnum, host) + await self.http_client.post_json_get_json(url, signed_body) + await self.store.update_replication_batch_for_host(host, batchnum) + logger.info( + "Successfully replicated profile batch %d to %s", batchnum, host + ) except Exception: # This will get retried when the looping call next comes around logger.exception( @@ -251,6 +262,9 @@ class BaseProfileHandler(BaseHandler): Codes.FORBIDDEN, ) + if not isinstance(new_displayname, str): + raise SynapseError(400, "Invalid displayname") + if len(new_displayname) > MAX_DISPLAYNAME_LEN: raise SynapseError( 400, "Displayname is too long (max %i)" % (MAX_DISPLAYNAME_LEN,) @@ -288,8 +302,7 @@ class BaseProfileHandler(BaseHandler): # start a profile replication push run_in_background(self._replicate_profiles) - @defer.inlineCallbacks - def set_active( + async def set_active( self, users: List[UserID], active: bool, hide: bool, ): """ @@ -312,19 +325,16 @@ class BaseProfileHandler(BaseHandler): hide: Whether to hide the user (withold from replication). If False and active is False, user will have their profile erased - - Returns: - Deferred """ if len(self.replicate_user_profiles_to) > 0: cur_batchnum = ( - yield self.store.get_latest_profile_replication_batch_number() + await self.store.get_latest_profile_replication_batch_number() ) new_batchnum = 0 if cur_batchnum is None else cur_batchnum + 1 else: new_batchnum = None - yield self.store.set_profiles_active(users, active, hide, new_batchnum) + await self.store.set_profiles_active(users, active, hide, new_batchnum) # start a profile replication push run_in_background(self._replicate_profiles) @@ -358,8 +368,14 @@ class BaseProfileHandler(BaseHandler): async def set_avatar_url( self, target_user, requester, new_avatar_url, by_admin=False ): - """target_user is the user whose avatar_url is to be changed; - auth_user is the user attempting to make this change.""" + """Set a new avatar URL for a user. + + Args: + target_user (UserID): the user whose avatar URL is to be changed. + requester (Requester): The user attempting to make this change. + new_avatar_url (str): The avatar URL to give this user. + by_admin (bool): Whether this change was made by an administrator. + """ if not self.hs.is_mine(target_user): raise SynapseError(400, "User is not hosted on this homeserver") @@ -373,6 +389,9 @@ class BaseProfileHandler(BaseHandler): 400, "Changing avatar is disabled on this server", Codes.FORBIDDEN ) + if not isinstance(new_avatar_url, str): + raise SynapseError(400, "Invalid displayname") + if len(new_avatar_url) > MAX_AVATAR_URL_LEN: raise SynapseError( 400, "Avatar URL is too long (max %i)" % (MAX_AVATAR_URL_LEN,) @@ -480,6 +499,12 @@ class BaseProfileHandler(BaseHandler): await self.ratelimit(requester) + # Do not actually update the room state for shadow-banned users. + if requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + return + room_ids = await self.store.get_rooms_for_user(target_user.to_string()) for room_id in room_ids: @@ -552,7 +577,7 @@ class MasterProfileHandler(BaseProfileHandler): PROFILE_UPDATE_EVERY_MS = 24 * 60 * 60 * 1000 def __init__(self, hs): - super(MasterProfileHandler, self).__init__(hs) + super().__init__(hs) assert hs.config.worker_app is None diff --git a/synapse/handlers/read_marker.py b/synapse/handlers/read_marker.py
index e3b528d271..c32f314a1c 100644 --- a/synapse/handlers/read_marker.py +++ b/synapse/handlers/read_marker.py
@@ -24,7 +24,7 @@ logger = logging.getLogger(__name__) class ReadMarkerHandler(BaseHandler): def __init__(self, hs): - super(ReadMarkerHandler, self).__init__(hs) + super().__init__(hs) self.server_name = hs.config.server_name self.store = hs.get_datastore() self.read_marker_linearizer = Linearizer(name="read_marker") diff --git a/synapse/handlers/receipts.py b/synapse/handlers/receipts.py
index f922d8a545..7225923757 100644 --- a/synapse/handlers/receipts.py +++ b/synapse/handlers/receipts.py
@@ -23,7 +23,7 @@ logger = logging.getLogger(__name__) class ReceiptsHandler(BaseHandler): def __init__(self, hs): - super(ReceiptsHandler, self).__init__(hs) + super().__init__(hs) self.server_name = hs.config.server_name self.store = hs.get_datastore() @@ -123,7 +123,7 @@ class ReceiptsHandler(BaseHandler): await self.federation.send_read_receipt(receipt) -class ReceiptEventSource(object): +class ReceiptEventSource: def __init__(self, hs): self.store = hs.get_datastore() @@ -142,18 +142,3 @@ class ReceiptEventSource(object): def get_current_key(self, direction="f"): return self.store.get_max_receipt_stream_id() - - async def get_pagination_rows(self, user, config, key): - to_key = int(config.from_key) - - if config.to_key: - from_key = int(config.to_key) - else: - from_key = None - - room_ids = await self.store.get_rooms_for_user(user.to_string()) - events = await self.store.get_linearized_receipts_for_rooms( - room_ids, from_key=from_key, to_key=to_key - ) - - return (events, to_key) diff --git a/synapse/handlers/register.py b/synapse/handlers/register.py
index d00b9dc537..a26d821b53 100644 --- a/synapse/handlers/register.py +++ b/synapse/handlers/register.py
@@ -26,6 +26,7 @@ from synapse.replication.http.register import ( ReplicationPostRegisterActionsServlet, ReplicationRegisterServlet, ) +from synapse.spam_checker_api import RegistrationBehaviour from synapse.storage.state import StateFilter from synapse.types import RoomAlias, UserID, create_requester @@ -41,7 +42,7 @@ class RegistrationHandler(BaseHandler): Args: hs (synapse.server.HomeServer): """ - super(RegistrationHandler, self).__init__(hs) + super().__init__(hs) self.hs = hs self.auth = hs.get_auth() self._auth_handler = hs.get_auth_handler() @@ -53,6 +54,8 @@ class RegistrationHandler(BaseHandler): self.macaroon_gen = hs.get_macaroon_generator() self._server_notices_mxid = hs.config.server_notices_mxid + self.spam_checker = hs.get_spam_checker() + self._show_in_user_directory = self.hs.config.show_users_in_user_directory if hs.config.worker_app: @@ -139,7 +142,9 @@ class RegistrationHandler(BaseHandler): try: int(localpart) raise SynapseError( - 400, "Numeric user IDs are reserved for guest users." + 400, + "Numeric user IDs are reserved for guest users.", + errcode=Codes.INVALID_USERNAME, ) except ValueError: pass @@ -157,6 +162,7 @@ class RegistrationHandler(BaseHandler): address=None, bind_emails=[], by_admin=False, + user_agent_ips=None, ): """Registers a new client on the server. @@ -174,6 +180,8 @@ class RegistrationHandler(BaseHandler): bind_emails (List[str]): list of emails to bind to this account. by_admin (bool): True if this registration is being made via the admin api, otherwise False. + user_agent_ips (List[(str, str)]): Tuples of IP addresses and user-agents used + during the registration process. Returns: str: user_id Raises: @@ -181,6 +189,24 @@ class RegistrationHandler(BaseHandler): """ self.check_registration_ratelimit(address) + result = self.spam_checker.check_registration_for_spam( + threepid, localpart, user_agent_ips or [], + ) + + if result == RegistrationBehaviour.DENY: + logger.info( + "Blocked registration of %r", localpart, + ) + # We return a 429 to make it not obvious that they've been + # denied. + raise SynapseError(429, "Rate limited") + + shadow_banned = result == RegistrationBehaviour.SHADOW_BAN + if shadow_banned: + logger.info( + "Shadow banning registration of %r", localpart, + ) + # do not check_auth_blocking if the call is coming through the Admin API if not by_admin: await self.auth.check_auth_blocking(threepid=threepid) @@ -209,6 +235,7 @@ class RegistrationHandler(BaseHandler): admin=admin, user_type=user_type, address=address, + shadow_banned=shadow_banned, ) if default_display_name: @@ -244,6 +271,7 @@ class RegistrationHandler(BaseHandler): make_guest=make_guest, create_profile_with_displayname=default_display_name, address=address, + shadow_banned=shadow_banned, ) await self.profile_handler.set_displayname( @@ -620,6 +648,7 @@ class RegistrationHandler(BaseHandler): admin=False, user_type=None, address=None, + shadow_banned=False, ): """Register user in the datastore. @@ -637,9 +666,10 @@ class RegistrationHandler(BaseHandler): user_type (str|None): type of user. One of the values from api.constants.UserTypes, or None for a normal user. address (str|None): the IP address used to perform the registration. + shadow_banned (bool): Whether to shadow-ban the user Returns: - Deferred + Awaitable """ if self.hs.config.worker_app: return self._register_client( @@ -652,6 +682,7 @@ class RegistrationHandler(BaseHandler): admin=admin, user_type=user_type, address=address, + shadow_banned=shadow_banned, ) else: return self.store.register_user( @@ -663,6 +694,7 @@ class RegistrationHandler(BaseHandler): create_profile_with_displayname=create_profile_with_displayname, admin=admin, user_type=user_type, + shadow_banned=shadow_banned, ) async def register_device( diff --git a/synapse/handlers/room.py b/synapse/handlers/room.py
index 4ab229963a..1333999782 100644 --- a/synapse/handlers/room.py +++ b/synapse/handlers/room.py
@@ -20,9 +20,10 @@ import itertools import logging import math +import random import string from collections import OrderedDict -from typing import Optional, Tuple +from typing import TYPE_CHECKING, Any, Awaitable, Dict, List, Optional, Tuple from synapse.api.constants import ( EventTypes, @@ -32,11 +33,15 @@ from synapse.api.constants import ( RoomEncryptionAlgorithms, ) from synapse.api.errors import AuthError, Codes, NotFoundError, StoreError, SynapseError +from synapse.api.filtering import Filter from synapse.api.room_versions import KNOWN_ROOM_VERSIONS, RoomVersion +from synapse.events import EventBase from synapse.events.utils import copy_power_levels_contents from synapse.http.endpoint import parse_and_validate_server_name from synapse.storage.state import StateFilter from synapse.types import ( + JsonDict, + MutableStateMap, Requester, RoomAlias, RoomID, @@ -47,12 +52,15 @@ from synapse.types import ( create_requester, ) from synapse.util import stringutils -from synapse.util.async_helpers import Linearizer, maybe_awaitable +from synapse.util.async_helpers import Linearizer from synapse.util.caches.response_cache import ResponseCache from synapse.visibility import filter_events_for_client from ._base import BaseHandler +if TYPE_CHECKING: + from synapse.server import HomeServer + logger = logging.getLogger(__name__) id_server_scheme = "https://" @@ -61,8 +69,8 @@ FIVE_MINUTES_IN_MS = 5 * 60 * 1000 class RoomCreationHandler(BaseHandler): - def __init__(self, hs): - super(RoomCreationHandler, self).__init__(hs) + def __init__(self, hs: "HomeServer"): + super().__init__(hs) self.spam_checker = hs.get_spam_checker() self.event_creation_handler = hs.get_event_creation_handler() @@ -92,7 +100,7 @@ class RoomCreationHandler(BaseHandler): "guest_can_join": False, "power_level_content_override": {}, }, - } + } # type: Dict[str, Dict[str, Any]] # Modify presets to selectively enable encryption by default per homeserver config for preset_name, preset_config in self._presets_dict.items(): @@ -129,6 +137,9 @@ class RoomCreationHandler(BaseHandler): Returns: the new room id + + Raises: + ShadowBanError if the requester is shadow-banned. """ await self.ratelimit(requester) @@ -164,6 +175,15 @@ class RoomCreationHandler(BaseHandler): async def _upgrade_room( self, requester: Requester, old_room_id: str, new_version: RoomVersion ): + """ + Args: + requester: the user requesting the upgrade + old_room_id: the id of the room to be replaced + new_versions: the version to upgrade the room to + + Raises: + ShadowBanError if the requester is shadow-banned. + """ user_id = requester.user.to_string() # start by allocating a new room id @@ -215,6 +235,9 @@ class RoomCreationHandler(BaseHandler): old_room_state = await tombstone_context.get_current_state_ids() + # We know the tombstone event isn't an outlier so it has current state. + assert old_room_state is not None + # update any aliases await self._move_aliases_to_new_room( requester, old_room_id, new_room_id, old_room_state @@ -247,6 +270,9 @@ class RoomCreationHandler(BaseHandler): old_room_id: the id of the room to be replaced new_room_id: the id of the replacement room old_room_state: the state map for the old room + + Raises: + ShadowBanError if the requester is shadow-banned. """ old_room_pl_event_id = old_room_state.get((EventTypes.PowerLevels, "")) @@ -437,7 +463,7 @@ class RoomCreationHandler(BaseHandler): old_room_member_state_events = await self.store.get_events( old_room_member_state_ids.values() ) - for k, old_event in old_room_member_state_events.items(): + for old_event in old_room_member_state_events.values(): # Only transfer ban events if ( "membership" in old_event.content @@ -540,17 +566,21 @@ class RoomCreationHandler(BaseHandler): logger.error("Unable to send updated alias events in new room: %s", e) async def create_room( - self, requester, config, ratelimit=True, creator_join_profile=None + self, + requester: Requester, + config: JsonDict, + ratelimit: bool = True, + creator_join_profile: Optional[JsonDict] = None, ) -> Tuple[dict, int]: """ Creates a new room. Args: - requester (synapse.types.Requester): + requester: The user who requested the room creation. - config (dict) : A dict of configuration options. - ratelimit (bool): set to False to disable the rate limiter + config : A dict of configuration options. + ratelimit: set to False to disable the rate limiter - creator_join_profile (dict|None): + creator_join_profile: Set to override the displayname and avatar for the creating user in this room. If unset, displayname and avatar will be derived from the user's profile. If set, should contain the @@ -619,6 +649,7 @@ class RoomCreationHandler(BaseHandler): Codes.UNSUPPORTED_ROOM_VERSION, ) + room_alias = None if "room_alias_name" in config: for wchar in string.whitespace: if wchar in config["room_alias_name"]: @@ -629,9 +660,9 @@ class RoomCreationHandler(BaseHandler): if mapping: raise SynapseError(400, "Room alias already taken", Codes.ROOM_IN_USE) - else: - room_alias = None + invite_3pid_list = config.get("invite_3pid", []) + invite_list = config.get("invite", []) for i in invite_list: try: uid = UserID.from_string(i) @@ -639,6 +670,14 @@ class RoomCreationHandler(BaseHandler): except Exception: raise SynapseError(400, "Invalid user_id: %s" % (i,)) + if (invite_list or invite_3pid_list) and requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + + # Allow the request to go through, but remove any associated invites. + invite_3pid_list = [] + invite_list = [] + await self.event_creation_handler.assert_accepted_privacy_policy(requester) power_level_content_override = config.get("power_level_content_override") @@ -756,6 +795,8 @@ class RoomCreationHandler(BaseHandler): if is_direct: content["is_direct"] = is_direct + # Note that update_membership with an action of "invite" can raise a + # ShadowBanError, but this was handled above by emptying invite_list. _, last_stream_id = await self.room_member_handler.update_membership( requester, UserID.from_string(invitee), @@ -771,6 +812,8 @@ class RoomCreationHandler(BaseHandler): id_access_token = invite_3pid.get("id_access_token") # optional address = invite_3pid["address"] medium = invite_3pid["medium"] + # Note that do_3pid_invite can raise a ShadowBanError, but this was + # handled above by emptying invite_3pid_list. last_stream_id = await self.hs.get_room_member_handler().do_3pid_invite( room_id, requester.user, @@ -790,30 +833,39 @@ class RoomCreationHandler(BaseHandler): # Always wait for room creation to progate before returning await self._replication.wait_for_stream_position( - self.hs.config.worker.writers.events, "events", last_stream_id + self.hs.config.worker.events_shard_config.get_instance(room_id), + "events", + last_stream_id, ) return result, last_stream_id async def _send_events_for_new_room( self, - creator, # A Requester object. - room_id, - preset_config, - invite_list, - initial_state, - creation_content, - room_alias=None, - power_level_content_override=None, # Doesn't apply when initial state has power level state event content - creator_join_profile=None, + creator: Requester, + room_id: str, + preset_config: str, + invite_list: List[str], + initial_state: MutableStateMap, + creation_content: JsonDict, + room_alias: Optional[RoomAlias] = None, + power_level_content_override: Optional[JsonDict] = None, + creator_join_profile: Optional[JsonDict] = None, ) -> int: """Sends the initial events into a new room. + `power_level_content_override` doesn't apply when initial state has + power level state event content. + Returns: The stream_id of the last event persisted. """ - def create(etype, content, **kwargs): + creator_id = creator.user.to_string() + + event_keys = {"room_id": room_id, "sender": creator_id, "state_key": ""} + + def create(etype: str, content: JsonDict, **kwargs) -> JsonDict: e = {"type": etype, "content": content} e.update(event_keys) @@ -821,23 +873,21 @@ class RoomCreationHandler(BaseHandler): return e - async def send(etype, content, **kwargs) -> int: + async def send(etype: str, content: JsonDict, **kwargs) -> int: event = create(etype, content, **kwargs) logger.debug("Sending %s in new room", etype) + # Allow these events to be sent even if the user is shadow-banned to + # allow the room creation to complete. ( _, last_stream_id, ) = await self.event_creation_handler.create_and_send_nonmember_event( - creator, event, ratelimit=False + creator, event, ratelimit=False, ignore_shadow_ban=True, ) return last_stream_id config = self._presets_dict[preset_config] - creator_id = creator.user.to_string() - - event_keys = {"room_id": room_id, "sender": creator_id, "state_key": ""} - creation_content.update({"creator": creator_id}) await send(etype=EventTypes.Create, content=creation_content) @@ -879,7 +929,7 @@ class RoomCreationHandler(BaseHandler): "kick": 50, "redact": 50, "invite": 50, - } + } # type: JsonDict if config["original_invitees_have_ops"]: for invitee in invite_list: @@ -933,7 +983,7 @@ class RoomCreationHandler(BaseHandler): return last_sent_stream_id async def _generate_room_id( - self, creator_id: str, is_public: str, room_version: RoomVersion, + self, creator_id: str, is_public: bool, room_version: RoomVersion, ): # autogen room IDs and try to create it. We may clash, so just # try a few times till one goes through, giving up eventually. @@ -956,24 +1006,31 @@ class RoomCreationHandler(BaseHandler): raise StoreError(500, "Couldn't generate a room ID.") -class RoomContextHandler(object): - def __init__(self, hs): +class RoomContextHandler: + def __init__(self, hs: "HomeServer"): self.hs = hs self.store = hs.get_datastore() self.storage = hs.get_storage() self.state_store = self.storage.state - async def get_event_context(self, user, room_id, event_id, limit, event_filter): + async def get_event_context( + self, + user: UserID, + room_id: str, + event_id: str, + limit: int, + event_filter: Optional[Filter], + ) -> Optional[JsonDict]: """Retrieves events, pagination tokens and state around a given event in a room. Args: - user (UserID) - room_id (str) - event_id (str) - limit (int): The maximum number of events to return in total + user + room_id + event_id + limit: The maximum number of events to return in total (excluding state). - event_filter (Filter|None): the filter to apply to the events returned + event_filter: the filter to apply to the events returned (excluding the target event_id) Returns: @@ -1050,30 +1107,37 @@ class RoomContextHandler(object): # the token, which we replace. token = StreamToken.START - results["start"] = token.copy_and_replace( + results["start"] = await token.copy_and_replace( "room_key", results["start"] - ).to_string() + ).to_string(self.store) - results["end"] = token.copy_and_replace("room_key", results["end"]).to_string() + results["end"] = await token.copy_and_replace( + "room_key", results["end"] + ).to_string(self.store) return results -class RoomEventSource(object): - def __init__(self, hs): +class RoomEventSource: + def __init__(self, hs: "HomeServer"): self.store = hs.get_datastore() async def get_new_events( - self, user, from_key, limit, room_ids, is_guest, explicit_room_id=None - ): + self, + user: UserID, + from_key: RoomStreamToken, + limit: int, + room_ids: List[str], + is_guest: bool, + explicit_room_id: Optional[str] = None, + ) -> Tuple[List[EventBase], RoomStreamToken]: # We just ignore the key for now. - to_key = await self.get_current_key() + to_key = self.get_current_key() - from_token = RoomStreamToken.parse(from_key) - if from_token.topological: + if from_key.topological: logger.warning("Stream has topological part!!!! %r", from_key) - from_key = "s%s" % (from_token.stream,) + from_key = RoomStreamToken(None, from_key.stream) app_service = self.store.get_app_service_by_user_id(user.to_string()) if app_service: @@ -1108,14 +1172,14 @@ class RoomEventSource(object): return (events, end_key) - def get_current_key(self): - return self.store.get_room_events_max_id() + def get_current_key(self) -> RoomStreamToken: + return self.store.get_room_max_token() - def get_current_key_for_room(self, room_id): + def get_current_key_for_room(self, room_id: str) -> Awaitable[str]: return self.store.get_room_events_max_id(room_id) -class RoomShutdownHandler(object): +class RoomShutdownHandler: DEFAULT_MESSAGE = ( "Sharing illegal content on this server is not permitted and rooms in" @@ -1123,7 +1187,7 @@ class RoomShutdownHandler(object): ) DEFAULT_ROOM_NAME = "Content Violation Notification" - def __init__(self, hs): + def __init__(self, hs: "HomeServer"): self.hs = hs self.room_member_handler = hs.get_room_member_handler() self._room_creation_handler = hs.get_room_creation_handler() @@ -1229,10 +1293,10 @@ class RoomShutdownHandler(object): # We now wait for the create room to come back in via replication so # that we can assume that all the joins/invites have propogated before # we try and auto join below. - # - # TODO: Currently the events stream is written to from master await self._replication.wait_for_stream_position( - self.hs.config.worker.writers.events, "events", stream_id + self.hs.config.worker.events_shard_config.get_instance(new_room_id), + "events", + stream_id, ) else: new_room_id = None @@ -1262,7 +1326,9 @@ class RoomShutdownHandler(object): # Wait for leave to come in over replication before trying to forget. await self._replication.wait_for_stream_position( - self.hs.config.worker.writers.events, "events", stream_id + self.hs.config.worker.events_shard_config.get_instance(room_id), + "events", + stream_id, ) await self.room_member_handler.forget(target_requester.user, room_id) @@ -1299,9 +1365,7 @@ class RoomShutdownHandler(object): ratelimit=False, ) - aliases_for_room = await maybe_awaitable( - self.store.get_aliases_for_room(room_id) - ) + aliases_for_room = await self.store.get_aliases_for_room(room_id) await self.store.update_aliases_for_room( room_id, new_room_id, requester_user_id diff --git a/synapse/handlers/room_list.py b/synapse/handlers/room_list.py
index 5dd7b28391..4a13c8e912 100644 --- a/synapse/handlers/room_list.py +++ b/synapse/handlers/room_list.py
@@ -38,7 +38,7 @@ EMPTY_THIRD_PARTY_ID = ThirdPartyInstanceID(None, None) class RoomListHandler(BaseHandler): def __init__(self, hs): - super(RoomListHandler, self).__init__(hs) + super().__init__(hs) self.enable_room_list_search = hs.config.enable_room_list_search self.response_cache = ResponseCache(hs, "room_list") self.remote_response_cache = ResponseCache( diff --git a/synapse/handlers/room_member.py b/synapse/handlers/room_member.py
index c3fee116c2..5a8120db57 100644 --- a/synapse/handlers/room_member.py +++ b/synapse/handlers/room_member.py
@@ -15,14 +15,22 @@ import abc import logging +import random from http import HTTPStatus -from typing import Dict, Iterable, List, Optional, Tuple, Union +from typing import TYPE_CHECKING, Iterable, List, Optional, Tuple, Union from unpaddedbase64 import encode_base64 from synapse import types from synapse.api.constants import MAX_DEPTH, EventTypes, Membership -from synapse.api.errors import AuthError, Codes, SynapseError +from synapse.api.errors import ( + AuthError, + Codes, + LimitExceededError, + ShadowBanError, + SynapseError, +) +from synapse.api.ratelimiting import Ratelimiter from synapse.api.room_versions import EventFormatVersions from synapse.crypto.event_signing import compute_event_reference_hash from synapse.events import EventBase @@ -30,24 +38,26 @@ from synapse.events.builder import create_local_event_from_event_dict from synapse.events.snapshot import EventContext from synapse.events.validator import EventValidator from synapse.storage.roommember import RoomsForUser -from synapse.types import Collection, JsonDict, Requester, RoomAlias, RoomID, UserID +from synapse.types import JsonDict, Requester, RoomAlias, RoomID, StateMap, UserID from synapse.util.async_helpers import Linearizer -from synapse.util.distributor import user_joined_room, user_left_room +from synapse.util.distributor import user_left_room from ._base import BaseHandler +if TYPE_CHECKING: + from synapse.server import HomeServer + + logger = logging.getLogger(__name__) -class RoomMemberHandler(object): +class RoomMemberHandler(metaclass=abc.ABCMeta): # TODO(paul): This handler currently contains a messy conflation of # low-level API that works on UserID objects and so on, and REST-level # API that takes ID strings and returns pagination chunks. These concerns # ought to be separated out a lot better. - __metaclass__ = abc.ABCMeta - - def __init__(self, hs): + def __init__(self, hs: "HomeServer"): self.hs = hs self.store = hs.get_datastore() self.auth = hs.get_auth() @@ -71,12 +81,16 @@ class RoomMemberHandler(object): self._enable_lookup = hs.config.enable_3pid_lookup self.allow_per_room_profiles = self.config.allow_per_room_profiles - self._event_stream_writer_instance = hs.config.worker.writers.events - self._is_on_event_persistence_instance = ( - self._event_stream_writer_instance == hs.get_instance_name() + self._join_rate_limiter_local = Ratelimiter( + clock=self.clock, + rate_hz=hs.config.ratelimiting.rc_joins_local.per_second, + burst_count=hs.config.ratelimiting.rc_joins_local.burst_count, + ) + self._join_rate_limiter_remote = Ratelimiter( + clock=self.clock, + rate_hz=hs.config.ratelimiting.rc_joins_remote.per_second, + burst_count=hs.config.ratelimiting.rc_joins_remote.burst_count, ) - if self._is_on_event_persistence_instance: - self.persist_event_storage = hs.get_storage().persistence # This is only used to get at ratelimit function, and # maybe_kick_guest_users. It's fine there are multiple of these as @@ -127,17 +141,6 @@ class RoomMemberHandler(object): raise NotImplementedError() @abc.abstractmethod - async def _user_joined_room(self, target: UserID, room_id: str) -> None: - """Notifies distributor on master process that the user has joined the - room. - - Args: - target - room_id - """ - raise NotImplementedError() - - @abc.abstractmethod async def _user_left_room(self, target: UserID, room_id: str) -> None: """Notifies distributor on master process that the user has left the room. @@ -154,7 +157,7 @@ class RoomMemberHandler(object): target: UserID, room_id: str, membership: str, - prev_event_ids: Collection[str], + prev_event_ids: List[str], txn_id: Optional[str] = None, ratelimit: bool = True, content: Optional[dict] = None, @@ -195,25 +198,35 @@ class RoomMemberHandler(object): _, stream_id = await self.store.get_event_ordering(duplicate.event_id) return duplicate.event_id, stream_id - stream_id = await self.event_creation_handler.handle_new_client_event( - requester, event, context, extra_users=[target], ratelimit=ratelimit - ) - prev_state_ids = await context.get_prev_state_ids() prev_member_event_id = prev_state_ids.get((EventTypes.Member, user_id), None) if event.membership == Membership.JOIN: - # Only fire user_joined_room if the user has actually joined the - # room. Don't bother if the user is just changing their profile - # info. newly_joined = True if prev_member_event_id: prev_member_event = await self.store.get_event(prev_member_event_id) newly_joined = prev_member_event.membership != Membership.JOIN + + # Only rate-limit if the user actually joined the room, otherwise we'll end + # up blocking profile updates. if newly_joined: - await self._user_joined_room(target, room_id) - elif event.membership == Membership.LEAVE: + time_now_s = self.clock.time() + ( + allowed, + time_allowed, + ) = self._join_rate_limiter_local.can_requester_do_action(requester) + + if not allowed: + raise LimitExceededError( + retry_after_ms=int(1000 * (time_allowed - time_now_s)) + ) + + stream_id = await self.event_creation_handler.handle_new_client_event( + requester, event, context, extra_users=[target], ratelimit=ratelimit, + ) + + if event.membership == Membership.LEAVE: if prev_member_event_id: prev_member_event = await self.store.get_event(prev_member_event_id) if prev_member_event.membership == Membership.JOIN: @@ -271,7 +284,31 @@ class RoomMemberHandler(object): new_room: bool = False, require_consent: bool = True, ) -> Tuple[str, int]: - """Update a user's membership in a room""" + """Update a user's membership in a room. + + Params: + requester: The user who is performing the update. + target: The user whose membership is being updated. + room_id: The room ID whose membership is being updated. + action: The membership change, see synapse.api.constants.Membership. + txn_id: The transaction ID, if given. + remote_room_hosts: Remote servers to send the update to. + third_party_signed: Information from a 3PID invite. + ratelimit: Whether to rate limit the request. + content: The content of the created event. + require_consent: Whether consent is required. + + Returns: + A tuple of the new event ID and stream ID. + + Raises: + ShadowBanError if a shadow-banned requester attempts to send an invite. + """ + if action == Membership.INVITE and requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() + key = (room_id,) with (await self.member_linearizer.queue(key)): @@ -313,7 +350,7 @@ class RoomMemberHandler(object): # later on. content = dict(content) - if not self.allow_per_room_profiles: + if not self.allow_per_room_profiles or requester.shadow_banned: # Strip profile data, knowing that new profile data will be added to the # event's content in event_creation_handler.create_event() using the target's # global profile. @@ -473,6 +510,17 @@ class RoomMemberHandler(object): raise SynapseError(403, "Not allowed to join this room") if not is_host_in_room: + time_now_s = self.clock.time() + ( + allowed, + time_allowed, + ) = self._join_rate_limiter_remote.can_requester_do_action(requester,) + + if not allowed: + raise LimitExceededError( + retry_after_ms=int(1000 * (time_allowed - time_now_s)) + ) + inviter = await self._get_inviter(target.to_string(), room_id) if inviter and not self.hs.is_mine(inviter): remote_room_hosts.append(inviter.domain) @@ -500,26 +548,39 @@ class RoomMemberHandler(object): user_id=target.to_string(), room_id=room_id ) # type: Optional[RoomsForUser] if not invite: + logger.info( + "%s sent a leave request to %s, but that is not an active room " + "on this server, and there is no pending invite", + target, + room_id, + ) + raise SynapseError(404, "Not a known room") logger.info( "%s rejects invite to %s from %s", target, room_id, invite.sender ) - if self.hs.is_mine_id(invite.sender): - # the inviter was on our server, but has now left. Carry on - # with the normal rejection codepath. - # - # This is a bit of a hack, because the room might still be - # active on other servers. - pass - else: + if not self.hs.is_mine_id(invite.sender): # send the rejection to the inviter's HS (with fallback to # local event) return await self.remote_reject_invite( invite.event_id, txn_id, requester, content, ) + # the inviter was on our server, but has now left. Carry on + # with the normal rejection codepath, which will also send the + # rejection out to any other servers we believe are still in the room. + + # thanks to overzealous cleaning up of event_forward_extremities in + # `delete_old_current_state_events`, it's possible to end up with no + # forward extremities here. If that happens, let's just hang the + # rejection off the invite event. + # + # see: https://github.com/matrix-org/synapse/issues/7139 + if len(latest_event_ids) == 0: + latest_event_ids = [invite.event_id] + return await self._local_membership_update( requester=requester, target=target, @@ -669,25 +730,13 @@ class RoomMemberHandler(object): (EventTypes.Member, event.state_key), None ) - if event.membership == Membership.JOIN: - # Only fire user_joined_room if the user has actually joined the - # room. Don't bother if the user is just changing their profile - # info. - newly_joined = True - if prev_member_event_id: - prev_member_event = await self.store.get_event(prev_member_event_id) - newly_joined = prev_member_event.membership != Membership.JOIN - if newly_joined: - await self._user_joined_room(target_user, room_id) - elif event.membership == Membership.LEAVE: + if event.membership == Membership.LEAVE: if prev_member_event_id: prev_member_event = await self.store.get_event(prev_member_event_id) if prev_member_event.membership == Membership.JOIN: await self._user_left_room(target_user, room_id) - async def _can_guest_join( - self, current_state_ids: Dict[Tuple[str, str], str] - ) -> bool: + async def _can_guest_join(self, current_state_ids: StateMap[str]) -> bool: """ Returns whether a guest can join a room based on its current state. """ @@ -697,7 +746,7 @@ class RoomMemberHandler(object): guest_access = await self.store.get_event(guest_access_id) - return ( + return bool( guest_access and guest_access.content and "guest_access" in guest_access.content @@ -755,6 +804,25 @@ class RoomMemberHandler(object): new_room: bool = False, id_access_token: Optional[str] = None, ) -> int: + """Invite a 3PID to a room. + + Args: + room_id: The room to invite the 3PID to. + inviter: The user sending the invite. + medium: The 3PID's medium. + address: The 3PID's address. + id_server: The identity server to use. + requester: The user making the request. + txn_id: The transaction ID this is part of, or None if this is not + part of a transaction. + id_access_token: The optional identity server access token. + + Returns: + The new stream ID. + + Raises: + ShadowBanError if the requester has been shadow-banned. + """ if self.config.block_non_admin_invites: is_requester_admin = await self.auth.is_server_admin(requester.user) if not is_requester_admin: @@ -762,6 +830,11 @@ class RoomMemberHandler(object): 403, "Invites have been disabled on this server", Codes.FORBIDDEN ) + if requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() + # We need to rate limit *before* we send out any 3PID invites, so we # can't just rely on the standard ratelimiting of events. await self.base_handler.ratelimit(requester) @@ -809,6 +882,8 @@ class RoomMemberHandler(object): raise SynapseError(403, "Invites have been disabled on this server") if invitee: + # Note that update_membership with an action of "invite" can raise + # a ShadowBanError, but this was done above already. _, stream_id = await self.update_membership( requester, UserID.from_string(invitee), room_id, "invite", txn_id=txn_id ) @@ -914,9 +989,7 @@ class RoomMemberHandler(object): ) return stream_id - async def _is_host_in_room( - self, current_state_ids: Dict[Tuple[str, str], str] - ) -> bool: + async def _is_host_in_room(self, current_state_ids: StateMap[str]) -> bool: # Have we just created the room, and is this about to be the very # first member event? create_event_id = current_state_ids.get(("m.room.create", "")) @@ -947,10 +1020,9 @@ class RoomMemberHandler(object): class RoomMemberMasterHandler(RoomMemberHandler): def __init__(self, hs): - super(RoomMemberMasterHandler, self).__init__(hs) + super().__init__(hs) self.distributor = hs.get_distributor() - self.distributor.declare("user_joined_room") self.distributor.declare("user_left_room") async def _is_remote_room_too_complex( @@ -1007,7 +1079,11 @@ class RoomMemberMasterHandler(RoomMemberHandler): if len(remote_room_hosts) == 0: raise SynapseError(404, "No known servers") - if self.hs.config.limit_remote_rooms.enabled: + check_complexity = self.hs.config.limit_remote_rooms.enabled + if check_complexity and self.hs.config.limit_remote_rooms.admins_can_join: + check_complexity = not await self.auth.is_server_admin(user) + + if check_complexity: # Fetch the room complexity too_complex = await self._is_remote_room_too_complex( room_id, remote_room_hosts @@ -1026,11 +1102,10 @@ class RoomMemberMasterHandler(RoomMemberHandler): event_id, stream_id = await self.federation_handler.do_invite_join( remote_room_hosts, room_id, user.to_string(), content ) - await self._user_joined_room(user, room_id) # Check the room we just joined wasn't too large, if we didn't fetch the # complexity of it before. - if self.hs.config.limit_remote_rooms.enabled: + if check_complexity: if too_complex is False: # We checked, and we're under the limit. return event_id, stream_id @@ -1043,7 +1118,7 @@ class RoomMemberMasterHandler(RoomMemberHandler): return event_id, stream_id # The room is too large. Leave. - requester = types.create_requester(user, None, False, None) + requester = types.create_requester(user, None, False, False, None) await self.update_membership( requester=requester, target=user, room_id=room_id, action="leave" ) @@ -1169,11 +1244,6 @@ class RoomMemberMasterHandler(RoomMemberHandler): ) return event.event_id, stream_id - async def _user_joined_room(self, target: UserID, room_id: str) -> None: - """Implements RoomMemberHandler._user_joined_room - """ - user_joined_room(self.distributor, target, room_id) - async def _user_left_room(self, target: UserID, room_id: str) -> None: """Implements RoomMemberHandler._user_left_room """ diff --git a/synapse/handlers/room_member_worker.py b/synapse/handlers/room_member_worker.py
index 897338fd54..f2e88f6a5b 100644 --- a/synapse/handlers/room_member_worker.py +++ b/synapse/handlers/room_member_worker.py
@@ -30,7 +30,7 @@ logger = logging.getLogger(__name__) class RoomMemberWorkerHandler(RoomMemberHandler): def __init__(self, hs): - super(RoomMemberWorkerHandler, self).__init__(hs) + super().__init__(hs) self._remote_join_client = ReplRemoteJoin.make_client(hs) self._remote_reject_client = ReplRejectInvite.make_client(hs) @@ -57,8 +57,6 @@ class RoomMemberWorkerHandler(RoomMemberHandler): content=content, ) - await self._user_joined_room(user, room_id) - return ret["event_id"], ret["stream_id"] async def remote_reject_invite( @@ -81,13 +79,6 @@ class RoomMemberWorkerHandler(RoomMemberHandler): ) return ret["event_id"], ret["stream_id"] - async def _user_joined_room(self, target: UserID, room_id: str) -> None: - """Implements RoomMemberHandler._user_joined_room - """ - await self._notify_change_client( - user_id=target.to_string(), room_id=room_id, change="joined" - ) - async def _user_left_room(self, target: UserID, room_id: str) -> None: """Implements RoomMemberHandler._user_left_room """ diff --git a/synapse/handlers/saml_handler.py b/synapse/handlers/saml_handler.py
index abecaa8313..285c481a96 100644 --- a/synapse/handlers/saml_handler.py +++ b/synapse/handlers/saml_handler.py
@@ -14,7 +14,7 @@ # limitations under the License. import logging import re -from typing import Callable, Dict, Optional, Set, Tuple +from typing import TYPE_CHECKING, Callable, Dict, Optional, Set, Tuple import attr import saml2 @@ -23,6 +23,8 @@ from saml2.client import Saml2Client from synapse.api.errors import SynapseError from synapse.config import ConfigError +from synapse.config.saml2_config import SamlAttributeRequirement +from synapse.http.server import respond_with_html from synapse.http.servlet import parse_string from synapse.http.site import SynapseRequest from synapse.module_api import ModuleApi @@ -34,10 +36,17 @@ from synapse.types import ( from synapse.util.async_helpers import Linearizer from synapse.util.iterutils import chunk_seq +if TYPE_CHECKING: + import synapse.server + logger = logging.getLogger(__name__) -@attr.s +class MappingException(Exception): + """Used to catch errors when mapping the SAML2 response to a user.""" + + +@attr.s(slots=True) class Saml2SessionData: """Data we track about SAML2 sessions""" @@ -49,7 +58,8 @@ class Saml2SessionData: class SamlHandler: - def __init__(self, hs): + def __init__(self, hs: "synapse.server.HomeServer"): + self.hs = hs self._saml_client = Saml2Client(hs.config.saml2_sp_config) self._auth = hs.get_auth() self._auth_handler = hs.get_auth_handler() @@ -62,6 +72,8 @@ class SamlHandler: self._grandfathered_mxid_source_attribute = ( hs.config.saml2_grandfathered_mxid_source_attribute ) + self._saml2_attribute_requirements = hs.config.saml2.attribute_requirements + self._error_template = hs.config.sso_error_template # plugin to do custom mapping from saml response to mxid self._user_mapping_provider = hs.config.saml2_user_mapping_provider_class( @@ -73,11 +85,30 @@ class SamlHandler: self._auth_provider_id = "saml" # a map from saml session id to Saml2SessionData object - self._outstanding_requests_dict = {} + self._outstanding_requests_dict = {} # type: Dict[str, Saml2SessionData] # a lock on the mappings self._mapping_lock = Linearizer(name="saml_mapping", clock=self._clock) + def _render_error( + self, request, error: str, error_description: Optional[str] = None + ) -> None: + """Render the error template and respond to the request with it. + + This is used to show errors to the user. The template of this page can + be found under `synapse/res/templates/sso_error.html`. + + Args: + request: The incoming request from the browser. + We'll respond with an HTML page describing the error. + error: A technical identifier for this error. + error_description: A human-readable description of the error. + """ + html = self._error_template.render( + error=error, error_description=error_description + ) + respond_with_html(request, 400, html) + def handle_redirect_request( self, client_redirect_url: bytes, ui_auth_session_id: Optional[str] = None ) -> bytes: @@ -96,6 +127,9 @@ class SamlHandler: relay_state=client_redirect_url ) + # Since SAML sessions timeout it is useful to log when they were created. + logger.info("Initiating a new SAML session: %s" % (reqid,)) + now = self._clock.time_msec() self._outstanding_requests_dict[reqid] = Saml2SessionData( creation_time=now, ui_auth_session_id=ui_auth_session_id, @@ -125,48 +159,35 @@ class SamlHandler: # the dict. self.expire_sessions() - user_id, current_session = await self._map_saml_response_to_user( - resp_bytes, relay_state - ) - - # Complete the interactive auth session or the login. - if current_session and current_session.ui_auth_session_id: - await self._auth_handler.complete_sso_ui_auth( - user_id, current_session.ui_auth_session_id, request - ) - - else: - await self._auth_handler.complete_sso_login(user_id, request, relay_state) - - async def _map_saml_response_to_user( - self, resp_bytes: str, client_redirect_url: str - ) -> Tuple[str, Optional[Saml2SessionData]]: - """ - Given a sample response, retrieve the cached session and user for it. - - Args: - resp_bytes: The SAML response. - client_redirect_url: The redirect URL passed in by the client. - - Returns: - Tuple of the user ID and SAML session associated with this response. - - Raises: - SynapseError if there was a problem with the response. - RedirectException: some mapping providers may raise this if they need - to redirect to an interstitial page. - """ try: saml2_auth = self._saml_client.parse_authn_request_response( resp_bytes, saml2.BINDING_HTTP_POST, outstanding=self._outstanding_requests_dict, ) + except saml2.response.UnsolicitedResponse as e: + # the pysaml2 library helpfully logs an ERROR here, but neglects to log + # the session ID. I don't really want to put the full text of the exception + # in the (user-visible) exception message, so let's log the exception here + # so we can track down the session IDs later. + logger.warning(str(e)) + self._render_error( + request, "unsolicited_response", "Unexpected SAML2 login." + ) + return except Exception as e: - raise SynapseError(400, "Unable to parse SAML2 response: %s" % (e,)) + self._render_error( + request, + "invalid_response", + "Unable to parse SAML2 response: %s." % (e,), + ) + return if saml2_auth.not_signed: - raise SynapseError(400, "SAML2 response was not signed") + self._render_error( + request, "unsigned_respond", "SAML2 response was not signed." + ) + return logger.debug("SAML2 response: %s", saml2_auth.origxml) for assertion in saml2_auth.assertions: @@ -185,12 +206,73 @@ class SamlHandler: saml2_auth.in_response_to, None ) + # Ensure that the attributes of the logged in user meet the required + # attributes. + for requirement in self._saml2_attribute_requirements: + if not _check_attribute_requirement(saml2_auth.ava, requirement): + self._render_error( + request, "unauthorised", "You are not authorised to log in here." + ) + return + + # Pull out the user-agent and IP from the request. + user_agent = request.requestHeaders.getRawHeaders(b"User-Agent", default=[b""])[ + 0 + ].decode("ascii", "surrogateescape") + ip_address = self.hs.get_ip_from_request(request) + + # Call the mapper to register/login the user + try: + user_id = await self._map_saml_response_to_user( + saml2_auth, relay_state, user_agent, ip_address + ) + except MappingException as e: + logger.exception("Could not map user") + self._render_error(request, "mapping_error", str(e)) + return + + # Complete the interactive auth session or the login. + if current_session and current_session.ui_auth_session_id: + await self._auth_handler.complete_sso_ui_auth( + user_id, current_session.ui_auth_session_id, request + ) + + else: + await self._auth_handler.complete_sso_login(user_id, request, relay_state) + + async def _map_saml_response_to_user( + self, + saml2_auth: saml2.response.AuthnResponse, + client_redirect_url: str, + user_agent: str, + ip_address: str, + ) -> str: + """ + Given a SAML response, retrieve the user ID for it and possibly register the user. + + Args: + saml2_auth: The parsed SAML2 response. + client_redirect_url: The redirect URL passed in by the client. + user_agent: The user agent of the client making the request. + ip_address: The IP address of the client making the request. + + Returns: + The user ID associated with this response. + + Raises: + MappingException if there was a problem mapping the response to a user. + RedirectException: some mapping providers may raise this if they need + to redirect to an interstitial page. + """ + remote_user_id = self._user_mapping_provider.get_remote_user_id( saml2_auth, client_redirect_url ) if not remote_user_id: - raise Exception("Failed to extract remote user id from SAML response") + raise MappingException( + "Failed to extract remote user id from SAML response" + ) with (await self._mapping_lock.queue(self._auth_provider_id)): # first of all, check if we already have a mapping for this user @@ -204,7 +286,7 @@ class SamlHandler: ) if registered_user_id is not None: logger.info("Found existing mapping %s", registered_user_id) - return registered_user_id, current_session + return registered_user_id # backwards-compatibility hack: see if there is an existing user with a # suitable mapping from the uid @@ -229,7 +311,7 @@ class SamlHandler: await self._datastore.record_user_external_id( self._auth_provider_id, remote_user_id, registered_user_id ) - return registered_user_id, current_session + return registered_user_id # Map saml response to user attributes using the configured mapping provider for i in range(1000): @@ -246,7 +328,7 @@ class SamlHandler: localpart = attribute_dict.get("mxid_localpart") if not localpart: - raise Exception( + raise MappingException( "Error parsing SAML2 response: SAML mapping provider plugin " "did not return a mxid_localpart value" ) @@ -263,8 +345,8 @@ class SamlHandler: else: # Unable to generate a username in 1000 iterations # Break and return error to the user - raise SynapseError( - 500, "Unable to generate a Matrix ID from the SAML response" + raise MappingException( + "Unable to generate a Matrix ID from the SAML response" ) logger.info("Mapped SAML user to local part %s", localpart) @@ -273,12 +355,13 @@ class SamlHandler: localpart=localpart, default_display_name=displayname, bind_emails=emails, + user_agent_ips=(user_agent, ip_address), ) await self._datastore.record_user_external_id( self._auth_provider_id, remote_user_id, registered_user_id ) - return registered_user_id, current_session + return registered_user_id def expire_sessions(self): expire_before = self._clock.time_msec() - self._saml2_session_lifetime @@ -291,6 +374,21 @@ class SamlHandler: del self._outstanding_requests_dict[reqid] +def _check_attribute_requirement(ava: dict, req: SamlAttributeRequirement) -> bool: + values = ava.get(req.attribute, []) + for v in values: + if v == req.value: + return True + + logger.info( + "SAML2 attribute %s did not match required value '%s' (was '%s')", + req.attribute, + req.value, + values, + ) + return False + + DOT_REPLACE_PATTERN = re.compile( ("[^%s]" % (re.escape("".join(mxid_localpart_allowed_characters)),)) ) @@ -313,12 +411,12 @@ MXID_MAPPER_MAP = { @attr.s -class SamlConfig(object): +class SamlConfig: mxid_source_attribute = attr.ib() mxid_mapper = attr.ib() -class DefaultSamlMappingProvider(object): +class DefaultSamlMappingProvider: __version__ = "0.0.1" def __init__(self, parsed_config: SamlConfig, module_api: ModuleApi): @@ -343,7 +441,7 @@ class DefaultSamlMappingProvider(object): return saml_response.ava["uid"][0] except KeyError: logger.warning("SAML2 response lacks a 'uid' attestation") - raise SynapseError(400, "'uid' not in SAML2 response") + raise MappingException("'uid' not in SAML2 response") def saml_response_to_user_attributes( self, diff --git a/synapse/handlers/search.py b/synapse/handlers/search.py
index 9b312a1558..e9402e6e2e 100644 --- a/synapse/handlers/search.py +++ b/synapse/handlers/search.py
@@ -32,7 +32,7 @@ logger = logging.getLogger(__name__) class SearchHandler(BaseHandler): def __init__(self, hs): - super(SearchHandler, self).__init__(hs) + super().__init__(hs) self._event_serializer = hs.get_event_client_serializer() self.storage = hs.get_storage() self.state_store = self.storage.state @@ -340,7 +340,7 @@ class SearchHandler(BaseHandler): # If client has asked for "context" for each event (i.e. some surrounding # events and state), fetch that if event_context is not None: - now_token = await self.hs.get_event_sources().get_current_token() + now_token = self.hs.get_event_sources().get_current_token() contexts = {} for event in allowed_events: @@ -362,13 +362,13 @@ class SearchHandler(BaseHandler): self.storage, user.to_string(), res["events_after"] ) - res["start"] = now_token.copy_and_replace( + res["start"] = await now_token.copy_and_replace( "room_key", res["start"] - ).to_string() + ).to_string(self.store) - res["end"] = now_token.copy_and_replace( + res["end"] = await now_token.copy_and_replace( "room_key", res["end"] - ).to_string() + ).to_string(self.store) if include_profile: senders = { diff --git a/synapse/handlers/set_password.py b/synapse/handlers/set_password.py
index 5d34989f21..7713c3cf91 100644 --- a/synapse/handlers/set_password.py +++ b/synapse/handlers/set_password.py
@@ -28,7 +28,7 @@ class SetPasswordHandler(BaseHandler): """Handler which deals with changing user account passwords""" def __init__(self, hs): - super(SetPasswordHandler, self).__init__(hs) + super().__init__(hs) self._auth_handler = hs.get_auth_handler() self._device_handler = hs.get_device_handler() self._password_policy_handler = hs.get_password_policy_handler() diff --git a/synapse/handlers/state_deltas.py b/synapse/handlers/state_deltas.py
index 8590c1eff4..7a4ae0727a 100644 --- a/synapse/handlers/state_deltas.py +++ b/synapse/handlers/state_deltas.py
@@ -18,7 +18,7 @@ import logging logger = logging.getLogger(__name__) -class StateDeltasHandler(object): +class StateDeltasHandler: def __init__(self, hs): self.store = hs.get_datastore() diff --git a/synapse/handlers/stats.py b/synapse/handlers/stats.py
index 149f861239..249ffe2a55 100644 --- a/synapse/handlers/stats.py +++ b/synapse/handlers/stats.py
@@ -232,7 +232,7 @@ class StatsHandler: if membership == prev_membership: pass # noop - if membership == Membership.JOIN: + elif membership == Membership.JOIN: room_stats_delta["joined_members"] += 1 elif membership == Membership.INVITE: room_stats_delta["invited_members"] += 1 diff --git a/synapse/handlers/sync.py b/synapse/handlers/sync.py
index ebd3e98105..bfe2583002 100644 --- a/synapse/handlers/sync.py +++ b/synapse/handlers/sync.py
@@ -16,7 +16,7 @@ import itertools import logging -from typing import Any, Dict, FrozenSet, List, Optional, Set, Tuple +from typing import TYPE_CHECKING, Any, Dict, FrozenSet, List, Optional, Set, Tuple import attr from prometheus_client import Counter @@ -31,6 +31,7 @@ from synapse.storage.state import StateFilter from synapse.types import ( Collection, JsonDict, + MutableStateMap, RoomStreamToken, StateMap, StreamToken, @@ -43,6 +44,9 @@ from synapse.util.caches.response_cache import ResponseCache from synapse.util.metrics import Measure, measure_func from synapse.visibility import filter_events_for_client +if TYPE_CHECKING: + from synapse.server import HomeServer + logger = logging.getLogger(__name__) # Debug logger for https://github.com/matrix-org/synapse/issues/4422 @@ -85,16 +89,19 @@ class TimelineBatch: events = attr.ib(type=List[EventBase]) limited = attr.ib(bool) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: """Make the result appear empty if there are no updates. This is used to tell if room needs to be part of the sync result. """ return bool(self.events) - __bool__ = __nonzero__ # python3 - -@attr.s(slots=True, frozen=True) +# We can't freeze this class, because we need to update it after it's instantiated to +# update its unread count. This is because we calculate the unread count for a room only +# if there are updates for it, which we check after the instance has been created. +# This should not be a big deal because we update the notification counts afterwards as +# well anyway. +@attr.s(slots=True) class JoinedSyncResult: room_id = attr.ib(type=str) timeline = attr.ib(type=TimelineBatch) @@ -103,8 +110,9 @@ class JoinedSyncResult: account_data = attr.ib(type=List[JsonDict]) unread_notifications = attr.ib(type=JsonDict) summary = attr.ib(type=Optional[JsonDict]) + unread_count = attr.ib(type=int) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: """Make the result appear empty if there are no updates. This is used to tell if room needs to be part of the sync result. """ @@ -117,8 +125,6 @@ class JoinedSyncResult: # else in the result, we don't need to send it. ) - __bool__ = __nonzero__ # python3 - @attr.s(slots=True, frozen=True) class ArchivedSyncResult: @@ -127,26 +133,22 @@ class ArchivedSyncResult: state = attr.ib(type=StateMap[EventBase]) account_data = attr.ib(type=List[JsonDict]) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: """Make the result appear empty if there are no updates. This is used to tell if room needs to be part of the sync result. """ return bool(self.timeline or self.state or self.account_data) - __bool__ = __nonzero__ # python3 - @attr.s(slots=True, frozen=True) class InvitedSyncResult: room_id = attr.ib(type=str) invite = attr.ib(type=EventBase) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: """Invited rooms should always be reported to the client""" return True - __bool__ = __nonzero__ # python3 - @attr.s(slots=True, frozen=True) class GroupsSyncResult: @@ -154,11 +156,9 @@ class GroupsSyncResult: invite = attr.ib(type=JsonDict) leave = attr.ib(type=JsonDict) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: return bool(self.join or self.invite or self.leave) - __bool__ = __nonzero__ # python3 - @attr.s(slots=True, frozen=True) class DeviceLists: @@ -171,13 +171,11 @@ class DeviceLists: changed = attr.ib(type=Collection[str]) left = attr.ib(type=Collection[str]) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: return bool(self.changed or self.left) - __bool__ = __nonzero__ # python3 - -@attr.s +@attr.s(slots=True) class _RoomChanges: """The set of room entries to include in the sync, plus the set of joined and left room IDs since last sync. @@ -217,7 +215,7 @@ class SyncResult: device_one_time_keys_count = attr.ib(type=JsonDict) groups = attr.ib(type=Optional[GroupsSyncResult]) - def __nonzero__(self) -> bool: + def __bool__(self) -> bool: """Make the result appear empty if there are no updates. This is used to tell if the notifier needs to wait for more events when polling for events. @@ -233,11 +231,9 @@ class SyncResult: or self.groups ) - __bool__ = __nonzero__ # python3 - -class SyncHandler(object): - def __init__(self, hs): +class SyncHandler: + def __init__(self, hs: "HomeServer"): self.hs_config = hs.config self.store = hs.get_datastore() self.notifier = hs.get_notifier() @@ -368,7 +364,7 @@ class SyncHandler(object): sync_config = sync_result_builder.sync_config with Measure(self.clock, "ephemeral_by_room"): - typing_key = since_token.typing_key if since_token else "0" + typing_key = since_token.typing_key if since_token else 0 room_ids = sync_result_builder.joined_room_ids @@ -392,7 +388,7 @@ class SyncHandler(object): event_copy = {k: v for (k, v) in event.items() if k != "room_id"} ephemeral_by_room.setdefault(room_id, []).append(event_copy) - receipt_key = since_token.receipt_key if since_token else "0" + receipt_key = since_token.receipt_key if since_token else 0 receipt_source = self.event_sources.sources["receipt"] receipts, receipt_key = await receipt_source.get_new_events( @@ -588,7 +584,7 @@ class SyncHandler(object): room_id: str, sync_config: SyncConfig, batch: TimelineBatch, - state: StateMap[EventBase], + state: MutableStateMap[EventBase], now_token: StreamToken, ) -> Optional[JsonDict]: """ Works out a room summary block for this room, summarising the number @@ -710,9 +706,8 @@ class SyncHandler(object): ] missing_hero_state = await self.store.get_events(missing_hero_event_ids) - missing_hero_state = missing_hero_state.values() - for s in missing_hero_state: + for s in missing_hero_state.values(): cache.set(s.state_key, s.event_id) state[(EventTypes.Member, s.state_key)] = s @@ -736,7 +731,7 @@ class SyncHandler(object): since_token: Optional[StreamToken], now_token: StreamToken, full_state: bool, - ) -> StateMap[EventBase]: + ) -> MutableStateMap[EventBase]: """ Works out the difference in state between the start of the timeline and the previous sync. @@ -930,7 +925,7 @@ class SyncHandler(object): async def unread_notifs_for_room_id( self, room_id: str, sync_config: SyncConfig - ) -> Optional[Dict[str, str]]: + ) -> Dict[str, int]: with Measure(self.clock, "unread_notifs_for_room_id"): last_unread_event_id = await self.store.get_last_receipt_event_id_for_user( user_id=sync_config.user.to_string(), @@ -938,15 +933,10 @@ class SyncHandler(object): receipt_type="m.read", ) - if last_unread_event_id: - notifs = await self.store.get_unread_event_push_actions_by_room_for_user( - room_id, sync_config.user.to_string(), last_unread_event_id - ) - return notifs - - # There is no new information in this period, so your notification - # count is whatever it was last time. - return None + notifs = await self.store.get_unread_event_push_actions_by_room_for_user( + room_id, sync_config.user.to_string(), last_unread_event_id + ) + return notifs async def generate_sync_result( self, @@ -960,7 +950,7 @@ class SyncHandler(object): # this is due to some of the underlying streams not supporting the ability # to query up to a given point. # Always use the `now_token` in `SyncResultBuilder` - now_token = await self.event_sources.get_current_token() + now_token = self.event_sources.get_current_token() logger.debug( "Calculating sync response for %r between %s and %s", @@ -977,7 +967,7 @@ class SyncHandler(object): raise NotImplementedError() else: joined_room_ids = await self.get_rooms_for_user_at( - user_id, now_token.room_stream_id + user_id, now_token.room_key ) sync_result_builder = SyncResultBuilder( sync_config, @@ -1306,12 +1296,11 @@ class SyncHandler(object): presence_source = self.event_sources.sources["presence"] since_token = sync_result_builder.since_token + presence_key = None + include_offline = False if since_token and not sync_result_builder.full_state: presence_key = since_token.presence_key include_offline = True - else: - presence_key = None - include_offline = False presence, presence_key = await presence_source.get_new_events( user=user, @@ -1319,6 +1308,7 @@ class SyncHandler(object): is_guest=sync_config.is_guest, include_offline=include_offline, ) + assert presence_key sync_result_builder.now_token = now_token.copy_and_replace( "presence_key", presence_key ) @@ -1481,7 +1471,7 @@ class SyncHandler(object): if rooms_changed: return True - stream_id = RoomStreamToken.parse_stream_token(since_token.room_key).stream + stream_id = since_token.room_key.stream for room_id in sync_result_builder.joined_room_ids: if self.store.has_room_changed_since(room_id, stream_id): return True @@ -1605,16 +1595,24 @@ class SyncHandler(object): if leave_events: leave_event = leave_events[-1] - leave_stream_token = await self.store.get_stream_token_for_event( + leave_position = await self.store.get_position_for_event( leave_event.event_id ) - leave_token = since_token.copy_and_replace( - "room_key", leave_stream_token - ) - if since_token and since_token.is_after(leave_token): + # If the leave event happened before the since token then we + # bail. + if since_token and not leave_position.persisted_after( + since_token.room_key + ): continue + # We can safely convert the position of the leave event into a + # stream token as it'll only be used in the context of this + # room. (c.f. the docstring of `to_room_stream_token`). + leave_token = since_token.copy_and_replace( + "room_key", leave_position.to_room_stream_token() + ) + # If this is an out of band message, like a remote invite # rejection, we include it in the recents batch. Otherwise, we # let _load_filtered_recents handle fetching the correct @@ -1747,7 +1745,7 @@ class SyncHandler(object): continue leave_token = now_token.copy_and_replace( - "room_key", "s%d" % (event.stream_ordering,) + "room_key", RoomStreamToken(None, event.stream_ordering) ) room_entries.append( RoomSyncResultBuilder( @@ -1769,7 +1767,7 @@ class SyncHandler(object): ignored_users: Set[str], room_builder: "RoomSyncResultBuilder", ephemeral: List[JsonDict], - tags: Optional[List[JsonDict]], + tags: Optional[Dict[str, Dict[str, Any]]], account_data: Dict[str, JsonDict], always_include: bool = False, ): @@ -1885,7 +1883,7 @@ class SyncHandler(object): ) if room_builder.rtype == "joined": - unread_notifications = {} # type: Dict[str, str] + unread_notifications = {} # type: Dict[str, int] room_sync = JoinedSyncResult( room_id=room_id, timeline=batch, @@ -1894,14 +1892,16 @@ class SyncHandler(object): account_data=account_data_events, unread_notifications=unread_notifications, summary=summary, + unread_count=0, ) if room_sync or always_include: notifs = await self.unread_notifs_for_room_id(room_id, sync_config) - if notifs is not None: - unread_notifications["notification_count"] = notifs["notify_count"] - unread_notifications["highlight_count"] = notifs["highlight_count"] + unread_notifications["notification_count"] = notifs["notify_count"] + unread_notifications["highlight_count"] = notifs["highlight_count"] + + room_sync.unread_count = notifs["unread_count"] sync_result_builder.joined.append(room_sync) @@ -1924,7 +1924,7 @@ class SyncHandler(object): raise Exception("Unrecognized rtype: %r", room_builder.rtype) async def get_rooms_for_user_at( - self, user_id: str, stream_ordering: int + self, user_id: str, room_key: RoomStreamToken ) -> FrozenSet[str]: """Get set of joined rooms for a user at the given stream ordering. @@ -1950,15 +1950,15 @@ class SyncHandler(object): # If the membership's stream ordering is after the given stream # ordering, we need to go and work out if the user was in the room # before. - for room_id, membership_stream_ordering in joined_rooms: - if membership_stream_ordering <= stream_ordering: + for room_id, event_pos in joined_rooms: + if not event_pos.persisted_after(room_key): joined_room_ids.add(room_id) continue logger.info("User joined room after current token: %s", room_id) extrems = await self.store.get_forward_extremeties_for_room( - room_id, stream_ordering + room_id, event_pos.stream ) users_in_room = await self.state.get_current_users_in_room(room_id, extrems) if user_id in users_in_room: @@ -2032,7 +2032,7 @@ def _calculate_state( return {event_id_to_key[e]: e for e in state_ids} -@attr.s +@attr.s(slots=True) class SyncResultBuilder: """Used to help build up a new SyncResult for a user @@ -2068,8 +2068,8 @@ class SyncResultBuilder: to_device = attr.ib(type=List[JsonDict], default=attr.Factory(list)) -@attr.s -class RoomSyncResultBuilder(object): +@attr.s(slots=True) +class RoomSyncResultBuilder: """Stores information needed to create either a `JoinedSyncResult` or `ArchivedSyncResult`. diff --git a/synapse/handlers/typing.py b/synapse/handlers/typing.py
index a86ac0150e..3cbfc2d780 100644 --- a/synapse/handlers/typing.py +++ b/synapse/handlers/typing.py
@@ -14,10 +14,11 @@ # limitations under the License. import logging +import random from collections import namedtuple from typing import TYPE_CHECKING, List, Set, Tuple -from synapse.api.errors import AuthError, SynapseError +from synapse.api.errors import AuthError, ShadowBanError, SynapseError from synapse.metrics.background_process_metrics import run_as_background_process from synapse.replication.tcp.streams import TypingStream from synapse.types import UserID, get_domain_from_id @@ -227,9 +228,9 @@ class TypingWriterHandler(FollowerTypingHandler): self._stopped_typing(member) return - async def started_typing(self, target_user, auth_user, room_id, timeout): + async def started_typing(self, target_user, requester, room_id, timeout): target_user_id = target_user.to_string() - auth_user_id = auth_user.to_string() + auth_user_id = requester.user.to_string() if not self.is_mine_id(target_user_id): raise SynapseError(400, "User is not hosted on this homeserver") @@ -237,6 +238,11 @@ class TypingWriterHandler(FollowerTypingHandler): if target_user_id != auth_user_id: raise AuthError(400, "Cannot set another user's typing state") + if requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() + await self.auth.check_user_in_room(room_id, target_user_id) logger.debug("%s has started typing in %s", target_user_id, room_id) @@ -256,9 +262,9 @@ class TypingWriterHandler(FollowerTypingHandler): self._push_update(member=member, typing=True) - async def stopped_typing(self, target_user, auth_user, room_id): + async def stopped_typing(self, target_user, requester, room_id): target_user_id = target_user.to_string() - auth_user_id = auth_user.to_string() + auth_user_id = requester.user.to_string() if not self.is_mine_id(target_user_id): raise SynapseError(400, "User is not hosted on this homeserver") @@ -266,6 +272,11 @@ class TypingWriterHandler(FollowerTypingHandler): if target_user_id != auth_user_id: raise AuthError(400, "Cannot set another user's typing state") + if requester.shadow_banned: + # We randomly sleep a bit just to annoy the requester. + await self.clock.sleep(random.randint(1, 10)) + raise ShadowBanError() + await self.auth.check_user_in_room(room_id, target_user_id) logger.debug("%s has stopped typing in %s", target_user_id, room_id) @@ -401,7 +412,7 @@ class TypingWriterHandler(FollowerTypingHandler): raise Exception("Typing writer instance got typing info over replication") -class TypingNotificationEventSource(object): +class TypingNotificationEventSource: def __init__(self, hs): self.hs = hs self.clock = hs.get_clock() diff --git a/synapse/handlers/ui_auth/checkers.py b/synapse/handlers/ui_auth/checkers.py
index a011e9fe29..9146dc1a3b 100644 --- a/synapse/handlers/ui_auth/checkers.py +++ b/synapse/handlers/ui_auth/checkers.py
@@ -16,13 +16,12 @@ import logging from typing import Any -from canonicaljson import json - from twisted.web.client import PartialDownloadError from synapse.api.constants import LoginType from synapse.api.errors import Codes, LoginError, SynapseError from synapse.config.emailconfig import ThreepidBehaviour +from synapse.util import json_decoder logger = logging.getLogger(__name__) @@ -117,7 +116,7 @@ class RecaptchaAuthChecker(UserInteractiveAuthChecker): except PartialDownloadError as pde: # Twisted is silly data = pde.response - resp_body = json.loads(data.decode("utf-8")) + resp_body = json_decoder.decode(data.decode("utf-8")) if "success" in resp_body: # Note that we do NOT check the hostname here: we explicitly diff --git a/synapse/handlers/user_directory.py b/synapse/handlers/user_directory.py
index 521b6d620d..79393c8829 100644 --- a/synapse/handlers/user_directory.py +++ b/synapse/handlers/user_directory.py
@@ -37,7 +37,7 @@ class UserDirectoryHandler(StateDeltasHandler): """ def __init__(self, hs): - super(UserDirectoryHandler, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.state = hs.get_state_handler() @@ -234,7 +234,7 @@ class UserDirectoryHandler(StateDeltasHandler): async def _handle_room_publicity_change( self, room_id, prev_event_id, event_id, typ ): - """Handle a room having potentially changed from/to world_readable/publically + """Handle a room having potentially changed from/to world_readable/publicly joinable. Args: @@ -388,9 +388,15 @@ class UserDirectoryHandler(StateDeltasHandler): prev_name = prev_event.content.get("displayname") new_name = event.content.get("displayname") + # If the new name is an unexpected form, do not update the directory. + if not isinstance(new_name, str): + new_name = prev_name prev_avatar = prev_event.content.get("avatar_url") new_avatar = event.content.get("avatar_url") + # If the new avatar is an unexpected form, do not update the directory. + if not isinstance(new_avatar, str): + new_avatar = prev_avatar if prev_name != new_name or prev_avatar != new_avatar: await self.store.update_profile_in_user_dir(user_id, new_name, new_avatar) diff --git a/synapse/http/__init__.py b/synapse/http/__init__.py
index 3880ce0d94..59b01b812c 100644 --- a/synapse/http/__init__.py +++ b/synapse/http/__init__.py
@@ -16,8 +16,6 @@ import re from twisted.internet import task -from twisted.internet.defer import CancelledError -from twisted.python import failure from twisted.web.client import FileBodyProducer from synapse.api.errors import SynapseError @@ -26,19 +24,8 @@ from synapse.api.errors import SynapseError class RequestTimedOutError(SynapseError): """Exception representing timeout of an outbound request""" - def __init__(self): - super(RequestTimedOutError, self).__init__(504, "Timed out") - - -def cancelled_to_request_timed_out_error(value, timeout): - """Turns CancelledErrors into RequestTimedOutErrors. - - For use with async.add_timeout_to_deferred - """ - if isinstance(value, failure.Failure): - value.trap(CancelledError) - raise RequestTimedOutError() - return value + def __init__(self, msg): + super().__init__(504, msg) ACCESS_TOKEN_RE = re.compile(r"(\?.*access(_|%5[Ff])token=)[^&]*(.*)$") diff --git a/synapse/http/client.py b/synapse/http/client.py
index 6bc51202cd..8324632cb6 100644 --- a/synapse/http/client.py +++ b/synapse/http/client.py
@@ -13,20 +13,31 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import logging import urllib from io import BytesIO +from typing import ( + Any, + BinaryIO, + Dict, + Iterable, + List, + Mapping, + Optional, + Sequence, + Tuple, + Union, +) import treq -from canonicaljson import encode_canonical_json, json +from canonicaljson import encode_canonical_json from netaddr import IPAddress from prometheus_client import Counter from zope.interface import implementer, provider from OpenSSL import SSL from OpenSSL.SSL import VERIFY_NONE -from twisted.internet import defer, protocol, ssl +from twisted.internet import defer, error as twisted_error, protocol, ssl from twisted.internet.interfaces import ( IReactorPluggableNameResolver, IResolutionReceiver, @@ -34,19 +45,22 @@ from twisted.internet.interfaces import ( from twisted.internet.task import Cooperator from twisted.python.failure import Failure from twisted.web._newclient import ResponseDone -from twisted.web.client import Agent, HTTPConnectionPool, readBody +from twisted.web.client import ( + Agent, + HTTPConnectionPool, + ResponseNeverReceived, + readBody, +) from twisted.web.http import PotentialDataLoss from twisted.web.http_headers import Headers +from twisted.web.iweb import IResponse from synapse.api.errors import Codes, HttpResponseException, SynapseError -from synapse.http import ( - QuieterFileBodyProducer, - cancelled_to_request_timed_out_error, - redact_uri, -) +from synapse.http import QuieterFileBodyProducer, RequestTimedOutError, redact_uri from synapse.http.proxyagent import ProxyAgent from synapse.logging.context import make_deferred_yieldable from synapse.logging.opentracing import set_tag, start_active_span, tags +from synapse.util import json_decoder from synapse.util.async_helpers import timeout_deferred logger = logging.getLogger(__name__) @@ -56,6 +70,19 @@ incoming_responses_counter = Counter( "synapse_http_client_responses", "", ["method", "code"] ) +# the type of the headers list, to be passed to the t.w.h.Headers. +# Actually we can mix str and bytes keys, but Mapping treats 'key' as invariant so +# we simplify. +RawHeaders = Union[Mapping[str, "RawHeaderValue"], Mapping[bytes, "RawHeaderValue"]] + +# the value actually has to be a List, but List is invariant so we can't specify that +# the entries can either be Lists or bytes. +RawHeaderValue = Sequence[Union[str, bytes]] + +# the type of the query params, to be passed into `urlencode` +QueryParamValue = Union[str, bytes, Iterable[Union[str, bytes]]] +QueryParams = Union[Mapping[str, QueryParamValue], Mapping[bytes, QueryParamValue]] + def check_against_blacklist(ip_address, ip_whitelist, ip_blacklist): """ @@ -85,7 +112,7 @@ def _make_scheduler(reactor): return _scheduler -class IPBlacklistingResolver(object): +class IPBlacklistingResolver: """ A proxy for reactor.nameResolver which only produces non-blacklisted IP addresses, preventing DNS rebinding attacks on URL preview. @@ -132,7 +159,7 @@ class IPBlacklistingResolver(object): r.resolutionComplete() @provider(IResolutionReceiver) - class EndpointReceiver(object): + class EndpointReceiver: @staticmethod def resolutionBegan(resolutionInProgress): pass @@ -191,7 +218,7 @@ class BlacklistingAgentWrapper(Agent): ) -class SimpleHttpClient(object): +class SimpleHttpClient: """ A simple, no-frills HTTP client with methods that wrap up common ways of using HTTP in Matrix @@ -243,7 +270,7 @@ class SimpleHttpClient(object): ) @implementer(IReactorPluggableNameResolver) - class Reactor(object): + class Reactor: def __getattr__(_self, attr): if attr == "nameResolver": return nameResolver @@ -284,21 +311,31 @@ class SimpleHttpClient(object): ip_blacklist=self._ip_blacklist, ) - @defer.inlineCallbacks - def request(self, method, uri, data=None, headers=None): + async def request( + self, + method: str, + uri: str, + data: Optional[bytes] = None, + headers: Optional[Headers] = None, + ) -> IResponse: """ Args: - method (str): HTTP method to use. - uri (str): URI to query. - data (bytes): Data to send in the request body, if applicable. - headers (t.w.http_headers.Headers): Request headers. + method: HTTP method to use. + uri: URI to query. + data: Data to send in the request body, if applicable. + headers: Request headers. + + Returns: + Response object, once the headers have been read. + + Raises: + RequestTimedOutError if the request times out before the headers are read + """ - # A small wrapper around self.agent.request() so we can easily attach - # counters to it outgoing_requests_counter.labels(method).inc() # log request but strip `access_token` (AS requests for example include this) - logger.info("Sending request %s %s", method, redact_uri(uri)) + logger.debug("Sending request %s %s", method, redact_uri(uri)) with start_active_span( "outgoing-client-request", @@ -323,14 +360,18 @@ class SimpleHttpClient(object): data=body_producer, headers=headers, **self._extra_treq_args - ) + ) # type: defer.Deferred + + # we use our own timeout mechanism rather than treq's as a workaround + # for https://twistedmatrix.com/trac/ticket/9534. request_deferred = timeout_deferred( - request_deferred, - 60, - self.hs.get_reactor(), - cancelled_to_request_timed_out_error, + request_deferred, 60, self.hs.get_reactor(), ) - response = yield make_deferred_yieldable(request_deferred) + + # turn timeouts into RequestTimedOutErrors + request_deferred.addErrback(_timeout_to_request_timed_out_error) + + response = await make_deferred_yieldable(request_deferred) incoming_responses_counter.labels(method, response.code).inc() logger.info( @@ -353,19 +394,26 @@ class SimpleHttpClient(object): set_tag("error_reason", e.args[0]) raise - @defer.inlineCallbacks - def post_urlencoded_get_json(self, uri, args={}, headers=None): + async def post_urlencoded_get_json( + self, + uri: str, + args: Mapping[str, Union[str, List[str]]] = {}, + headers: Optional[RawHeaders] = None, + ) -> Any: """ Args: - uri (str): - args (dict[str, str|List[str]]): query params - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + uri: uri to query + args: parameters to be url-encoded in the body + headers: a map from header name to a list of values for that header Returns: - Deferred[object]: parsed json + parsed json Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + HttpResponseException: On a non-2xx HTTP response. ValueError: if the response was not JSON @@ -386,31 +434,37 @@ class SimpleHttpClient(object): if headers: actual_headers.update(headers) - response = yield self.request( + response = await self.request( "POST", uri, headers=Headers(actual_headers), data=query_bytes ) - body = yield make_deferred_yieldable(readBody(response)) + body = await make_deferred_yieldable(readBody(response)) if 200 <= response.code < 300: - return json.loads(body.decode("utf-8")) + return json_decoder.decode(body.decode("utf-8")) else: - raise HttpResponseException(response.code, response.phrase, body) + raise HttpResponseException( + response.code, response.phrase.decode("ascii", errors="replace"), body + ) - @defer.inlineCallbacks - def post_json_get_json(self, uri, post_json, headers=None): + async def post_json_get_json( + self, uri: str, post_json: Any, headers: Optional[RawHeaders] = None + ) -> Any: """ Args: - uri (str): - post_json (object): - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + uri: URI to query. + post_json: request body, to be encoded as json + headers: a map from header name to a list of values for that header Returns: - Deferred[object]: parsed json + parsed json Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + HttpResponseException: On a non-2xx HTTP response. ValueError: if the response was not JSON @@ -427,33 +481,35 @@ class SimpleHttpClient(object): if headers: actual_headers.update(headers) - response = yield self.request( + response = await self.request( "POST", uri, headers=Headers(actual_headers), data=json_str ) - body = yield make_deferred_yieldable(readBody(response)) + body = await make_deferred_yieldable(readBody(response)) if 200 <= response.code < 300: - return json.loads(body.decode("utf-8")) + return json_decoder.decode(body.decode("utf-8")) else: - raise HttpResponseException(response.code, response.phrase, body) + raise HttpResponseException( + response.code, response.phrase.decode("ascii", errors="replace"), body + ) - @defer.inlineCallbacks - def get_json(self, uri, args={}, headers=None): - """ Gets some json from the given URI. + async def get_json( + self, uri: str, args: QueryParams = {}, headers: Optional[RawHeaders] = None, + ) -> Any: + """Gets some json from the given URI. Args: - uri (str): The URI to request, not including query parameters - args (dict): A dictionary used to create query strings, defaults to - None. - **Note**: The value of each key is assumed to be an iterable - and *not* a string. - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + uri: The URI to request, not including query parameters + args: A dictionary used to create query string + headers: a map from header name to a list of values for that header Returns: - Deferred: Succeeds when we get *any* 2xx HTTP response, with the - HTTP body as JSON. + Succeeds when we get a 2xx HTTP response, with the HTTP body as JSON. Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + HttpResponseException On a non-2xx HTTP response. ValueError: if the response was not JSON @@ -462,26 +518,30 @@ class SimpleHttpClient(object): if headers: actual_headers.update(headers) - body = yield self.get_raw(uri, args, headers=headers) - return json.loads(body.decode("utf-8")) + body = await self.get_raw(uri, args, headers=headers) + return json_decoder.decode(body.decode("utf-8")) - @defer.inlineCallbacks - def put_json(self, uri, json_body, args={}, headers=None): - """ Puts some json to the given URI. + async def put_json( + self, + uri: str, + json_body: Any, + args: QueryParams = {}, + headers: RawHeaders = None, + ) -> Any: + """Puts some json to the given URI. Args: - uri (str): The URI to request, not including query parameters - json_body (dict): The JSON to put in the HTTP body, - args (dict): A dictionary used to create query strings, defaults to - None. - **Note**: The value of each key is assumed to be an iterable - and *not* a string. - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + uri: The URI to request, not including query parameters + json_body: The JSON to put in the HTTP body, + args: A dictionary used to create query strings + headers: a map from header name to a list of values for that header Returns: - Deferred: Succeeds when we get *any* 2xx HTTP response, with the - HTTP body as JSON. + Succeeds when we get a 2xx HTTP response, with the HTTP body as JSON. Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + HttpResponseException On a non-2xx HTTP response. ValueError: if the response was not JSON @@ -500,33 +560,36 @@ class SimpleHttpClient(object): if headers: actual_headers.update(headers) - response = yield self.request( + response = await self.request( "PUT", uri, headers=Headers(actual_headers), data=json_str ) - body = yield make_deferred_yieldable(readBody(response)) + body = await make_deferred_yieldable(readBody(response)) if 200 <= response.code < 300: - return json.loads(body.decode("utf-8")) + return json_decoder.decode(body.decode("utf-8")) else: - raise HttpResponseException(response.code, response.phrase, body) + raise HttpResponseException( + response.code, response.phrase.decode("ascii", errors="replace"), body + ) - @defer.inlineCallbacks - def get_raw(self, uri, args={}, headers=None): - """ Gets raw text from the given URI. + async def get_raw( + self, uri: str, args: QueryParams = {}, headers: Optional[RawHeaders] = None + ) -> bytes: + """Gets raw text from the given URI. Args: - uri (str): The URI to request, not including query parameters - args (dict): A dictionary used to create query strings, defaults to - None. - **Note**: The value of each key is assumed to be an iterable - and *not* a string. - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + uri: The URI to request, not including query parameters + args: A dictionary used to create query strings + headers: a map from header name to a list of values for that header Returns: - Deferred: Succeeds when we get *any* 2xx HTTP response, with the + Succeeds when we get a 2xx HTTP response, with the HTTP body as bytes. Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + HttpResponseException on a non-2xx HTTP response. """ if len(args): @@ -537,36 +600,50 @@ class SimpleHttpClient(object): if headers: actual_headers.update(headers) - response = yield self.request("GET", uri, headers=Headers(actual_headers)) + response = await self.request("GET", uri, headers=Headers(actual_headers)) - body = yield make_deferred_yieldable(readBody(response)) + body = await make_deferred_yieldable(readBody(response)) if 200 <= response.code < 300: return body else: - raise HttpResponseException(response.code, response.phrase, body) + raise HttpResponseException( + response.code, response.phrase.decode("ascii", errors="replace"), body + ) # XXX: FIXME: This is horribly copy-pasted from matrixfederationclient. # The two should be factored out. - @defer.inlineCallbacks - def get_file(self, url, output_stream, max_size=None, headers=None): + async def get_file( + self, + url: str, + output_stream: BinaryIO, + max_size: Optional[int] = None, + headers: Optional[RawHeaders] = None, + ) -> Tuple[int, Dict[bytes, List[bytes]], str, int]: """GETs a file from a given URL Args: - url (str): The URL to GET - output_stream (file): File to write the response body to. - headers (dict[str|bytes, List[str|bytes]]|None): If not None, a map from - header name to a list of values for that header + url: The URL to GET + output_stream: File to write the response body to. + headers: A map from header name to a list of values for that header Returns: - A (int,dict,string,int) tuple of the file length, dict of the response + A tuple of the file length, dict of the response headers, absolute URI of the response and HTTP response code. + + Raises: + RequestTimedOutError: if there is a timeout before the response headers + are received. Note there is currently no timeout on reading the response + body. + + SynapseError: if the response is not a 2xx, the remote file is too large, or + another exception happens during the download. """ actual_headers = {b"User-Agent": [self.user_agent]} if headers: actual_headers.update(headers) - response = yield self.request("GET", url, headers=Headers(actual_headers)) + response = await self.request("GET", url, headers=Headers(actual_headers)) resp_headers = dict(response.headers.getAllRawHeaders()) @@ -590,7 +667,7 @@ class SimpleHttpClient(object): # straight back in again try: - length = yield make_deferred_yieldable( + length = await make_deferred_yieldable( _readBodyToFile(response, output_stream, max_size) ) except SynapseError: @@ -607,6 +684,18 @@ class SimpleHttpClient(object): ) +def _timeout_to_request_timed_out_error(f: Failure): + if f.check(twisted_error.TimeoutError, twisted_error.ConnectingCancelledError): + # The TCP connection has its own timeout (set by the 'connectTimeout' param + # on the Agent), which raises twisted_error.TimeoutError exception. + raise RequestTimedOutError("Timeout connecting to remote server") + elif f.check(defer.TimeoutError, ResponseNeverReceived): + # this one means that we hit our overall timeout on the request + raise RequestTimedOutError("Timeout waiting for response from remote server") + + return f + + # XXX: FIXME: This is horribly copy-pasted from matrixfederationclient. # The two should be factored out. diff --git a/synapse/http/connectproxyclient.py b/synapse/http/connectproxyclient.py
index be7b2ceb8e..856e28454f 100644 --- a/synapse/http/connectproxyclient.py +++ b/synapse/http/connectproxyclient.py
@@ -31,7 +31,7 @@ class ProxyConnectError(ConnectError): @implementer(IStreamClientEndpoint) -class HTTPConnectProxyEndpoint(object): +class HTTPConnectProxyEndpoint: """An Endpoint implementation which will send a CONNECT request to an http proxy Wraps an existing HostnameEndpoint for the proxy. diff --git a/synapse/http/federation/matrix_federation_agent.py b/synapse/http/federation/matrix_federation_agent.py
index 0c02648015..83d6196d4a 100644 --- a/synapse/http/federation/matrix_federation_agent.py +++ b/synapse/http/federation/matrix_federation_agent.py
@@ -36,7 +36,7 @@ logger = logging.getLogger(__name__) @implementer(IAgent) -class MatrixFederationAgent(object): +class MatrixFederationAgent: """An Agent-like thing which provides a `request` method which correctly handles resolving matrix server names when using matrix://. Handles standard https URIs as normal. @@ -134,8 +134,8 @@ class MatrixFederationAgent(object): and not _is_ip_literal(parsed_uri.hostname) and not parsed_uri.port ): - well_known_result = yield self._well_known_resolver.get_well_known( - parsed_uri.hostname + well_known_result = yield defer.ensureDeferred( + self._well_known_resolver.get_well_known(parsed_uri.hostname) ) delegated_server = well_known_result.delegated_server @@ -175,7 +175,7 @@ class MatrixFederationAgent(object): @implementer(IAgentEndpointFactory) -class MatrixHostnameEndpointFactory(object): +class MatrixHostnameEndpointFactory: """Factory for MatrixHostnameEndpoint for parsing to an Agent. """ @@ -198,7 +198,7 @@ class MatrixHostnameEndpointFactory(object): @implementer(IStreamClientEndpoint) -class MatrixHostnameEndpoint(object): +class MatrixHostnameEndpoint: """An endpoint that resolves matrix:// URLs using Matrix server name resolution (i.e. via SRV). Does not check for well-known delegation. @@ -247,7 +247,7 @@ class MatrixHostnameEndpoint(object): port = server.port try: - logger.info("Connecting to %s:%i", host.decode("ascii"), port) + logger.debug("Connecting to %s:%i", host.decode("ascii"), port) endpoint = HostnameEndpoint(self._reactor, host, port) if self._tls_options: endpoint = wrapClientTLS(self._tls_options, endpoint) diff --git a/synapse/http/federation/srv_resolver.py b/synapse/http/federation/srv_resolver.py
index 2ede90a9b1..d9620032d2 100644 --- a/synapse/http/federation/srv_resolver.py +++ b/synapse/http/federation/srv_resolver.py
@@ -33,7 +33,7 @@ SERVER_CACHE = {} @attr.s(slots=True, frozen=True) -class Server(object): +class Server: """ Our record of an individual server which can be tried to reach a destination. @@ -96,7 +96,7 @@ def _sort_server_list(server_list): return results -class SrvResolver(object): +class SrvResolver: """Interface to the dns client to do SRV lookups, with result caching. The default resolver in twisted.names doesn't do any caching (it has a CacheResolver, diff --git a/synapse/http/federation/well_known_resolver.py b/synapse/http/federation/well_known_resolver.py
index 89a3b041ce..a306faa267 100644 --- a/synapse/http/federation/well_known_resolver.py +++ b/synapse/http/federation/well_known_resolver.py
@@ -13,10 +13,10 @@ # See the License for the specific language governing permissions and # limitations under the License. -import json import logging import random import time +from typing import Callable, Dict, Optional, Tuple import attr @@ -24,9 +24,10 @@ from twisted.internet import defer from twisted.web.client import RedirectAgent, readBody from twisted.web.http import stringToDatetime from twisted.web.http_headers import Headers +from twisted.web.iweb import IResponse from synapse.logging.context import make_deferred_yieldable -from synapse.util import Clock +from synapse.util import Clock, json_decoder from synapse.util.caches.ttlcache import TTLCache from synapse.util.metrics import Measure @@ -70,11 +71,11 @@ _had_valid_well_known_cache = TTLCache("had-valid-well-known") @attr.s(slots=True, frozen=True) -class WellKnownLookupResult(object): +class WellKnownLookupResult: delegated_server = attr.ib() -class WellKnownResolver(object): +class WellKnownResolver: """Handles well-known lookups for matrix servers. """ @@ -100,15 +101,14 @@ class WellKnownResolver(object): self._well_known_agent = RedirectAgent(agent) self.user_agent = user_agent - @defer.inlineCallbacks - def get_well_known(self, server_name): + async def get_well_known(self, server_name: bytes) -> WellKnownLookupResult: """Attempt to fetch and parse a .well-known file for the given server Args: - server_name (bytes): name of the server, from the requested url + server_name: name of the server, from the requested url Returns: - Deferred[WellKnownLookupResult]: The result of the lookup + The result of the lookup """ try: prev_result, expiry, ttl = self._well_known_cache.get_with_expiry( @@ -125,7 +125,9 @@ class WellKnownResolver(object): # requests for the same server in parallel? try: with Measure(self._clock, "get_well_known"): - result, cache_period = yield self._fetch_well_known(server_name) + result, cache_period = await self._fetch_well_known( + server_name + ) # type: Tuple[Optional[bytes], float] except _FetchWellKnownFailure as e: if prev_result and e.temporary: @@ -154,18 +156,17 @@ class WellKnownResolver(object): return WellKnownLookupResult(delegated_server=result) - @defer.inlineCallbacks - def _fetch_well_known(self, server_name): + async def _fetch_well_known(self, server_name: bytes) -> Tuple[bytes, float]: """Actually fetch and parse a .well-known, without checking the cache Args: - server_name (bytes): name of the server, from the requested url + server_name: name of the server, from the requested url Raises: _FetchWellKnownFailure if we fail to lookup a result Returns: - Deferred[Tuple[bytes,int]]: The lookup result and cache period. + The lookup result and cache period. """ had_valid_well_known = self._had_valid_well_known_cache.get(server_name, False) @@ -173,7 +174,7 @@ class WellKnownResolver(object): # We do this in two steps to differentiate between possibly transient # errors (e.g. can't connect to host, 503 response) and more permenant # errors (such as getting a 404 response). - response, body = yield self._make_well_known_request( + response, body = await self._make_well_known_request( server_name, retry=had_valid_well_known ) @@ -181,7 +182,7 @@ class WellKnownResolver(object): if response.code != 200: raise Exception("Non-200 response %s" % (response.code,)) - parsed_body = json.loads(body.decode("utf-8")) + parsed_body = json_decoder.decode(body.decode("utf-8")) logger.info("Response from .well-known: %s", parsed_body) result = parsed_body["m.server"].encode("ascii") @@ -216,20 +217,20 @@ class WellKnownResolver(object): return result, cache_period - @defer.inlineCallbacks - def _make_well_known_request(self, server_name, retry): + async def _make_well_known_request( + self, server_name: bytes, retry: bool + ) -> Tuple[IResponse, bytes]: """Make the well known request. This will retry the request if requested and it fails (with unable to connect or receives a 5xx error). Args: - server_name (bytes) - retry (bool): Whether to retry the request if it fails. + server_name: name of the server, from the requested url + retry: Whether to retry the request if it fails. Returns: - Deferred[tuple[IResponse, bytes]] Returns the response object and - body. Response may be a non-200 response. + Returns the response object and body. Response may be a non-200 response. """ uri = b"https://%s/.well-known/matrix/server" % (server_name,) uri_str = uri.decode("ascii") @@ -244,12 +245,12 @@ class WellKnownResolver(object): logger.info("Fetching %s", uri_str) try: - response = yield make_deferred_yieldable( + response = await make_deferred_yieldable( self._well_known_agent.request( b"GET", uri, headers=Headers(headers) ) ) - body = yield make_deferred_yieldable(readBody(response)) + body = await make_deferred_yieldable(readBody(response)) if 500 <= response.code < 600: raise Exception("Non-200 response %s" % (response.code,)) @@ -266,21 +267,24 @@ class WellKnownResolver(object): logger.info("Error fetching %s: %s. Retrying", uri_str, e) # Sleep briefly in the hopes that they come back up - yield self._clock.sleep(0.5) + await self._clock.sleep(0.5) -def _cache_period_from_headers(headers, time_now=time.time): +def _cache_period_from_headers( + headers: Headers, time_now: Callable[[], float] = time.time +) -> Optional[float]: cache_controls = _parse_cache_control(headers) if b"no-store" in cache_controls: return 0 if b"max-age" in cache_controls: - try: - max_age = int(cache_controls[b"max-age"]) - return max_age - except ValueError: - pass + max_age = cache_controls[b"max-age"] + if max_age: + try: + return int(max_age) + except ValueError: + pass expires = headers.getRawHeaders(b"expires") if expires is not None: @@ -296,7 +300,7 @@ def _cache_period_from_headers(headers, time_now=time.time): return None -def _parse_cache_control(headers): +def _parse_cache_control(headers: Headers) -> Dict[bytes, Optional[bytes]]: cache_controls = {} for hdr in headers.getRawHeaders(b"cache-control", []): for directive in hdr.split(b","): @@ -307,7 +311,7 @@ def _parse_cache_control(headers): return cache_controls -@attr.s() +@attr.s(slots=True) class _FetchWellKnownFailure(Exception): # True if we didn't get a non-5xx HTTP response, i.e. this may or may not be # a temporary failure. diff --git a/synapse/http/matrixfederationclient.py b/synapse/http/matrixfederationclient.py
index 148eeb19dc..c23a4d7c0c 100644 --- a/synapse/http/matrixfederationclient.py +++ b/synapse/http/matrixfederationclient.py
@@ -29,10 +29,11 @@ from zope.interface import implementer from twisted.internet import defer, protocol from twisted.internet.error import DNSLookupError -from twisted.internet.interfaces import IReactorPluggableNameResolver +from twisted.internet.interfaces import IReactorPluggableNameResolver, IReactorTime from twisted.internet.task import _EPSILON, Cooperator from twisted.web._newclient import ResponseDone from twisted.web.http_headers import Headers +from twisted.web.iweb import IResponse import synapse.metrics import synapse.util.retryutils @@ -53,6 +54,7 @@ from synapse.logging.opentracing import ( start_active_span, tags, ) +from synapse.util import json_decoder from synapse.util.async_helpers import timeout_deferred from synapse.util.metrics import Measure @@ -74,8 +76,8 @@ MAXINT = sys.maxsize _next_id = 1 -@attr.s -class MatrixFederationRequest(object): +@attr.s(slots=True, frozen=True) +class MatrixFederationRequest: method = attr.ib() """HTTP method :type: str @@ -110,27 +112,52 @@ class MatrixFederationRequest(object): :type: str|None """ + uri = attr.ib(init=False, type=bytes) + """The URI of this request + """ + def __attrs_post_init__(self): global _next_id - self.txn_id = "%s-O-%s" % (self.method, _next_id) + txn_id = "%s-O-%s" % (self.method, _next_id) _next_id = (_next_id + 1) % (MAXINT - 1) + object.__setattr__(self, "txn_id", txn_id) + + destination_bytes = self.destination.encode("ascii") + path_bytes = self.path.encode("ascii") + if self.query: + query_bytes = encode_query_args(self.query) + else: + query_bytes = b"" + + # The object is frozen so we can pre-compute this. + uri = urllib.parse.urlunparse( + (b"matrix", destination_bytes, path_bytes, None, query_bytes, b"") + ) + object.__setattr__(self, "uri", uri) + def get_json(self): if self.json_callback: return self.json_callback() return self.json -@defer.inlineCallbacks -def _handle_json_response(reactor, timeout_sec, request, response): +async def _handle_json_response( + reactor: IReactorTime, + timeout_sec: float, + request: MatrixFederationRequest, + response: IResponse, + start_ms: int, +): """ Reads the JSON body of a response, with a timeout Args: - reactor (IReactor): twisted reactor, for the timeout - timeout_sec (float): number of seconds to wait for response to complete - request (MatrixFederationRequest): the request that triggered the response - response (IResponse): response to the request + reactor: twisted reactor, for the timeout + timeout_sec: number of seconds to wait for response to complete + request: the request that triggered the response + response: response to the request + start_ms: Timestamp when request was made Returns: dict: parsed JSON response @@ -138,34 +165,48 @@ def _handle_json_response(reactor, timeout_sec, request, response): try: check_content_type_is_json(response.headers) - d = treq.json_content(response) + # Use the custom JSON decoder (partially re-implements treq.json_content). + d = treq.text_content(response, encoding="utf-8") + d.addCallback(json_decoder.decode) d = timeout_deferred(d, timeout=timeout_sec, reactor=reactor) - body = yield make_deferred_yieldable(d) - except TimeoutError as e: + body = await make_deferred_yieldable(d) + except defer.TimeoutError as e: logger.warning( - "{%s} [%s] Timed out reading response", request.txn_id, request.destination, + "{%s} [%s] Timed out reading response - %s %s", + request.txn_id, + request.destination, + request.method, + request.uri.decode("ascii"), ) raise RequestSendFailed(e, can_retry=True) from e except Exception as e: logger.warning( - "{%s} [%s] Error reading response: %s", + "{%s} [%s] Error reading response %s %s: %s", request.txn_id, request.destination, + request.method, + request.uri.decode("ascii"), e, ) raise + + time_taken_secs = reactor.seconds() - start_ms / 1000 + logger.info( - "{%s} [%s] Completed: %d %s", + "{%s} [%s] Completed request: %d %s in %.2f secs - %s %s", request.txn_id, request.destination, response.code, response.phrase.decode("ascii", errors="replace"), + time_taken_secs, + request.method, + request.uri.decode("ascii"), ) return body -class MatrixFederationHttpClient(object): +class MatrixFederationHttpClient: """HTTP client used to talk to other homeservers over the federation protocol. Send client certificates and signs requests. @@ -188,7 +229,7 @@ class MatrixFederationHttpClient(object): ) @implementer(IReactorPluggableNameResolver) - class Reactor(object): + class Reactor: def __getattr__(_self, attr): if attr == "nameResolver": return nameResolver @@ -224,8 +265,7 @@ class MatrixFederationHttpClient(object): self._cooperator = Cooperator(scheduler=schedule) - @defer.inlineCallbacks - def _send_request_with_optional_trailing_slash( + async def _send_request_with_optional_trailing_slash( self, request, try_trailing_slash_on_400=False, **send_request_args ): """Wrapper for _send_request which can optionally retry the request @@ -246,10 +286,10 @@ class MatrixFederationHttpClient(object): (except 429). Returns: - Deferred[Dict]: Parsed JSON response body. + Dict: Parsed JSON response body. """ try: - response = yield self._send_request(request, **send_request_args) + response = await self._send_request(request, **send_request_args) except HttpResponseException as e: # Received an HTTP error > 300. Check if it meets the requirements # to retry with a trailing slash @@ -263,14 +303,15 @@ class MatrixFederationHttpClient(object): # 'M_UNRECOGNIZED' which some endpoints can return when omitting a # trailing slash on Synapse <= v0.99.3. logger.info("Retrying request with trailing slash") - request.path += "/" - response = yield self._send_request(request, **send_request_args) + # Request is frozen so we create a new instance + request = attr.evolve(request, path=request.path + "/") + + response = await self._send_request(request, **send_request_args) return response - @defer.inlineCallbacks - def _send_request( + async def _send_request( self, request, retry_on_dns_fail=True, @@ -311,7 +352,7 @@ class MatrixFederationHttpClient(object): backoff_on_404 (bool): Back off if we get a 404 Returns: - Deferred[twisted.web.client.Response]: resolves with the HTTP + twisted.web.client.Response: resolves with the HTTP response object on success. Raises: @@ -335,7 +376,7 @@ class MatrixFederationHttpClient(object): ): raise FederationDeniedError(request.destination) - limiter = yield synapse.util.retryutils.get_retry_limiter( + limiter = await synapse.util.retryutils.get_retry_limiter( request.destination, self.clock, self._store, @@ -376,9 +417,7 @@ class MatrixFederationHttpClient(object): else: retries_left = MAX_SHORT_RETRIES - url_bytes = urllib.parse.urlunparse( - (b"matrix", destination_bytes, path_bytes, None, query_bytes, b"") - ) + url_bytes = request.uri url_str = url_bytes.decode("ascii") url_to_sign_bytes = urllib.parse.urlunparse( @@ -405,7 +444,7 @@ class MatrixFederationHttpClient(object): headers_dict[b"Authorization"] = auth_headers - logger.info( + logger.debug( "{%s} [%s] Sending request: %s %s; timeout %fs", request.txn_id, request.destination, @@ -433,13 +472,10 @@ class MatrixFederationHttpClient(object): reactor=self.reactor, ) - response = yield request_deferred - except TimeoutError as e: - raise RequestSendFailed(e, can_retry=True) from e + response = await request_deferred except DNSLookupError as e: raise RequestSendFailed(e, can_retry=retry_on_dns_fail) from e except Exception as e: - logger.info("Failed to send request: %s", e) raise RequestSendFailed(e, can_retry=True) from e incoming_responses_counter.labels( @@ -447,6 +483,7 @@ class MatrixFederationHttpClient(object): ).inc() set_tag(tags.HTTP_STATUS_CODE, response.code) + response_phrase = response.phrase.decode("ascii", errors="replace") if 200 <= response.code < 300: logger.debug( @@ -454,7 +491,7 @@ class MatrixFederationHttpClient(object): request.txn_id, request.destination, response.code, - response.phrase.decode("ascii", errors="replace"), + response_phrase, ) pass else: @@ -463,7 +500,7 @@ class MatrixFederationHttpClient(object): request.txn_id, request.destination, response.code, - response.phrase.decode("ascii", errors="replace"), + response_phrase, ) # :'( # Update transactions table? @@ -473,7 +510,7 @@ class MatrixFederationHttpClient(object): ) try: - body = yield make_deferred_yieldable(d) + body = await make_deferred_yieldable(d) except Exception as e: # Eh, we're already going to raise an exception so lets # ignore if this fails. @@ -487,7 +524,7 @@ class MatrixFederationHttpClient(object): ) body = None - e = HttpResponseException(response.code, response.phrase, body) + e = HttpResponseException(response.code, response_phrase, body) # Retry if the error is a 429 (Too Many Requests), # otherwise just raise a standard HttpResponseException @@ -498,7 +535,7 @@ class MatrixFederationHttpClient(object): break except RequestSendFailed as e: - logger.warning( + logger.info( "{%s} [%s] Request failed: %s %s: %s", request.txn_id, request.destination, @@ -527,7 +564,7 @@ class MatrixFederationHttpClient(object): delay, ) - yield self.clock.sleep(delay) + await self.clock.sleep(delay) retries_left -= 1 else: raise @@ -590,8 +627,7 @@ class MatrixFederationHttpClient(object): ) return auth_headers - @defer.inlineCallbacks - def put_json( + async def put_json( self, destination, path, @@ -619,10 +655,14 @@ class MatrixFederationHttpClient(object): long_retries (bool): whether to use the long retry algorithm. See docs on _send_request for details. - timeout (int|None): number of milliseconds to wait for the response headers - (including connecting to the server), *for each attempt*. + timeout (int|None): number of milliseconds to wait for the response. self._default_timeout (60s) by default. + Note that we may make several attempts to send the request; this + timeout applies to the time spent waiting for response headers for + *each* attempt (including connection time) as well as the time spent + reading the response body after a 200 response. + ignore_backoff (bool): true to ignore the historical backoff data and try the request anyway. backoff_on_404 (bool): True if we should count a 404 response as @@ -635,7 +675,7 @@ class MatrixFederationHttpClient(object): enabled. Returns: - Deferred[dict|list]: Succeeds when we get a 2xx HTTP response. The + dict|list: Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body. Raises: @@ -657,7 +697,9 @@ class MatrixFederationHttpClient(object): json=data, ) - response = yield self._send_request_with_optional_trailing_slash( + start_ms = self.clock.time_msec() + + response = await self._send_request_with_optional_trailing_slash( request, try_trailing_slash_on_400, backoff_on_404=backoff_on_404, @@ -666,14 +708,18 @@ class MatrixFederationHttpClient(object): timeout=timeout, ) - body = yield _handle_json_response( - self.reactor, self.default_timeout, request, response + if timeout is not None: + _sec_timeout = timeout / 1000 + else: + _sec_timeout = self.default_timeout + + body = await _handle_json_response( + self.reactor, _sec_timeout, request, response, start_ms ) return body - @defer.inlineCallbacks - def post_json( + async def post_json( self, destination, path, @@ -697,16 +743,20 @@ class MatrixFederationHttpClient(object): long_retries (bool): whether to use the long retry algorithm. See docs on _send_request for details. - timeout (int|None): number of milliseconds to wait for the response headers - (including connecting to the server), *for each attempt*. + timeout (int|None): number of milliseconds to wait for the response. self._default_timeout (60s) by default. + Note that we may make several attempts to send the request; this + timeout applies to the time spent waiting for response headers for + *each* attempt (including connection time) as well as the time spent + reading the response body after a 200 response. + ignore_backoff (bool): true to ignore the historical backoff data and try the request anyway. args (dict): query params Returns: - Deferred[dict|list]: Succeeds when we get a 2xx HTTP response. The + dict|list: Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body. Raises: @@ -724,7 +774,9 @@ class MatrixFederationHttpClient(object): method="POST", destination=destination, path=path, query=args, json=data ) - response = yield self._send_request( + start_ms = self.clock.time_msec() + + response = await self._send_request( request, long_retries=long_retries, timeout=timeout, @@ -736,13 +788,12 @@ class MatrixFederationHttpClient(object): else: _sec_timeout = self.default_timeout - body = yield _handle_json_response( - self.reactor, _sec_timeout, request, response + body = await _handle_json_response( + self.reactor, _sec_timeout, request, response, start_ms, ) return body - @defer.inlineCallbacks - def get_json( + async def get_json( self, destination, path, @@ -763,10 +814,14 @@ class MatrixFederationHttpClient(object): args (dict|None): A dictionary used to create query strings, defaults to None. - timeout (int|None): number of milliseconds to wait for the response headers - (including connecting to the server), *for each attempt*. + timeout (int|None): number of milliseconds to wait for the response. self._default_timeout (60s) by default. + Note that we may make several attempts to send the request; this + timeout applies to the time spent waiting for response headers for + *each* attempt (including connection time) as well as the time spent + reading the response body after a 200 response. + ignore_backoff (bool): true to ignore the historical backoff data and try the request anyway. @@ -774,7 +829,7 @@ class MatrixFederationHttpClient(object): response we should try appending a trailing slash to the end of the request. Workaround for #3622 in Synapse <= v0.99.3. Returns: - Deferred[dict|list]: Succeeds when we get a 2xx HTTP response. The + dict|list: Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body. Raises: @@ -791,7 +846,9 @@ class MatrixFederationHttpClient(object): method="GET", destination=destination, path=path, query=args ) - response = yield self._send_request_with_optional_trailing_slash( + start_ms = self.clock.time_msec() + + response = await self._send_request_with_optional_trailing_slash( request, try_trailing_slash_on_400, backoff_on_404=False, @@ -800,14 +857,18 @@ class MatrixFederationHttpClient(object): timeout=timeout, ) - body = yield _handle_json_response( - self.reactor, self.default_timeout, request, response + if timeout is not None: + _sec_timeout = timeout / 1000 + else: + _sec_timeout = self.default_timeout + + body = await _handle_json_response( + self.reactor, _sec_timeout, request, response, start_ms ) return body - @defer.inlineCallbacks - def delete_json( + async def delete_json( self, destination, path, @@ -826,16 +887,20 @@ class MatrixFederationHttpClient(object): long_retries (bool): whether to use the long retry algorithm. See docs on _send_request for details. - timeout (int|None): number of milliseconds to wait for the response headers - (including connecting to the server), *for each attempt*. + timeout (int|None): number of milliseconds to wait for the response. self._default_timeout (60s) by default. + Note that we may make several attempts to send the request; this + timeout applies to the time spent waiting for response headers for + *each* attempt (including connection time) as well as the time spent + reading the response body after a 200 response. + ignore_backoff (bool): true to ignore the historical backoff data and try the request anyway. args (dict): query params Returns: - Deferred[dict|list]: Succeeds when we get a 2xx HTTP response. The + dict|list: Succeeds when we get a 2xx HTTP response. The result will be the decoded JSON body. Raises: @@ -852,20 +917,26 @@ class MatrixFederationHttpClient(object): method="DELETE", destination=destination, path=path, query=args ) - response = yield self._send_request( + start_ms = self.clock.time_msec() + + response = await self._send_request( request, long_retries=long_retries, timeout=timeout, ignore_backoff=ignore_backoff, ) - body = yield _handle_json_response( - self.reactor, self.default_timeout, request, response + if timeout is not None: + _sec_timeout = timeout / 1000 + else: + _sec_timeout = self.default_timeout + + body = await _handle_json_response( + self.reactor, _sec_timeout, request, response, start_ms ) return body - @defer.inlineCallbacks - def get_file( + async def get_file( self, destination, path, @@ -885,7 +956,7 @@ class MatrixFederationHttpClient(object): and try the request anyway. Returns: - Deferred[tuple[int, dict]]: Resolves with an (int,dict) tuple of + tuple[int, dict]: Resolves with an (int,dict) tuple of the file length and a dict of the response headers. Raises: @@ -902,7 +973,7 @@ class MatrixFederationHttpClient(object): method="GET", destination=destination, path=path, query=args ) - response = yield self._send_request( + response = await self._send_request( request, retry_on_dns_fail=retry_on_dns_fail, ignore_backoff=ignore_backoff ) @@ -911,7 +982,7 @@ class MatrixFederationHttpClient(object): try: d = _readBodyToFile(response, output_stream, max_size) d.addTimeout(self.default_timeout, self.reactor) - length = yield make_deferred_yieldable(d) + length = await make_deferred_yieldable(d) except Exception as e: logger.warning( "{%s} [%s] Error reading response: %s", @@ -921,12 +992,14 @@ class MatrixFederationHttpClient(object): ) raise logger.info( - "{%s} [%s] Completed: %d %s [%d bytes]", + "{%s} [%s] Completed: %d %s [%d bytes] %s %s", request.txn_id, request.destination, response.code, response.phrase.decode("ascii", errors="replace"), length, + request.method, + request.uri.decode("ascii"), ) return (length, headers) diff --git a/synapse/http/proxyagent.py b/synapse/http/proxyagent.py
index 332da02a8d..e32d3f43e0 100644 --- a/synapse/http/proxyagent.py +++ b/synapse/http/proxyagent.py
@@ -44,8 +44,11 @@ class ProxyAgent(_AgentBase): `BrowserLikePolicyForHTTPS`, so unless you have special requirements you can leave this as-is. - connectTimeout (float): The amount of time that this Agent will wait - for the peer to accept a connection. + connectTimeout (Optional[float]): The amount of time that this Agent will wait + for the peer to accept a connection, in seconds. If 'None', + HostnameEndpoint's default (30s) will be used. + + This is used for connections to both proxies and destination servers. bindAddress (bytes): The local address for client sockets to bind to. @@ -108,6 +111,15 @@ class ProxyAgent(_AgentBase): Returns: Deferred[IResponse]: completes when the header of the response has been received (regardless of the response status code). + + Can fail with: + SchemeNotSupported: if the uri is not http or https + + twisted.internet.error.TimeoutError if the server we are connecting + to (proxy or destination) does not accept a connection before + connectTimeout. + + ... other things too. """ uri = uri.strip() if not _VALID_URI.match(uri): diff --git a/synapse/http/request_metrics.py b/synapse/http/request_metrics.py
index b58ae3d9db..cd94e789e8 100644 --- a/synapse/http/request_metrics.py +++ b/synapse/http/request_metrics.py
@@ -145,7 +145,7 @@ LaterGauge( ) -class RequestMetrics(object): +class RequestMetrics: def start(self, time_sec, name, method): self.start = time_sec self.start_context = current_context() diff --git a/synapse/http/server.py b/synapse/http/server.py
index d4f9ad6e67..996a31a9ec 100644 --- a/synapse/http/server.py +++ b/synapse/http/server.py
@@ -22,12 +22,13 @@ import types import urllib from http import HTTPStatus from io import BytesIO -from typing import Any, Callable, Dict, Tuple, Union +from typing import Any, Callable, Dict, Iterator, List, Tuple, Union import jinja2 -from canonicaljson import encode_canonical_json, encode_pretty_printed_json, json +from canonicaljson import iterencode_canonical_json, iterencode_pretty_printed_json +from zope.interface import implementer -from twisted.internet import defer +from twisted.internet import defer, interfaces from twisted.python import failure from twisted.web import resource from twisted.web.server import NOT_DONE_YET, Request @@ -46,6 +47,7 @@ from synapse.api.errors import ( from synapse.http.site import SynapseRequest from synapse.logging.context import preserve_fn from synapse.logging.opentracing import trace_servlet +from synapse.util import json_encoder from synapse.util.caches import intern_dict logger = logging.getLogger(__name__) @@ -172,7 +174,7 @@ def wrap_async_request_handler(h): return preserve_fn(wrapped_async_request_handler) -class HttpServer(object): +class HttpServer: """ Interface for registering callbacks on a HTTP server """ @@ -242,10 +244,12 @@ class _AsyncResource(resource.Resource, metaclass=abc.ABCMeta): no appropriate method exists. Can be overriden in sub classes for different routing. """ + # Treat HEAD requests as GET requests. + request_method = request.method.decode("ascii") + if request_method == "HEAD": + request_method = "GET" - method_handler = getattr( - self, "_async_render_%s" % (request.method.decode("ascii"),), None - ) + method_handler = getattr(self, "_async_render_%s" % (request_method,), None) if method_handler: raw_callback_return = method_handler(request) @@ -362,11 +366,15 @@ class JsonResource(DirectServeJsonResource): A tuple of the callback to use, the name of the servlet, and the key word arguments to pass to the callback """ + # Treat HEAD requests as GET requests. request_path = request.path.decode("ascii") + request_method = request.method + if request_method == b"HEAD": + request_method = b"GET" # Loop through all the registered callbacks to check if the method # and path regex match - for path_entry in self.path_regexs.get(request.method, []): + for path_entry in self.path_regexs.get(request_method, []): m = path_entry.pattern.match(request_path) if m: # We found a match! @@ -492,6 +500,90 @@ class RootOptionsRedirectResource(OptionsResource, RootRedirect): pass +@implementer(interfaces.IPushProducer) +class _ByteProducer: + """ + Iteratively write bytes to the request. + """ + + # The minimum number of bytes for each chunk. Note that the last chunk will + # usually be smaller than this. + min_chunk_size = 1024 + + def __init__( + self, request: Request, iterator: Iterator[bytes], + ): + self._request = request + self._iterator = iterator + self._paused = False + + # Register the producer and start producing data. + self._request.registerProducer(self, True) + self.resumeProducing() + + def _send_data(self, data: List[bytes]) -> None: + """ + Send a list of bytes as a chunk of a response. + """ + if not data: + return + self._request.write(b"".join(data)) + + def pauseProducing(self) -> None: + self._paused = True + + def resumeProducing(self) -> None: + # We've stopped producing in the meantime (note that this might be + # re-entrant after calling write). + if not self._request: + return + + self._paused = False + + # Write until there's backpressure telling us to stop. + while not self._paused: + # Get the next chunk and write it to the request. + # + # The output of the JSON encoder is buffered and coalesced until + # min_chunk_size is reached. This is because JSON encoders produce + # very small output per iteration and the Request object converts + # each call to write() to a separate chunk. Without this there would + # be an explosion in bytes written (e.g. b"{" becoming "1\r\n{\r\n"). + # + # Note that buffer stores a list of bytes (instead of appending to + # bytes) to hopefully avoid many allocations. + buffer = [] + buffered_bytes = 0 + while buffered_bytes < self.min_chunk_size: + try: + data = next(self._iterator) + buffer.append(data) + buffered_bytes += len(data) + except StopIteration: + # The entire JSON object has been serialized, write any + # remaining data, finalize the producer and the request, and + # clean-up any references. + self._send_data(buffer) + self._request.unregisterProducer() + self._request.finish() + self.stopProducing() + return + + self._send_data(buffer) + + def stopProducing(self) -> None: + # Clear a circular reference. + self._request = None + + +def _encode_json_bytes(json_object: Any) -> Iterator[bytes]: + """ + Encode an object into JSON. Returns an iterator of bytes. + """ + for chunk in json_encoder.iterencode(json_object): + yield chunk.encode("utf-8") + + def respond_with_json( request: Request, code: int, @@ -526,15 +618,22 @@ def respond_with_json( return None if pretty_print: - json_bytes = encode_pretty_printed_json(json_object) + b"\n" + encoder = iterencode_pretty_printed_json else: if canonical_json or synapse.events.USE_FROZEN_DICTS: - # canonicaljson already encodes to bytes - json_bytes = encode_canonical_json(json_object) + encoder = iterencode_canonical_json else: - json_bytes = json.dumps(json_object).encode("utf-8") + encoder = _encode_json_bytes - return respond_with_json_bytes(request, code, json_bytes, send_cors=send_cors) + request.setResponseCode(code) + request.setHeader(b"Content-Type", b"application/json") + request.setHeader(b"Cache-Control", b"no-cache, no-store, must-revalidate") + + if send_cors: + set_cors_headers(request) + + _ByteProducer(request, encoder(json_object)) + return NOT_DONE_YET def respond_with_json_bytes( @@ -579,7 +678,7 @@ def set_cors_headers(request: Request): """ request.setHeader(b"Access-Control-Allow-Origin", b"*") request.setHeader( - b"Access-Control-Allow-Methods", b"GET, POST, PUT, DELETE, OPTIONS" + b"Access-Control-Allow-Methods", b"GET, HEAD, POST, PUT, DELETE, OPTIONS" ) request.setHeader( b"Access-Control-Allow-Headers", diff --git a/synapse/http/servlet.py b/synapse/http/servlet.py
index a34e5ead88..fd90ba7828 100644 --- a/synapse/http/servlet.py +++ b/synapse/http/servlet.py
@@ -17,9 +17,8 @@ import logging -from canonicaljson import json - from synapse.api.errors import Codes, SynapseError +from synapse.util import json_decoder logger = logging.getLogger(__name__) @@ -215,7 +214,7 @@ def parse_json_value_from_request(request, allow_empty_body=False): return None try: - content = json.loads(content_bytes.decode("utf-8")) + content = json_decoder.decode(content_bytes.decode("utf-8")) except Exception as e: logger.warning("Unable to parse JSON: %s", e) raise SynapseError(400, "Content not JSON.", errcode=Codes.NOT_JSON) @@ -257,7 +256,7 @@ def assert_params_in_dict(body, required): raise SynapseError(400, "Missing params: %r" % absent, Codes.MISSING_PARAM) -class RestServlet(object): +class RestServlet: """ A Synapse REST Servlet. diff --git a/synapse/http/site.py b/synapse/http/site.py
index 6f3b2258cc..6e79b47828 100644 --- a/synapse/http/site.py +++ b/synapse/http/site.py
@@ -146,10 +146,9 @@ class SynapseRequest(Request): Returns a context manager; the correct way to use this is: - @defer.inlineCallbacks - def handle_request(request): + async def handle_request(request): with request.processing("FooServlet"): - yield really_handle_the_request() + await really_handle_the_request() Once the context manager is closed, the completion of the request will be logged, and the various metrics will be updated. @@ -287,7 +286,9 @@ class SynapseRequest(Request): # the connection dropped) code += "!" - self.site.access_logger.info( + log_level = logging.INFO if self._should_log_request() else logging.DEBUG + self.site.access_logger.log( + log_level, "%s - %s - {%s}" " Processed request: %.3fsec/%.3fsec (%.3fsec, %.3fsec) (%.3fsec/%.3fsec/%d)" ' %sB %s "%s %s %s" "%s" [%d dbevts]', @@ -315,6 +316,17 @@ class SynapseRequest(Request): except Exception as e: logger.warning("Failed to stop metrics: %r", e) + def _should_log_request(self) -> bool: + """Whether we should log at INFO that we processed the request. + """ + if self.path == b"/health": + return False + + if self.method == b"OPTIONS": + return False + + return True + class XForwardedForRequest(SynapseRequest): def __init__(self, *args, **kw): diff --git a/synapse/logging/_structured.py b/synapse/logging/_structured.py
index 7372450b45..144506c8f2 100644 --- a/synapse/logging/_structured.py +++ b/synapse/logging/_structured.py
@@ -55,7 +55,7 @@ def stdlib_log_level_to_twisted(level: str) -> LogLevel: @attr.s @implementer(ILogObserver) -class LogContextObserver(object): +class LogContextObserver: """ An ILogObserver which adds Synapse-specific log context information. @@ -169,7 +169,7 @@ class OutputPipeType(Values): @attr.s -class DrainConfiguration(object): +class DrainConfiguration: name = attr.ib() type = attr.ib() location = attr.ib() @@ -177,7 +177,7 @@ class DrainConfiguration(object): @attr.s -class NetworkJSONTerseOptions(object): +class NetworkJSONTerseOptions: maximum_buffer = attr.ib(type=int) diff --git a/synapse/logging/_terse_json.py b/synapse/logging/_terse_json.py
index c0b9384189..1b8916cfa2 100644 --- a/synapse/logging/_terse_json.py +++ b/synapse/logging/_terse_json.py
@@ -152,7 +152,7 @@ def TerseJSONToConsoleLogObserver(outFile: IO[str], metadata: dict) -> FileLogOb @attr.s @implementer(IPushProducer) -class LogProducer(object): +class LogProducer: """ An IPushProducer that writes logs from its buffer to its transport when it is resumed. @@ -190,7 +190,7 @@ class LogProducer(object): @attr.s @implementer(ILogObserver) -class TerseJSONToTCPLogObserver(object): +class TerseJSONToTCPLogObserver: """ An IObserver that writes JSON logs to a TCP target. diff --git a/synapse/logging/context.py b/synapse/logging/context.py
index cbeeb870cb..ca0c774cc5 100644 --- a/synapse/logging/context.py +++ b/synapse/logging/context.py
@@ -65,6 +65,11 @@ except Exception: return None +# a hook which can be set during testing to assert that we aren't abusing logcontexts. +def logcontext_error(msg: str): + logger.warning(msg) + + # get an id for the current thread. # # threading.get_ident doesn't actually return an OS-level tid, and annoyingly, @@ -74,7 +79,7 @@ except Exception: get_thread_id = threading.get_ident -class ContextResourceUsage(object): +class ContextResourceUsage: """Object for tracking the resources used by a log context Attributes: @@ -179,7 +184,7 @@ class ContextResourceUsage(object): LoggingContextOrSentinel = Union["LoggingContext", "_Sentinel"] -class _Sentinel(object): +class _Sentinel: """Sentinel to represent the root context""" __slots__ = ["previous_context", "finished", "request", "scope", "tag"] @@ -217,16 +222,14 @@ class _Sentinel(object): def record_event_fetch(self, event_count): pass - def __nonzero__(self): + def __bool__(self): return False - __bool__ = __nonzero__ # python3 - SENTINEL_CONTEXT = _Sentinel() -class LoggingContext(object): +class LoggingContext: """Additional context for log formatting. Contexts are scoped within a "with" block. @@ -332,10 +335,9 @@ class LoggingContext(object): """Enters this logging context into thread local storage""" old_context = set_current_context(self) if self.previous_context != old_context: - logger.warning( - "Expected previous context %r, found %r", - self.previous_context, - old_context, + logcontext_error( + "Expected previous context %r, found %r" + % (self.previous_context, old_context,) ) return self @@ -348,10 +350,10 @@ class LoggingContext(object): current = set_current_context(self.previous_context) if current is not self: if current is SENTINEL_CONTEXT: - logger.warning("Expected logging context %s was lost", self) + logcontext_error("Expected logging context %s was lost" % (self,)) else: - logger.warning( - "Expected logging context %s but found %s", self, current + logcontext_error( + "Expected logging context %s but found %s" % (self, current) ) # the fact that we are here suggests that the caller thinks that everything @@ -389,16 +391,16 @@ class LoggingContext(object): support getrusuage. """ if get_thread_id() != self.main_thread: - logger.warning("Started logcontext %s on different thread", self) + logcontext_error("Started logcontext %s on different thread" % (self,)) return if self.finished: - logger.warning("Re-starting finished log context %s", self) + logcontext_error("Re-starting finished log context %s" % (self,)) # If we haven't already started record the thread resource usage so # far if self.usage_start: - logger.warning("Re-starting already-active log context %s", self) + logcontext_error("Re-starting already-active log context %s" % (self,)) else: self.usage_start = rusage @@ -416,7 +418,7 @@ class LoggingContext(object): try: if get_thread_id() != self.main_thread: - logger.warning("Stopped logcontext %s on different thread", self) + logcontext_error("Stopped logcontext %s on different thread" % (self,)) return if not rusage: @@ -424,9 +426,9 @@ class LoggingContext(object): # Record the cpu used since we started if not self.usage_start: - logger.warning( - "Called stop on logcontext %s without recording a start rusage", - self, + logcontext_error( + "Called stop on logcontext %s without recording a start rusage" + % (self,) ) return @@ -586,14 +588,13 @@ class PreserveLoggingContext: if context != self._new_context: if not context: - logger.warning( - "Expected logging context %s was lost", self._new_context + logcontext_error( + "Expected logging context %s was lost" % (self._new_context,) ) else: - logger.warning( - "Expected logging context %s but found %s", - self._new_context, - context, + logcontext_error( + "Expected logging context %s but found %s" + % (self._new_context, context,) ) diff --git a/synapse/logging/formatter.py b/synapse/logging/formatter.py
index d736ad5b9b..11f60a77f7 100644 --- a/synapse/logging/formatter.py +++ b/synapse/logging/formatter.py
@@ -30,7 +30,7 @@ class LogFormatter(logging.Formatter): """ def __init__(self, *args, **kwargs): - super(LogFormatter, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) def formatException(self, ei): sio = StringIO() diff --git a/synapse/logging/opentracing.py b/synapse/logging/opentracing.py
index 21dbd9f415..e58850faff 100644 --- a/synapse/logging/opentracing.py +++ b/synapse/logging/opentracing.py
@@ -172,11 +172,11 @@ from functools import wraps from typing import TYPE_CHECKING, Dict, Optional, Type import attr -from canonicaljson import json from twisted.internet import defer from synapse.config import ConfigError +from synapse.util import json_decoder, json_encoder if TYPE_CHECKING: from synapse.http.site import SynapseRequest @@ -185,7 +185,7 @@ if TYPE_CHECKING: # Helper class -class _DummyTagNames(object): +class _DummyTagNames: """wrapper of opentracings tags. We need to have them if we want to reference them without opentracing around. Clearly they should never actually show up in a trace. `set_tags` overwrites @@ -499,7 +499,9 @@ def start_active_span_from_edu( if opentracing is None: return _noop_context_manager() - carrier = json.loads(edu_content.get("context", "{}")).get("opentracing", {}) + carrier = json_decoder.decode(edu_content.get("context", "{}")).get( + "opentracing", {} + ) context = opentracing.tracer.extract(opentracing.Format.TEXT_MAP, carrier) _references = [ opentracing.child_of(span_context_from_string(x)) @@ -507,7 +509,7 @@ def start_active_span_from_edu( ] # For some reason jaeger decided not to support the visualization of multiple parent - # spans or explicitely show references. I include the span context as a tag here as + # spans or explicitly show references. I include the span context as a tag here as # an aid to people debugging but it's really not an ideal solution. references += _references @@ -690,7 +692,7 @@ def active_span_context_as_string(): opentracing.tracer.inject( opentracing.tracer.active_span, opentracing.Format.TEXT_MAP, carrier ) - return json.dumps(carrier) + return json_encoder.encode(carrier) @only_if_tracing @@ -699,7 +701,7 @@ def span_context_from_string(carrier): Returns: The active span context decoded from a string. """ - carrier = json.loads(carrier) + carrier = json_decoder.decode(carrier) return opentracing.tracer.extract(opentracing.Format.TEXT_MAP, carrier) diff --git a/synapse/logging/scopecontextmanager.py b/synapse/logging/scopecontextmanager.py
index 026854b4c7..7b9c657456 100644 --- a/synapse/logging/scopecontextmanager.py +++ b/synapse/logging/scopecontextmanager.py
@@ -107,7 +107,7 @@ class _LogContextScope(Scope): finish_on_close (Boolean): if True finish the span when the scope is closed """ - super(_LogContextScope, self).__init__(manager, span) + super().__init__(manager, span) self.logcontext = logcontext self._finish_on_close = finish_on_close self._enter_logcontext = enter_logcontext @@ -120,9 +120,9 @@ class _LogContextScope(Scope): def __exit__(self, type, value, traceback): if type == twisted.internet.defer._DefGen_Return: - super(_LogContextScope, self).__exit__(None, None, None) + super().__exit__(None, None, None) else: - super(_LogContextScope, self).__exit__(type, value, traceback) + super().__exit__(type, value, traceback) if self._enter_logcontext: self.logcontext.__exit__(type, value, traceback) else: # the logcontext existed before the creation of the scope diff --git a/synapse/logging/utils.py b/synapse/logging/utils.py
index fea774e2e5..becf66dd86 100644 --- a/synapse/logging/utils.py +++ b/synapse/logging/utils.py
@@ -29,11 +29,11 @@ def _log_debug_as_f(f, msg, msg_args): lineno = f.__code__.co_firstlineno pathname = f.__code__.co_filename - record = logging.LogRecord( + record = logger.makeRecord( name=name, level=logging.DEBUG, - pathname=pathname, - lineno=lineno, + fn=pathname, + lno=lineno, msg=msg, args=msg_args, exc_info=None, diff --git a/synapse/metrics/__init__.py b/synapse/metrics/__init__.py
index 6035672698..b8d2a8e8a9 100644 --- a/synapse/metrics/__init__.py +++ b/synapse/metrics/__init__.py
@@ -15,6 +15,7 @@ import functools import gc +import itertools import logging import os import platform @@ -27,8 +28,8 @@ from prometheus_client import Counter, Gauge, Histogram from prometheus_client.core import ( REGISTRY, CounterMetricFamily, + GaugeHistogramMetricFamily, GaugeMetricFamily, - HistogramMetricFamily, ) from twisted.internet import reactor @@ -46,12 +47,12 @@ logger = logging.getLogger(__name__) METRICS_PREFIX = "/_synapse/metrics" running_on_pypy = platform.python_implementation() == "PyPy" -all_gauges = {} # type: Dict[str, Union[LaterGauge, InFlightGauge, BucketCollector]] +all_gauges = {} # type: Dict[str, Union[LaterGauge, InFlightGauge]] HAVE_PROC_SELF_STAT = os.path.exists("/proc/self/stat") -class RegistryProxy(object): +class RegistryProxy: @staticmethod def collect(): for metric in REGISTRY.collect(): @@ -59,8 +60,8 @@ class RegistryProxy(object): yield metric -@attr.s(hash=True) -class LaterGauge(object): +@attr.s(slots=True, hash=True) +class LaterGauge: name = attr.ib(type=str) desc = attr.ib(type=str) @@ -100,7 +101,7 @@ class LaterGauge(object): all_gauges[self.name] = self -class InFlightGauge(object): +class InFlightGauge: """Tracks number of things (e.g. requests, Measure blocks, etc) in flight at any given time. @@ -205,63 +206,83 @@ class InFlightGauge(object): all_gauges[self.name] = self -@attr.s(hash=True) -class BucketCollector(object): - """ - Like a Histogram, but allows buckets to be point-in-time instead of - incrementally added to. +class GaugeBucketCollector: + """Like a Histogram, but the buckets are Gauges which are updated atomically. - Args: - name (str): Base name of metric to be exported to Prometheus. - data_collector (callable -> dict): A synchronous callable that - returns a dict mapping bucket to number of items in the - bucket. If these buckets are not the same as the buckets - given to this class, they will be remapped into them. - buckets (list[float]): List of floats/ints of the buckets to - give to Prometheus. +Inf is ignored, if given. + The data is updated by calling `update_data` with an iterable of measurements. + We assume that the data is updated less frequently than it is reported to + Prometheus, and optimise for that case. """ - name = attr.ib() - data_collector = attr.ib() - buckets = attr.ib() + __slots__ = ("_name", "_documentation", "_bucket_bounds", "_metric") - def collect(self): + def __init__( + self, + name: str, + documentation: str, + buckets: Iterable[float], + registry=REGISTRY, + ): + """ + Args: + name: base name of metric to be exported to Prometheus. (a _bucket suffix + will be added.) + documentation: help text for the metric + buckets: The top bounds of the buckets to report + registry: metric registry to register with + """ + self._name = name + self._documentation = documentation - # Fetch the data -- this must be synchronous! - data = self.data_collector() + # the tops of the buckets + self._bucket_bounds = [float(b) for b in buckets] + if self._bucket_bounds != sorted(self._bucket_bounds): + raise ValueError("Buckets not in sorted order") - buckets = {} # type: Dict[float, int] + if self._bucket_bounds[-1] != float("inf"): + self._bucket_bounds.append(float("inf")) - res = [] - for x in data.keys(): - for i, bound in enumerate(self.buckets): - if x <= bound: - buckets[bound] = buckets.get(bound, 0) + data[x] + self._metric = self._values_to_metric([]) + registry.register(self) - for i in self.buckets: - res.append([str(i), buckets.get(i, 0)]) + def collect(self): + yield self._metric - res.append(["+Inf", sum(data.values())]) + def update_data(self, values: Iterable[float]): + """Update the data to be reported by the metric - metric = HistogramMetricFamily( - self.name, "", buckets=res, sum_value=sum(x * y for x, y in data.items()) + The existing data is cleared, and each measurement in the input is assigned + to the relevant bucket. + """ + self._metric = self._values_to_metric(values) + + def _values_to_metric(self, values: Iterable[float]) -> GaugeHistogramMetricFamily: + total = 0.0 + bucket_values = [0 for _ in self._bucket_bounds] + + for v in values: + # assign each value to a bucket + for i, bound in enumerate(self._bucket_bounds): + if v <= bound: + bucket_values[i] += 1 + break + + # ... and increment the sum + total += v + + # now, aggregate the bucket values so that they count the number of entries in + # that bucket or below. + accumulated_values = itertools.accumulate(bucket_values) + + return GaugeHistogramMetricFamily( + self._name, + self._documentation, + buckets=list( + zip((str(b) for b in self._bucket_bounds), accumulated_values) + ), + gsum_value=total, ) - yield metric - - def __attrs_post_init__(self): - self.buckets = [float(x) for x in self.buckets if x != "+Inf"] - if self.buckets != sorted(self.buckets): - raise ValueError("Buckets not sorted") - - self.buckets = tuple(self.buckets) - - if self.name in all_gauges.keys(): - logger.warning("%s already registered, reregistering" % (self.name,)) - REGISTRY.unregister(all_gauges.pop(self.name)) - - REGISTRY.register(self) - all_gauges[self.name] = self # @@ -269,7 +290,7 @@ class BucketCollector(object): # -class CPUMetrics(object): +class CPUMetrics: def __init__(self): ticks_per_sec = 100 try: @@ -329,7 +350,7 @@ gc_time = Histogram( ) -class GCCounts(object): +class GCCounts: def collect(self): cm = GaugeMetricFamily("python_gc_counts", "GC object counts", labels=["gen"]) for n, m in enumerate(gc.get_count()): @@ -347,7 +368,7 @@ if not running_on_pypy: # -class PyPyGCStats(object): +class PyPyGCStats: def collect(self): # @stats is a pretty-printer object with __str__() returning a nice table, @@ -482,7 +503,7 @@ build_info.labels( last_ticked = time.time() -class ReactorLastSeenMetric(object): +class ReactorLastSeenMetric: def collect(self): cm = GaugeMetricFamily( "python_twisted_reactor_last_seen", diff --git a/synapse/metrics/_exposition.py b/synapse/metrics/_exposition.py
index 4304c60d56..734271e765 100644 --- a/synapse/metrics/_exposition.py +++ b/synapse/metrics/_exposition.py
@@ -24,9 +24,9 @@ expect, and the newer "best practice" version of the up-to-date official client. import math import threading -from collections import namedtuple from http.server import BaseHTTPRequestHandler, HTTPServer from socketserver import ThreadingMixIn +from typing import Dict, List from urllib.parse import parse_qs, urlparse from prometheus_client import REGISTRY @@ -35,14 +35,6 @@ from twisted.web.resource import Resource from synapse.util import caches -try: - from prometheus_client.samples import Sample -except ImportError: - Sample = namedtuple( # type: ignore[no-redef] # noqa - "Sample", ["name", "labels", "value", "timestamp", "exemplar"] - ) - - CONTENT_TYPE_LATEST = str("text/plain; version=0.0.4; charset=utf-8") @@ -93,17 +85,6 @@ def sample_line(line, name): ) -def nameify_sample(sample): - """ - If we get a prometheus_client<0.4.0 sample as a tuple, transform it into a - namedtuple which has the names we expect. - """ - if not isinstance(sample, Sample): - sample = Sample(*sample, None, None) - - return sample - - def generate_latest(registry, emit_help=False): # Trigger the cache metrics to be rescraped, which updates the common @@ -144,16 +125,33 @@ def generate_latest(registry, emit_help=False): ) ) output.append("# TYPE {0} {1}\n".format(mname, mtype)) - for sample in map(nameify_sample, metric.samples): - # Get rid of the OpenMetrics specific samples + + om_samples = {} # type: Dict[str, List[str]] + for s in metric.samples: for suffix in ["_created", "_gsum", "_gcount"]: - if sample.name.endswith(suffix): + if s.name == metric.name + suffix: + # OpenMetrics specific sample, put in a gauge at the end. + # (these come from gaugehistograms which don't get renamed, + # so no need to faff with mnewname) + om_samples.setdefault(suffix, []).append(sample_line(s, s.name)) break else: - newname = sample.name.replace(mnewname, mname) + newname = s.name.replace(mnewname, mname) if ":" in newname and newname.endswith("_total"): newname = newname[: -len("_total")] - output.append(sample_line(sample, newname)) + output.append(sample_line(s, newname)) + + for suffix, lines in sorted(om_samples.items()): + if emit_help: + output.append( + "# HELP {0}{1} {2}\n".format( + metric.name, + suffix, + metric.documentation.replace("\\", r"\\").replace("\n", r"\n"), + ) + ) + output.append("# TYPE {0}{1} gauge\n".format(metric.name, suffix)) + output.extend(lines) # Get rid of the weird colon things while we're at it if mtype == "counter": @@ -172,16 +170,16 @@ def generate_latest(registry, emit_help=False): ) ) output.append("# TYPE {0} {1}\n".format(mnewname, mtype)) - for sample in map(nameify_sample, metric.samples): - # Get rid of the OpenMetrics specific samples + + for s in metric.samples: + # Get rid of the OpenMetrics specific samples (we should already have + # dealt with them above anyway.) for suffix in ["_created", "_gsum", "_gcount"]: - if sample.name.endswith(suffix): + if s.name == metric.name + suffix: break else: output.append( - sample_line( - sample, sample.name.replace(":total", "").replace(":", "_") - ) + sample_line(s, s.name.replace(":total", "").replace(":", "_")) ) return "".join(output).encode("utf-8") diff --git a/synapse/metrics/background_process_metrics.py b/synapse/metrics/background_process_metrics.py
index a9269196b3..5b73463504 100644 --- a/synapse/metrics/background_process_metrics.py +++ b/synapse/metrics/background_process_metrics.py
@@ -13,16 +13,15 @@ # See the License for the specific language governing permissions and # limitations under the License. +import inspect import logging import threading -from asyncio import iscoroutine from functools import wraps from typing import TYPE_CHECKING, Dict, Optional, Set from prometheus_client.core import REGISTRY, Counter, Gauge from twisted.internet import defer -from twisted.python.failure import Failure from synapse.logging.context import LoggingContext, PreserveLoggingContext @@ -106,7 +105,7 @@ _background_processes_active_since_last_scrape = set() # type: Set[_BackgroundP _bg_metrics_lock = threading.Lock() -class _Collector(object): +class _Collector: """A custom metrics collector for the background process metrics. Ensures that all of the metrics are up-to-date with any in-flight processes @@ -141,7 +140,7 @@ class _Collector(object): REGISTRY.register(_Collector()) -class _BackgroundProcess(object): +class _BackgroundProcess: def __init__(self, desc, ctx): self.desc = desc self._context = ctx @@ -167,7 +166,7 @@ class _BackgroundProcess(object): ) -def run_as_background_process(desc, func, *args, **kwargs): +def run_as_background_process(desc: str, func, *args, **kwargs): """Run the given function in its own logcontext, with resource metrics This should be used to wrap processes which are fired off to run in the @@ -176,10 +175,10 @@ def run_as_background_process(desc, func, *args, **kwargs): It returns a Deferred which completes when the function completes, but it doesn't follow the synapse logcontext rules, which makes it appropriate for passing to clock.looping_call and friends (or for firing-and-forgetting in the middle of a - normal synapse inlineCallbacks function). + normal synapse async function). Args: - desc (str): a description for this background process type + desc: a description for this background process type func: a function, which may return a Deferred or a coroutine args: positional args for func kwargs: keyword args for func @@ -188,8 +187,7 @@ def run_as_background_process(desc, func, *args, **kwargs): follow the synapse logcontext rules. """ - @defer.inlineCallbacks - def run(): + async def run(): with _bg_metrics_lock: count = _background_process_counts.get(desc, 0) _background_process_counts[desc] = count + 1 @@ -203,29 +201,21 @@ def run_as_background_process(desc, func, *args, **kwargs): try: result = func(*args, **kwargs) - # We probably don't have an ensureDeferred in our call stack to handle - # coroutine results, so we need to ensureDeferred here. - # - # But we need this check because ensureDeferred doesn't like being - # called on immediate values (as opposed to Deferreds or coroutines). - if iscoroutine(result): - result = defer.ensureDeferred(result) + if inspect.isawaitable(result): + result = await result - return (yield result) + return result except Exception: - # failure.Failure() fishes the original Failure out of our stack, and - # thus gives us a sensible stack trace. - f = Failure() - logger.error( - "Background process '%s' threw an exception", - desc, - exc_info=(f.type, f.value, f.getTracebackObject()), + logger.exception( + "Background process '%s' threw an exception", desc, ) finally: _background_process_in_flight_count.labels(desc).dec() with PreserveLoggingContext(): - return run() + # Note that we return a Deferred here so that it can be used in a + # looping_call and other places that expect a Deferred. + return defer.ensureDeferred(run()) def wrap_as_background_process(desc): diff --git a/synapse/module_api/__init__.py b/synapse/module_api/__init__.py
index a885076282..16ca86b7bc 100644 --- a/synapse/module_api/__init__.py +++ b/synapse/module_api/__init__.py
@@ -81,7 +81,7 @@ class PublicRoomListManager: await self._store.set_room_is_public(room_id, False) -class ModuleApi(object): +class ModuleApi: """A proxy object that gets passed to various plugin modules so they can register new users etc if necessary. """ @@ -220,8 +220,10 @@ class ModuleApi(object): external_id: id on that system user_id: complete mxid that it is mapped to """ - return self._store.record_user_external_id( - auth_provider_id, remote_user_id, registered_user_id + return defer.ensureDeferred( + self._store.record_user_external_id( + auth_provider_id, remote_user_id, registered_user_id + ) ) def generate_short_term_login_token( @@ -247,12 +249,16 @@ class ModuleApi(object): synapse.api.errors.AuthError: the access token is invalid """ # see if the access token corresponds to a device - user_info = yield self._auth.get_user_by_access_token(access_token) + user_info = yield defer.ensureDeferred( + self._auth.get_user_by_access_token(access_token) + ) device_id = user_info.get("device_id") user_id = user_info["user"].to_string() if device_id: # delete the device, which will also delete its access tokens - yield self._hs.get_device_handler().delete_device(user_id, device_id) + yield defer.ensureDeferred( + self._hs.get_device_handler().delete_device(user_id, device_id) + ) else: # no associated device. Just delete the access token. yield defer.ensureDeferred( @@ -272,7 +278,9 @@ class ModuleApi(object): Returns: Deferred[object]: result of func """ - return self._store.db.runInteraction(desc, func, *args, **kwargs) + return defer.ensureDeferred( + self._store.db_pool.runInteraction(desc, func, *args, **kwargs) + ) def complete_sso_login( self, registered_user_id: str, request: SynapseRequest, client_redirect_url: str diff --git a/synapse/notifier.py b/synapse/notifier.py
index bd41f77852..59415f6f88 100644 --- a/synapse/notifier.py +++ b/synapse/notifier.py
@@ -15,7 +15,18 @@ import logging from collections import namedtuple -from typing import Callable, Iterable, List, TypeVar +from typing import ( + Awaitable, + Callable, + Dict, + Iterable, + List, + Optional, + Set, + Tuple, + TypeVar, + Union, +) from prometheus_client import Counter @@ -24,12 +35,20 @@ from twisted.internet import defer import synapse.server from synapse.api.constants import EventTypes, Membership from synapse.api.errors import AuthError +from synapse.events import EventBase from synapse.handlers.presence import format_user_presence_state from synapse.logging.context import PreserveLoggingContext from synapse.logging.utils import log_function from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.types import StreamToken +from synapse.streams.config import PaginationConfig +from synapse.types import ( + Collection, + PersistedEventPosition, + RoomStreamToken, + StreamToken, + UserID, +) from synapse.util.async_helpers import ObservableDeferred, timeout_deferred from synapse.util.metrics import Measure from synapse.visibility import filter_events_for_client @@ -55,7 +74,7 @@ def count(func: Callable[[T], bool], it: Iterable[T]) -> int: return n -class _NotificationListener(object): +class _NotificationListener: """ This represents a single client connection to the events stream. The events stream handler will have yielded to the deferred, so to notify the handler it is sufficient to resolve the deferred. @@ -67,7 +86,7 @@ class _NotificationListener(object): self.deferred = deferred -class _NotifierUserStream(object): +class _NotifierUserStream: """This represents a user connected to the event stream. It tracks the most recent stream token for that user. At a given point a user may have a number of streams listening for @@ -77,7 +96,13 @@ class _NotifierUserStream(object): so that it can remove itself from the indexes in the Notifier class. """ - def __init__(self, user_id, rooms, current_token, time_now_ms): + def __init__( + self, + user_id: str, + rooms: Collection[str], + current_token: StreamToken, + time_now_ms: int, + ): self.user_id = user_id self.rooms = set(rooms) self.current_token = current_token @@ -93,13 +118,15 @@ class _NotifierUserStream(object): with PreserveLoggingContext(): self.notify_deferred = ObservableDeferred(defer.Deferred()) - def notify(self, stream_key, stream_id, time_now_ms): + def notify( + self, stream_key: str, stream_id: Union[int, RoomStreamToken], time_now_ms: int, + ): """Notify any listeners for this user of a new event from an event source. Args: - stream_key(str): The stream the event came from. - stream_id(str): The new id for the stream the event came from. - time_now_ms(int): The current time in milliseconds. + stream_key: The stream the event came from. + stream_id: The new id for the stream the event came from. + time_now_ms: The current time in milliseconds. """ self.current_token = self.current_token.copy_and_advance(stream_key, stream_id) self.last_notified_token = self.current_token @@ -112,7 +139,7 @@ class _NotifierUserStream(object): self.notify_deferred = ObservableDeferred(defer.Deferred()) noify_deferred.callback(self.current_token) - def remove(self, notifier): + def remove(self, notifier: "Notifier"): """ Remove this listener from all the indexes in the Notifier it knows about. """ @@ -123,10 +150,10 @@ class _NotifierUserStream(object): notifier.user_to_user_stream.pop(self.user_id) - def count_listeners(self): + def count_listeners(self) -> int: return len(self.notify_deferred.observers()) - def new_listener(self, token): + def new_listener(self, token: StreamToken) -> _NotificationListener: """Returns a deferred that is resolved when there is a new token greater than the given token. @@ -136,20 +163,18 @@ class _NotifierUserStream(object): """ # Immediately wake up stream if something has already since happened # since their last token. - if self.last_notified_token.is_after(token): + if self.last_notified_token != token: return _NotificationListener(defer.succeed(self.current_token)) else: return _NotificationListener(self.notify_deferred.observe()) class EventStreamResult(namedtuple("EventStreamResult", ("events", "tokens"))): - def __nonzero__(self): + def __bool__(self): return bool(self.events) - __bool__ = __nonzero__ # python3 - -class Notifier(object): +class Notifier: """ This class is responsible for notifying any listeners when there are new events available for it. @@ -159,14 +184,16 @@ class Notifier(object): UNUSED_STREAM_EXPIRY_MS = 10 * 60 * 1000 def __init__(self, hs: "synapse.server.HomeServer"): - self.user_to_user_stream = {} - self.room_to_user_streams = {} + self.user_to_user_stream = {} # type: Dict[str, _NotifierUserStream] + self.room_to_user_streams = {} # type: Dict[str, Set[_NotifierUserStream]] self.hs = hs self.storage = hs.get_storage() self.event_sources = hs.get_event_sources() self.store = hs.get_datastore() - self.pending_new_room_events = [] + self.pending_new_room_events = ( + [] + ) # type: List[Tuple[PersistedEventPosition, EventBase, Collection[UserID]]] # Called when there are new things to stream over replication self.replication_callbacks = [] # type: List[Callable[[], None]] @@ -177,11 +204,11 @@ class Notifier(object): self.clock = hs.get_clock() self.appservice_handler = hs.get_application_service_handler() + self._pusher_pool = hs.get_pusherpool() + self.federation_sender = None if hs.should_send_federation(): self.federation_sender = hs.get_federation_sender() - else: - self.federation_sender = None self.state_handler = hs.get_state_handler() @@ -193,12 +220,12 @@ class Notifier(object): # when rendering the metrics page, which is likely once per minute at # most when scraping it. def count_listeners(): - all_user_streams = set() + all_user_streams = set() # type: Set[_NotifierUserStream] - for x in list(self.room_to_user_streams.values()): - all_user_streams |= x - for x in list(self.user_to_user_stream.values()): - all_user_streams.add(x) + for streams in list(self.room_to_user_streams.values()): + all_user_streams |= streams + for stream in list(self.user_to_user_stream.values()): + all_user_streams.add(stream) return sum(stream.count_listeners() for stream in all_user_streams) @@ -223,7 +250,11 @@ class Notifier(object): self.replication_callbacks.append(cb) def on_new_room_event( - self, event, room_stream_id, max_room_stream_id, extra_users=[] + self, + event: EventBase, + event_pos: PersistedEventPosition, + max_room_stream_token: RoomStreamToken, + extra_users: Collection[UserID] = [], ): """ Used by handlers to inform the notifier something has happened in the room, room event wise. @@ -236,52 +267,81 @@ class Notifier(object): until all previous events have been persisted before notifying the client streams. """ - self.pending_new_room_events.append((room_stream_id, event, extra_users)) - self._notify_pending_new_room_events(max_room_stream_id) + self.pending_new_room_events.append((event_pos, event, extra_users)) + self._notify_pending_new_room_events(max_room_stream_token) self.notify_replication() - def _notify_pending_new_room_events(self, max_room_stream_id): + def _notify_pending_new_room_events(self, max_room_stream_token: RoomStreamToken): """Notify for the room events that were queued waiting for a previous event to be persisted. Args: - max_room_stream_id(int): The highest stream_id below which all + max_room_stream_token: The highest stream_id below which all events have been persisted. """ pending = self.pending_new_room_events self.pending_new_room_events = [] - for room_stream_id, event, extra_users in pending: - if room_stream_id > max_room_stream_id: - self.pending_new_room_events.append( - (room_stream_id, event, extra_users) - ) + + users = set() # type: Set[UserID] + rooms = set() # type: Set[str] + + for event_pos, event, extra_users in pending: + if event_pos.persisted_after(max_room_stream_token): + self.pending_new_room_events.append((event_pos, event, extra_users)) else: - self._on_new_room_event(event, room_stream_id, extra_users) + if ( + event.type == EventTypes.Member + and event.membership == Membership.JOIN + ): + self._user_joined_room(event.state_key, event.room_id) + + users.update(extra_users) + rooms.add(event.room_id) + + if users or rooms: + self.on_new_event( + "room_key", max_room_stream_token, users=users, rooms=rooms, + ) + self._on_updated_room_token(max_room_stream_token) + + def _on_updated_room_token(self, max_room_stream_token: RoomStreamToken): + """Poke services that might care that the room position has been + updated. + """ - def _on_new_room_event(self, event, room_stream_id, extra_users=[]): - """Notify any user streams that are interested in this room event""" # poke any interested application service. run_as_background_process( - "notify_app_services", self._notify_app_services, room_stream_id + "_notify_app_services", self._notify_app_services, max_room_stream_token ) - if self.federation_sender: - self.federation_sender.notify_new_events(room_stream_id) - - if event.type == EventTypes.Member and event.membership == Membership.JOIN: - self._user_joined_room(event.state_key, event.room_id) - - self.on_new_event( - "room_key", room_stream_id, users=extra_users, rooms=[event.room_id] + run_as_background_process( + "_notify_pusher_pool", self._notify_pusher_pool, max_room_stream_token ) - async def _notify_app_services(self, room_stream_id): + if self.federation_sender: + self.federation_sender.notify_new_events(max_room_stream_token.stream) + + async def _notify_app_services(self, max_room_stream_token: RoomStreamToken): try: - await self.appservice_handler.notify_interested_services(room_stream_id) + await self.appservice_handler.notify_interested_services( + max_room_stream_token.stream + ) except Exception: logger.exception("Error notifying application services of event") - def on_new_event(self, stream_key, new_token, users=[], rooms=[]): + async def _notify_pusher_pool(self, max_room_stream_token: RoomStreamToken): + try: + await self._pusher_pool.on_new_notifications(max_room_stream_token.stream) + except Exception: + logger.exception("Error pusher pool of event") + + def on_new_event( + self, + stream_key: str, + new_token: Union[int, RoomStreamToken], + users: Collection[UserID] = [], + rooms: Collection[str] = [], + ): """ Used to inform listeners that something has happened event wise. Will wake up all listeners for the given users and rooms. @@ -307,20 +367,25 @@ class Notifier(object): self.notify_replication() - def on_new_replication_data(self): + def on_new_replication_data(self) -> None: """Used to inform replication listeners that something has happend without waking up any of the normal user event streams""" self.notify_replication() async def wait_for_events( - self, user_id, timeout, callback, room_ids=None, from_token=StreamToken.START - ): + self, + user_id: str, + timeout: int, + callback: Callable[[StreamToken, StreamToken], Awaitable[T]], + room_ids=None, + from_token=StreamToken.START, + ) -> T: """Wait until the callback returns a non empty response or the timeout fires. """ user_stream = self.user_to_user_stream.get(user_id) if user_stream is None: - current_token = await self.event_sources.get_current_token() + current_token = self.event_sources.get_current_token() if room_ids is None: room_ids = await self.store.get_rooms_for_user(user_id) user_stream = _NotifierUserStream( @@ -377,38 +442,38 @@ class Notifier(object): async def get_events_for( self, - user, - pagination_config, - timeout, - only_keys=None, - is_guest=False, - explicit_room_id=None, - ): + user: UserID, + pagination_config: PaginationConfig, + timeout: int, + is_guest: bool = False, + explicit_room_id: str = None, + ) -> EventStreamResult: """ For the given user and rooms, return any new events for them. If there are no new events wait for up to `timeout` milliseconds for any new events to happen before returning. - If `only_keys` is not None, events from keys will be sent down. - If explicit_room_id is not set, the user's joined rooms will be polled for events. If explicit_room_id is set, that room will be polled for events only if it is world readable or the user has joined the room. """ - from_token = pagination_config.from_token - if not from_token: - from_token = await self.event_sources.get_current_token() + if pagination_config.from_token: + from_token = pagination_config.from_token + else: + from_token = self.event_sources.get_current_token() limit = pagination_config.limit room_ids, is_joined = await self._get_room_ids(user, explicit_room_id) is_peeking = not is_joined - async def check_for_updates(before_token, after_token): - if not after_token.is_after(before_token): + async def check_for_updates( + before_token: StreamToken, after_token: StreamToken + ) -> EventStreamResult: + if after_token == before_token: return EventStreamResult([], (from_token, from_token)) - events = [] + events = [] # type: List[EventBase] end_token = from_token for name, source in self.event_sources.sources.items(): @@ -417,8 +482,6 @@ class Notifier(object): after_id = getattr(after_token, keyname) if before_id == after_id: continue - if only_keys and name not in only_keys: - continue new_events, new_key = await source.get_new_events( user=user, @@ -476,7 +539,9 @@ class Notifier(object): return result - async def _get_room_ids(self, user, explicit_room_id): + async def _get_room_ids( + self, user: UserID, explicit_room_id: Optional[str] + ) -> Tuple[Collection[str], bool]: joined_room_ids = await self.store.get_rooms_for_user(user.to_string()) if explicit_room_id: if explicit_room_id in joined_room_ids: @@ -486,7 +551,7 @@ class Notifier(object): raise AuthError(403, "Non-joined access not allowed") return joined_room_ids, True - async def _is_world_readable(self, room_id): + async def _is_world_readable(self, room_id: str) -> bool: state = await self.state_handler.get_current_state( room_id, EventTypes.RoomHistoryVisibility, "" ) @@ -496,7 +561,7 @@ class Notifier(object): return False @log_function - def remove_expired_streams(self): + def remove_expired_streams(self) -> None: time_now_ms = self.clock.time_msec() expired_streams = [] expire_before_ts = time_now_ms - self.UNUSED_STREAM_EXPIRY_MS @@ -510,21 +575,21 @@ class Notifier(object): expired_stream.remove(self) @log_function - def _register_with_keys(self, user_stream): + def _register_with_keys(self, user_stream: _NotifierUserStream): self.user_to_user_stream[user_stream.user_id] = user_stream for room in user_stream.rooms: s = self.room_to_user_streams.setdefault(room, set()) s.add(user_stream) - def _user_joined_room(self, user_id, room_id): + def _user_joined_room(self, user_id: str, room_id: str): new_user_stream = self.user_to_user_stream.get(user_id) if new_user_stream is not None: room_streams = self.room_to_user_streams.setdefault(room_id, set()) room_streams.add(new_user_stream) new_user_stream.rooms.add(room_id) - def notify_replication(self): + def notify_replication(self) -> None: """Notify the any replication listeners that there's a new event""" for cb in self.replication_callbacks: cb() diff --git a/synapse/push/__init__.py b/synapse/push/__init__.py
index edf45dc599..5a437f9810 100644 --- a/synapse/push/__init__.py +++ b/synapse/push/__init__.py
@@ -16,4 +16,4 @@ class PusherConfigException(Exception): def __init__(self, msg): - super(PusherConfigException, self).__init__(msg) + super().__init__(msg) diff --git a/synapse/push/action_generator.py b/synapse/push/action_generator.py
index 1ffd5e2df3..fabc9ba126 100644 --- a/synapse/push/action_generator.py +++ b/synapse/push/action_generator.py
@@ -15,8 +15,6 @@ import logging -from twisted.internet import defer - from synapse.util.metrics import Measure from .bulk_push_rule_evaluator import BulkPushRuleEvaluator @@ -24,7 +22,7 @@ from .bulk_push_rule_evaluator import BulkPushRuleEvaluator logger = logging.getLogger(__name__) -class ActionGenerator(object): +class ActionGenerator: def __init__(self, hs): self.hs = hs self.clock = hs.get_clock() @@ -37,7 +35,6 @@ class ActionGenerator(object): # event stream, so we just run the rules for a client with no profile # tag (ie. we just need all the users). - @defer.inlineCallbacks - def handle_push_actions_for_event(self, event, context): + async def handle_push_actions_for_event(self, event, context): with Measure(self.clock, "action_for_event_by_user"): - yield self.bulk_evaluator.action_for_event_by_user(event, context) + await self.bulk_evaluator.action_for_event_by_user(event, context) diff --git a/synapse/push/baserules.py b/synapse/push/baserules.py
index 204d2d83f6..172af1a5a4 100644 --- a/synapse/push/baserules.py +++ b/synapse/push/baserules.py
@@ -19,11 +19,13 @@ import copy from synapse.push.rulekinds import PRIORITY_CLASS_INVERSE_MAP, PRIORITY_CLASS_MAP -def list_with_base_rules(rawrules): +def list_with_base_rules(rawrules, use_new_defaults=False): """Combine the list of rules set by the user with the default push rules Args: rawrules(list): The rules the user has modified or set. + use_new_defaults(bool): Whether to use the new experimental default rules when + appending or prepending default rules. Returns: A new list with the rules set by the user combined with the defaults. @@ -43,7 +45,9 @@ def list_with_base_rules(rawrules): ruleslist.extend( make_base_prepend_rules( - PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules + PRIORITY_CLASS_INVERSE_MAP[current_prio_class], + modified_base_rules, + use_new_defaults, ) ) @@ -54,6 +58,7 @@ def list_with_base_rules(rawrules): make_base_append_rules( PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules, + use_new_defaults, ) ) current_prio_class -= 1 @@ -62,6 +67,7 @@ def list_with_base_rules(rawrules): make_base_prepend_rules( PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules, + use_new_defaults, ) ) @@ -70,27 +76,39 @@ def list_with_base_rules(rawrules): while current_prio_class > 0: ruleslist.extend( make_base_append_rules( - PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules + PRIORITY_CLASS_INVERSE_MAP[current_prio_class], + modified_base_rules, + use_new_defaults, ) ) current_prio_class -= 1 if current_prio_class > 0: ruleslist.extend( make_base_prepend_rules( - PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules + PRIORITY_CLASS_INVERSE_MAP[current_prio_class], + modified_base_rules, + use_new_defaults, ) ) return ruleslist -def make_base_append_rules(kind, modified_base_rules): +def make_base_append_rules(kind, modified_base_rules, use_new_defaults=False): rules = [] if kind == "override": - rules = BASE_APPEND_OVERRIDE_RULES + rules = ( + NEW_APPEND_OVERRIDE_RULES + if use_new_defaults + else BASE_APPEND_OVERRIDE_RULES + ) elif kind == "underride": - rules = BASE_APPEND_UNDERRIDE_RULES + rules = ( + NEW_APPEND_UNDERRIDE_RULES + if use_new_defaults + else BASE_APPEND_UNDERRIDE_RULES + ) elif kind == "content": rules = BASE_APPEND_CONTENT_RULES @@ -105,7 +123,7 @@ def make_base_append_rules(kind, modified_base_rules): return rules -def make_base_prepend_rules(kind, modified_base_rules): +def make_base_prepend_rules(kind, modified_base_rules, use_new_defaults=False): rules = [] if kind == "override": @@ -270,6 +288,135 @@ BASE_APPEND_OVERRIDE_RULES = [ ] +NEW_APPEND_OVERRIDE_RULES = [ + { + "rule_id": "global/override/.m.rule.encrypted", + "conditions": [ + { + "kind": "event_match", + "key": "type", + "pattern": "m.room.encrypted", + "_id": "_encrypted", + } + ], + "actions": ["notify"], + }, + { + "rule_id": "global/override/.m.rule.suppress_notices", + "conditions": [ + { + "kind": "event_match", + "key": "type", + "pattern": "m.room.message", + "_id": "_suppress_notices_type", + }, + { + "kind": "event_match", + "key": "content.msgtype", + "pattern": "m.notice", + "_id": "_suppress_notices", + }, + ], + "actions": [], + }, + { + "rule_id": "global/underride/.m.rule.suppress_edits", + "conditions": [ + { + "kind": "event_match", + "key": "m.relates_to.m.rel_type", + "pattern": "m.replace", + "_id": "_suppress_edits", + } + ], + "actions": [], + }, + { + "rule_id": "global/override/.m.rule.invite_for_me", + "conditions": [ + { + "kind": "event_match", + "key": "type", + "pattern": "m.room.member", + "_id": "_member", + }, + { + "kind": "event_match", + "key": "content.membership", + "pattern": "invite", + "_id": "_invite_member", + }, + {"kind": "event_match", "key": "state_key", "pattern_type": "user_id"}, + ], + "actions": ["notify", {"set_tweak": "sound", "value": "default"}], + }, + { + "rule_id": "global/override/.m.rule.contains_display_name", + "conditions": [{"kind": "contains_display_name"}], + "actions": [ + "notify", + {"set_tweak": "sound", "value": "default"}, + {"set_tweak": "highlight"}, + ], + }, + { + "rule_id": "global/override/.m.rule.tombstone", + "conditions": [ + { + "kind": "event_match", + "key": "type", + "pattern": "m.room.tombstone", + "_id": "_tombstone", + }, + { + "kind": "event_match", + "key": "state_key", + "pattern": "", + "_id": "_tombstone_statekey", + }, + ], + "actions": [ + "notify", + {"set_tweak": "sound", "value": "default"}, + {"set_tweak": "highlight"}, + ], + }, + { + "rule_id": "global/override/.m.rule.roomnotif", + "conditions": [ + { + "kind": "event_match", + "key": "content.body", + "pattern": "@room", + "_id": "_roomnotif_content", + }, + { + "kind": "sender_notification_permission", + "key": "room", + "_id": "_roomnotif_pl", + }, + ], + "actions": [ + "notify", + {"set_tweak": "highlight"}, + {"set_tweak": "sound", "value": "default"}, + ], + }, + { + "rule_id": "global/override/.m.rule.call", + "conditions": [ + { + "kind": "event_match", + "key": "type", + "pattern": "m.call.invite", + "_id": "_call", + } + ], + "actions": ["notify", {"set_tweak": "sound", "value": "ring"}], + }, +] + + BASE_APPEND_UNDERRIDE_RULES = [ { "rule_id": "global/underride/.m.rule.call", @@ -362,6 +509,36 @@ BASE_APPEND_UNDERRIDE_RULES = [ ] +NEW_APPEND_UNDERRIDE_RULES = [ + { + "rule_id": "global/underride/.m.rule.room_one_to_one", + "conditions": [ + {"kind": "room_member_count", "is": "2", "_id": "member_count"}, + { + "kind": "event_match", + "key": "content.body", + "pattern": "*", + "_id": "body", + }, + ], + "actions": ["notify", {"set_tweak": "sound", "value": "default"}], + }, + { + "rule_id": "global/underride/.m.rule.message", + "conditions": [ + { + "kind": "event_match", + "key": "content.body", + "pattern": "*", + "_id": "body", + }, + ], + "actions": ["notify"], + "enabled": False, + }, +] + + BASE_RULE_IDS = set() for r in BASE_APPEND_CONTENT_RULES: @@ -383,3 +560,26 @@ for r in BASE_APPEND_UNDERRIDE_RULES: r["priority_class"] = PRIORITY_CLASS_MAP["underride"] r["default"] = True BASE_RULE_IDS.add(r["rule_id"]) + + +NEW_RULE_IDS = set() + +for r in BASE_APPEND_CONTENT_RULES: + r["priority_class"] = PRIORITY_CLASS_MAP["content"] + r["default"] = True + NEW_RULE_IDS.add(r["rule_id"]) + +for r in BASE_PREPEND_OVERRIDE_RULES: + r["priority_class"] = PRIORITY_CLASS_MAP["override"] + r["default"] = True + NEW_RULE_IDS.add(r["rule_id"]) + +for r in NEW_APPEND_OVERRIDE_RULES: + r["priority_class"] = PRIORITY_CLASS_MAP["override"] + r["default"] = True + NEW_RULE_IDS.add(r["rule_id"]) + +for r in NEW_APPEND_UNDERRIDE_RULES: + r["priority_class"] = PRIORITY_CLASS_MAP["underride"] + r["default"] = True + NEW_RULE_IDS.add(r["rule_id"]) diff --git a/synapse/push/bulk_push_rule_evaluator.py b/synapse/push/bulk_push_rule_evaluator.py
index 472ddf9f7d..c440f2545c 100644 --- a/synapse/push/bulk_push_rule_evaluator.py +++ b/synapse/push/bulk_push_rule_evaluator.py
@@ -19,10 +19,10 @@ from collections import namedtuple from prometheus_client import Counter -from twisted.internet import defer - -from synapse.api.constants import EventTypes, Membership +from synapse.api.constants import EventTypes, Membership, RelationTypes from synapse.event_auth import get_user_power_level +from synapse.events import EventBase +from synapse.events.snapshot import EventContext from synapse.state import POWER_KEY from synapse.util.async_helpers import Linearizer from synapse.util.caches import register_cache @@ -53,7 +53,49 @@ push_rules_delta_state_cache_metric = register_cache( ) -class BulkPushRuleEvaluator(object): +STATE_EVENT_TYPES_TO_MARK_UNREAD = { + EventTypes.Topic, + EventTypes.Name, + EventTypes.RoomAvatar, + EventTypes.Tombstone, +} + + +def _should_count_as_unread(event: EventBase, context: EventContext) -> bool: + # Exclude rejected and soft-failed events. + if context.rejected or event.internal_metadata.is_soft_failed(): + return False + + # Exclude notices. + if ( + not event.is_state() + and event.type == EventTypes.Message + and event.content.get("msgtype") == "m.notice" + ): + return False + + # Exclude edits. + relates_to = event.content.get("m.relates_to", {}) + if relates_to.get("rel_type") == RelationTypes.REPLACE: + return False + + # Mark events that have a non-empty string body as unread. + body = event.content.get("body") + if isinstance(body, str) and body: + return True + + # Mark some state events as unread. + if event.is_state() and event.type in STATE_EVENT_TYPES_TO_MARK_UNREAD: + return True + + # Mark encrypted events as unread. + if not event.is_state() and event.type == EventTypes.Encrypted: + return True + + return False + + +class BulkPushRuleEvaluator: """Calculates the outcome of push rules for an event for all users in the room at once. """ @@ -70,8 +112,7 @@ class BulkPushRuleEvaluator(object): resizable=False, ) - @defer.inlineCallbacks - def _get_rules_for_event(self, event, context): + async def _get_rules_for_event(self, event, context): """This gets the rules for all users in the room at the time of the event, as well as the push rules for the invitee if the event is an invite. @@ -79,19 +120,19 @@ class BulkPushRuleEvaluator(object): dict of user_id -> push_rules """ room_id = event.room_id - rules_for_room = yield self._get_rules_for_room(room_id) + rules_for_room = await self._get_rules_for_room(room_id) - rules_by_user = yield rules_for_room.get_rules(event, context) + rules_by_user = await rules_for_room.get_rules(event, context) # if this event is an invite event, we may need to run rules for the user # who's been invited, otherwise they won't get told they've been invited if event.type == "m.room.member" and event.content["membership"] == "invite": invited = event.state_key if invited and self.hs.is_mine_id(invited): - has_pusher = yield self.store.user_has_pusher(invited) + has_pusher = await self.store.user_has_pusher(invited) if has_pusher: rules_by_user = dict(rules_by_user) - rules_by_user[invited] = yield self.store.get_push_rules_for_user( + rules_by_user[invited] = await self.store.get_push_rules_for_user( invited ) @@ -114,20 +155,19 @@ class BulkPushRuleEvaluator(object): self.room_push_rule_cache_metrics, ) - @defer.inlineCallbacks - def _get_power_levels_and_sender_level(self, event, context): - prev_state_ids = yield context.get_prev_state_ids() + async def _get_power_levels_and_sender_level(self, event, context): + prev_state_ids = await context.get_prev_state_ids() pl_event_id = prev_state_ids.get(POWER_KEY) if pl_event_id: # fastpath: if there's a power level event, that's all we need, and # not having a power level event is an extreme edge case - pl_event = yield self.store.get_event(pl_event_id) + pl_event = await self.store.get_event(pl_event_id) auth_events = {POWER_KEY: pl_event} else: - auth_events_ids = yield self.auth.compute_auth_events( + auth_events_ids = self.auth.compute_auth_events( event, prev_state_ids, for_verification=False ) - auth_events = yield self.store.get_events(auth_events_ids) + auth_events = await self.store.get_events(auth_events_ids) auth_events = {(e.type, e.state_key): e for e in auth_events.values()} sender_level = get_user_power_level(event.sender, auth_events) @@ -136,23 +176,22 @@ class BulkPushRuleEvaluator(object): return pl_event.content if pl_event else {}, sender_level - @defer.inlineCallbacks - def action_for_event_by_user(self, event, context): - """Given an event and context, evaluate the push rules and insert the - results into the event_push_actions_staging table. - - Returns: - Deferred + async def action_for_event_by_user(self, event, context) -> None: + """Given an event and context, evaluate the push rules, check if the message + should increment the unread count, and insert the results into the + event_push_actions_staging table. """ - rules_by_user = yield self._get_rules_for_event(event, context) + count_as_unread = _should_count_as_unread(event, context) + + rules_by_user = await self._get_rules_for_event(event, context) actions_by_user = {} - room_members = yield self.store.get_joined_users_from_context(event, context) + room_members = await self.store.get_joined_users_from_context(event, context) ( power_levels, sender_power_level, - ) = yield self._get_power_levels_and_sender_level(event, context) + ) = await self._get_power_levels_and_sender_level(event, context) evaluator = PushRuleEvaluatorForEvent( event, len(room_members), sender_power_level, power_levels @@ -165,7 +204,7 @@ class BulkPushRuleEvaluator(object): continue if not event.is_state(): - is_ignored = yield self.store.is_ignored_by(event.sender, uid) + is_ignored = await self.store.is_ignored_by(event.sender, uid) if is_ignored: continue @@ -180,6 +219,13 @@ class BulkPushRuleEvaluator(object): if event.type == EventTypes.Member and event.state_key == uid: display_name = event.content.get("displayname", None) + if count_as_unread: + # Add an element for the current user if the event needs to be marked as + # unread, so that add_push_actions_to_staging iterates over it. + # If the event shouldn't be marked as unread but should notify the + # current user, it'll be added to the dict later. + actions_by_user[uid] = [] + for rule in rules: if "enabled" in rule and not rule["enabled"]: continue @@ -197,7 +243,9 @@ class BulkPushRuleEvaluator(object): # Mark in the DB staging area the push actions for users who should be # notified for this event. (This will then get handled when we persist # the event) - yield self.store.add_push_actions_to_staging(event.event_id, actions_by_user) + await self.store.add_push_actions_to_staging( + event.event_id, actions_by_user, count_as_unread, + ) def _condition_checker(evaluator, conditions, uid, display_name, cache): @@ -220,7 +268,7 @@ def _condition_checker(evaluator, conditions, uid, display_name, cache): return True -class RulesForRoom(object): +class RulesForRoom: """Caches push rules for users in a room. This efficiently handles users joining/leaving the room by not invalidating @@ -274,8 +322,7 @@ class RulesForRoom(object): # to self around in the callback. self.invalidate_all_cb = _Invalidation(rules_for_room_cache, room_id) - @defer.inlineCallbacks - def get_rules(self, event, context): + async def get_rules(self, event, context): """Given an event context return the rules for all users who are currently in the room. """ @@ -286,7 +333,7 @@ class RulesForRoom(object): self.room_push_rule_cache_metrics.inc_hits() return self.rules_by_user - with (yield self.linearizer.queue(())): + with (await self.linearizer.queue(())): if state_group and self.state_group == state_group: logger.debug("Using cached rules for %r", self.room_id) self.room_push_rule_cache_metrics.inc_hits() @@ -304,9 +351,7 @@ class RulesForRoom(object): push_rules_delta_state_cache_metric.inc_hits() else: - current_state_ids = yield defer.ensureDeferred( - context.get_current_state_ids() - ) + current_state_ids = await context.get_current_state_ids() push_rules_delta_state_cache_metric.inc_misses() push_rules_state_size_counter.inc(len(current_state_ids)) @@ -353,7 +398,7 @@ class RulesForRoom(object): # If we have some memebr events we haven't seen, look them up # and fetch push rules for them if appropriate. logger.debug("Found new member events %r", missing_member_event_ids) - yield self._update_rules_with_member_event_ids( + await self._update_rules_with_member_event_ids( ret_rules_by_user, missing_member_event_ids, state_group, event ) else: @@ -371,8 +416,7 @@ class RulesForRoom(object): ) return ret_rules_by_user - @defer.inlineCallbacks - def _update_rules_with_member_event_ids( + async def _update_rules_with_member_event_ids( self, ret_rules_by_user, member_event_ids, state_group, event ): """Update the partially filled rules_by_user dict by fetching rules for @@ -381,14 +425,14 @@ class RulesForRoom(object): Args: ret_rules_by_user (dict): Partiallly filled dict of push rules. Gets updated with any new rules. - member_event_ids (list): List of event ids for membership events that - have happened since the last time we filled rules_by_user + member_event_ids (dict): Dict of user id to event id for membership events + that have happened since the last time we filled rules_by_user state_group: The state group we are currently computing push rules for. Used when updating the cache. """ sequence = self.sequence - rows = yield self.store.get_membership_from_event_ids(member_event_ids.values()) + rows = await self.store.get_membership_from_event_ids(member_event_ids.values()) members = {row["event_id"]: (row["user_id"], row["membership"]) for row in rows} @@ -402,36 +446,21 @@ class RulesForRoom(object): if logger.isEnabledFor(logging.DEBUG): logger.debug("Found members %r: %r", self.room_id, members.values()) - interested_in_user_ids = { + user_ids = { user_id for user_id, membership in members.values() if membership == Membership.JOIN } - logger.debug("Joined: %r", interested_in_user_ids) - - if_users_with_pushers = yield self.store.get_if_users_have_pushers( - interested_in_user_ids, on_invalidate=self.invalidate_all_cb - ) - - user_ids = { - uid for uid, have_pusher in if_users_with_pushers.items() if have_pusher - } - - logger.debug("With pushers: %r", user_ids) - - users_with_receipts = yield self.store.get_users_with_read_receipts_in_room( - self.room_id, on_invalidate=self.invalidate_all_cb - ) - - logger.debug("With receipts: %r", users_with_receipts) + logger.debug("Joined: %r", user_ids) - # any users with pushers must be ours: they have pushers - for uid in users_with_receipts: - if uid in interested_in_user_ids: - user_ids.add(uid) + # Previously we only considered users with pushers or read receipts in that + # room. We can't do this anymore because we use push actions to calculate unread + # counts, which don't rely on the user having pushers or sent a read receipt into + # the room. Therefore we just need to filter for local users here. + user_ids = list(filter(self.is_mine_id, user_ids)) - rules_by_user = yield self.store.bulk_get_push_rules( + rules_by_user = await self.store.bulk_get_push_rules( user_ids, on_invalidate=self.invalidate_all_cb ) diff --git a/synapse/push/emailpusher.py b/synapse/push/emailpusher.py
index 568c13eaea..28bd8ab748 100644 --- a/synapse/push/emailpusher.py +++ b/synapse/push/emailpusher.py
@@ -45,7 +45,7 @@ THROTTLE_RESET_AFTER_MS = 12 * 60 * 60 * 1000 INCLUDE_ALL_UNREAD_NOTIFS = False -class EmailPusher(object): +class EmailPusher: """ A pusher that sends email notifications about events (approximately) when they happen. @@ -91,7 +91,7 @@ class EmailPusher(object): pass self.timed_call = None - def on_new_notifications(self, min_stream_ordering, max_stream_ordering): + def on_new_notifications(self, max_stream_ordering): if self.max_stream_ordering: self.max_stream_ordering = max( max_stream_ordering, self.max_stream_ordering diff --git a/synapse/push/httppusher.py b/synapse/push/httppusher.py
index 2fac07593b..26706bf3e1 100644 --- a/synapse/push/httppusher.py +++ b/synapse/push/httppusher.py
@@ -17,7 +17,6 @@ import logging from prometheus_client import Counter -from twisted.internet import defer from twisted.internet.error import AlreadyCalled, AlreadyCancelled from synapse.api.constants import EventTypes @@ -50,7 +49,7 @@ http_badges_failed_counter = Counter( ) -class HttpPusher(object): +class HttpPusher: INITIAL_BACKOFF_SEC = 1 # in seconds because that's what Twisted takes MAX_BACKOFF_SEC = 60 * 60 @@ -115,7 +114,7 @@ class HttpPusher(object): if should_check_for_notifs: self._start_processing() - def on_new_notifications(self, min_stream_ordering, max_stream_ordering): + def on_new_notifications(self, max_stream_ordering): self.max_stream_ordering = max( max_stream_ordering, self.max_stream_ordering or 0 ) @@ -128,12 +127,11 @@ class HttpPusher(object): # but currently that's the only type of receipt anyway... run_as_background_process("http_pusher.on_new_receipts", self._update_badge) - @defer.inlineCallbacks - def _update_badge(self): + async def _update_badge(self): # XXX as per https://github.com/matrix-org/matrix-doc/issues/2627, this seems # to be largely redundant. perhaps we can remove it. - badge = yield push_tools.get_badge_count(self.hs.get_datastore(), self.user_id) - yield self._send_badge(badge) + badge = await push_tools.get_badge_count(self.hs.get_datastore(), self.user_id) + await self._send_badge(badge) def on_timer(self): self._start_processing() @@ -152,8 +150,7 @@ class HttpPusher(object): run_as_background_process("httppush.process", self._process) - @defer.inlineCallbacks - def _process(self): + async def _process(self): # we should never get here if we are already processing assert not self._is_processing @@ -164,7 +161,7 @@ class HttpPusher(object): while True: starting_max_ordering = self.max_stream_ordering try: - yield self._unsafe_process() + await self._unsafe_process() except Exception: logger.exception("Exception processing notifs") if self.max_stream_ordering == starting_max_ordering: @@ -172,8 +169,7 @@ class HttpPusher(object): finally: self._is_processing = False - @defer.inlineCallbacks - def _unsafe_process(self): + async def _unsafe_process(self): """ Looks for unset notifications and dispatch them, in order Never call this directly: use _process which will only allow this to @@ -181,7 +177,7 @@ class HttpPusher(object): """ fn = self.store.get_unread_push_actions_for_user_in_range_for_http - unprocessed = yield fn( + unprocessed = await fn( self.user_id, self.last_stream_ordering, self.max_stream_ordering ) @@ -203,13 +199,13 @@ class HttpPusher(object): "app_display_name": self.app_display_name, }, ): - processed = yield self._process_one(push_action) + processed = await self._process_one(push_action) if processed: http_push_processed_counter.inc() self.backoff_delay = HttpPusher.INITIAL_BACKOFF_SEC self.last_stream_ordering = push_action["stream_ordering"] - pusher_still_exists = yield self.store.update_pusher_last_stream_ordering_and_success( + pusher_still_exists = await self.store.update_pusher_last_stream_ordering_and_success( self.app_id, self.pushkey, self.user_id, @@ -224,14 +220,14 @@ class HttpPusher(object): if self.failing_since: self.failing_since = None - yield self.store.update_pusher_failing_since( + await self.store.update_pusher_failing_since( self.app_id, self.pushkey, self.user_id, self.failing_since ) else: http_push_failed_counter.inc() if not self.failing_since: self.failing_since = self.clock.time_msec() - yield self.store.update_pusher_failing_since( + await self.store.update_pusher_failing_since( self.app_id, self.pushkey, self.user_id, self.failing_since ) @@ -250,7 +246,7 @@ class HttpPusher(object): ) self.backoff_delay = HttpPusher.INITIAL_BACKOFF_SEC self.last_stream_ordering = push_action["stream_ordering"] - pusher_still_exists = yield self.store.update_pusher_last_stream_ordering( + pusher_still_exists = await self.store.update_pusher_last_stream_ordering( self.app_id, self.pushkey, self.user_id, @@ -263,7 +259,7 @@ class HttpPusher(object): return self.failing_since = None - yield self.store.update_pusher_failing_since( + await self.store.update_pusher_failing_since( self.app_id, self.pushkey, self.user_id, self.failing_since ) else: @@ -276,18 +272,17 @@ class HttpPusher(object): ) break - @defer.inlineCallbacks - def _process_one(self, push_action): + async def _process_one(self, push_action): if "notify" not in push_action["actions"]: return True tweaks = push_rule_evaluator.tweaks_for_actions(push_action["actions"]) - badge = yield push_tools.get_badge_count(self.hs.get_datastore(), self.user_id) + badge = await push_tools.get_badge_count(self.hs.get_datastore(), self.user_id) - event = yield self.store.get_event(push_action["event_id"], allow_none=True) + event = await self.store.get_event(push_action["event_id"], allow_none=True) if event is None: return True # It's been redacted - rejected = yield self.dispatch_push(event, tweaks, badge) + rejected = await self.dispatch_push(event, tweaks, badge) if rejected is False: return False @@ -301,11 +296,10 @@ class HttpPusher(object): ) else: logger.info("Pushkey %s was rejected: removing", pk) - yield self.hs.remove_pusher(self.app_id, pk, self.user_id) + await self.hs.remove_pusher(self.app_id, pk, self.user_id) return True - @defer.inlineCallbacks - def _build_notification_dict(self, event, tweaks, badge): + async def _build_notification_dict(self, event, tweaks, badge): priority = "low" if ( event.type == EventTypes.Encrypted @@ -335,7 +329,7 @@ class HttpPusher(object): } return d - ctx = yield push_tools.get_context_for_event( + ctx = await push_tools.get_context_for_event( self.storage, self.state_handler, event, self.user_id ) @@ -377,13 +371,12 @@ class HttpPusher(object): return d - @defer.inlineCallbacks - def dispatch_push(self, event, tweaks, badge): - notification_dict = yield self._build_notification_dict(event, tweaks, badge) + async def dispatch_push(self, event, tweaks, badge): + notification_dict = await self._build_notification_dict(event, tweaks, badge) if not notification_dict: return [] try: - resp = yield self.http_client.post_json_get_json( + resp = await self.http_client.post_json_get_json( self.url, notification_dict ) except Exception as e: @@ -400,8 +393,7 @@ class HttpPusher(object): rejected = resp["rejected"] return rejected - @defer.inlineCallbacks - def _send_badge(self, badge): + async def _send_badge(self, badge): """ Args: badge (int): number of unread messages @@ -424,7 +416,7 @@ class HttpPusher(object): } } try: - yield self.http_client.post_json_get_json(self.url, d) + await self.http_client.post_json_get_json(self.url, d) http_badges_processed_counter.inc() except Exception as e: logger.warning( diff --git a/synapse/push/mailer.py b/synapse/push/mailer.py
index af117fddf9..455a1acb46 100644 --- a/synapse/push/mailer.py +++ b/synapse/push/mailer.py
@@ -16,8 +16,7 @@ import email.mime.multipart import email.utils import logging -import time -import urllib +import urllib.parse from email.mime.multipart import MIMEMultipart from email.mime.text import MIMEText from typing import Iterable, List, TypeVar @@ -93,7 +92,7 @@ ALLOWED_ATTRS = { # ALLOWED_SCHEMES = ["http", "https", "ftp", "mailto"] -class Mailer(object): +class Mailer: def __init__(self, hs, app_name, template_html, template_text): self.hs = hs self.template_html = template_html @@ -124,7 +123,7 @@ class Mailer(object): params = {"token": token, "client_secret": client_secret, "sid": sid} link = ( self.hs.config.public_baseurl - + "_matrix/client/unstable/password_reset/email/submit_token?%s" + + "_synapse/client/password_reset/email/submit_token?%s" % urllib.parse.urlencode(params) ) @@ -640,72 +639,3 @@ def string_ordinal_total(s): for c in s: tot += ord(c) return tot - - -def format_ts_filter(value, format): - return time.strftime(format, time.localtime(value / 1000)) - - -def load_jinja2_templates( - template_dir, - template_filenames, - apply_format_ts_filter=False, - apply_mxc_to_http_filter=False, - public_baseurl=None, -): - """Loads and returns one or more jinja2 templates and applies optional filters - - Args: - template_dir (str): The directory where templates are stored - template_filenames (list[str]): A list of template filenames - apply_format_ts_filter (bool): Whether to apply a template filter that formats - timestamps - apply_mxc_to_http_filter (bool): Whether to apply a template filter that converts - mxc urls to http urls - public_baseurl (str|None): The public baseurl of the server. Required for - apply_mxc_to_http_filter to be enabled - - Returns: - A list of jinja2 templates corresponding to the given list of filenames, - with order preserved - """ - logger.info( - "loading email templates %s from '%s'", template_filenames, template_dir - ) - loader = jinja2.FileSystemLoader(template_dir) - env = jinja2.Environment(loader=loader) - - if apply_format_ts_filter: - env.filters["format_ts"] = format_ts_filter - - if apply_mxc_to_http_filter and public_baseurl: - env.filters["mxc_to_http"] = _create_mxc_to_http_filter(public_baseurl) - - templates = [] - for template_filename in template_filenames: - template = env.get_template(template_filename) - templates.append(template) - - return templates - - -def _create_mxc_to_http_filter(public_baseurl): - def mxc_to_http_filter(value, width, height, resize_method="crop"): - if value[0:6] != "mxc://": - return "" - - serverAndMediaId = value[6:] - fragment = None - if "#" in serverAndMediaId: - (serverAndMediaId, fragment) = serverAndMediaId.split("#", 1) - fragment = "#" + fragment - - params = {"width": width, "height": height, "method": resize_method} - return "%s_matrix/media/v1/thumbnail/%s?%s%s" % ( - public_baseurl, - serverAndMediaId, - urllib.parse.urlencode(params), - fragment or "", - ) - - return mxc_to_http_filter diff --git a/synapse/push/presentable_names.py b/synapse/push/presentable_names.py
index 0644a13cfc..d8f4a453cd 100644 --- a/synapse/push/presentable_names.py +++ b/synapse/push/presentable_names.py
@@ -16,8 +16,6 @@ import logging import re -from twisted.internet import defer - from synapse.api.constants import EventTypes logger = logging.getLogger(__name__) @@ -29,8 +27,7 @@ ALIAS_RE = re.compile(r"^#.*:.+$") ALL_ALONE = "Empty Room" -@defer.inlineCallbacks -def calculate_room_name( +async def calculate_room_name( store, room_state_ids, user_id, @@ -53,7 +50,7 @@ def calculate_room_name( """ # does it have a name? if (EventTypes.Name, "") in room_state_ids: - m_room_name = yield store.get_event( + m_room_name = await store.get_event( room_state_ids[(EventTypes.Name, "")], allow_none=True ) if m_room_name and m_room_name.content and m_room_name.content["name"]: @@ -61,7 +58,7 @@ def calculate_room_name( # does it have a canonical alias? if (EventTypes.CanonicalAlias, "") in room_state_ids: - canon_alias = yield store.get_event( + canon_alias = await store.get_event( room_state_ids[(EventTypes.CanonicalAlias, "")], allow_none=True ) if ( @@ -81,7 +78,7 @@ def calculate_room_name( my_member_event = None if (EventTypes.Member, user_id) in room_state_ids: - my_member_event = yield store.get_event( + my_member_event = await store.get_event( room_state_ids[(EventTypes.Member, user_id)], allow_none=True ) @@ -90,7 +87,7 @@ def calculate_room_name( and my_member_event.content["membership"] == "invite" ): if (EventTypes.Member, my_member_event.sender) in room_state_ids: - inviter_member_event = yield store.get_event( + inviter_member_event = await store.get_event( room_state_ids[(EventTypes.Member, my_member_event.sender)], allow_none=True, ) @@ -107,7 +104,7 @@ def calculate_room_name( # we're going to have to generate a name based on who's in the room, # so find out who is in the room that isn't the user. if EventTypes.Member in room_state_bytype_ids: - member_events = yield store.get_events( + member_events = await store.get_events( list(room_state_bytype_ids[EventTypes.Member].values()) ) all_members = [ diff --git a/synapse/push/push_rule_evaluator.py b/synapse/push/push_rule_evaluator.py
index 2d79ada189..709ace01e5 100644 --- a/synapse/push/push_rule_evaluator.py +++ b/synapse/push/push_rule_evaluator.py
@@ -105,7 +105,7 @@ def tweaks_for_actions(actions: List[Union[str, Dict]]) -> Dict[str, Any]: return tweaks -class PushRuleEvaluatorForEvent(object): +class PushRuleEvaluatorForEvent: def __init__( self, event: EventBase, diff --git a/synapse/push/push_tools.py b/synapse/push/push_tools.py
index 5dae4648c0..d0145666bf 100644 --- a/synapse/push/push_tools.py +++ b/synapse/push/push_tools.py
@@ -13,18 +13,15 @@ # See the License for the specific language governing permissions and # limitations under the License. -from twisted.internet import defer - from synapse.push.presentable_names import calculate_room_name, name_from_member_event from synapse.storage import Storage -@defer.inlineCallbacks -def get_badge_count(store, user_id): - invites = yield store.get_invited_rooms_for_local_user(user_id) - joins = yield store.get_rooms_for_user(user_id) +async def get_badge_count(store, user_id): + invites = await store.get_invited_rooms_for_local_user(user_id) + joins = await store.get_rooms_for_user(user_id) - my_receipts_by_room = yield store.get_receipts_for_user(user_id, "m.read") + my_receipts_by_room = await store.get_receipts_for_user(user_id, "m.read") badge = len(invites) @@ -32,7 +29,7 @@ def get_badge_count(store, user_id): if room_id in my_receipts_by_room: last_unread_event_id = my_receipts_by_room[room_id] - notifs = yield ( + notifs = await ( store.get_unread_event_push_actions_by_room_for_user( room_id, user_id, last_unread_event_id ) @@ -43,23 +40,22 @@ def get_badge_count(store, user_id): return badge -@defer.inlineCallbacks -def get_context_for_event(storage: Storage, state_handler, ev, user_id): +async def get_context_for_event(storage: Storage, state_handler, ev, user_id): ctx = {} - room_state_ids = yield storage.state.get_state_ids_for_event(ev.event_id) + room_state_ids = await storage.state.get_state_ids_for_event(ev.event_id) # we no longer bother setting room_alias, and make room_name the # human-readable name instead, be that m.room.name, an alias or # a list of people in the room - name = yield calculate_room_name( + name = await calculate_room_name( storage.main, room_state_ids, user_id, fallback_to_single_member=False ) if name: ctx["name"] = name sender_state_event_id = room_state_ids[("m.room.member", ev.sender)] - sender_state_event = yield storage.main.get_event(sender_state_event_id) + sender_state_event = await storage.main.get_event(sender_state_event_id) ctx["sender_display_name"] = name_from_member_event(sender_state_event) return ctx diff --git a/synapse/push/pusher.py b/synapse/push/pusher.py
index 8ad0bf5936..2a52e226e3 100644 --- a/synapse/push/pusher.py +++ b/synapse/push/pusher.py
@@ -15,24 +15,15 @@ import logging +from synapse.push.emailpusher import EmailPusher +from synapse.push.mailer import Mailer + from .httppusher import HttpPusher logger = logging.getLogger(__name__) -# We try importing this if we can (it will fail if we don't -# have the optional email dependencies installed). We don't -# yet have the config to know if we need the email pusher, -# but importing this after daemonizing seems to fail -# (even though a simple test of importing from a daemonized -# process works fine) -try: - from synapse.push.emailpusher import EmailPusher - from synapse.push.mailer import Mailer, load_jinja2_templates -except Exception: - pass - -class PusherFactory(object): +class PusherFactory: def __init__(self, hs): self.hs = hs self.config = hs.config @@ -43,16 +34,8 @@ class PusherFactory(object): if hs.config.email_enable_notifs: self.mailers = {} # app_name -> Mailer - self.notif_template_html, self.notif_template_text = load_jinja2_templates( - self.config.email_template_dir, - [ - self.config.email_notif_template_html, - self.config.email_notif_template_text, - ], - apply_format_ts_filter=True, - apply_mxc_to_http_filter=True, - public_baseurl=self.config.public_baseurl, - ) + self._notif_template_html = hs.config.email_notif_template_html + self._notif_template_text = hs.config.email_notif_template_text self.pusher_types["email"] = self._create_email_pusher @@ -73,8 +56,8 @@ class PusherFactory(object): mailer = Mailer( hs=self.hs, app_name=app_name, - template_html=self.notif_template_html, - template_text=self.notif_template_text, + template_html=self._notif_template_html, + template_text=self._notif_template_text, ) self.mailers[app_name] = mailer return EmailPusher(self.hs, pusherdict, mailer) diff --git a/synapse/push/pusherpool.py b/synapse/push/pusherpool.py
index 7c5e47bc81..76150e117b 100644 --- a/synapse/push/pusherpool.py +++ b/synapse/push/pusherpool.py
@@ -19,8 +19,6 @@ from typing import TYPE_CHECKING, Dict, Union from prometheus_client import Gauge -from twisted.internet import defer - from synapse.metrics.background_process_metrics import run_as_background_process from synapse.push import PusherConfigException from synapse.push.emailpusher import EmailPusher @@ -52,7 +50,7 @@ class PusherPool: Note that it is expected that each pusher will have its own 'processing' loop which will send out the notifications in the background, rather than blocking until the notifications are sent; accordingly Pusher.on_started, Pusher.on_new_notifications and - Pusher.on_new_receipts are not expected to return deferreds. + Pusher.on_new_receipts are not expected to return awaitables. """ def __init__(self, hs: "HomeServer"): @@ -62,11 +60,17 @@ class PusherPool: self.store = self.hs.get_datastore() self.clock = self.hs.get_clock() + self._account_validity = hs.config.account_validity + # We shard the handling of push notifications by user ID. self._pusher_shard_config = hs.config.push.pusher_shard_config self._instance_name = hs.get_instance_name() - self._account_validity = hs.config.account_validity + # Record the last stream ID that we were poked about so we can get + # changes since then. We set this to the current max stream ID on + # startup as every individual pusher will have checked for changes on + # startup. + self._last_room_stream_id_seen = self.store.get_room_max_stream_ordering() # map from user id to app_id:pushkey to pusher self.pushers = {} # type: Dict[str, Dict[str, Union[HttpPusher, EmailPusher]]] @@ -79,8 +83,7 @@ class PusherPool: return run_as_background_process("start_pushers", self._start_pushers) - @defer.inlineCallbacks - def add_pusher( + async def add_pusher( self, user_id, access_token, @@ -96,7 +99,7 @@ class PusherPool: """Creates a new pusher and adds it to the pool Returns: - Deferred[EmailPusher|HttpPusher] + EmailPusher|HttpPusher """ time_now_msec = self.clock.time_msec() @@ -126,9 +129,9 @@ class PusherPool: # create the pusher setting last_stream_ordering to the current maximum # stream ordering in event_push_actions, so it will process # pushes from this point onwards. - last_stream_ordering = yield self.store.get_latest_push_action_stream_ordering() + last_stream_ordering = await self.store.get_latest_push_action_stream_ordering() - yield self.store.add_pusher( + await self.store.add_pusher( user_id=user_id, access_token=access_token, kind=kind, @@ -142,15 +145,14 @@ class PusherPool: last_stream_ordering=last_stream_ordering, profile_tag=profile_tag, ) - pusher = yield self.start_pusher_by_id(app_id, pushkey, user_id) + pusher = await self.start_pusher_by_id(app_id, pushkey, user_id) return pusher - @defer.inlineCallbacks - def remove_pushers_by_app_id_and_pushkey_not_user( + async def remove_pushers_by_app_id_and_pushkey_not_user( self, app_id, pushkey, not_user_id ): - to_remove = yield self.store.get_pushers_by_app_id_and_pushkey(app_id, pushkey) + to_remove = await self.store.get_pushers_by_app_id_and_pushkey(app_id, pushkey) for p in to_remove: if p["user_name"] != not_user_id: logger.info( @@ -159,10 +161,9 @@ class PusherPool: pushkey, p["user_name"], ) - yield self.remove_pusher(p["app_id"], p["pushkey"], p["user_name"]) + await self.remove_pusher(p["app_id"], p["pushkey"], p["user_name"]) - @defer.inlineCallbacks - def remove_pushers_by_access_token(self, user_id, access_tokens): + async def remove_pushers_by_access_token(self, user_id, access_tokens): """Remove the pushers for a given user corresponding to a set of access_tokens. @@ -175,7 +176,7 @@ class PusherPool: return tokens = set(access_tokens) - for p in (yield self.store.get_pushers_by_user_id(user_id)): + for p in await self.store.get_pushers_by_user_id(user_id): if p["access_token"] in tokens: logger.info( "Removing pusher for app id %s, pushkey %s, user %s", @@ -183,37 +184,42 @@ class PusherPool: p["pushkey"], p["user_name"], ) - yield self.remove_pusher(p["app_id"], p["pushkey"], p["user_name"]) + await self.remove_pusher(p["app_id"], p["pushkey"], p["user_name"]) - @defer.inlineCallbacks - def on_new_notifications(self, min_stream_id, max_stream_id): + async def on_new_notifications(self, max_stream_id: int): if not self.pushers: # nothing to do here. return + if max_stream_id < self._last_room_stream_id_seen: + # Nothing to do + return + + prev_stream_id = self._last_room_stream_id_seen + self._last_room_stream_id_seen = max_stream_id + try: - users_affected = yield self.store.get_push_action_users_in_range( - min_stream_id, max_stream_id + users_affected = await self.store.get_push_action_users_in_range( + prev_stream_id, max_stream_id ) for u in users_affected: - if u in self.pushers: - # Don't push if the user account has expired - if self._account_validity.enabled: - expired = yield self.store.is_account_expired( - u, self.clock.time_msec() - ) - if expired: - continue + # Don't push if the user account has expired + if self._account_validity.enabled: + expired = await self.store.is_account_expired( + u, self.clock.time_msec() + ) + if expired: + continue + if u in self.pushers: for p in self.pushers[u].values(): - p.on_new_notifications(min_stream_id, max_stream_id) + p.on_new_notifications(max_stream_id) except Exception: logger.exception("Exception in pusher on_new_notifications") - @defer.inlineCallbacks - def on_new_receipts(self, min_stream_id, max_stream_id, affected_room_ids): + async def on_new_receipts(self, min_stream_id, max_stream_id, affected_room_ids): if not self.pushers: # nothing to do here. return @@ -221,32 +227,31 @@ class PusherPool: try: # Need to subtract 1 from the minimum because the lower bound here # is not inclusive - users_affected = yield self.store.get_users_sent_receipts_between( + users_affected = await self.store.get_users_sent_receipts_between( min_stream_id - 1, max_stream_id ) for u in users_affected: - if u in self.pushers: - # Don't push if the user account has expired - if self._account_validity.enabled: - expired = yield self.store.is_account_expired( - u, self.clock.time_msec() - ) - if expired: - continue + # Don't push if the user account has expired + if self._account_validity.enabled: + expired = await self.store.is_account_expired( + u, self.clock.time_msec() + ) + if expired: + continue + if u in self.pushers: for p in self.pushers[u].values(): p.on_new_receipts(min_stream_id, max_stream_id) except Exception: logger.exception("Exception in pusher on_new_receipts") - @defer.inlineCallbacks - def start_pusher_by_id(self, app_id, pushkey, user_id): + async def start_pusher_by_id(self, app_id, pushkey, user_id): """Look up the details for the given pusher, and start it Returns: - Deferred[EmailPusher|HttpPusher|None]: The pusher started, if any + EmailPusher|HttpPusher|None: The pusher started, if any """ if not self._should_start_pushers: return @@ -254,7 +259,7 @@ class PusherPool: if not self._pusher_shard_config.should_handle(self._instance_name, user_id): return - resultlist = yield self.store.get_pushers_by_app_id_and_pushkey(app_id, pushkey) + resultlist = await self.store.get_pushers_by_app_id_and_pushkey(app_id, pushkey) pusher_dict = None for r in resultlist: @@ -263,34 +268,29 @@ class PusherPool: pusher = None if pusher_dict: - pusher = yield self._start_pusher(pusher_dict) + pusher = await self._start_pusher(pusher_dict) return pusher - @defer.inlineCallbacks - def _start_pushers(self): + async def _start_pushers(self) -> None: """Start all the pushers - - Returns: - Deferred """ - pushers = yield self.store.get_all_pushers() + pushers = await self.store.get_all_pushers() # Stagger starting up the pushers so we don't completely drown the # process on start up. - yield concurrently_execute(self._start_pusher, pushers, 10) + await concurrently_execute(self._start_pusher, pushers, 10) logger.info("Started pushers") - @defer.inlineCallbacks - def _start_pusher(self, pusherdict): + async def _start_pusher(self, pusherdict): """Start the given pusher Args: pusherdict (dict): dict with the values pulled from the db table Returns: - Deferred[EmailPusher|HttpPusher] + EmailPusher|HttpPusher """ if not self._pusher_shard_config.should_handle( self._instance_name, pusherdict["user_name"] @@ -333,7 +333,7 @@ class PusherPool: user_id = pusherdict["user_name"] last_stream_ordering = pusherdict["last_stream_ordering"] if last_stream_ordering: - have_notifs = yield self.store.get_if_maybe_push_in_range_for_user( + have_notifs = await self.store.get_if_maybe_push_in_range_for_user( user_id, last_stream_ordering ) else: @@ -345,8 +345,7 @@ class PusherPool: return p - @defer.inlineCallbacks - def remove_pusher(self, app_id, pushkey, user_id): + async def remove_pusher(self, app_id, pushkey, user_id): appid_pushkey = "%s:%s" % (app_id, pushkey) byuser = self.pushers.get(user_id, {}) @@ -358,6 +357,6 @@ class PusherPool: synapse_pushers.labels(type(pusher).__name__, pusher.app_id).dec() - yield self.store.delete_pusher_by_app_id_pushkey_user_id( + await self.store.delete_pusher_by_app_id_pushkey_user_id( app_id, pushkey, user_id ) diff --git a/synapse/python_dependencies.py b/synapse/python_dependencies.py
index 8cfcdb0573..0ddead8a0f 100644 --- a/synapse/python_dependencies.py +++ b/synapse/python_dependencies.py
@@ -37,13 +37,16 @@ logger = logging.getLogger(__name__) # installed when that optional dependency requirement is specified. It is passed # to setup() as extras_require in setup.py # +# Note that these both represent runtime dependencies (and the versions +# installed are checked at runtime). +# # [1] https://pip.pypa.io/en/stable/reference/pip_install/#requirement-specifiers. REQUIREMENTS = [ "jsonschema>=2.5.1", "frozendict>=1", "unpaddedbase64>=1.1.0", - "canonicaljson>=1.1.3", + "canonicaljson>=1.4.0", # we use the type definitions added in signedjson 1.1. "signedjson>=1.1.0", "pynacl>=1.2.1", @@ -59,15 +62,20 @@ REQUIREMENTS = [ "pyyaml>=3.11", "pyasn1>=0.1.9", "pyasn1-modules>=0.0.7", - "daemonize>=2.3.1", "bcrypt>=3.1.0", "pillow>=4.3.0", "sortedcontainers>=1.4.4", "pymacaroons>=0.13.0", "msgpack>=0.5.2", "phonenumbers>=8.2.0", - "prometheus_client>=0.0.18,<0.9.0", - # we use attr.validators.deep_iterable, which arrived in 19.1.0 + # we use GaugeHistogramMetric, which was added in prom-client 0.4.0. + # prom-client has a history of breaking backwards compatibility between + # minor versions (https://github.com/prometheus/client_python/issues/317), + # so we also pin the minor version. + "prometheus_client>=0.4.0,<0.9.0", + # we use attr.validators.deep_iterable, which arrived in 19.1.0 (Note: + # Fedora 31 only has 19.1, so if we want to upgrade we should wait until 33 + # is out in November.) "attrs>=19.1.0", "netaddr>=0.7.18", "Jinja2>=2.9", @@ -79,8 +87,6 @@ CONDITIONAL_REQUIREMENTS = { "matrix-synapse-ldap3": ["matrix-synapse-ldap3>=0.1"], # we use execute_batch, which arrived in psycopg 2.7. "postgres": ["psycopg2>=2.7"], - # ConsentResource uses select_autoescape, which arrived in jinja 2.9 - "resources.consent": ["Jinja2>=2.9"], # ACME support is required to provision TLS certificates from authorities # that use the protocol, such as Let's Encrypt. "acme": [ @@ -93,12 +99,6 @@ CONDITIONAL_REQUIREMENTS = { "oidc": ["authlib>=0.14.0"], "systemd": ["systemd-python>=231"], "url_preview": ["lxml>=3.5.0"], - # Dependencies which are exclusively required by unit test code. This is - # NOT a list of all modules that are necessary to run the unit tests. - # Tests assume that all optional dependencies are installed. - # - # parameterized_class decorator was introduced in parameterized 0.7.0 - "test": ["mock>=2.0", "parameterized>=0.7.0"], "sentry": ["sentry-sdk>=0.7.2"], "opentracing": ["jaeger-client>=4.0.0", "opentracing>=2.2.0"], "jwt": ["pyjwt>=1.6.4"], @@ -111,6 +111,7 @@ ALL_OPTIONAL_REQUIREMENTS = set() # type: Set[str] for name, optional_deps in CONDITIONAL_REQUIREMENTS.items(): # Exclude systemd as it's a system-based requirement. + # Exclude lint as it's a dev-based requirement. if name not in ["systemd"]: ALL_OPTIONAL_REQUIREMENTS = set(optional_deps) | ALL_OPTIONAL_REQUIREMENTS diff --git a/synapse/replication/http/_base.py b/synapse/replication/http/_base.py
index fb0dd04f88..64edadb624 100644 --- a/synapse/replication/http/_base.py +++ b/synapse/replication/http/_base.py
@@ -20,22 +20,30 @@ import urllib from inspect import signature from typing import Dict, List, Tuple -from twisted.internet import defer +from prometheus_client import Counter, Gauge -from synapse.api.errors import ( - CodeMessageException, - HttpResponseException, - RequestSendFailed, - SynapseError, -) +from synapse.api.errors import HttpResponseException, SynapseError +from synapse.http import RequestTimedOutError from synapse.logging.opentracing import inject_active_span_byte_dict, trace from synapse.util.caches.response_cache import ResponseCache from synapse.util.stringutils import random_string logger = logging.getLogger(__name__) +_pending_outgoing_requests = Gauge( + "synapse_pending_outgoing_replication_requests", + "Number of active outgoing replication requests, by replication method name", + ["name"], +) + +_outgoing_request_counter = Counter( + "synapse_outgoing_replication_requests", + "Number of outgoing replication requests, by replication method name and result", + ["name", "code"], +) -class ReplicationEndpoint(object): + +class ReplicationEndpoint(metaclass=abc.ABCMeta): """Helper base class for defining new replication HTTP endpoints. This creates an endpoint under `/_synapse/replication/:NAME/:PATH_ARGS..` @@ -74,8 +82,6 @@ class ReplicationEndpoint(object): is received. """ - __metaclass__ = abc.ABCMeta - NAME = abc.abstractproperty() # type: str # type: ignore PATH_ARGS = abc.abstractproperty() # type: Tuple[str, ...] # type: ignore METHOD = "POST" @@ -101,7 +107,7 @@ class ReplicationEndpoint(object): assert self.METHOD in ("PUT", "POST", "GET") @abc.abstractmethod - def _serialize_payload(**kwargs): + async def _serialize_payload(**kwargs): """Static method that is called when creating a request. Concrete implementations should have explicit parameters (rather than @@ -110,9 +116,8 @@ class ReplicationEndpoint(object): argument list. Returns: - Deferred[dict]|dict: If POST/PUT request then dictionary must be - JSON serialisable, otherwise must be appropriate for adding as - query args. + dict: If POST/PUT request then dictionary must be JSON serialisable, + otherwise must be appropriate for adding as query args. """ return {} @@ -143,9 +148,11 @@ class ReplicationEndpoint(object): instance_map = hs.config.worker.instance_map + outgoing_gauge = _pending_outgoing_requests.labels(cls.NAME) + @trace(opname="outgoing_replication_request") - @defer.inlineCallbacks - def send_request(instance_name="master", **kwargs): + @outgoing_gauge.track_inprogress() + async def send_request(instance_name="master", **kwargs): if instance_name == local_instance_name: raise Exception("Trying to send HTTP request to self") if instance_name == "master": @@ -159,7 +166,7 @@ class ReplicationEndpoint(object): "Instance %r not in 'instance_map' config" % (instance_name,) ) - data = yield cls._serialize_payload(**kwargs) + data = await cls._serialize_payload(**kwargs) url_args = [ urllib.parse.quote(kwargs[name], safe="") for name in cls.PATH_ARGS @@ -197,25 +204,28 @@ class ReplicationEndpoint(object): headers = {} # type: Dict[bytes, List[bytes]] inject_active_span_byte_dict(headers, None, check_destination=False) try: - result = yield request_func(uri, data, headers=headers) + result = await request_func(uri, data, headers=headers) break - except CodeMessageException as e: - if e.code != 504 or not cls.RETRY_ON_TIMEOUT: + except RequestTimedOutError: + if not cls.RETRY_ON_TIMEOUT: raise - logger.warning("%s request timed out", cls.NAME) + logger.warning("%s request timed out; retrying", cls.NAME) # If we timed out we probably don't need to worry about backing # off too much, but lets just wait a little anyway. - yield clock.sleep(1) + await clock.sleep(1) except HttpResponseException as e: # We convert to SynapseError as we know that it was a SynapseError - # on the master process that we should send to the client. (And + # on the main process that we should send to the client. (And # importantly, not stack traces everywhere) + _outgoing_request_counter.labels(cls.NAME, e.code).inc() raise e.to_synapse_error() - except RequestSendFailed as e: - raise SynapseError(502, "Failed to talk to master") from e + except Exception as e: + _outgoing_request_counter.labels(cls.NAME, "ERR").inc() + raise SynapseError(502, "Failed to talk to main process") from e + _outgoing_request_counter.labels(cls.NAME, 200).inc() return result return send_request diff --git a/synapse/replication/http/devices.py b/synapse/replication/http/devices.py
index e32aac0a25..807b85d2e1 100644 --- a/synapse/replication/http/devices.py +++ b/synapse/replication/http/devices.py
@@ -53,14 +53,14 @@ class ReplicationUserDevicesResyncRestServlet(ReplicationEndpoint): CACHE = False def __init__(self, hs): - super(ReplicationUserDevicesResyncRestServlet, self).__init__(hs) + super().__init__(hs) self.device_list_updater = hs.get_device_handler().device_list_updater self.store = hs.get_datastore() self.clock = hs.get_clock() @staticmethod - def _serialize_payload(user_id): + async def _serialize_payload(user_id): return {} async def _handle_request(self, request, user_id): diff --git a/synapse/replication/http/federation.py b/synapse/replication/http/federation.py
index c287c4e269..5393b9a9e7 100644 --- a/synapse/replication/http/federation.py +++ b/synapse/replication/http/federation.py
@@ -15,8 +15,6 @@ import logging -from twisted.internet import defer - from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.events import make_event_from_dict from synapse.events.snapshot import EventContext @@ -59,7 +57,7 @@ class ReplicationFederationSendEventsRestServlet(ReplicationEndpoint): PATH_ARGS = () def __init__(self, hs): - super(ReplicationFederationSendEventsRestServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.storage = hs.get_storage() @@ -67,18 +65,18 @@ class ReplicationFederationSendEventsRestServlet(ReplicationEndpoint): self.federation_handler = hs.get_handlers().federation_handler @staticmethod - @defer.inlineCallbacks - def _serialize_payload(store, event_and_contexts, backfilled): + async def _serialize_payload(store, room_id, event_and_contexts, backfilled): """ Args: store + room_id (str) event_and_contexts (list[tuple[FrozenEvent, EventContext]]) backfilled (bool): Whether or not the events are the result of backfilling """ event_payloads = [] for event, context in event_and_contexts: - serialized_context = yield context.serialize(event, store) + serialized_context = await context.serialize(event, store) event_payloads.append( { @@ -91,7 +89,11 @@ class ReplicationFederationSendEventsRestServlet(ReplicationEndpoint): } ) - payload = {"events": event_payloads, "backfilled": backfilled} + payload = { + "events": event_payloads, + "backfilled": backfilled, + "room_id": room_id, + } return payload @@ -99,6 +101,7 @@ class ReplicationFederationSendEventsRestServlet(ReplicationEndpoint): with Measure(self.clock, "repl_fed_send_events_parse"): content = parse_json_object_from_request(request) + room_id = content["room_id"] backfilled = content["backfilled"] event_payloads = content["events"] @@ -123,7 +126,7 @@ class ReplicationFederationSendEventsRestServlet(ReplicationEndpoint): logger.info("Got %d events from federation", len(event_and_contexts)) max_stream_id = await self.federation_handler.persist_events_and_notify( - event_and_contexts, backfilled + room_id, event_and_contexts, backfilled ) return 200, {"max_stream_id": max_stream_id} @@ -147,14 +150,14 @@ class ReplicationFederationSendEduRestServlet(ReplicationEndpoint): PATH_ARGS = ("edu_type",) def __init__(self, hs): - super(ReplicationFederationSendEduRestServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.clock = hs.get_clock() self.registry = hs.get_federation_registry() @staticmethod - def _serialize_payload(edu_type, origin, content): + async def _serialize_payload(edu_type, origin, content): return {"origin": origin, "content": content} async def _handle_request(self, request, edu_type): @@ -190,14 +193,14 @@ class ReplicationGetQueryRestServlet(ReplicationEndpoint): CACHE = False def __init__(self, hs): - super(ReplicationGetQueryRestServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.clock = hs.get_clock() self.registry = hs.get_federation_registry() @staticmethod - def _serialize_payload(query_type, args): + async def _serialize_payload(query_type, args): """ Args: query_type (str) @@ -233,12 +236,12 @@ class ReplicationCleanRoomRestServlet(ReplicationEndpoint): PATH_ARGS = ("room_id",) def __init__(self, hs): - super(ReplicationCleanRoomRestServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() @staticmethod - def _serialize_payload(room_id, args): + async def _serialize_payload(room_id, args): """ Args: room_id (str) @@ -273,7 +276,7 @@ class ReplicationStoreRoomOnInviteRestServlet(ReplicationEndpoint): self.store = hs.get_datastore() @staticmethod - def _serialize_payload(room_id, room_version): + async def _serialize_payload(room_id, room_version): return {"room_version": room_version.identifier} async def _handle_request(self, request, room_id): diff --git a/synapse/replication/http/login.py b/synapse/replication/http/login.py
index 798b9d3af5..4c81e2d784 100644 --- a/synapse/replication/http/login.py +++ b/synapse/replication/http/login.py
@@ -32,11 +32,11 @@ class RegisterDeviceReplicationServlet(ReplicationEndpoint): PATH_ARGS = ("user_id",) def __init__(self, hs): - super(RegisterDeviceReplicationServlet, self).__init__(hs) + super().__init__(hs) self.registration_handler = hs.get_registration_handler() @staticmethod - def _serialize_payload(user_id, device_id, initial_display_name, is_guest): + async def _serialize_payload(user_id, device_id, initial_display_name, is_guest): """ Args: device_id (str|None): Device ID to use, if None a new one is diff --git a/synapse/replication/http/membership.py b/synapse/replication/http/membership.py
index 63ef6eb7be..30680baee8 100644 --- a/synapse/replication/http/membership.py +++ b/synapse/replication/http/membership.py
@@ -19,7 +19,7 @@ from typing import TYPE_CHECKING, Optional from synapse.http.servlet import parse_json_object_from_request from synapse.replication.http._base import ReplicationEndpoint from synapse.types import JsonDict, Requester, UserID -from synapse.util.distributor import user_joined_room, user_left_room +from synapse.util.distributor import user_left_room if TYPE_CHECKING: from synapse.server import HomeServer @@ -45,14 +45,16 @@ class ReplicationRemoteJoinRestServlet(ReplicationEndpoint): PATH_ARGS = ("room_id", "user_id") def __init__(self, hs): - super(ReplicationRemoteJoinRestServlet, self).__init__(hs) + super().__init__(hs) self.federation_handler = hs.get_handlers().federation_handler self.store = hs.get_datastore() self.clock = hs.get_clock() @staticmethod - def _serialize_payload(requester, room_id, user_id, remote_room_hosts, content): + async def _serialize_payload( + requester, room_id, user_id, remote_room_hosts, content + ): """ Args: requester(Requester) @@ -105,14 +107,14 @@ class ReplicationRemoteRejectInviteRestServlet(ReplicationEndpoint): PATH_ARGS = ("invite_event_id",) def __init__(self, hs: "HomeServer"): - super(ReplicationRemoteRejectInviteRestServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.clock = hs.get_clock() self.member_handler = hs.get_room_member_handler() @staticmethod - def _serialize_payload( # type: ignore + async def _serialize_payload( # type: ignore invite_event_id: str, txn_id: Optional[str], requester: Requester, @@ -166,7 +168,7 @@ class ReplicationUserJoinedLeftRoomRestServlet(ReplicationEndpoint): CACHE = False # No point caching as should return instantly. def __init__(self, hs): - super(ReplicationUserJoinedLeftRoomRestServlet, self).__init__(hs) + super().__init__(hs) self.registeration_handler = hs.get_registration_handler() self.store = hs.get_datastore() @@ -174,14 +176,14 @@ class ReplicationUserJoinedLeftRoomRestServlet(ReplicationEndpoint): self.distributor = hs.get_distributor() @staticmethod - def _serialize_payload(room_id, user_id, change): + async def _serialize_payload(room_id, user_id, change): """ Args: room_id (str) user_id (str) - change (str): Either "joined" or "left" + change (str): "left" """ - assert change in ("joined", "left") + assert change == "left" return {} @@ -190,9 +192,7 @@ class ReplicationUserJoinedLeftRoomRestServlet(ReplicationEndpoint): user = UserID.from_string(user_id) - if change == "joined": - user_joined_room(self.distributor, user, room_id) - elif change == "left": + if change == "left": user_left_room(self.distributor, user, room_id) else: raise Exception("Unrecognized change: %r", change) diff --git a/synapse/replication/http/presence.py b/synapse/replication/http/presence.py
index ea1b33331b..bc9aa82cb4 100644 --- a/synapse/replication/http/presence.py +++ b/synapse/replication/http/presence.py
@@ -50,7 +50,7 @@ class ReplicationBumpPresenceActiveTime(ReplicationEndpoint): self._presence_handler = hs.get_presence_handler() @staticmethod - def _serialize_payload(user_id): + async def _serialize_payload(user_id): return {} async def _handle_request(self, request, user_id): @@ -92,7 +92,7 @@ class ReplicationPresenceSetState(ReplicationEndpoint): self._presence_handler = hs.get_presence_handler() @staticmethod - def _serialize_payload(user_id, state, ignore_status_msg=False): + async def _serialize_payload(user_id, state, ignore_status_msg=False): return { "state": state, "ignore_status_msg": ignore_status_msg, diff --git a/synapse/replication/http/register.py b/synapse/replication/http/register.py
index 0c4aca1291..7b12ec9060 100644 --- a/synapse/replication/http/register.py +++ b/synapse/replication/http/register.py
@@ -29,12 +29,12 @@ class ReplicationRegisterServlet(ReplicationEndpoint): PATH_ARGS = ("user_id",) def __init__(self, hs): - super(ReplicationRegisterServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.registration_handler = hs.get_registration_handler() @staticmethod - def _serialize_payload( + async def _serialize_payload( user_id, password_hash, was_guest, @@ -44,6 +44,7 @@ class ReplicationRegisterServlet(ReplicationEndpoint): admin, user_type, address, + shadow_banned, ): """ Args: @@ -60,6 +61,7 @@ class ReplicationRegisterServlet(ReplicationEndpoint): user_type (str|None): type of user. One of the values from api.constants.UserTypes, or None for a normal user. address (str|None): the IP address used to perform the regitration. + shadow_banned (bool): Whether to shadow-ban the user """ return { "password_hash": password_hash, @@ -70,6 +72,7 @@ class ReplicationRegisterServlet(ReplicationEndpoint): "admin": admin, "user_type": user_type, "address": address, + "shadow_banned": shadow_banned, } async def _handle_request(self, request, user_id): @@ -87,6 +90,7 @@ class ReplicationRegisterServlet(ReplicationEndpoint): admin=content["admin"], user_type=content["user_type"], address=content["address"], + shadow_banned=content["shadow_banned"], ) return 200, {} @@ -100,12 +104,12 @@ class ReplicationPostRegisterActionsServlet(ReplicationEndpoint): PATH_ARGS = ("user_id",) def __init__(self, hs): - super(ReplicationPostRegisterActionsServlet, self).__init__(hs) + super().__init__(hs) self.store = hs.get_datastore() self.registration_handler = hs.get_registration_handler() @staticmethod - def _serialize_payload(user_id, auth_result, access_token): + async def _serialize_payload(user_id, auth_result, access_token): """ Args: user_id (str): The user ID that consented diff --git a/synapse/replication/http/send_event.py b/synapse/replication/http/send_event.py
index c981723c1a..9a3a694d5d 100644 --- a/synapse/replication/http/send_event.py +++ b/synapse/replication/http/send_event.py
@@ -15,8 +15,6 @@ import logging -from twisted.internet import defer - from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.events import make_event_from_dict from synapse.events.snapshot import EventContext @@ -54,7 +52,7 @@ class ReplicationSendEventRestServlet(ReplicationEndpoint): PATH_ARGS = ("event_id",) def __init__(self, hs): - super(ReplicationSendEventRestServlet, self).__init__(hs) + super().__init__(hs) self.event_creation_handler = hs.get_event_creation_handler() self.store = hs.get_datastore() @@ -62,8 +60,7 @@ class ReplicationSendEventRestServlet(ReplicationEndpoint): self.clock = hs.get_clock() @staticmethod - @defer.inlineCallbacks - def _serialize_payload( + async def _serialize_payload( event_id, store, event, context, requester, ratelimit, extra_users ): """ @@ -77,7 +74,7 @@ class ReplicationSendEventRestServlet(ReplicationEndpoint): extra_users (list(UserID)): Any extra users to notify about event """ - serialized_context = yield context.serialize(event, store) + serialized_context = await context.serialize(event, store) payload = { "event": event.get_pdu_json(), diff --git a/synapse/replication/http/streams.py b/synapse/replication/http/streams.py
index bde97eef32..309159e304 100644 --- a/synapse/replication/http/streams.py +++ b/synapse/replication/http/streams.py
@@ -54,7 +54,7 @@ class ReplicationGetStreamUpdates(ReplicationEndpoint): self.streams = hs.get_replication_streams() @staticmethod - def _serialize_payload(stream_name, from_token, upto_token): + async def _serialize_payload(stream_name, from_token, upto_token): return {"from_token": from_token, "upto_token": upto_token} async def _handle_request(self, request, stream_name): diff --git a/synapse/replication/slave/storage/_base.py b/synapse/replication/slave/storage/_base.py
index f9e2533e96..d0089fe06c 100644 --- a/synapse/replication/slave/storage/_base.py +++ b/synapse/replication/slave/storage/_base.py
@@ -16,8 +16,8 @@ import logging from typing import Optional -from synapse.storage.data_stores.main.cache import CacheInvalidationWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.cache import CacheInvalidationWorkerStore from synapse.storage.engines import PostgresEngine from synapse.storage.util.id_generators import MultiWriterIdGenerator @@ -25,17 +25,19 @@ logger = logging.getLogger(__name__) class BaseSlavedStore(CacheInvalidationWorkerStore): - def __init__(self, database: Database, db_conn, hs): - super(BaseSlavedStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) if isinstance(self.database_engine, PostgresEngine): self._cache_id_gen = MultiWriterIdGenerator( db_conn, database, + stream_name="caches", instance_name=hs.get_instance_name(), table="cache_invalidation_stream_by_instance", instance_column="instance_name", id_column="stream_id", sequence_name="cache_invalidation_stream_seq", + writers=[], ) # type: Optional[MultiWriterIdGenerator] else: self._cache_id_gen = None diff --git a/synapse/replication/slave/storage/_slaved_id_tracker.py b/synapse/replication/slave/storage/_slaved_id_tracker.py
index 9d1d173b2f..eb74903d68 100644 --- a/synapse/replication/slave/storage/_slaved_id_tracker.py +++ b/synapse/replication/slave/storage/_slaved_id_tracker.py
@@ -16,14 +16,14 @@ from synapse.storage.util.id_generators import _load_current_id -class SlavedIdTracker(object): +class SlavedIdTracker: def __init__(self, db_conn, table, column, extra_tables=[], step=1): self.step = step self._current = _load_current_id(db_conn, table, column, step) for table, column in extra_tables: - self.advance(_load_current_id(db_conn, table, column)) + self.advance(None, _load_current_id(db_conn, table, column)) - def advance(self, new_id): + def advance(self, instance_name, new_id): self._current = (max if self.step > 0 else min)(self._current, new_id) def get_current_token(self): @@ -33,3 +33,11 @@ class SlavedIdTracker(object): int """ return self._current + + def get_current_token_for_writer(self, instance_name: str) -> int: + """Returns the position of the given writer. + + For streams with single writers this is equivalent to + `get_current_token`. + """ + return self.get_current_token() diff --git a/synapse/replication/slave/storage/account_data.py b/synapse/replication/slave/storage/account_data.py
index 525b94fd87..4268565fc8 100644 --- a/synapse/replication/slave/storage/account_data.py +++ b/synapse/replication/slave/storage/account_data.py
@@ -17,13 +17,13 @@ from synapse.replication.slave.storage._base import BaseSlavedStore from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams import AccountDataStream, TagAccountDataStream -from synapse.storage.data_stores.main.account_data import AccountDataWorkerStore -from synapse.storage.data_stores.main.tags import TagsWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.account_data import AccountDataWorkerStore +from synapse.storage.databases.main.tags import TagsWorkerStore class SlavedAccountDataStore(TagsWorkerStore, AccountDataWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self._account_data_id_gen = SlavedIdTracker( db_conn, "account_data", @@ -34,19 +34,19 @@ class SlavedAccountDataStore(TagsWorkerStore, AccountDataWorkerStore, BaseSlaved ], ) - super(SlavedAccountDataStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) def get_max_account_data_stream_id(self): return self._account_data_id_gen.get_current_token() def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == TagAccountDataStream.NAME: - self._account_data_id_gen.advance(token) + self._account_data_id_gen.advance(instance_name, token) for row in rows: self.get_tags_for_user.invalidate((row.user_id,)) self._account_data_stream_cache.entity_has_changed(row.user_id, token) elif stream_name == AccountDataStream.NAME: - self._account_data_id_gen.advance(token) + self._account_data_id_gen.advance(instance_name, token) for row in rows: if not row.room_id: self.get_global_account_data_by_type_for_user.invalidate( diff --git a/synapse/replication/slave/storage/appservice.py b/synapse/replication/slave/storage/appservice.py
index a67fbeffb7..0f8d7037bd 100644 --- a/synapse/replication/slave/storage/appservice.py +++ b/synapse/replication/slave/storage/appservice.py
@@ -14,7 +14,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.appservice import ( +from synapse.storage.databases.main.appservice import ( ApplicationServiceTransactionWorkerStore, ApplicationServiceWorkerStore, ) diff --git a/synapse/replication/slave/storage/client_ips.py b/synapse/replication/slave/storage/client_ips.py
index 1a38f53dfb..1f8dafe7ea 100644 --- a/synapse/replication/slave/storage/client_ips.py +++ b/synapse/replication/slave/storage/client_ips.py
@@ -13,22 +13,22 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.client_ips import LAST_SEEN_GRANULARITY -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.client_ips import LAST_SEEN_GRANULARITY from synapse.util.caches.descriptors import Cache from ._base import BaseSlavedStore class SlavedClientIpStore(BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedClientIpStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.client_ip_last_seen = Cache( name="client_ip_last_seen", keylen=4, max_entries=50000 ) - def insert_client_ip(self, user_id, access_token, ip, user_agent, device_id): + async def insert_client_ip(self, user_id, access_token, ip, user_agent, device_id): now = int(self._clock.time_msec()) key = (user_id, access_token, ip) diff --git a/synapse/replication/slave/storage/deviceinbox.py b/synapse/replication/slave/storage/deviceinbox.py
index a8a16dbc71..5b045bed02 100644 --- a/synapse/replication/slave/storage/deviceinbox.py +++ b/synapse/replication/slave/storage/deviceinbox.py
@@ -16,15 +16,15 @@ from synapse.replication.slave.storage._base import BaseSlavedStore from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams import ToDeviceStream -from synapse.storage.data_stores.main.deviceinbox import DeviceInboxWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.deviceinbox import DeviceInboxWorkerStore from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.caches.stream_change_cache import StreamChangeCache class SlavedDeviceInboxStore(DeviceInboxWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedDeviceInboxStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._device_inbox_id_gen = SlavedIdTracker( db_conn, "device_inbox", "stream_id" ) @@ -46,7 +46,7 @@ class SlavedDeviceInboxStore(DeviceInboxWorkerStore, BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == ToDeviceStream.NAME: - self._device_inbox_id_gen.advance(token) + self._device_inbox_id_gen.advance(instance_name, token) for row in rows: if row.entity.startswith("@"): self._device_inbox_stream_cache.entity_has_changed( diff --git a/synapse/replication/slave/storage/devices.py b/synapse/replication/slave/storage/devices.py
index 9d8067342f..e0d86240dd 100644 --- a/synapse/replication/slave/storage/devices.py +++ b/synapse/replication/slave/storage/devices.py
@@ -16,15 +16,15 @@ from synapse.replication.slave.storage._base import BaseSlavedStore from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams._base import DeviceListsStream, UserSignatureStream -from synapse.storage.data_stores.main.devices import DeviceWorkerStore -from synapse.storage.data_stores.main.end_to_end_keys import EndToEndKeyWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.devices import DeviceWorkerStore +from synapse.storage.databases.main.end_to_end_keys import EndToEndKeyWorkerStore from synapse.util.caches.stream_change_cache import StreamChangeCache class SlavedDeviceStore(EndToEndKeyWorkerStore, DeviceWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedDeviceStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.hs = hs @@ -48,12 +48,15 @@ class SlavedDeviceStore(EndToEndKeyWorkerStore, DeviceWorkerStore, BaseSlavedSto "DeviceListFederationStreamChangeCache", device_list_max ) + def get_device_stream_token(self) -> int: + return self._device_list_id_gen.get_current_token() + def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == DeviceListsStream.NAME: - self._device_list_id_gen.advance(token) + self._device_list_id_gen.advance(instance_name, token) self._invalidate_caches_for_devices(token, rows) elif stream_name == UserSignatureStream.NAME: - self._device_list_id_gen.advance(token) + self._device_list_id_gen.advance(instance_name, token) for row in rows: self._user_signature_stream_cache.entity_has_changed(row.user_id, token) return super().process_replication_rows(stream_name, instance_name, token, rows) diff --git a/synapse/replication/slave/storage/directory.py b/synapse/replication/slave/storage/directory.py
index 8b9717c46f..1945bcf9a8 100644 --- a/synapse/replication/slave/storage/directory.py +++ b/synapse/replication/slave/storage/directory.py
@@ -13,7 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.directory import DirectoryWorkerStore +from synapse.storage.databases.main.directory import DirectoryWorkerStore from ._base import BaseSlavedStore diff --git a/synapse/replication/slave/storage/events.py b/synapse/replication/slave/storage/events.py
index 1a1a50a24f..fbffe6d85c 100644 --- a/synapse/replication/slave/storage/events.py +++ b/synapse/replication/slave/storage/events.py
@@ -15,18 +15,18 @@ # limitations under the License. import logging -from synapse.storage.data_stores.main.event_federation import EventFederationWorkerStore -from synapse.storage.data_stores.main.event_push_actions import ( +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.event_federation import EventFederationWorkerStore +from synapse.storage.databases.main.event_push_actions import ( EventPushActionsWorkerStore, ) -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.data_stores.main.relations import RelationsWorkerStore -from synapse.storage.data_stores.main.roommember import RoomMemberWorkerStore -from synapse.storage.data_stores.main.signatures import SignatureWorkerStore -from synapse.storage.data_stores.main.state import StateGroupWorkerStore -from synapse.storage.data_stores.main.stream import StreamWorkerStore -from synapse.storage.data_stores.main.user_erasure_store import UserErasureWorkerStore -from synapse.storage.database import Database +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.storage.databases.main.relations import RelationsWorkerStore +from synapse.storage.databases.main.roommember import RoomMemberWorkerStore +from synapse.storage.databases.main.signatures import SignatureWorkerStore +from synapse.storage.databases.main.state import StateGroupWorkerStore +from synapse.storage.databases.main.stream import StreamWorkerStore +from synapse.storage.databases.main.user_erasure_store import UserErasureWorkerStore from synapse.util.caches.stream_change_cache import StreamChangeCache from ._base import BaseSlavedStore @@ -55,11 +55,11 @@ class SlavedEventStore( RelationsWorkerStore, BaseSlavedStore, ): - def __init__(self, database: Database, db_conn, hs): - super(SlavedEventStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) events_max = self._stream_id_gen.get_current_token() - curr_state_delta_prefill, min_curr_state_delta_id = self.db.get_cache_dict( + curr_state_delta_prefill, min_curr_state_delta_id = self.db_pool.get_cache_dict( db_conn, "current_state_delta_stream", entity_column="room_id", diff --git a/synapse/replication/slave/storage/filtering.py b/synapse/replication/slave/storage/filtering.py
index bcb0688954..6a23252861 100644 --- a/synapse/replication/slave/storage/filtering.py +++ b/synapse/replication/slave/storage/filtering.py
@@ -13,15 +13,15 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.filtering import FilteringStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.filtering import FilteringStore from ._base import BaseSlavedStore class SlavedFilteringStore(BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedFilteringStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # Filters are immutable so this cache doesn't need to be expired get_user_filter = FilteringStore.__dict__["get_user_filter"] diff --git a/synapse/replication/slave/storage/groups.py b/synapse/replication/slave/storage/groups.py
index 5d210fa3a1..30955bcbfe 100644 --- a/synapse/replication/slave/storage/groups.py +++ b/synapse/replication/slave/storage/groups.py
@@ -16,14 +16,14 @@ from synapse.replication.slave.storage._base import BaseSlavedStore from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams import GroupServerStream -from synapse.storage.data_stores.main.group_server import GroupServerWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.group_server import GroupServerWorkerStore from synapse.util.caches.stream_change_cache import StreamChangeCache class SlavedGroupServerStore(GroupServerWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedGroupServerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.hs = hs @@ -40,7 +40,7 @@ class SlavedGroupServerStore(GroupServerWorkerStore, BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == GroupServerStream.NAME: - self._group_updates_id_gen.advance(token) + self._group_updates_id_gen.advance(instance_name, token) for row in rows: self._group_updates_stream_cache.entity_has_changed(row.user_id, token) diff --git a/synapse/replication/slave/storage/keys.py b/synapse/replication/slave/storage/keys.py
index 3def367ae9..961579751c 100644 --- a/synapse/replication/slave/storage/keys.py +++ b/synapse/replication/slave/storage/keys.py
@@ -13,7 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.keys import KeyStore +from synapse.storage.databases.main.keys import KeyStore # KeyStore isn't really safe to use from a worker, but for now we do so and hope that # the races it creates aren't too bad. diff --git a/synapse/replication/slave/storage/presence.py b/synapse/replication/slave/storage/presence.py
index 2938cb8e43..55620c03d8 100644 --- a/synapse/replication/slave/storage/presence.py +++ b/synapse/replication/slave/storage/presence.py
@@ -15,8 +15,8 @@ from synapse.replication.tcp.streams import PresenceStream from synapse.storage import DataStore -from synapse.storage.data_stores.main.presence import PresenceStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.presence import PresenceStore from synapse.util.caches.stream_change_cache import StreamChangeCache from ._base import BaseSlavedStore @@ -24,8 +24,8 @@ from ._slaved_id_tracker import SlavedIdTracker class SlavedPresenceStore(BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedPresenceStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._presence_id_gen = SlavedIdTracker(db_conn, "presence_stream", "stream_id") self._presence_on_startup = self._get_active_presence(db_conn) # type: ignore @@ -44,7 +44,7 @@ class SlavedPresenceStore(BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == PresenceStream.NAME: - self._presence_id_gen.advance(token) + self._presence_id_gen.advance(instance_name, token) for row in rows: self.presence_stream_cache.entity_has_changed(row.user_id, token) self._get_presence_for_user.invalidate((row.user_id,)) diff --git a/synapse/replication/slave/storage/profile.py b/synapse/replication/slave/storage/profile.py
index 28c508aad3..f85b20a071 100644 --- a/synapse/replication/slave/storage/profile.py +++ b/synapse/replication/slave/storage/profile.py
@@ -14,7 +14,7 @@ # limitations under the License. from synapse.replication.slave.storage._base import BaseSlavedStore -from synapse.storage.data_stores.main.profile import ProfileWorkerStore +from synapse.storage.databases.main.profile import ProfileWorkerStore class SlavedProfileStore(ProfileWorkerStore, BaseSlavedStore): diff --git a/synapse/replication/slave/storage/push_rule.py b/synapse/replication/slave/storage/push_rule.py
index 23ec1c5b11..de904c943c 100644 --- a/synapse/replication/slave/storage/push_rule.py +++ b/synapse/replication/slave/storage/push_rule.py
@@ -14,25 +14,23 @@ # See the License for the specific language governing permissions and # limitations under the License. +from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams import PushRulesStream -from synapse.storage.data_stores.main.push_rule import PushRulesWorkerStore +from synapse.storage.databases.main.push_rule import PushRulesWorkerStore from .events import SlavedEventStore class SlavedPushRuleStore(SlavedEventStore, PushRulesWorkerStore): - def get_push_rules_stream_token(self): - return ( - self._push_rules_stream_id_gen.get_current_token(), - self._stream_id_gen.get_current_token(), - ) - def get_max_push_rules_stream_id(self): return self._push_rules_stream_id_gen.get_current_token() def process_replication_rows(self, stream_name, instance_name, token, rows): + # We assert this for the benefit of mypy + assert isinstance(self._push_rules_stream_id_gen, SlavedIdTracker) + if stream_name == PushRulesStream.NAME: - self._push_rules_stream_id_gen.advance(token) + self._push_rules_stream_id_gen.advance(instance_name, token) for row in rows: self.get_push_rules_for_user.invalidate((row.user_id,)) self.get_push_rules_enabled_for_user.invalidate((row.user_id,)) diff --git a/synapse/replication/slave/storage/pushers.py b/synapse/replication/slave/storage/pushers.py
index ff449f3658..c418730ba8 100644 --- a/synapse/replication/slave/storage/pushers.py +++ b/synapse/replication/slave/storage/pushers.py
@@ -15,16 +15,16 @@ # limitations under the License. from synapse.replication.tcp.streams import PushersStream -from synapse.storage.data_stores.main.pusher import PusherWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.pusher import PusherWorkerStore from ._base import BaseSlavedStore from ._slaved_id_tracker import SlavedIdTracker class SlavedPusherStore(PusherWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(SlavedPusherStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._pushers_id_gen = SlavedIdTracker( db_conn, "pushers", "id", extra_tables=[("deleted_pushers", "stream_id")] ) @@ -34,5 +34,5 @@ class SlavedPusherStore(PusherWorkerStore, BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == PushersStream.NAME: - self._pushers_id_gen.advance(token) + self._pushers_id_gen.advance(instance_name, token) return super().process_replication_rows(stream_name, instance_name, token, rows) diff --git a/synapse/replication/slave/storage/receipts.py b/synapse/replication/slave/storage/receipts.py
index 6982686eb5..6195917376 100644 --- a/synapse/replication/slave/storage/receipts.py +++ b/synapse/replication/slave/storage/receipts.py
@@ -15,22 +15,22 @@ # limitations under the License. from synapse.replication.tcp.streams import ReceiptsStream -from synapse.storage.data_stores.main.receipts import ReceiptsWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.receipts import ReceiptsWorkerStore from ._base import BaseSlavedStore from ._slaved_id_tracker import SlavedIdTracker class SlavedReceiptsStore(ReceiptsWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): # We instantiate this first as the ReceiptsWorkerStore constructor # needs to be able to call get_max_receipt_stream_id self._receipts_id_gen = SlavedIdTracker( db_conn, "receipts_linearized", "stream_id" ) - super(SlavedReceiptsStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) def get_max_receipt_stream_id(self): return self._receipts_id_gen.get_current_token() @@ -46,7 +46,7 @@ class SlavedReceiptsStore(ReceiptsWorkerStore, BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == ReceiptsStream.NAME: - self._receipts_id_gen.advance(token) + self._receipts_id_gen.advance(instance_name, token) for row in rows: self.invalidate_caches_for_receipt( row.room_id, row.receipt_type, row.user_id diff --git a/synapse/replication/slave/storage/registration.py b/synapse/replication/slave/storage/registration.py
index 4b8553e250..a40f064e2b 100644 --- a/synapse/replication/slave/storage/registration.py +++ b/synapse/replication/slave/storage/registration.py
@@ -13,7 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.registration import RegistrationWorkerStore +from synapse.storage.databases.main.registration import RegistrationWorkerStore from ._base import BaseSlavedStore diff --git a/synapse/replication/slave/storage/room.py b/synapse/replication/slave/storage/room.py
index 8710207ada..109ac6bea1 100644 --- a/synapse/replication/slave/storage/room.py +++ b/synapse/replication/slave/storage/room.py
@@ -14,16 +14,16 @@ # limitations under the License. from synapse.replication.tcp.streams import PublicRoomsStream -from synapse.storage.data_stores.main.room import RoomWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.room import RoomWorkerStore from ._base import BaseSlavedStore from ._slaved_id_tracker import SlavedIdTracker class RoomStore(RoomWorkerStore, BaseSlavedStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._public_room_id_gen = SlavedIdTracker( db_conn, "public_room_list_stream", "stream_id" ) @@ -33,6 +33,6 @@ class RoomStore(RoomWorkerStore, BaseSlavedStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == PublicRoomsStream.NAME: - self._public_room_id_gen.advance(token) + self._public_room_id_gen.advance(instance_name, token) return super().process_replication_rows(stream_name, instance_name, token, rows) diff --git a/synapse/replication/slave/storage/transactions.py b/synapse/replication/slave/storage/transactions.py
index ac88e6b8c3..2091ac0df6 100644 --- a/synapse/replication/slave/storage/transactions.py +++ b/synapse/replication/slave/storage/transactions.py
@@ -13,7 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.main.transactions import TransactionStore +from synapse.storage.databases.main.transactions import TransactionStore from ._base import BaseSlavedStore diff --git a/synapse/replication/tcp/client.py b/synapse/replication/tcp/client.py
index fcf8ebf1e7..e165429cad 100644 --- a/synapse/replication/tcp/client.py +++ b/synapse/replication/tcp/client.py
@@ -14,7 +14,6 @@ # limitations under the License. """A replication client for use by synapse workers. """ -import heapq import logging from typing import TYPE_CHECKING, Dict, List, Tuple @@ -30,6 +29,7 @@ from synapse.replication.tcp.streams.events import ( EventsStreamEventRow, EventsStreamRow, ) +from synapse.types import PersistedEventPosition, UserID from synapse.util.async_helpers import timeout_deferred from synapse.util.metrics import Measure @@ -99,7 +99,6 @@ class ReplicationDataHandler: def __init__(self, hs: "HomeServer"): self.store = hs.get_datastore() - self.pusher_pool = hs.get_pusherpool() self.notifier = hs.get_notifier() self._reactor = hs.get_reactor() self._clock = hs.get_clock() @@ -149,13 +148,15 @@ class ReplicationDataHandler: if event.rejected_reason: continue - extra_users = () # type: Tuple[str, ...] + extra_users = () # type: Tuple[UserID, ...] if event.type == EventTypes.Member: - extra_users = (event.state_key,) - max_token = self.store.get_room_max_stream_ordering() - self.notifier.on_new_room_event(event, token, max_token, extra_users) + extra_users = (UserID.from_string(event.state_key),) - await self.pusher_pool.on_new_notifications(token, token) + max_token = self.store.get_room_max_token() + event_pos = PersistedEventPosition(instance_name, token) + self.notifier.on_new_room_event( + event, event_pos, max_token, extra_users + ) # Notify any waiting deferreds. The list is ordered by position so we # just iterate through the list until we reach a position that is @@ -219,9 +220,8 @@ class ReplicationDataHandler: waiting_list = self._streams_to_waiters.setdefault(stream_name, []) - # We insert into the list using heapq as it is more efficient than - # pushing then resorting each time. - heapq.heappush(waiting_list, (position, deferred)) + waiting_list.append((position, deferred)) + waiting_list.sort(key=lambda t: t[0]) # We measure here to get in flight counts and average waiting time. with Measure(self._clock, "repl.wait_for_stream_position"): diff --git a/synapse/replication/tcp/commands.py b/synapse/replication/tcp/commands.py
index f33801f883..8cd47770c1 100644 --- a/synapse/replication/tcp/commands.py +++ b/synapse/replication/tcp/commands.py
@@ -18,11 +18,10 @@ The VALID_SERVER_COMMANDS and VALID_CLIENT_COMMANDS define which commands are allowed to be sent by which side. """ import abc -import json import logging from typing import Tuple, Type -_json_encoder = json.JSONEncoder() +from synapse.util import json_decoder, json_encoder logger = logging.getLogger(__name__) @@ -124,7 +123,7 @@ class RdataCommand(Command): stream_name, instance_name, None if token == "batch" else int(token), - json.loads(row_json), + json_decoder.decode(row_json), ) def to_line(self): @@ -133,7 +132,7 @@ class RdataCommand(Command): self.stream_name, self.instance_name, str(self.token) if self.token is not None else "batch", - _json_encoder.encode(self.row), + json_encoder.encode(self.row), ) ) @@ -358,7 +357,7 @@ class UserIpCommand(Command): def from_line(cls, line): user_id, jsn = line.split(" ", 1) - access_token, ip, user_agent, device_id, last_seen = json.loads(jsn) + access_token, ip, user_agent, device_id, last_seen = json_decoder.decode(jsn) return cls(user_id, access_token, ip, user_agent, device_id, last_seen) @@ -366,7 +365,7 @@ class UserIpCommand(Command): return ( self.user_id + " " - + _json_encoder.encode( + + json_encoder.encode( ( self.access_token, self.ip, diff --git a/synapse/replication/tcp/handler.py b/synapse/replication/tcp/handler.py
index 1c303f3a46..b323841f73 100644 --- a/synapse/replication/tcp/handler.py +++ b/synapse/replication/tcp/handler.py
@@ -109,7 +109,7 @@ class ReplicationCommandHandler: if isinstance(stream, (EventsStream, BackfillStream)): # Only add EventStream and BackfillStream as a source on the # instance in charge of event persistence. - if hs.config.worker.writers.events == hs.get_instance_name(): + if hs.get_instance_name() in hs.config.worker.writers.events: self._streams_to_replicate.append(stream) continue diff --git a/synapse/replication/tcp/protocol.py b/synapse/replication/tcp/protocol.py
index 0350923898..0b0d204e64 100644 --- a/synapse/replication/tcp/protocol.py +++ b/synapse/replication/tcp/protocol.py
@@ -113,7 +113,7 @@ PING_TIMEOUT_MULTIPLIER = 5 PING_TIMEOUT_MS = PING_TIME * PING_TIMEOUT_MULTIPLIER -class ConnectionStates(object): +class ConnectionStates: CONNECTING = "connecting" ESTABLISHED = "established" PAUSED = "paused" diff --git a/synapse/replication/tcp/resource.py b/synapse/replication/tcp/resource.py
index 41569305df..687984e7a8 100644 --- a/synapse/replication/tcp/resource.py +++ b/synapse/replication/tcp/resource.py
@@ -58,7 +58,7 @@ class ReplicationStreamProtocolFactory(Factory): ) -class ReplicationStreamer(object): +class ReplicationStreamer: """Handles replication connections. This needs to be poked when new replication data may be available. When new @@ -93,7 +93,7 @@ class ReplicationStreamer(object): """ if not self.command_handler.connected(): # Don't bother if nothing is listening. We still need to advance - # the stream tokens otherwise they'll fall beihind forever + # the stream tokens otherwise they'll fall behind forever for stream in self.streams: stream.discard_updates_and_advance() return diff --git a/synapse/replication/tcp/streams/_base.py b/synapse/replication/tcp/streams/_base.py
index 7a42de3f7d..54dccd15a6 100644 --- a/synapse/replication/tcp/streams/_base.py +++ b/synapse/replication/tcp/streams/_base.py
@@ -79,7 +79,7 @@ StreamUpdateResult = Tuple[List[Tuple[Token, StreamRow]], Token, bool] UpdateFunction = Callable[[str, Token, Token, int], Awaitable[StreamUpdateResult]] -class Stream(object): +class Stream: """Base class for the streams. Provides a `get_updates()` function that returns new updates since the last @@ -345,14 +345,14 @@ class PushRulesStream(Stream): def __init__(self, hs): self.store = hs.get_datastore() - super(PushRulesStream, self).__init__( + super().__init__( hs.get_instance_name(), self._current_token, self.store.get_all_push_rule_updates, ) def _current_token(self, instance_name: str) -> int: - push_rules_token, _ = self.store.get_push_rules_stream_token() + push_rules_token = self.store.get_max_push_rules_stream_id() return push_rules_token @@ -383,7 +383,7 @@ class CachesStream(Stream): the cache on the workers """ - @attr.s + @attr.s(slots=True) class CachesStreamRow: """Stream to inform workers they should invalidate their cache. @@ -405,7 +405,7 @@ class CachesStream(Stream): store = hs.get_datastore() super().__init__( hs.get_instance_name(), - store.get_cache_stream_token, + store.get_cache_stream_token_for_writer, store.get_all_updated_caches, ) @@ -441,7 +441,7 @@ class DeviceListsStream(Stream): told about a device update. """ - @attr.s + @attr.s(slots=True) class DeviceListsStreamRow: entity = attr.ib(type=str) diff --git a/synapse/replication/tcp/streams/events.py b/synapse/replication/tcp/streams/events.py
index 16c63ff4ec..ccc7ca30d8 100644 --- a/synapse/replication/tcp/streams/events.py +++ b/synapse/replication/tcp/streams/events.py
@@ -19,7 +19,7 @@ from typing import List, Tuple, Type import attr -from ._base import Stream, StreamUpdateResult, Token, current_token_without_instance +from ._base import Stream, StreamUpdateResult, Token """Handling of the 'events' replication stream @@ -49,14 +49,14 @@ data part are: @attr.s(slots=True, frozen=True) -class EventsStreamRow(object): +class EventsStreamRow: """A parsed row from the events replication stream""" type = attr.ib() # str: the TypeId of one of the *EventsStreamRows data = attr.ib() # BaseEventsStreamRow -class BaseEventsStreamRow(object): +class BaseEventsStreamRow: """Base class for rows to be sent in the events stream. Specifies how to identify, serialize and deserialize the different types. @@ -117,7 +117,7 @@ class EventsStream(Stream): self._store = hs.get_datastore() super().__init__( hs.get_instance_name(), - current_token_without_instance(self._store.get_current_events_token), + self._store._stream_id_gen.get_current_token_for_writer, self._update_function, ) diff --git a/synapse/res/templates/auth_success.html b/synapse/res/templates/auth_success.html new file mode 100644
index 0000000000..baf4633142 --- /dev/null +++ b/synapse/res/templates/auth_success.html
@@ -0,0 +1,21 @@ +<html> +<head> +<title>Success!</title> +<meta name='viewport' content='width=device-width, initial-scale=1, + user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> +<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> +<script> +if (window.onAuthDone) { + window.onAuthDone(); +} else if (window.opener && window.opener.postMessage) { + window.opener.postMessage("authDone", "*"); +} +</script> +</head> +<body> + <div> + <p>Thank you</p> + <p>You may now close this window and return to the application</p> + </div> +</body> +</html> diff --git a/synapse/res/templates/password_reset_confirmation.html b/synapse/res/templates/password_reset_confirmation.html new file mode 100644
index 0000000000..def4b5162b --- /dev/null +++ b/synapse/res/templates/password_reset_confirmation.html
@@ -0,0 +1,16 @@ +<html> +<head></head> +<body> +<!--Use a hidden form to resubmit the information necessary to reset the password--> +<form method="post"> + <input type="hidden" name="sid" value="{{ sid }}"> + <input type="hidden" name="token" value="{{ token }}"> + <input type="hidden" name="client_secret" value="{{ client_secret }}"> + + <p>You have requested to <strong>reset your Matrix account password</strong>. Click the link below to confirm this action. <br /><br /> + If you did not mean to do this, please close this page and your password will not be changed.</p> + <p><button type="submit">Confirm changing my password</button></p> +</form> +</body> +</html> + diff --git a/synapse/res/templates/recaptcha.html b/synapse/res/templates/recaptcha.html new file mode 100644
index 0000000000..63944dc608 --- /dev/null +++ b/synapse/res/templates/recaptcha.html
@@ -0,0 +1,38 @@ +<html> +<head> +<title>Authentication</title> +<meta name='viewport' content='width=device-width, initial-scale=1, + user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> +<script src="https://www.recaptcha.net/recaptcha/api.js" + async defer></script> +<script src="//code.jquery.com/jquery-1.11.2.min.js"></script> +<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> +<script> +function captchaDone() { + $('#registrationForm').submit(); +} +</script> +</head> +<body> +<form id="registrationForm" method="post" action="{{ myurl }}"> + <div> + <p> + Hello! We need to prevent computer programs and other automated + things from creating accounts on this server. + </p> + <p> + Please verify that you're not a robot. + </p> + <input type="hidden" name="session" value="{{ session }}" /> + <div class="g-recaptcha" + data-sitekey="{{ sitekey }}" + data-callback="captchaDone"> + </div> + <noscript> + <input type="submit" value="All Done" /> + </noscript> + </div> + </div> +</form> +</body> +</html> diff --git a/synapse/res/templates/saml_error.html b/synapse/res/templates/saml_error.html deleted file mode 100644
index bfd6449c5d..0000000000 --- a/synapse/res/templates/saml_error.html +++ /dev/null
@@ -1,45 +0,0 @@ -<!DOCTYPE html> -<html lang="en"> -<head> - <meta charset="UTF-8"> - <title>SSO error</title> -</head> -<body> - <p>Oops! Something went wrong during authentication<span id="errormsg"></span>.</p> - <p> - If you are seeing this page after clicking a link sent to you via email, make - sure you only click the confirmation link once, and that you open the - validation link in the same client you're logging in from. - </p> - <p> - Try logging in again from your Matrix client and if the problem persists - please contact the server's administrator. - </p> - - <script type="text/javascript"> - // Error handling to support Auth0 errors that we might get through a GET request - // to the validation endpoint. If an error is provided, it's either going to be - // located in the query string or in a query string-like URI fragment. - // We try to locate the error from any of these two locations, but if we can't - // we just don't print anything specific. - let searchStr = ""; - if (window.location.search) { - // window.location.searchParams isn't always defined when - // window.location.search is, so it's more reliable to parse the latter. - searchStr = window.location.search; - } else if (window.location.hash) { - // Replace the # with a ? so that URLSearchParams does the right thing and - // doesn't parse the first parameter incorrectly. - searchStr = window.location.hash.replace("#", "?"); - } - - // We might end up with no error in the URL, so we need to check if we have one - // to print one. - let errorDesc = new URLSearchParams(searchStr).get("error_description") - if (errorDesc) { - - document.getElementById("errormsg").innerText = ` ("${errorDesc}")`; - } - </script> -</body> -</html> \ No newline at end of file diff --git a/synapse/res/templates/sso_error.html b/synapse/res/templates/sso_error.html
index 43a211386b..944bc9c9ca 100644 --- a/synapse/res/templates/sso_error.html +++ b/synapse/res/templates/sso_error.html
@@ -5,14 +5,49 @@ <title>SSO error</title> </head> <body> - <p>Oops! Something went wrong during authentication.</p> +{# If an error of unauthorised is returned it means we have actively rejected their login #} +{% if error == "unauthorised" %} + <p>You are not allowed to log in here.</p> +{% else %} + <p> + There was an error during authentication: + </p> + <div id="errormsg" style="margin:20px 80px">{{ error_description | e }}</div> + <p> + If you are seeing this page after clicking a link sent to you via email, make + sure you only click the confirmation link once, and that you open the + validation link in the same client you're logging in from. + </p> <p> Try logging in again from your Matrix client and if the problem persists please contact the server's administrator. </p> <p>Error: <code>{{ error }}</code></p> - {% if error_description %} - <pre><code>{{ error_description }}</code></pre> - {% endif %} + + <script type="text/javascript"> + // Error handling to support Auth0 errors that we might get through a GET request + // to the validation endpoint. If an error is provided, it's either going to be + // located in the query string or in a query string-like URI fragment. + // We try to locate the error from any of these two locations, but if we can't + // we just don't print anything specific. + let searchStr = ""; + if (window.location.search) { + // window.location.searchParams isn't always defined when + // window.location.search is, so it's more reliable to parse the latter. + searchStr = window.location.search; + } else if (window.location.hash) { + // Replace the # with a ? so that URLSearchParams does the right thing and + // doesn't parse the first parameter incorrectly. + searchStr = window.location.hash.replace("#", "?"); + } + + // We might end up with no error in the URL, so we need to check if we have one + // to print one. + let errorDesc = new URLSearchParams(searchStr).get("error_description") + if (errorDesc) { + document.getElementById("errormsg").innerText = errorDesc; + } + </script> +{% endif %} </body> </html> diff --git a/synapse/res/templates/terms.html b/synapse/res/templates/terms.html new file mode 100644
index 0000000000..dfef9897ee --- /dev/null +++ b/synapse/res/templates/terms.html
@@ -0,0 +1,20 @@ +<html> +<head> +<title>Authentication</title> +<meta name='viewport' content='width=device-width, initial-scale=1, + user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> +<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> +</head> +<body> +<form id="registrationForm" method="post" action="{{ myurl }}"> + <div> + <p> + Please click the button below if you agree to the + <a href="{{ terms_url }}">privacy policy of this homeserver.</a> + </p> + <input type="hidden" name="session" value="{{ session }}" /> + <input type="submit" value="Agree" /> + </div> +</form> +</body> +</html> diff --git a/synapse/rest/__init__.py b/synapse/rest/__init__.py
index 2e81eeff65..5d9cdf4bde 100644 --- a/synapse/rest/__init__.py +++ b/synapse/rest/__init__.py
@@ -13,8 +13,8 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import synapse.rest.admin from synapse.http.server import JsonResource +from synapse.rest import admin from synapse.rest.client import versions from synapse.rest.client.v1 import ( directory, @@ -50,6 +50,7 @@ from synapse.rest.client.v2_alpha import ( room_keys, room_upgrade_rest_servlet, sendtodevice, + shared_rooms, sync, tags, thirdparty, @@ -123,6 +124,7 @@ class ClientRestResource(JsonResource): password_policy.register_servlets(hs, client_resource) # moving to /_synapse/admin - synapse.rest.admin.register_servlets_for_client_rest_resource( - hs, client_resource - ) + admin.register_servlets_for_client_rest_resource(hs, client_resource) + + # unstable + shared_rooms.register_servlets(hs, client_resource) diff --git a/synapse/rest/admin/__init__.py b/synapse/rest/admin/__init__.py
index 1c88c93f38..57cac22252 100644 --- a/synapse/rest/admin/__init__.py +++ b/synapse/rest/admin/__init__.py
@@ -16,13 +16,13 @@ import logging import platform -import re import synapse from synapse.api.errors import Codes, NotFoundError, SynapseError from synapse.http.server import JsonResource from synapse.http.servlet import RestServlet, parse_json_object_from_request from synapse.rest.admin._base import ( + admin_patterns, assert_requester_is_admin, historical_admin_path_patterns, ) @@ -31,6 +31,7 @@ from synapse.rest.admin.devices import ( DeviceRestServlet, DevicesRestServlet, ) +from synapse.rest.admin.event_reports import EventReportsRestServlet from synapse.rest.admin.groups import DeleteGroupAdminRestServlet from synapse.rest.admin.media import ListMediaInRoom, register_servlets_for_media_repo from synapse.rest.admin.purge_room_servlet import PurgeRoomServlet @@ -49,6 +50,7 @@ from synapse.rest.admin.users import ( ResetPasswordRestServlet, SearchUsersRestServlet, UserAdminServlet, + UserMembershipRestServlet, UserRegisterServlet, UserRestServletV2, UsersRestServlet, @@ -61,7 +63,7 @@ logger = logging.getLogger(__name__) class VersionServlet(RestServlet): - PATTERNS = (re.compile("^/_synapse/admin/v1/server_version$"),) + PATTERNS = admin_patterns("/server_version$") def __init__(self, hs): self.res = { @@ -107,7 +109,8 @@ class PurgeHistoryRestServlet(RestServlet): if event.room_id != room_id: raise SynapseError(400, "Event is for wrong room.") - token = await self.store.get_topological_token_for_event(event_id) + room_token = await self.store.get_topological_token_for_event(event_id) + token = await room_token.to_string(self.store) logger.info("[purge] purging up to token %s (event_id %s)", token, event_id) elif "purge_up_to_ts" in body: @@ -209,11 +212,13 @@ def register_servlets(hs, http_server): SendServerNoticeServlet(hs).register(http_server) VersionServlet(hs).register(http_server) UserAdminServlet(hs).register(http_server) + UserMembershipRestServlet(hs).register(http_server) UserRestServletV2(hs).register(http_server) UsersRestServletV2(hs).register(http_server) DeviceRestServlet(hs).register(http_server) DevicesRestServlet(hs).register(http_server) DeleteDevicesRestServlet(hs).register(http_server) + EventReportsRestServlet(hs).register(http_server) def register_servlets_for_client_rest_resource(hs, http_server): diff --git a/synapse/rest/admin/_base.py b/synapse/rest/admin/_base.py
index d82eaf5e38..db9fea263a 100644 --- a/synapse/rest/admin/_base.py +++ b/synapse/rest/admin/_base.py
@@ -44,7 +44,7 @@ def historical_admin_path_patterns(path_regex): ] -def admin_patterns(path_regex: str): +def admin_patterns(path_regex: str, version: str = "v1"): """Returns the list of patterns for an admin endpoint Args: @@ -54,7 +54,7 @@ def admin_patterns(path_regex: str): Returns: A list of regex patterns. """ - admin_prefix = "^/_synapse/admin/v1" + admin_prefix = "^/_synapse/admin/" + version patterns = [re.compile(admin_prefix + path_regex)] return patterns diff --git a/synapse/rest/admin/devices.py b/synapse/rest/admin/devices.py
index 8d32677339..a163863322 100644 --- a/synapse/rest/admin/devices.py +++ b/synapse/rest/admin/devices.py
@@ -13,7 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging -import re from synapse.api.errors import NotFoundError, SynapseError from synapse.http.servlet import ( @@ -21,7 +20,7 @@ from synapse.http.servlet import ( assert_params_in_dict, parse_json_object_from_request, ) -from synapse.rest.admin._base import assert_requester_is_admin +from synapse.rest.admin._base import admin_patterns, assert_requester_is_admin from synapse.types import UserID logger = logging.getLogger(__name__) @@ -32,14 +31,12 @@ class DeviceRestServlet(RestServlet): Get, update or delete the given user's device """ - PATTERNS = ( - re.compile( - "^/_synapse/admin/v2/users/(?P<user_id>[^/]*)/devices/(?P<device_id>[^/]*)$" - ), + PATTERNS = admin_patterns( + "/users/(?P<user_id>[^/]*)/devices/(?P<device_id>[^/]*)$", "v2" ) def __init__(self, hs): - super(DeviceRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.device_handler = hs.get_device_handler() @@ -98,7 +95,7 @@ class DevicesRestServlet(RestServlet): Retrieve the given user's devices """ - PATTERNS = (re.compile("^/_synapse/admin/v2/users/(?P<user_id>[^/]*)/devices$"),) + PATTERNS = admin_patterns("/users/(?P<user_id>[^/]*)/devices$", "v2") def __init__(self, hs): """ @@ -131,9 +128,7 @@ class DeleteDevicesRestServlet(RestServlet): key which lists the device_ids to delete. """ - PATTERNS = ( - re.compile("^/_synapse/admin/v2/users/(?P<user_id>[^/]*)/delete_devices$"), - ) + PATTERNS = admin_patterns("/users/(?P<user_id>[^/]*)/delete_devices$", "v2") def __init__(self, hs): self.hs = hs diff --git a/synapse/rest/admin/event_reports.py b/synapse/rest/admin/event_reports.py new file mode 100644
index 0000000000..5b8d0594cd --- /dev/null +++ b/synapse/rest/admin/event_reports.py
@@ -0,0 +1,88 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 Dirk Klimpel +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import logging + +from synapse.api.errors import Codes, SynapseError +from synapse.http.servlet import RestServlet, parse_integer, parse_string +from synapse.rest.admin._base import admin_patterns, assert_requester_is_admin + +logger = logging.getLogger(__name__) + + +class EventReportsRestServlet(RestServlet): + """ + List all reported events that are known to the homeserver. Results are returned + in a dictionary containing report information. Supports pagination. + The requester must have administrator access in Synapse. + + GET /_synapse/admin/v1/event_reports + returns: + 200 OK with list of reports if success otherwise an error. + + Args: + The parameters `from` and `limit` are required only for pagination. + By default, a `limit` of 100 is used. + The parameter `dir` can be used to define the order of results. + The parameter `user_id` can be used to filter by user id. + The parameter `room_id` can be used to filter by room id. + Returns: + A list of reported events and an integer representing the total number of + reported events that exist given this query + """ + + PATTERNS = admin_patterns("/event_reports$") + + def __init__(self, hs): + self.hs = hs + self.auth = hs.get_auth() + self.store = hs.get_datastore() + + async def on_GET(self, request): + await assert_requester_is_admin(self.auth, request) + + start = parse_integer(request, "from", default=0) + limit = parse_integer(request, "limit", default=100) + direction = parse_string(request, "dir", default="b") + user_id = parse_string(request, "user_id") + room_id = parse_string(request, "room_id") + + if start < 0: + raise SynapseError( + 400, + "The start parameter must be a positive integer.", + errcode=Codes.INVALID_PARAM, + ) + + if limit < 0: + raise SynapseError( + 400, + "The limit parameter must be a positive integer.", + errcode=Codes.INVALID_PARAM, + ) + + if direction not in ("f", "b"): + raise SynapseError( + 400, "Unknown direction: %s" % (direction,), errcode=Codes.INVALID_PARAM + ) + + event_reports, total = await self.store.get_event_reports_paginate( + start, limit, direction, user_id, room_id + ) + ret = {"event_reports": event_reports, "total": total} + if (start + limit) < total: + ret["next_token"] = start + len(event_reports) + + return 200, ret diff --git a/synapse/rest/admin/purge_room_servlet.py b/synapse/rest/admin/purge_room_servlet.py
index f474066542..8b7bb6d44e 100644 --- a/synapse/rest/admin/purge_room_servlet.py +++ b/synapse/rest/admin/purge_room_servlet.py
@@ -12,14 +12,13 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import re - from synapse.http.servlet import ( RestServlet, assert_params_in_dict, parse_json_object_from_request, ) from synapse.rest.admin import assert_requester_is_admin +from synapse.rest.admin._base import admin_patterns class PurgeRoomServlet(RestServlet): @@ -35,7 +34,7 @@ class PurgeRoomServlet(RestServlet): {} """ - PATTERNS = (re.compile("^/_synapse/admin/v1/purge_room$"),) + PATTERNS = admin_patterns("/purge_room$") def __init__(self, hs): """ diff --git a/synapse/rest/admin/rooms.py b/synapse/rest/admin/rooms.py
index b8c95d045a..09726d52d6 100644 --- a/synapse/rest/admin/rooms.py +++ b/synapse/rest/admin/rooms.py
@@ -31,7 +31,7 @@ from synapse.rest.admin._base import ( assert_user_is_admin, historical_admin_path_patterns, ) -from synapse.storage.data_stores.main.room import RoomSortOrder +from synapse.storage.databases.main.room import RoomSortOrder from synapse.types import RoomAlias, RoomID, UserID, create_requester logger = logging.getLogger(__name__) @@ -103,6 +103,14 @@ class DeleteRoomRestServlet(RestServlet): Codes.BAD_JSON, ) + purge = content.get("purge", True) + if not isinstance(purge, bool): + raise SynapseError( + HTTPStatus.BAD_REQUEST, + "Param 'purge' must be a boolean, if given", + Codes.BAD_JSON, + ) + ret = await self.room_shutdown_handler.shutdown_room( room_id=room_id, new_room_user_id=content.get("new_room_user_id"), @@ -113,7 +121,8 @@ class DeleteRoomRestServlet(RestServlet): ) # Purge room - await self.pagination_handler.purge_room(room_id) + if purge: + await self.pagination_handler.purge_room(room_id) return (200, ret) @@ -307,6 +316,9 @@ class JoinRoomAliasServlet(RestServlet): join_rules_event = room_state.get((EventTypes.JoinRules, "")) if join_rules_event: if not (join_rules_event.content.get("join_rule") == JoinRules.PUBLIC): + # update_membership with an action of "invite" can raise a + # ShadowBanError. This is not handled since it is assumed that + # an admin isn't going to call this API with a shadow-banned user. await self.room_member_handler.update_membership( requester=requester, target=fake_requester.user, diff --git a/synapse/rest/admin/server_notice_servlet.py b/synapse/rest/admin/server_notice_servlet.py
index 6e9a874121..375d055445 100644 --- a/synapse/rest/admin/server_notice_servlet.py +++ b/synapse/rest/admin/server_notice_servlet.py
@@ -12,8 +12,6 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import re - from synapse.api.constants import EventTypes from synapse.api.errors import SynapseError from synapse.http.servlet import ( @@ -22,6 +20,7 @@ from synapse.http.servlet import ( parse_json_object_from_request, ) from synapse.rest.admin import assert_requester_is_admin +from synapse.rest.admin._base import admin_patterns from synapse.rest.client.transactions import HttpTransactionCache from synapse.types import UserID @@ -56,13 +55,13 @@ class SendServerNoticeServlet(RestServlet): self.snm = hs.get_server_notices_manager() def register(self, json_resource): - PATTERN = "^/_synapse/admin/v1/send_server_notice" + PATTERN = "/send_server_notice" json_resource.register_paths( - "POST", (re.compile(PATTERN + "$"),), self.on_POST, self.__class__.__name__ + "POST", admin_patterns(PATTERN + "$"), self.on_POST, self.__class__.__name__ ) json_resource.register_paths( "PUT", - (re.compile(PATTERN + "/(?P<txn_id>[^/]*)$"),), + admin_patterns(PATTERN + "/(?P<txn_id>[^/]*)$"), self.on_PUT, self.__class__.__name__, ) diff --git a/synapse/rest/admin/users.py b/synapse/rest/admin/users.py
index cc0bdfa5c9..20dc1d0e05 100644 --- a/synapse/rest/admin/users.py +++ b/synapse/rest/admin/users.py
@@ -15,7 +15,6 @@ import hashlib import hmac import logging -import re from http import HTTPStatus from synapse.api.constants import UserTypes @@ -29,6 +28,7 @@ from synapse.http.servlet import ( parse_string, ) from synapse.rest.admin._base import ( + admin_patterns, assert_requester_is_admin, assert_user_is_admin, historical_admin_path_patterns, @@ -60,7 +60,7 @@ class UsersRestServlet(RestServlet): class UsersRestServletV2(RestServlet): - PATTERNS = (re.compile("^/_synapse/admin/v2/users$"),) + PATTERNS = admin_patterns("/users$", "v2") """Get request to list all local users. This needs user to have administrator access in Synapse. @@ -73,6 +73,7 @@ class UsersRestServletV2(RestServlet): The parameters `from` and `limit` are required only for pagination. By default, a `limit` of 100 is used. The parameter `user_id` can be used to filter by user id. + The parameter `name` can be used to filter by user id or display name. The parameter `guests` can be used to exclude guest users. The parameter `deactivated` can be used to include deactivated users. """ @@ -89,11 +90,12 @@ class UsersRestServletV2(RestServlet): start = parse_integer(request, "from", default=0) limit = parse_integer(request, "limit", default=100) user_id = parse_string(request, "user_id", default=None) + name = parse_string(request, "name", default=None) guests = parse_boolean(request, "guests", default=True) deactivated = parse_boolean(request, "deactivated", default=False) users, total = await self.store.get_users_paginate( - start, limit, user_id, guests, deactivated + start, limit, user_id, name, guests, deactivated ) ret = {"users": users, "total": total} if len(users) >= limit: @@ -103,7 +105,7 @@ class UsersRestServletV2(RestServlet): class UserRestServletV2(RestServlet): - PATTERNS = (re.compile("^/_synapse/admin/v2/users/(?P<user_id>[^/]+)$"),) + PATTERNS = admin_patterns("/users/(?P<user_id>[^/]+)$", "v2") """Get request to list user details. This needs user to have administrator access in Synapse. @@ -640,7 +642,7 @@ class UserAdminServlet(RestServlet): {} """ - PATTERNS = (re.compile("^/_synapse/admin/v1/users/(?P<user_id>[^/]*)/admin$"),) + PATTERNS = admin_patterns("/users/(?P<user_id>[^/]*)/admin$") def __init__(self, hs): self.hs = hs @@ -681,3 +683,29 @@ class UserAdminServlet(RestServlet): await self.store.set_server_admin(target_user, set_admin_to) return 200, {} + + +class UserMembershipRestServlet(RestServlet): + """ + Get room list of an user. + """ + + PATTERNS = admin_patterns("/users/(?P<user_id>[^/]+)/joined_rooms$") + + def __init__(self, hs): + self.is_mine = hs.is_mine + self.auth = hs.get_auth() + self.store = hs.get_datastore() + + async def on_GET(self, request, user_id): + await assert_requester_is_admin(self.auth, request) + + if not self.is_mine(UserID.from_string(user_id)): + raise SynapseError(400, "Can only lookup local users") + + room_ids = await self.store.get_rooms_for_user(user_id) + if not room_ids: + raise NotFoundError("User not found") + + ret = {"joined_rooms": list(room_ids), "total": len(room_ids)} + return 200, ret diff --git a/synapse/rest/client/transactions.py b/synapse/rest/client/transactions.py
index 6da71dc46f..7be5c0fb88 100644 --- a/synapse/rest/client/transactions.py +++ b/synapse/rest/client/transactions.py
@@ -25,7 +25,7 @@ logger = logging.getLogger(__name__) CLEANUP_PERIOD_MS = 1000 * 60 * 30 # 30 mins -class HttpTransactionCache(object): +class HttpTransactionCache: def __init__(self, hs): self.hs = hs self.auth = self.hs.get_auth() diff --git a/synapse/rest/client/v1/directory.py b/synapse/rest/client/v1/directory.py
index 5934b1fe8b..faabeeb91c 100644 --- a/synapse/rest/client/v1/directory.py +++ b/synapse/rest/client/v1/directory.py
@@ -40,7 +40,7 @@ class ClientDirectoryServer(RestServlet): PATTERNS = client_patterns("/directory/room/(?P<room_alias>[^/]*)$", v1=True) def __init__(self, hs): - super(ClientDirectoryServer, self).__init__() + super().__init__() self.store = hs.get_datastore() self.handlers = hs.get_handlers() self.auth = hs.get_auth() @@ -89,7 +89,7 @@ class ClientDirectoryServer(RestServlet): dir_handler = self.handlers.directory_handler try: - service = await self.auth.get_appservice_by_req(request) + service = self.auth.get_appservice_by_req(request) room_alias = RoomAlias.from_string(room_alias) await dir_handler.delete_appservice_association(service, room_alias) logger.info( @@ -120,7 +120,7 @@ class ClientDirectoryListServer(RestServlet): PATTERNS = client_patterns("/directory/list/room/(?P<room_id>[^/]*)$", v1=True) def __init__(self, hs): - super(ClientDirectoryListServer, self).__init__() + super().__init__() self.store = hs.get_datastore() self.handlers = hs.get_handlers() self.auth = hs.get_auth() @@ -160,7 +160,7 @@ class ClientAppserviceDirectoryListServer(RestServlet): ) def __init__(self, hs): - super(ClientAppserviceDirectoryListServer, self).__init__() + super().__init__() self.store = hs.get_datastore() self.handlers = hs.get_handlers() self.auth = hs.get_auth() diff --git a/synapse/rest/client/v1/events.py b/synapse/rest/client/v1/events.py
index 25effd0261..1ecb77aa26 100644 --- a/synapse/rest/client/v1/events.py +++ b/synapse/rest/client/v1/events.py
@@ -30,9 +30,10 @@ class EventStreamRestServlet(RestServlet): DEFAULT_LONGPOLL_TIME_MS = 30000 def __init__(self, hs): - super(EventStreamRestServlet, self).__init__() + super().__init__() self.event_stream_handler = hs.get_event_stream_handler() self.auth = hs.get_auth() + self.store = hs.get_datastore() async def on_GET(self, request): requester = await self.auth.get_user_by_req(request, allow_guest=True) @@ -44,7 +45,7 @@ class EventStreamRestServlet(RestServlet): if b"room_id" in request.args: room_id = request.args[b"room_id"][0].decode("ascii") - pagin_config = PaginationConfig.from_request(request) + pagin_config = await PaginationConfig.from_request(self.store, request) timeout = EventStreamRestServlet.DEFAULT_LONGPOLL_TIME_MS if b"timeout" in request.args: try: @@ -74,7 +75,7 @@ class EventRestServlet(RestServlet): PATTERNS = client_patterns("/events/(?P<event_id>[^/]*)$", v1=True) def __init__(self, hs): - super(EventRestServlet, self).__init__() + super().__init__() self.clock = hs.get_clock() self.event_handler = hs.get_event_handler() self.auth = hs.get_auth() diff --git a/synapse/rest/client/v1/initial_sync.py b/synapse/rest/client/v1/initial_sync.py
index 910b3b4eeb..91da0ee573 100644 --- a/synapse/rest/client/v1/initial_sync.py +++ b/synapse/rest/client/v1/initial_sync.py
@@ -24,14 +24,15 @@ class InitialSyncRestServlet(RestServlet): PATTERNS = client_patterns("/initialSync$", v1=True) def __init__(self, hs): - super(InitialSyncRestServlet, self).__init__() + super().__init__() self.initial_sync_handler = hs.get_initial_sync_handler() self.auth = hs.get_auth() + self.store = hs.get_datastore() async def on_GET(self, request): requester = await self.auth.get_user_by_req(request) as_client_event = b"raw" not in request.args - pagination_config = PaginationConfig.from_request(request) + pagination_config = await PaginationConfig.from_request(self.store, request) include_archived = parse_boolean(request, "archived", default=False) content = await self.initial_sync_handler.snapshot_all_rooms( user_id=requester.user.to_string(), diff --git a/synapse/rest/client/v1/login.py b/synapse/rest/client/v1/login.py
index 379f668d6f..3d1693d7ac 100644 --- a/synapse/rest/client/v1/login.py +++ b/synapse/rest/client/v1/login.py
@@ -18,6 +18,11 @@ from typing import Awaitable, Callable, Dict, Optional from synapse.api.errors import Codes, LoginError, SynapseError from synapse.api.ratelimiting import Ratelimiter +from synapse.appservice import ApplicationService +from synapse.handlers.auth import ( + convert_client_dict_legacy_fields_to_identifier, + login_id_phone_to_thirdparty, +) from synapse.http.server import finish_request from synapse.http.servlet import ( RestServlet, @@ -28,56 +33,11 @@ from synapse.http.site import SynapseRequest from synapse.rest.client.v2_alpha._base import client_patterns from synapse.rest.well_known import WellKnownBuilder from synapse.types import JsonDict, UserID -from synapse.util.msisdn import phone_number_to_msisdn from synapse.util.threepids import canonicalise_email logger = logging.getLogger(__name__) -def login_submission_legacy_convert(submission): - """ - If the input login submission is an old style object - (ie. with top-level user / medium / address) convert it - to a typed object. - """ - if "user" in submission: - submission["identifier"] = {"type": "m.id.user", "user": submission["user"]} - del submission["user"] - - if "medium" in submission and "address" in submission: - submission["identifier"] = { - "type": "m.id.thirdparty", - "medium": submission["medium"], - "address": submission["address"], - } - del submission["medium"] - del submission["address"] - - -def login_id_thirdparty_from_phone(identifier): - """ - Convert a phone login identifier type to a generic threepid identifier - Args: - identifier(dict): Login identifier dict of type 'm.id.phone' - - Returns: Login identifier dict of type 'm.id.threepid' - """ - if "country" not in identifier or ( - # The specification requires a "phone" field, while Synapse used to require a "number" - # field. Accept both for backwards compatibility. - "phone" not in identifier - and "number" not in identifier - ): - raise SynapseError(400, "Invalid phone-type identifier") - - # Accept both "phone" and "number" as valid keys in m.id.phone - phone_number = identifier.get("phone", identifier["number"]) - - msisdn = phone_number_to_msisdn(identifier["country"], phone_number) - - return {"type": "m.id.thirdparty", "medium": "msisdn", "address": msisdn} - - class LoginRestServlet(RestServlet): PATTERNS = client_patterns("/login$", v1=True) CAS_TYPE = "m.login.cas" @@ -85,9 +45,10 @@ class LoginRestServlet(RestServlet): TOKEN_TYPE = "m.login.token" JWT_TYPE = "org.matrix.login.jwt" JWT_TYPE_DEPRECATED = "m.login.jwt" + APPSERVICE_TYPE = "uk.half-shot.msc2778.login.application_service" def __init__(self, hs): - super(LoginRestServlet, self).__init__() + super().__init__() self.hs = hs # JWT configuration variables. @@ -102,6 +63,8 @@ class LoginRestServlet(RestServlet): self.cas_enabled = hs.config.cas_enabled self.oidc_enabled = hs.config.oidc_enabled + self.auth = hs.get_auth() + self.auth_handler = self.hs.get_auth_handler() self.registration_handler = hs.get_registration_handler() self.handlers = hs.get_handlers() @@ -157,8 +120,12 @@ class LoginRestServlet(RestServlet): self._address_ratelimiter.ratelimit(request.getClientIP()) login_submission = parse_json_object_from_request(request) + try: - if self.jwt_enabled and ( + if login_submission["type"] == LoginRestServlet.APPSERVICE_TYPE: + appservice = self.auth.get_appservice_by_req(request) + result = await self._do_appservice_login(login_submission, appservice) + elif self.jwt_enabled and ( login_submission["type"] == LoginRestServlet.JWT_TYPE or login_submission["type"] == LoginRestServlet.JWT_TYPE_DEPRECATED ): @@ -175,6 +142,33 @@ class LoginRestServlet(RestServlet): result["well_known"] = well_known_data return 200, result + def _get_qualified_user_id(self, identifier): + if identifier["type"] != "m.id.user": + raise SynapseError(400, "Unknown login identifier type") + if "user" not in identifier: + raise SynapseError(400, "User identifier is missing 'user' key") + + if identifier["user"].startswith("@"): + return identifier["user"] + else: + return UserID(identifier["user"], self.hs.hostname).to_string() + + async def _do_appservice_login( + self, login_submission: JsonDict, appservice: ApplicationService + ): + logger.info( + "Got appservice login request with identifier: %r", + login_submission.get("identifier"), + ) + + identifier = convert_client_dict_legacy_fields_to_identifier(login_submission) + qualified_user_id = self._get_qualified_user_id(identifier) + + if not appservice.is_interested_in_user(qualified_user_id): + raise LoginError(403, "Invalid access_token", errcode=Codes.FORBIDDEN) + + return await self._complete_login(qualified_user_id, login_submission) + async def _do_other_login(self, login_submission: JsonDict) -> Dict[str, str]: """Handle non-token/saml/jwt logins @@ -194,18 +188,11 @@ class LoginRestServlet(RestServlet): login_submission.get("address"), login_submission.get("user"), ) - login_submission_legacy_convert(login_submission) - - if "identifier" not in login_submission: - raise SynapseError(400, "Missing param: identifier") - - identifier = login_submission["identifier"] - if "type" not in identifier: - raise SynapseError(400, "Login identifier has no type") + identifier = convert_client_dict_legacy_fields_to_identifier(login_submission) # convert phone type identifiers to generic threepids if identifier["type"] == "m.id.phone": - identifier = login_id_thirdparty_from_phone(identifier) + identifier = login_id_phone_to_thirdparty(identifier) # convert threepid identifiers to user IDs if identifier["type"] == "m.id.thirdparty": @@ -267,15 +254,7 @@ class LoginRestServlet(RestServlet): # by this point, the identifier should be an m.id.user: if it's anything # else, we haven't understood it. - if identifier["type"] != "m.id.user": - raise SynapseError(400, "Unknown login identifier type") - if "user" not in identifier: - raise SynapseError(400, "User identifier is missing 'user' key") - - if identifier["user"].startswith("@"): - qualified_user_id = identifier["user"] - else: - qualified_user_id = UserID(identifier["user"], self.hs.hostname).to_string() + qualified_user_id = self._get_qualified_user_id(identifier) # Check if we've hit the failed ratelimit (but don't update it) self._failed_attempts_ratelimiter.ratelimit( @@ -303,9 +282,7 @@ class LoginRestServlet(RestServlet): self, user_id: str, login_submission: JsonDict, - callback: Optional[ - Callable[[Dict[str, str]], Awaitable[Dict[str, str]]] - ] = None, + callback: Optional[Callable[[Dict[str, str]], Awaitable[None]]] = None, create_non_existent_users: bool = False, ) -> Dict[str, str]: """Called when we've successfully authed the user and now need to @@ -318,12 +295,12 @@ class LoginRestServlet(RestServlet): Args: user_id: ID of the user to register. login_submission: Dictionary of login information. - callback: Callback function to run after registration. + callback: Callback function to run after login. create_non_existent_users: Whether to create the user if they don't exist. Defaults to False. Returns: - result: Dictionary of account information after successful registration. + result: Dictionary of account information after successful login. """ # Before we actually log them in we check if they've already logged in @@ -358,14 +335,24 @@ class LoginRestServlet(RestServlet): return result async def _do_token_login(self, login_submission: JsonDict) -> Dict[str, str]: + """ + Handle the final stage of SSO login. + + Args: + login_submission: The JSON request body. + + Returns: + The body of the JSON response. + """ token = login_submission["token"] auth_handler = self.auth_handler user_id = await auth_handler.validate_short_term_login_token_and_get_user_id( token ) - result = await self._complete_login(user_id, login_submission) - return result + return await self._complete_login( + user_id, login_submission, self.auth_handler._sso_login_callback + ) async def _do_jwt_login(self, login_submission: JsonDict) -> Dict[str, str]: token = login_submission.get("token", None) @@ -448,7 +435,7 @@ class CasTicketServlet(RestServlet): PATTERNS = client_patterns("/login/cas/ticket", v1=True) def __init__(self, hs): - super(CasTicketServlet, self).__init__() + super().__init__() self._cas_handler = hs.get_cas_handler() async def on_GET(self, request: SynapseRequest) -> None: diff --git a/synapse/rest/client/v1/logout.py b/synapse/rest/client/v1/logout.py
index b0c30b65be..f792b50cdc 100644 --- a/synapse/rest/client/v1/logout.py +++ b/synapse/rest/client/v1/logout.py
@@ -25,7 +25,7 @@ class LogoutRestServlet(RestServlet): PATTERNS = client_patterns("/logout$", v1=True) def __init__(self, hs): - super(LogoutRestServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self._auth_handler = hs.get_auth_handler() self._device_handler = hs.get_device_handler() @@ -53,7 +53,7 @@ class LogoutAllRestServlet(RestServlet): PATTERNS = client_patterns("/logout/all$", v1=True) def __init__(self, hs): - super(LogoutAllRestServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self._auth_handler = hs.get_auth_handler() self._device_handler = hs.get_device_handler() diff --git a/synapse/rest/client/v1/presence.py b/synapse/rest/client/v1/presence.py
index ceaa28c212..4796cdac05 100644 --- a/synapse/rest/client/v1/presence.py +++ b/synapse/rest/client/v1/presence.py
@@ -30,7 +30,7 @@ class PresenceStatusRestServlet(RestServlet): PATTERNS = client_patterns("/presence/(?P<user_id>[^/]*)/status", v1=True) def __init__(self, hs): - super(PresenceStatusRestServlet, self).__init__() + super().__init__() self.hs = hs self.presence_handler = hs.get_presence_handler() self.clock = hs.get_clock() diff --git a/synapse/rest/client/v1/profile.py b/synapse/rest/client/v1/profile.py
index 165313b572..204b2ec9e5 100644 --- a/synapse/rest/client/v1/profile.py +++ b/synapse/rest/client/v1/profile.py
@@ -26,7 +26,7 @@ class ProfileDisplaynameRestServlet(RestServlet): PATTERNS = client_patterns("/profile/(?P<user_id>[^/]*)/displayname", v1=True) def __init__(self, hs): - super(ProfileDisplaynameRestServlet, self).__init__() + super().__init__() self.hs = hs self.profile_handler = hs.get_profile_handler() self.http_client = hs.get_simple_http_client() @@ -91,7 +91,7 @@ class ProfileAvatarURLRestServlet(RestServlet): PATTERNS = client_patterns("/profile/(?P<user_id>[^/]*)/avatar_url", v1=True) def __init__(self, hs): - super(ProfileAvatarURLRestServlet, self).__init__() + super().__init__() self.hs = hs self.profile_handler = hs.get_profile_handler() self.http_client = hs.get_simple_http_client() @@ -159,7 +159,7 @@ class ProfileRestServlet(RestServlet): PATTERNS = client_patterns("/profile/(?P<user_id>[^/]*)", v1=True) def __init__(self, hs): - super(ProfileRestServlet, self).__init__() + super().__init__() self.hs = hs self.profile_handler = hs.get_profile_handler() self.auth = hs.get_auth() diff --git a/synapse/rest/client/v1/push_rule.py b/synapse/rest/client/v1/push_rule.py
index 9fd4908136..f9eecb7cf5 100644 --- a/synapse/rest/client/v1/push_rule.py +++ b/synapse/rest/client/v1/push_rule.py
@@ -13,7 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. - from synapse.api.errors import ( NotFoundError, StoreError, @@ -25,7 +24,7 @@ from synapse.http.servlet import ( parse_json_value_from_request, parse_string, ) -from synapse.push.baserules import BASE_RULE_IDS +from synapse.push.baserules import BASE_RULE_IDS, NEW_RULE_IDS from synapse.push.clientformat import format_push_rules_for_user from synapse.push.rulekinds import PRIORITY_CLASS_MAP from synapse.rest.client.v2_alpha._base import client_patterns @@ -39,12 +38,14 @@ class PushRuleRestServlet(RestServlet): ) def __init__(self, hs): - super(PushRuleRestServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.notifier = hs.get_notifier() self._is_worker = hs.config.worker_app is not None + self._users_new_default_push_rules = hs.config.users_new_default_push_rules + async def on_PUT(self, request, path): if self._is_worker: raise Exception("Cannot handle PUT /push_rules on worker") @@ -158,10 +159,22 @@ class PushRuleRestServlet(RestServlet): return 200, {} def notify_user(self, user_id): - stream_id, _ = self.store.get_push_rules_stream_token() + stream_id = self.store.get_max_push_rules_stream_id() self.notifier.on_new_event("push_rules_key", stream_id, users=[user_id]) - def set_rule_attr(self, user_id, spec, val): + async def set_rule_attr(self, user_id, spec, val): + if spec["attr"] not in ("enabled", "actions"): + # for the sake of potential future expansion, shouldn't report + # 404 in the case of an unknown request so check it corresponds to + # a known attribute first. + raise UnrecognizedRequestError() + + namespaced_rule_id = _namespaced_rule_id_from_spec(spec) + rule_id = spec["rule_id"] + is_default_rule = rule_id.startswith(".") + if is_default_rule: + if namespaced_rule_id not in BASE_RULE_IDS: + raise NotFoundError("Unknown rule %s" % (namespaced_rule_id,)) if spec["attr"] == "enabled": if isinstance(val, dict) and "enabled" in val: val = val["enabled"] @@ -170,8 +183,9 @@ class PushRuleRestServlet(RestServlet): # This should *actually* take a dict, but many clients pass # bools directly, so let's not break them. raise SynapseError(400, "Value for 'enabled' must be boolean") - namespaced_rule_id = _namespaced_rule_id_from_spec(spec) - return self.store.set_push_rule_enabled(user_id, namespaced_rule_id, val) + return await self.store.set_push_rule_enabled( + user_id, namespaced_rule_id, val, is_default_rule + ) elif spec["attr"] == "actions": actions = val.get("actions") _check_actions(actions) @@ -179,9 +193,14 @@ class PushRuleRestServlet(RestServlet): rule_id = spec["rule_id"] is_default_rule = rule_id.startswith(".") if is_default_rule: - if namespaced_rule_id not in BASE_RULE_IDS: + if user_id in self._users_new_default_push_rules: + rule_ids = NEW_RULE_IDS + else: + rule_ids = BASE_RULE_IDS + + if namespaced_rule_id not in rule_ids: raise SynapseError(404, "Unknown rule %r" % (namespaced_rule_id,)) - return self.store.set_push_rule_actions( + return await self.store.set_push_rule_actions( user_id, namespaced_rule_id, actions, is_default_rule ) else: diff --git a/synapse/rest/client/v1/pusher.py b/synapse/rest/client/v1/pusher.py
index 5f65cb7d83..28dabf1c7a 100644 --- a/synapse/rest/client/v1/pusher.py +++ b/synapse/rest/client/v1/pusher.py
@@ -44,7 +44,7 @@ class PushersRestServlet(RestServlet): PATTERNS = client_patterns("/pushers$", v1=True) def __init__(self, hs): - super(PushersRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() @@ -68,7 +68,7 @@ class PushersSetRestServlet(RestServlet): PATTERNS = client_patterns("/pushers/set$", v1=True) def __init__(self, hs): - super(PushersSetRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.notifier = hs.get_notifier() @@ -153,7 +153,7 @@ class PushersRemoveRestServlet(RestServlet): SUCCESS_HTML = b"<html><body>You have been unsubscribed</body><html>" def __init__(self, hs): - super(PushersRemoveRestServlet, self).__init__() + super().__init__() self.hs = hs self.notifier = hs.get_notifier() self.auth = hs.get_auth() diff --git a/synapse/rest/client/v1/room.py b/synapse/rest/client/v1/room.py
index 1a3398316d..b421fe855e 100644 --- a/synapse/rest/client/v1/room.py +++ b/synapse/rest/client/v1/room.py
@@ -21,14 +21,13 @@ import re from typing import List, Optional from urllib import parse as urlparse -from canonicaljson import json - from synapse.api.constants import EventTypes, Membership from synapse.api.errors import ( AuthError, Codes, HttpResponseException, InvalidClientCredentialsError, + ShadowBanError, SynapseError, ) from synapse.api.filtering import Filter @@ -46,6 +45,8 @@ from synapse.rest.client.v2_alpha._base import client_patterns from synapse.storage.state import StateFilter from synapse.streams.config import PaginationConfig from synapse.types import RoomAlias, RoomID, StreamToken, ThirdPartyInstanceID, UserID +from synapse.util import json_decoder +from synapse.util.stringutils import random_string MYPY = False if MYPY: @@ -56,7 +57,7 @@ logger = logging.getLogger(__name__) class TransactionRestServlet(RestServlet): def __init__(self, hs): - super(TransactionRestServlet, self).__init__() + super().__init__() self.txns = HttpTransactionCache(hs) @@ -64,7 +65,7 @@ class RoomCreateRestServlet(TransactionRestServlet): # No PATTERN; we have custom dispatch rules here def __init__(self, hs): - super(RoomCreateRestServlet, self).__init__(hs) + super().__init__(hs) self._room_creation_handler = hs.get_room_creation_handler() self.auth = hs.get_auth() @@ -110,7 +111,7 @@ class RoomCreateRestServlet(TransactionRestServlet): # TODO: Needs unit testing for generic events class RoomStateEventRestServlet(TransactionRestServlet): def __init__(self, hs): - super(RoomStateEventRestServlet, self).__init__(hs) + super().__init__(hs) self.handlers = hs.get_handlers() self.event_creation_handler = hs.get_event_creation_handler() self.room_member_handler = hs.get_room_member_handler() @@ -170,7 +171,6 @@ class RoomStateEventRestServlet(TransactionRestServlet): room_id=room_id, event_type=event_type, state_key=state_key, - is_guest=requester.is_guest, ) if not data: @@ -200,23 +200,26 @@ class RoomStateEventRestServlet(TransactionRestServlet): if state_key is not None: event_dict["state_key"] = state_key - if event_type == EventTypes.Member: - membership = content.get("membership", None) - event_id, _ = await self.room_member_handler.update_membership( - requester, - target=UserID.from_string(state_key), - room_id=room_id, - action=membership, - content=content, - ) - else: - ( - event, - _, - ) = await self.event_creation_handler.create_and_send_nonmember_event( - requester, event_dict, txn_id=txn_id - ) - event_id = event.event_id + try: + if event_type == EventTypes.Member: + membership = content.get("membership", None) + event_id, _ = await self.room_member_handler.update_membership( + requester, + target=UserID.from_string(state_key), + room_id=room_id, + action=membership, + content=content, + ) + else: + ( + event, + _, + ) = await self.event_creation_handler.create_and_send_nonmember_event( + requester, event_dict, txn_id=txn_id + ) + event_id = event.event_id + except ShadowBanError: + event_id = "$" + random_string(43) set_tag("event_id", event_id) ret = {"event_id": event_id} @@ -226,7 +229,7 @@ class RoomStateEventRestServlet(TransactionRestServlet): # TODO: Needs unit testing for generic events + feedback class RoomSendEventRestServlet(TransactionRestServlet): def __init__(self, hs): - super(RoomSendEventRestServlet, self).__init__(hs) + super().__init__(hs) self.event_creation_handler = hs.get_event_creation_handler() self.auth = hs.get_auth() @@ -249,12 +252,19 @@ class RoomSendEventRestServlet(TransactionRestServlet): if b"ts" in request.args and requester.app_service: event_dict["origin_server_ts"] = parse_integer(request, "ts", 0) - event, _ = await self.event_creation_handler.create_and_send_nonmember_event( - requester, event_dict, txn_id=txn_id - ) + try: + ( + event, + _, + ) = await self.event_creation_handler.create_and_send_nonmember_event( + requester, event_dict, txn_id=txn_id + ) + event_id = event.event_id + except ShadowBanError: + event_id = "$" + random_string(43) - set_tag("event_id", event.event_id) - return 200, {"event_id": event.event_id} + set_tag("event_id", event_id) + return 200, {"event_id": event_id} def on_GET(self, request, room_id, event_type, txn_id): return 200, "Not implemented" @@ -270,7 +280,7 @@ class RoomSendEventRestServlet(TransactionRestServlet): # TODO: Needs unit testing for room ID + alias joins class JoinRoomAliasServlet(TransactionRestServlet): def __init__(self, hs): - super(JoinRoomAliasServlet, self).__init__(hs) + super().__init__(hs) self.room_member_handler = hs.get_room_member_handler() self.auth = hs.get_auth() @@ -333,7 +343,7 @@ class PublicRoomListRestServlet(TransactionRestServlet): PATTERNS = client_patterns("/publicRooms$", v1=True) def __init__(self, hs): - super(PublicRoomListRestServlet, self).__init__(hs) + super().__init__(hs) self.hs = hs self.auth = hs.get_auth() @@ -438,13 +448,14 @@ class RoomMemberListRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/members$", v1=True) def __init__(self, hs): - super(RoomMemberListRestServlet, self).__init__() + super().__init__() self.message_handler = hs.get_message_handler() self.auth = hs.get_auth() + self.store = hs.get_datastore() async def on_GET(self, request, room_id): # TODO support Pagination stream API (limit/tokens) - requester = await self.auth.get_user_by_req(request) + requester = await self.auth.get_user_by_req(request, allow_guest=True) handler = self.message_handler # request the state as of a given event, as identified by a stream token, @@ -455,7 +466,7 @@ class RoomMemberListRestServlet(RestServlet): if at_token_string is None: at_token = None else: - at_token = StreamToken.from_string(at_token_string) + at_token = await StreamToken.from_string(self.store, at_token_string) # let you filter down on particular memberships. # XXX: this may not be the best shape for this API - we could pass in a filter @@ -489,7 +500,7 @@ class JoinedRoomMemberListRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/joined_members$", v1=True) def __init__(self, hs): - super(JoinedRoomMemberListRestServlet, self).__init__() + super().__init__() self.message_handler = hs.get_message_handler() self.auth = hs.get_auth() @@ -508,18 +519,23 @@ class RoomMessageListRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/messages$", v1=True) def __init__(self, hs): - super(RoomMessageListRestServlet, self).__init__() + super().__init__() self.pagination_handler = hs.get_pagination_handler() self.auth = hs.get_auth() + self.store = hs.get_datastore() async def on_GET(self, request, room_id): requester = await self.auth.get_user_by_req(request, allow_guest=True) - pagination_config = PaginationConfig.from_request(request, default_limit=10) + pagination_config = await PaginationConfig.from_request( + self.store, request, default_limit=10 + ) as_client_event = b"raw" not in request.args filter_str = parse_string(request, b"filter", encoding="utf-8") if filter_str: filter_json = urlparse.unquote(filter_str) - event_filter = Filter(json.loads(filter_json)) # type: Optional[Filter] + event_filter = Filter( + json_decoder.decode(filter_json) + ) # type: Optional[Filter] if ( event_filter and event_filter.filter_json.get("event_format", "client") @@ -545,7 +561,7 @@ class RoomStateRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/state$", v1=True) def __init__(self, hs): - super(RoomStateRestServlet, self).__init__() + super().__init__() self.message_handler = hs.get_message_handler() self.auth = hs.get_auth() @@ -565,13 +581,14 @@ class RoomInitialSyncRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/initialSync$", v1=True) def __init__(self, hs): - super(RoomInitialSyncRestServlet, self).__init__() + super().__init__() self.initial_sync_handler = hs.get_initial_sync_handler() self.auth = hs.get_auth() + self.store = hs.get_datastore() async def on_GET(self, request, room_id): requester = await self.auth.get_user_by_req(request, allow_guest=True) - pagination_config = PaginationConfig.from_request(request) + pagination_config = await PaginationConfig.from_request(self.store, request) content = await self.initial_sync_handler.room_initial_sync( room_id=room_id, requester=requester, pagin_config=pagination_config ) @@ -584,7 +601,7 @@ class RoomEventServlet(RestServlet): ) def __init__(self, hs): - super(RoomEventServlet, self).__init__() + super().__init__() self.clock = hs.get_clock() self.event_handler = hs.get_event_handler() self._event_serializer = hs.get_event_client_serializer() @@ -616,7 +633,7 @@ class RoomEventContextServlet(RestServlet): ) def __init__(self, hs): - super(RoomEventContextServlet, self).__init__() + super().__init__() self.clock = hs.get_clock() self.room_context_handler = hs.get_room_context_handler() self._event_serializer = hs.get_event_client_serializer() @@ -631,7 +648,9 @@ class RoomEventContextServlet(RestServlet): filter_str = parse_string(request, b"filter", encoding="utf-8") if filter_str: filter_json = urlparse.unquote(filter_str) - event_filter = Filter(json.loads(filter_json)) # type: Optional[Filter] + event_filter = Filter( + json_decoder.decode(filter_json) + ) # type: Optional[Filter] else: event_filter = None @@ -661,7 +680,7 @@ class RoomEventContextServlet(RestServlet): class RoomForgetRestServlet(TransactionRestServlet): def __init__(self, hs): - super(RoomForgetRestServlet, self).__init__(hs) + super().__init__(hs) self.room_member_handler = hs.get_room_member_handler() self.auth = hs.get_auth() @@ -687,7 +706,7 @@ class RoomForgetRestServlet(TransactionRestServlet): # TODO: Needs unit testing class RoomMembershipRestServlet(TransactionRestServlet): def __init__(self, hs): - super(RoomMembershipRestServlet, self).__init__(hs) + super().__init__(hs) self.room_member_handler = hs.get_room_member_handler() self.auth = hs.get_auth() @@ -716,17 +735,21 @@ class RoomMembershipRestServlet(TransactionRestServlet): content = {} if membership_action == "invite" and self._has_3pid_invite_keys(content): - await self.room_member_handler.do_3pid_invite( - room_id, - requester.user, - content["medium"], - content["address"], - content["id_server"], - requester, - txn_id, - new_room=False, - id_access_token=content.get("id_access_token"), - ) + try: + await self.room_member_handler.do_3pid_invite( + room_id, + requester.user, + content["medium"], + content["address"], + content["id_server"], + requester, + txn_id, + new_room=False, + id_access_token=content.get("id_access_token"), + ) + except ShadowBanError: + # Pretend the request succeeded. + pass return 200, {} target = requester.user @@ -738,15 +761,19 @@ class RoomMembershipRestServlet(TransactionRestServlet): if "reason" in content: event_content = {"reason": content["reason"]} - await self.room_member_handler.update_membership( - requester=requester, - target=target, - room_id=room_id, - action=membership_action, - txn_id=txn_id, - third_party_signed=content.get("third_party_signed", None), - content=event_content, - ) + try: + await self.room_member_handler.update_membership( + requester=requester, + target=target, + room_id=room_id, + action=membership_action, + txn_id=txn_id, + third_party_signed=content.get("third_party_signed", None), + content=event_content, + ) + except ShadowBanError: + # Pretend the request succeeded. + pass return_value = {} @@ -771,7 +798,7 @@ class RoomMembershipRestServlet(TransactionRestServlet): class RoomRedactEventRestServlet(TransactionRestServlet): def __init__(self, hs): - super(RoomRedactEventRestServlet, self).__init__(hs) + super().__init__(hs) self.handlers = hs.get_handlers() self.event_creation_handler = hs.get_event_creation_handler() self.auth = hs.get_auth() @@ -784,20 +811,27 @@ class RoomRedactEventRestServlet(TransactionRestServlet): requester = await self.auth.get_user_by_req(request) content = parse_json_object_from_request(request) - event, _ = await self.event_creation_handler.create_and_send_nonmember_event( - requester, - { - "type": EventTypes.Redaction, - "content": content, - "room_id": room_id, - "sender": requester.user.to_string(), - "redacts": event_id, - }, - txn_id=txn_id, - ) + try: + ( + event, + _, + ) = await self.event_creation_handler.create_and_send_nonmember_event( + requester, + { + "type": EventTypes.Redaction, + "content": content, + "room_id": room_id, + "sender": requester.user.to_string(), + "redacts": event_id, + }, + txn_id=txn_id, + ) + event_id = event.event_id + except ShadowBanError: + event_id = "$" + random_string(43) - set_tag("event_id", event.event_id) - return 200, {"event_id": event.event_id} + set_tag("event_id", event_id) + return 200, {"event_id": event_id} def on_PUT(self, request, room_id, event_id, txn_id): set_tag("txn_id", txn_id) @@ -813,7 +847,7 @@ class RoomTypingRestServlet(RestServlet): ) def __init__(self, hs): - super(RoomTypingRestServlet, self).__init__() + super().__init__() self.presence_handler = hs.get_presence_handler() self.typing_handler = hs.get_typing_handler() self.auth = hs.get_auth() @@ -840,17 +874,21 @@ class RoomTypingRestServlet(RestServlet): # Limit timeout to stop people from setting silly typing timeouts. timeout = min(content.get("timeout", 30000), 120000) - if content["typing"]: - await self.typing_handler.started_typing( - target_user=target_user, - auth_user=requester.user, - room_id=room_id, - timeout=timeout, - ) - else: - await self.typing_handler.stopped_typing( - target_user=target_user, auth_user=requester.user, room_id=room_id - ) + try: + if content["typing"]: + await self.typing_handler.started_typing( + target_user=target_user, + requester=requester, + room_id=room_id, + timeout=timeout, + ) + else: + await self.typing_handler.stopped_typing( + target_user=target_user, requester=requester, room_id=room_id + ) + except ShadowBanError: + # Pretend this worked without error. + pass return 200, {} @@ -882,7 +920,7 @@ class SearchRestServlet(RestServlet): PATTERNS = client_patterns("/search$", v1=True) def __init__(self, hs): - super(SearchRestServlet, self).__init__() + super().__init__() self.handlers = hs.get_handlers() self.auth = hs.get_auth() @@ -903,7 +941,7 @@ class JoinedRoomsRestServlet(RestServlet): PATTERNS = client_patterns("/joined_rooms$", v1=True) def __init__(self, hs): - super(JoinedRoomsRestServlet, self).__init__() + super().__init__() self.store = hs.get_datastore() self.auth = hs.get_auth() diff --git a/synapse/rest/client/v1/voip.py b/synapse/rest/client/v1/voip.py
index 50277c6cf6..b8d491ca5c 100644 --- a/synapse/rest/client/v1/voip.py +++ b/synapse/rest/client/v1/voip.py
@@ -25,7 +25,7 @@ class VoipRestServlet(RestServlet): PATTERNS = client_patterns("/voip/turnServer$", v1=True) def __init__(self, hs): - super(VoipRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() diff --git a/synapse/rest/client/v2_alpha/account.py b/synapse/rest/client/v2_alpha/account.py
index d4b1ee1e8c..1320aad8f6 100644 --- a/synapse/rest/client/v2_alpha/account.py +++ b/synapse/rest/client/v2_alpha/account.py
@@ -15,6 +15,7 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +import random import re from http import HTTPStatus from typing import TYPE_CHECKING @@ -23,10 +24,13 @@ from urllib.parse import urlparse if TYPE_CHECKING: from synapse.app.homeserver import HomeServer -from twisted.internet import defer - from synapse.api.constants import LoginType -from synapse.api.errors import Codes, SynapseError, ThreepidValidationError +from synapse.api.errors import ( + Codes, + InteractiveAuthIncompleteError, + SynapseError, + ThreepidValidationError, +) from synapse.config.emailconfig import ThreepidBehaviour from synapse.http.server import finish_request, respond_with_html from synapse.http.servlet import ( @@ -35,7 +39,7 @@ from synapse.http.servlet import ( parse_json_object_from_request, parse_string, ) -from synapse.push.mailer import Mailer, load_jinja2_templates +from synapse.push.mailer import Mailer from synapse.types import UserID from synapse.util.msisdn import phone_number_to_msisdn from synapse.util.stringutils import assert_valid_client_secret, random_string @@ -50,28 +54,18 @@ class EmailPasswordRequestTokenRestServlet(RestServlet): PATTERNS = client_patterns("/account/password/email/requestToken$") def __init__(self, hs): - super(EmailPasswordRequestTokenRestServlet, self).__init__() + super().__init__() self.hs = hs self.datastore = hs.get_datastore() self.config = hs.config self.identity_handler = hs.get_handlers().identity_handler if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - template_html, template_text = load_jinja2_templates( - self.config.email_template_dir, - [ - self.config.email_password_reset_template_html, - self.config.email_password_reset_template_text, - ], - apply_format_ts_filter=True, - apply_mxc_to_http_filter=True, - public_baseurl=self.config.public_baseurl, - ) self.mailer = Mailer( hs=self.hs, app_name=self.config.email_app_name, - template_html=template_html, - template_text=template_text, + template_html=self.config.email_password_reset_template_html, + template_text=self.config.email_password_reset_template_text, ) async def on_POST(self, request): @@ -104,13 +98,6 @@ class EmailPasswordRequestTokenRestServlet(RestServlet): send_attempt = body["send_attempt"] next_link = body.get("next_link") # Optional param - if not check_3pid_allowed(self.hs, "email", email): - raise SynapseError( - 403, - "Your email is not authorized on this server", - Codes.THREEPID_DENIED, - ) - if next_link: # Raise if the provided next_link value isn't valid assert_valid_next_link(self.hs, next_link) @@ -127,6 +114,9 @@ class EmailPasswordRequestTokenRestServlet(RestServlet): if self.config.request_token_inhibit_3pid_errors: # Make the client think the operation succeeded. See the rationale in the # comments for request_token_inhibit_3pid_errors. + # Also wait for some random amount of time between 100ms and 1s to make it + # look like we did something. + await self.hs.clock.sleep(random.randint(1, 10) / 10) return 200, {"sid": random_string(16)} raise SynapseError(400, "Email not found", Codes.THREEPID_NOT_FOUND) @@ -158,87 +148,11 @@ class EmailPasswordRequestTokenRestServlet(RestServlet): return 200, ret -class PasswordResetSubmitTokenServlet(RestServlet): - """Handles 3PID validation token submission""" - - PATTERNS = client_patterns( - "/password_reset/(?P<medium>[^/]*)/submit_token$", releases=(), unstable=True - ) - - def __init__(self, hs): - """ - Args: - hs (synapse.server.HomeServer): server - """ - super(PasswordResetSubmitTokenServlet, self).__init__() - self.hs = hs - self.auth = hs.get_auth() - self.config = hs.config - self.clock = hs.get_clock() - self.store = hs.get_datastore() - if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - (self.failure_email_template,) = load_jinja2_templates( - self.config.email_template_dir, - [self.config.email_password_reset_template_failure_html], - ) - - async def on_GET(self, request, medium): - # We currently only handle threepid token submissions for email - if medium != "email": - raise SynapseError( - 400, "This medium is currently not supported for password resets" - ) - if self.config.threepid_behaviour_email == ThreepidBehaviour.OFF: - if self.config.local_threepid_handling_disabled_due_to_email_config: - logger.warning( - "Password reset emails have been disabled due to lack of an email config" - ) - raise SynapseError( - 400, "Email-based password resets are disabled on this server" - ) - - sid = parse_string(request, "sid", required=True) - token = parse_string(request, "token", required=True) - client_secret = parse_string(request, "client_secret", required=True) - assert_valid_client_secret(client_secret) - - # Attempt to validate a 3PID session - try: - # Mark the session as valid - next_link = await self.store.validate_threepid_session( - sid, client_secret, token, self.clock.time_msec() - ) - - # Perform a 302 redirect if next_link is set - if next_link: - if next_link.startswith("file:///"): - logger.warning( - "Not redirecting to next_link as it is a local file: address" - ) - else: - request.setResponseCode(302) - request.setHeader("Location", next_link) - finish_request(request) - return None - - # Otherwise show the success template - html = self.config.email_password_reset_template_success_html - status_code = 200 - except ThreepidValidationError as e: - status_code = e.code - - # Show a failure page with a reason - template_vars = {"failure_reason": e.msg} - html = self.failure_email_template.render(**template_vars) - - respond_with_html(request, status_code, html) - - class PasswordRestServlet(RestServlet): PATTERNS = client_patterns("/account/password$") def __init__(self, hs): - super(PasswordRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.auth_handler = hs.get_auth_handler() @@ -253,18 +167,12 @@ class PasswordRestServlet(RestServlet): # we do basic sanity checks here because the auth layer will store these # in sessions. Pull out the new password provided to us. - if "new_password" in body: - new_password = body.pop("new_password") + new_password = body.pop("new_password", None) + if new_password is not None: if not isinstance(new_password, str) or len(new_password) > 512: raise SynapseError(400, "Invalid password") self.password_policy_handler.validate_password(new_password) - # If the password is valid, hash it and store it back on the body. - # This ensures that only the hashed password is handled everywhere. - if "new_password_hash" in body: - raise SynapseError(400, "Unexpected property: new_password_hash") - body["new_password_hash"] = await self.auth_handler.hash(new_password) - # there are two possibilities here. Either the user does not have an # access token, and needs to do a password reset; or they have one and # need to validate their identity. @@ -281,23 +189,52 @@ class PasswordRestServlet(RestServlet): if requester.app_service: params = body else: - params = await self.auth_handler.validate_user_via_ui_auth( - requester, + try: + ( + params, + session_id, + ) = await self.auth_handler.validate_user_via_ui_auth( + requester, + request, + body, + self.hs.get_ip_from_request(request), + "modify your account password", + ) + except InteractiveAuthIncompleteError as e: + # The user needs to provide more steps to complete auth, but + # they're not required to provide the password again. + # + # If a password is available now, hash the provided password and + # store it for later. + if new_password: + password_hash = await self.auth_handler.hash(new_password) + await self.auth_handler.set_session_data( + e.session_id, "password_hash", password_hash + ) + raise + user_id = requester.user.to_string() + else: + requester = None + try: + result, params, session_id = await self.auth_handler.check_ui_auth( + [[LoginType.EMAIL_IDENTITY]], request, body, self.hs.get_ip_from_request(request), "modify your account password", ) - user_id = requester.user.to_string() - else: - requester = None - result, params, _ = await self.auth_handler.check_auth( - [[LoginType.EMAIL_IDENTITY]], - request, - body, - self.hs.get_ip_from_request(request), - "modify your account password", - ) + except InteractiveAuthIncompleteError as e: + # The user needs to provide more steps to complete auth, but + # they're not required to provide the password again. + # + # If a password is available now, hash the provided password and + # store it for later. + if new_password: + password_hash = await self.auth_handler.hash(new_password) + await self.auth_handler.set_session_data( + e.session_id, "password_hash", password_hash + ) + raise if LoginType.EMAIL_IDENTITY in result: threepid = result[LoginType.EMAIL_IDENTITY] @@ -322,32 +259,40 @@ class PasswordRestServlet(RestServlet): logger.error("Auth succeeded but no known type! %r", result.keys()) raise SynapseError(500, "", Codes.UNKNOWN) - assert_params_in_dict(params, ["new_password_hash"]) - new_password_hash = params["new_password_hash"] + # If we have a password in this request, prefer it. Otherwise, there + # must be a password hash from an earlier request. + if new_password: + password_hash = await self.auth_handler.hash(new_password) + else: + password_hash = await self.auth_handler.get_session_data( + session_id, "password_hash", None + ) + if not password_hash: + raise SynapseError(400, "Missing params: password", Codes.MISSING_PARAM) + logout_devices = params.get("logout_devices", True) await self._set_password_handler.set_password( - user_id, new_password_hash, logout_devices, requester + user_id, password_hash, logout_devices, requester ) if self.hs.config.shadow_server: shadow_user = UserID( requester.user.localpart, self.hs.config.shadow_server.get("hs") ) - self.shadow_password(params, shadow_user.to_string()) + await self.shadow_password(params, shadow_user.to_string()) return 200, {} def on_OPTIONS(self, _): return 200, {} - @defer.inlineCallbacks - def shadow_password(self, body, user_id): + async def shadow_password(self, body, user_id): # TODO: retries shadow_hs_url = self.hs.config.shadow_server.get("hs_url") as_token = self.hs.config.shadow_server.get("as_token") - yield self.http_client.post_json_get_json( + await self.http_client.post_json_get_json( "%s/_matrix/client/r0/account/password?access_token=%s&user_id=%s" % (shadow_hs_url, as_token, user_id), body, @@ -358,7 +303,7 @@ class DeactivateAccountRestServlet(RestServlet): PATTERNS = client_patterns("/account/deactivate$") def __init__(self, hs): - super(DeactivateAccountRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.auth_handler = hs.get_auth_handler() @@ -377,7 +322,7 @@ class DeactivateAccountRestServlet(RestServlet): requester = await self.auth.get_user_by_req(request) - # allow ASes to dectivate their own users + # allow ASes to deactivate their own users if requester.app_service: await self._deactivate_account_handler.deactivate_account( requester.user.to_string(), erase @@ -406,26 +351,18 @@ class EmailThreepidRequestTokenRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid/email/requestToken$") def __init__(self, hs): - super(EmailThreepidRequestTokenRestServlet, self).__init__() + super().__init__() self.hs = hs self.config = hs.config self.identity_handler = hs.get_handlers().identity_handler self.store = self.hs.get_datastore() if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - template_html, template_text = load_jinja2_templates( - self.config.email_template_dir, - [ - self.config.email_add_threepid_template_html, - self.config.email_add_threepid_template_text, - ], - public_baseurl=self.config.public_baseurl, - ) self.mailer = Mailer( hs=self.hs, app_name=self.config.email_app_name, - template_html=template_html, - template_text=template_text, + template_html=self.config.email_add_threepid_template_html, + template_text=self.config.email_add_threepid_template_text, ) async def on_POST(self, request): @@ -473,6 +410,9 @@ class EmailThreepidRequestTokenRestServlet(RestServlet): if self.config.request_token_inhibit_3pid_errors: # Make the client think the operation succeeded. See the rationale in the # comments for request_token_inhibit_3pid_errors. + # Also wait for some random amount of time between 100ms and 1s to make it + # look like we did something. + await self.hs.clock.sleep(random.randint(1, 10) / 10) return 200, {"sid": random_string(16)} raise SynapseError(400, "Email is already in use", Codes.THREEPID_IN_USE) @@ -509,7 +449,7 @@ class MsisdnThreepidRequestTokenRestServlet(RestServlet): def __init__(self, hs): self.hs = hs - super(MsisdnThreepidRequestTokenRestServlet, self).__init__() + super().__init__() self.store = self.hs.get_datastore() self.identity_handler = hs.get_handlers().identity_handler @@ -545,6 +485,9 @@ class MsisdnThreepidRequestTokenRestServlet(RestServlet): if self.hs.config.request_token_inhibit_3pid_errors: # Make the client think the operation succeeded. See the rationale in the # comments for request_token_inhibit_3pid_errors. + # Also wait for some random amount of time between 100ms and 1s to make it + # look like we did something. + await self.hs.clock.sleep(random.randint(1, 10) / 10) return 200, {"sid": random_string(16)} raise SynapseError(400, "MSISDN is already in use", Codes.THREEPID_IN_USE) @@ -588,9 +531,8 @@ class AddThreepidEmailSubmitTokenServlet(RestServlet): self.clock = hs.get_clock() self.store = hs.get_datastore() if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - (self.failure_email_template,) = load_jinja2_templates( - self.config.email_template_dir, - [self.config.email_add_threepid_template_failure_html], + self._failure_email_template = ( + self.config.email_add_threepid_template_failure_html ) async def on_GET(self, request): @@ -636,7 +578,7 @@ class AddThreepidEmailSubmitTokenServlet(RestServlet): # Show a failure page with a reason template_vars = {"failure_reason": e.msg} - html = self.failure_email_template.render(**template_vars) + html = self._failure_email_template.render(**template_vars) respond_with_html(request, status_code, html) @@ -687,7 +629,7 @@ class ThreepidRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid$") def __init__(self, hs): - super(ThreepidRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler self.auth = hs.get_auth() @@ -731,7 +673,7 @@ class ThreepidRestServlet(RestServlet): shadow_user = UserID( requester.user.localpart, self.hs.config.shadow_server.get("hs") ) - self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) + await self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) return 200, {} @@ -766,7 +708,7 @@ class ThreepidRestServlet(RestServlet): "address": validation_session["address"], "validated_at": validation_session["validated_at"], } - self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) + await self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) return 200, {} @@ -774,13 +716,12 @@ class ThreepidRestServlet(RestServlet): 400, "No validated 3pid session found", Codes.THREEPID_AUTH_FAILED ) - @defer.inlineCallbacks - def shadow_3pid(self, body, user_id): + async def shadow_3pid(self, body, user_id): # TODO: retries shadow_hs_url = self.hs.config.shadow_server.get("hs_url") as_token = self.hs.config.shadow_server.get("as_token") - yield self.http_client.post_json_get_json( + await self.http_client.post_json_get_json( "%s/_matrix/client/r0/account/3pid?access_token=%s&user_id=%s" % (shadow_hs_url, as_token, user_id), body, @@ -791,7 +732,7 @@ class ThreepidAddRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid/add$") def __init__(self, hs): - super(ThreepidAddRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler self.auth = hs.get_auth() @@ -841,20 +782,19 @@ class ThreepidAddRestServlet(RestServlet): "address": validation_session["address"], "validated_at": validation_session["validated_at"], } - self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) + await self.shadow_3pid({"threepid": threepid}, shadow_user.to_string()) return 200, {} raise SynapseError( 400, "No validated 3pid session found", Codes.THREEPID_AUTH_FAILED ) - @defer.inlineCallbacks - def shadow_3pid(self, body, user_id): + async def shadow_3pid(self, body, user_id): # TODO: retries shadow_hs_url = self.hs.config.shadow_server.get("hs_url") as_token = self.hs.config.shadow_server.get("as_token") - yield self.http_client.post_json_get_json( + await self.http_client.post_json_get_json( "%s/_matrix/client/r0/account/3pid?access_token=%s&user_id=%s" % (shadow_hs_url, as_token, user_id), body, @@ -865,7 +805,7 @@ class ThreepidBindRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid/bind$") def __init__(self, hs): - super(ThreepidBindRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler self.auth = hs.get_auth() @@ -894,7 +834,7 @@ class ThreepidUnbindRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid/unbind$") def __init__(self, hs): - super(ThreepidUnbindRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler self.auth = hs.get_auth() @@ -925,7 +865,7 @@ class ThreepidDeleteRestServlet(RestServlet): PATTERNS = client_patterns("/account/3pid/delete$") def __init__(self, hs): - super(ThreepidDeleteRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.auth_handler = hs.get_auth_handler() @@ -958,7 +898,7 @@ class ThreepidDeleteRestServlet(RestServlet): shadow_user = UserID( requester.user.localpart, self.hs.config.shadow_server.get("hs") ) - self.shadow_3pid_delete(body, shadow_user.to_string()) + await self.shadow_3pid_delete(body, shadow_user.to_string()) if ret: id_server_unbind_result = "success" @@ -967,13 +907,12 @@ class ThreepidDeleteRestServlet(RestServlet): return 200, {"id_server_unbind_result": id_server_unbind_result} - @defer.inlineCallbacks - def shadow_3pid_delete(self, body, user_id): + async def shadow_3pid_delete(self, body, user_id): # TODO: retries shadow_hs_url = self.hs.config.shadow_server.get("hs_url") as_token = self.hs.config.shadow_server.get("as_token") - yield self.http_client.post_json_get_json( + await self.http_client.post_json_get_json( "%s/_matrix/client/r0/account/3pid/delete?access_token=%s&user_id=%s" % (shadow_hs_url, as_token, user_id), body, @@ -988,12 +927,11 @@ class ThreepidLookupRestServlet(RestServlet): self.auth = hs.get_auth() self.identity_handler = hs.get_handlers().identity_handler - @defer.inlineCallbacks - def on_GET(self, request): + async def on_GET(self, request): """Proxy a /_matrix/identity/api/v1/lookup request to an identity server """ - yield self.auth.get_user_by_req(request) + await self.auth.get_user_by_req(request) # Verify query parameters query_params = request.args @@ -1006,9 +944,9 @@ class ThreepidLookupRestServlet(RestServlet): # Proxy the request to the identity server. lookup_3pid handles checking # if the lookup is allowed so we don't need to do it here. - ret = yield self.identity_handler.proxy_lookup_3pid(id_server, medium, address) + ret = await self.identity_handler.proxy_lookup_3pid(id_server, medium, address) - defer.returnValue((200, ret)) + return 200, ret class ThreepidBulkLookupRestServlet(RestServlet): @@ -1019,12 +957,11 @@ class ThreepidBulkLookupRestServlet(RestServlet): self.auth = hs.get_auth() self.identity_handler = hs.get_handlers().identity_handler - @defer.inlineCallbacks - def on_POST(self, request): + async def on_POST(self, request): """Proxy a /_matrix/identity/api/v1/bulk_lookup request to an identity server """ - yield self.auth.get_user_by_req(request) + await self.auth.get_user_by_req(request) body = parse_json_object_from_request(request) @@ -1032,11 +969,11 @@ class ThreepidBulkLookupRestServlet(RestServlet): # Proxy the request to the identity server. lookup_3pid handles checking # if the lookup is allowed so we don't need to do it here. - ret = yield self.identity_handler.proxy_bulk_lookup_3pid( + ret = await self.identity_handler.proxy_bulk_lookup_3pid( body["id_server"], body["threepids"] ) - defer.returnValue((200, ret)) + return 200, ret def assert_valid_next_link(hs: "HomeServer", next_link: str): @@ -1082,7 +1019,7 @@ class WhoamiRestServlet(RestServlet): PATTERNS = client_patterns("/account/whoami$") def __init__(self, hs): - super(WhoamiRestServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() async def on_GET(self, request): @@ -1093,7 +1030,6 @@ class WhoamiRestServlet(RestServlet): def register_servlets(hs, http_server): EmailPasswordRequestTokenRestServlet(hs).register(http_server) - PasswordResetSubmitTokenServlet(hs).register(http_server) PasswordRestServlet(hs).register(http_server) DeactivateAccountRestServlet(hs).register(http_server) EmailThreepidRequestTokenRestServlet(hs).register(http_server) diff --git a/synapse/rest/client/v2_alpha/account_data.py b/synapse/rest/client/v2_alpha/account_data.py
index d31ec7c29d..617ee6d62a 100644 --- a/synapse/rest/client/v2_alpha/account_data.py +++ b/synapse/rest/client/v2_alpha/account_data.py
@@ -35,7 +35,7 @@ class AccountDataServlet(RestServlet): ) def __init__(self, hs): - super(AccountDataServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.notifier = hs.get_notifier() @@ -93,7 +93,7 @@ class RoomAccountDataServlet(RestServlet): ) def __init__(self, hs): - super(RoomAccountDataServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.notifier = hs.get_notifier() diff --git a/synapse/rest/client/v2_alpha/account_validity.py b/synapse/rest/client/v2_alpha/account_validity.py
index d06336ceea..bd7f9ae203 100644 --- a/synapse/rest/client/v2_alpha/account_validity.py +++ b/synapse/rest/client/v2_alpha/account_validity.py
@@ -32,7 +32,7 @@ class AccountValidityRenewServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(AccountValidityRenewServlet, self).__init__() + super().__init__() self.hs = hs self.account_activity_handler = hs.get_account_validity_handler() @@ -67,7 +67,7 @@ class AccountValiditySendMailServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(AccountValiditySendMailServlet, self).__init__() + super().__init__() self.hs = hs self.account_activity_handler = hs.get_account_validity_handler() diff --git a/synapse/rest/client/v2_alpha/auth.py b/synapse/rest/client/v2_alpha/auth.py
index 8e585e9153..5fbfae5991 100644 --- a/synapse/rest/client/v2_alpha/auth.py +++ b/synapse/rest/client/v2_alpha/auth.py
@@ -25,94 +25,6 @@ from ._base import client_patterns logger = logging.getLogger(__name__) -RECAPTCHA_TEMPLATE = """ -<html> -<head> -<title>Authentication</title> -<meta name='viewport' content='width=device-width, initial-scale=1, - user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> -<script src="https://www.recaptcha.net/recaptcha/api.js" - async defer></script> -<script src="//code.jquery.com/jquery-1.11.2.min.js"></script> -<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> -<script> -function captchaDone() { - $('#registrationForm').submit(); -} -</script> -</head> -<body> -<form id="registrationForm" method="post" action="%(myurl)s"> - <div> - <p> - Hello! We need to prevent computer programs and other automated - things from creating accounts on this server. - </p> - <p> - Please verify that you're not a robot. - </p> - <input type="hidden" name="session" value="%(session)s" /> - <div class="g-recaptcha" - data-sitekey="%(sitekey)s" - data-callback="captchaDone"> - </div> - <noscript> - <input type="submit" value="All Done" /> - </noscript> - </div> - </div> -</form> -</body> -</html> -""" - -TERMS_TEMPLATE = """ -<html> -<head> -<title>Authentication</title> -<meta name='viewport' content='width=device-width, initial-scale=1, - user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> -<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> -</head> -<body> -<form id="registrationForm" method="post" action="%(myurl)s"> - <div> - <p> - Please click the button below if you agree to the - <a href="%(terms_url)s">privacy policy of this homeserver.</a> - </p> - <input type="hidden" name="session" value="%(session)s" /> - <input type="submit" value="Agree" /> - </div> -</form> -</body> -</html> -""" - -SUCCESS_TEMPLATE = """ -<html> -<head> -<title>Success!</title> -<meta name='viewport' content='width=device-width, initial-scale=1, - user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'> -<link rel="stylesheet" href="/_matrix/static/client/register/style.css"> -<script> -if (window.onAuthDone) { - window.onAuthDone(); -} else if (window.opener && window.opener.postMessage) { - window.opener.postMessage("authDone", "*"); -} -</script> -</head> -<body> - <div> - <p>Thank you</p> - <p>You may now close this window and return to the application</p> - </div> -</body> -</html> -""" - class AuthRestServlet(RestServlet): """ @@ -124,7 +36,7 @@ class AuthRestServlet(RestServlet): PATTERNS = client_patterns(r"/auth/(?P<stagetype>[\w\.]*)/fallback/web") def __init__(self, hs): - super(AuthRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.auth_handler = hs.get_auth_handler() @@ -145,26 +57,30 @@ class AuthRestServlet(RestServlet): self._cas_server_url = hs.config.cas_server_url self._cas_service_url = hs.config.cas_service_url + self.recaptcha_template = hs.config.recaptcha_template + self.terms_template = hs.config.terms_template + self.success_template = hs.config.fallback_success_template + async def on_GET(self, request, stagetype): session = parse_string(request, "session") if not session: raise SynapseError(400, "No session supplied") if stagetype == LoginType.RECAPTCHA: - html = RECAPTCHA_TEMPLATE % { - "session": session, - "myurl": "%s/r0/auth/%s/fallback/web" + html = self.recaptcha_template.render( + session=session, + myurl="%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.RECAPTCHA), - "sitekey": self.hs.config.recaptcha_public_key, - } + sitekey=self.hs.config.recaptcha_public_key, + ) elif stagetype == LoginType.TERMS: - html = TERMS_TEMPLATE % { - "session": session, - "terms_url": "%s_matrix/consent?v=%s" + html = self.terms_template.render( + session=session, + terms_url="%s_matrix/consent?v=%s" % (self.hs.config.public_baseurl, self.hs.config.user_consent_version), - "myurl": "%s/r0/auth/%s/fallback/web" + myurl="%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.TERMS), - } + ) elif stagetype == LoginType.SSO: # Display a confirmation page which prompts the user to @@ -222,14 +138,14 @@ class AuthRestServlet(RestServlet): ) if success: - html = SUCCESS_TEMPLATE + html = self.success_template.render() else: - html = RECAPTCHA_TEMPLATE % { - "session": session, - "myurl": "%s/r0/auth/%s/fallback/web" + html = self.recaptcha_template.render( + session=session, + myurl="%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.RECAPTCHA), - "sitekey": self.hs.config.recaptcha_public_key, - } + sitekey=self.hs.config.recaptcha_public_key, + ) elif stagetype == LoginType.TERMS: authdict = {"session": session} @@ -238,18 +154,18 @@ class AuthRestServlet(RestServlet): ) if success: - html = SUCCESS_TEMPLATE + html = self.success_template.render() else: - html = TERMS_TEMPLATE % { - "session": session, - "terms_url": "%s_matrix/consent?v=%s" + html = self.terms_template.render( + session=session, + terms_url="%s_matrix/consent?v=%s" % ( self.hs.config.public_baseurl, self.hs.config.user_consent_version, ), - "myurl": "%s/r0/auth/%s/fallback/web" + myurl="%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.TERMS), - } + ) elif stagetype == LoginType.SSO: # The SSO fallback workflow should not post here, raise SynapseError(404, "Fallback SSO auth does not support POST requests.") diff --git a/synapse/rest/client/v2_alpha/capabilities.py b/synapse/rest/client/v2_alpha/capabilities.py
index fe9d019c44..76879ac559 100644 --- a/synapse/rest/client/v2_alpha/capabilities.py +++ b/synapse/rest/client/v2_alpha/capabilities.py
@@ -32,7 +32,7 @@ class CapabilitiesRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(CapabilitiesRestServlet, self).__init__() + super().__init__() self.hs = hs self.config = hs.config self.auth = hs.get_auth() diff --git a/synapse/rest/client/v2_alpha/devices.py b/synapse/rest/client/v2_alpha/devices.py
index c0714fcfb1..7e174de692 100644 --- a/synapse/rest/client/v2_alpha/devices.py +++ b/synapse/rest/client/v2_alpha/devices.py
@@ -35,7 +35,7 @@ class DevicesRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(DevicesRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.device_handler = hs.get_device_handler() @@ -57,7 +57,7 @@ class DeleteDevicesRestServlet(RestServlet): PATTERNS = client_patterns("/delete_devices") def __init__(self, hs): - super(DeleteDevicesRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.device_handler = hs.get_device_handler() @@ -102,7 +102,7 @@ class DeviceRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(DeviceRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.device_handler = hs.get_device_handler() diff --git a/synapse/rest/client/v2_alpha/filter.py b/synapse/rest/client/v2_alpha/filter.py
index b28da017cd..7cc692643b 100644 --- a/synapse/rest/client/v2_alpha/filter.py +++ b/synapse/rest/client/v2_alpha/filter.py
@@ -28,7 +28,7 @@ class GetFilterRestServlet(RestServlet): PATTERNS = client_patterns("/user/(?P<user_id>[^/]*)/filter/(?P<filter_id>[^/]*)") def __init__(self, hs): - super(GetFilterRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.filtering = hs.get_filtering() @@ -64,7 +64,7 @@ class CreateFilterRestServlet(RestServlet): PATTERNS = client_patterns("/user/(?P<user_id>[^/]*)/filter") def __init__(self, hs): - super(CreateFilterRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.filtering = hs.get_filtering() diff --git a/synapse/rest/client/v2_alpha/groups.py b/synapse/rest/client/v2_alpha/groups.py
index d84a6d7e11..a3bb095c2d 100644 --- a/synapse/rest/client/v2_alpha/groups.py +++ b/synapse/rest/client/v2_alpha/groups.py
@@ -16,6 +16,7 @@ import logging +from synapse.api.errors import SynapseError from synapse.http.servlet import RestServlet, parse_json_object_from_request from synapse.types import GroupID @@ -31,7 +32,7 @@ class GroupServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/profile$") def __init__(self, hs): - super(GroupServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -65,7 +66,7 @@ class GroupSummaryServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/summary$") def __init__(self, hs): - super(GroupSummaryServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -96,7 +97,7 @@ class GroupSummaryRoomsCatServlet(RestServlet): ) def __init__(self, hs): - super(GroupSummaryRoomsCatServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -136,7 +137,7 @@ class GroupCategoryServlet(RestServlet): ) def __init__(self, hs): - super(GroupCategoryServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -180,7 +181,7 @@ class GroupCategoriesServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/categories/$") def __init__(self, hs): - super(GroupCategoriesServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -203,7 +204,7 @@ class GroupRoleServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/roles/(?P<role_id>[^/]+)$") def __init__(self, hs): - super(GroupRoleServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -247,7 +248,7 @@ class GroupRolesServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/roles/$") def __init__(self, hs): - super(GroupRolesServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -278,7 +279,7 @@ class GroupSummaryUsersRoleServlet(RestServlet): ) def __init__(self, hs): - super(GroupSummaryUsersRoleServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -316,7 +317,7 @@ class GroupRoomServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/rooms$") def __init__(self, hs): - super(GroupRoomServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -325,6 +326,9 @@ class GroupRoomServlet(RestServlet): requester = await self.auth.get_user_by_req(request, allow_guest=True) requester_user_id = requester.user.to_string() + if not GroupID.is_valid(group_id): + raise SynapseError(400, "%s was not legal group ID" % (group_id,)) + result = await self.groups_handler.get_rooms_in_group( group_id, requester_user_id ) @@ -339,7 +343,7 @@ class GroupUsersServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/users$") def __init__(self, hs): - super(GroupUsersServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -362,7 +366,7 @@ class GroupInvitedUsersServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/invited_users$") def __init__(self, hs): - super(GroupInvitedUsersServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -385,7 +389,7 @@ class GroupSettingJoinPolicyServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/settings/m.join_policy$") def __init__(self, hs): - super(GroupSettingJoinPolicyServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.groups_handler = hs.get_groups_local_handler() @@ -409,7 +413,7 @@ class GroupCreateServlet(RestServlet): PATTERNS = client_patterns("/create_group$") def __init__(self, hs): - super(GroupCreateServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -440,7 +444,7 @@ class GroupAdminRoomsServlet(RestServlet): ) def __init__(self, hs): - super(GroupAdminRoomsServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -477,7 +481,7 @@ class GroupAdminRoomsConfigServlet(RestServlet): ) def __init__(self, hs): - super(GroupAdminRoomsConfigServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -503,7 +507,7 @@ class GroupAdminUsersInviteServlet(RestServlet): ) def __init__(self, hs): - super(GroupAdminUsersInviteServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -532,7 +536,7 @@ class GroupAdminUsersKickServlet(RestServlet): ) def __init__(self, hs): - super(GroupAdminUsersKickServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -556,7 +560,7 @@ class GroupSelfLeaveServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/self/leave$") def __init__(self, hs): - super(GroupSelfLeaveServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -580,7 +584,7 @@ class GroupSelfJoinServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/self/join$") def __init__(self, hs): - super(GroupSelfJoinServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -604,7 +608,7 @@ class GroupSelfAcceptInviteServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/self/accept_invite$") def __init__(self, hs): - super(GroupSelfAcceptInviteServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() @@ -628,7 +632,7 @@ class GroupSelfUpdatePublicityServlet(RestServlet): PATTERNS = client_patterns("/groups/(?P<group_id>[^/]*)/self/update_publicity$") def __init__(self, hs): - super(GroupSelfUpdatePublicityServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.store = hs.get_datastore() @@ -651,7 +655,7 @@ class PublicisedGroupsForUserServlet(RestServlet): PATTERNS = client_patterns("/publicised_groups/(?P<user_id>[^/]*)$") def __init__(self, hs): - super(PublicisedGroupsForUserServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.store = hs.get_datastore() @@ -672,7 +676,7 @@ class PublicisedGroupsForUsersServlet(RestServlet): PATTERNS = client_patterns("/publicised_groups$") def __init__(self, hs): - super(PublicisedGroupsForUsersServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.store = hs.get_datastore() @@ -696,7 +700,7 @@ class GroupsForUserServlet(RestServlet): PATTERNS = client_patterns("/joined_groups$") def __init__(self, hs): - super(GroupsForUserServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.clock = hs.get_clock() self.groups_handler = hs.get_groups_local_handler() diff --git a/synapse/rest/client/v2_alpha/keys.py b/synapse/rest/client/v2_alpha/keys.py
index 24bb090822..55c4606569 100644 --- a/synapse/rest/client/v2_alpha/keys.py +++ b/synapse/rest/client/v2_alpha/keys.py
@@ -64,7 +64,7 @@ class KeyUploadServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(KeyUploadServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_keys_handler = hs.get_e2e_keys_handler() @@ -147,7 +147,7 @@ class KeyQueryServlet(RestServlet): Args: hs (synapse.server.HomeServer): """ - super(KeyQueryServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_keys_handler = hs.get_e2e_keys_handler() @@ -177,9 +177,10 @@ class KeyChangesServlet(RestServlet): Args: hs (synapse.server.HomeServer): """ - super(KeyChangesServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.device_handler = hs.get_device_handler() + self.store = hs.get_datastore() async def on_GET(self, request): requester = await self.auth.get_user_by_req(request, allow_guest=True) @@ -191,7 +192,7 @@ class KeyChangesServlet(RestServlet): # changes after the "to" as well as before. set_tag("to", parse_string(request, "to")) - from_token = StreamToken.from_string(from_token_string) + from_token = await StreamToken.from_string(self.store, from_token_string) user_id = requester.user.to_string() @@ -222,7 +223,7 @@ class OneTimeKeyServlet(RestServlet): PATTERNS = client_patterns("/keys/claim$") def __init__(self, hs): - super(OneTimeKeyServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_keys_handler = hs.get_e2e_keys_handler() @@ -250,7 +251,7 @@ class SigningKeyUploadServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(SigningKeyUploadServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.e2e_keys_handler = hs.get_e2e_keys_handler() @@ -308,7 +309,7 @@ class SignaturesUploadServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(SignaturesUploadServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_keys_handler = hs.get_e2e_keys_handler() diff --git a/synapse/rest/client/v2_alpha/notifications.py b/synapse/rest/client/v2_alpha/notifications.py
index aa911d75ee..87063ec8b1 100644 --- a/synapse/rest/client/v2_alpha/notifications.py +++ b/synapse/rest/client/v2_alpha/notifications.py
@@ -27,7 +27,7 @@ class NotificationsServlet(RestServlet): PATTERNS = client_patterns("/notifications$") def __init__(self, hs): - super(NotificationsServlet, self).__init__() + super().__init__() self.store = hs.get_datastore() self.auth = hs.get_auth() self.clock = hs.get_clock() diff --git a/synapse/rest/client/v2_alpha/openid.py b/synapse/rest/client/v2_alpha/openid.py
index 6ae9a5a8e9..5b996e2d63 100644 --- a/synapse/rest/client/v2_alpha/openid.py +++ b/synapse/rest/client/v2_alpha/openid.py
@@ -60,7 +60,7 @@ class IdTokenServlet(RestServlet): EXPIRES_MS = 3600 * 1000 def __init__(self, hs): - super(IdTokenServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.clock = hs.get_clock() diff --git a/synapse/rest/client/v2_alpha/password_policy.py b/synapse/rest/client/v2_alpha/password_policy.py
index 968403cca4..68b27ff23a 100644 --- a/synapse/rest/client/v2_alpha/password_policy.py +++ b/synapse/rest/client/v2_alpha/password_policy.py
@@ -30,7 +30,7 @@ class PasswordPolicyServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(PasswordPolicyServlet, self).__init__() + super().__init__() self.policy = hs.config.password_policy self.enabled = hs.config.password_policy_enabled diff --git a/synapse/rest/client/v2_alpha/read_marker.py b/synapse/rest/client/v2_alpha/read_marker.py
index 67cbc37312..55c6688f52 100644 --- a/synapse/rest/client/v2_alpha/read_marker.py +++ b/synapse/rest/client/v2_alpha/read_marker.py
@@ -26,7 +26,7 @@ class ReadMarkerRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/read_markers$") def __init__(self, hs): - super(ReadMarkerRestServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.receipts_handler = hs.get_receipts_handler() self.read_marker_handler = hs.get_read_marker_handler() diff --git a/synapse/rest/client/v2_alpha/receipts.py b/synapse/rest/client/v2_alpha/receipts.py
index 92555bd4a9..6f7246a394 100644 --- a/synapse/rest/client/v2_alpha/receipts.py +++ b/synapse/rest/client/v2_alpha/receipts.py
@@ -31,7 +31,7 @@ class ReceiptRestServlet(RestServlet): ) def __init__(self, hs): - super(ReceiptRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.receipts_handler = hs.get_receipts_handler() diff --git a/synapse/rest/client/v2_alpha/register.py b/synapse/rest/client/v2_alpha/register.py
index 001f49fb3e..91ea76bc20 100644 --- a/synapse/rest/client/v2_alpha/register.py +++ b/synapse/rest/client/v2_alpha/register.py
@@ -17,6 +17,7 @@ import hmac import logging +import random import re from typing import List, Union @@ -26,6 +27,7 @@ import synapse.types from synapse.api.constants import LoginType from synapse.api.errors import ( Codes, + InteractiveAuthIncompleteError, SynapseError, ThreepidValidationError, UnrecognizedRequestError, @@ -45,7 +47,7 @@ from synapse.http.servlet import ( parse_json_object_from_request, parse_string, ) -from synapse.push.mailer import load_jinja2_templates +from synapse.push.mailer import Mailer from synapse.util.msisdn import phone_number_to_msisdn from synapse.util.ratelimitutils import FederationRateLimiter from synapse.util.stringutils import assert_valid_client_secret, random_string @@ -76,29 +78,17 @@ class EmailRegisterRequestTokenRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(EmailRegisterRequestTokenRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler self.config = hs.config if self.hs.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - from synapse.push.mailer import Mailer, load_jinja2_templates - - template_html, template_text = load_jinja2_templates( - self.config.email_template_dir, - [ - self.config.email_registration_template_html, - self.config.email_registration_template_text, - ], - apply_format_ts_filter=True, - apply_mxc_to_http_filter=True, - public_baseurl=self.config.public_baseurl, - ) self.mailer = Mailer( hs=self.hs, app_name=self.config.email_app_name, - template_html=template_html, - template_text=template_text, + template_html=self.config.email_registration_template_html, + template_text=self.config.email_registration_template_text, ) async def on_POST(self, request): @@ -144,6 +134,9 @@ class EmailRegisterRequestTokenRestServlet(RestServlet): if self.hs.config.request_token_inhibit_3pid_errors: # Make the client think the operation succeeded. See the rationale in the # comments for request_token_inhibit_3pid_errors. + # Also wait for some random amount of time between 100ms and 1s to make it + # look like we did something. + await self.hs.clock.sleep(random.randint(1, 10) / 10) return 200, {"sid": random_string(16)} raise SynapseError(400, "Email is already in use", Codes.THREEPID_IN_USE) @@ -183,7 +176,7 @@ class MsisdnRegisterRequestTokenRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(MsisdnRegisterRequestTokenRestServlet, self).__init__() + super().__init__() self.hs = hs self.identity_handler = hs.get_handlers().identity_handler @@ -218,6 +211,9 @@ class MsisdnRegisterRequestTokenRestServlet(RestServlet): if self.hs.config.request_token_inhibit_3pid_errors: # Make the client think the operation succeeded. See the rationale in the # comments for request_token_inhibit_3pid_errors. + # Also wait for some random amount of time between 100ms and 1s to make it + # look like we did something. + await self.hs.clock.sleep(random.randint(1, 10) / 10) return 200, {"sid": random_string(16)} raise SynapseError( @@ -257,7 +253,7 @@ class RegistrationSubmitTokenServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(RegistrationSubmitTokenServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.config = hs.config @@ -265,15 +261,8 @@ class RegistrationSubmitTokenServlet(RestServlet): self.store = hs.get_datastore() if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - (self.failure_email_template,) = load_jinja2_templates( - self.config.email_template_dir, - [self.config.email_registration_template_failure_html], - ) - - if self.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL: - (self.failure_email_template,) = load_jinja2_templates( - self.config.email_template_dir, - [self.config.email_registration_template_failure_html], + self._failure_email_template = ( + self.config.email_registration_template_failure_html ) async def on_GET(self, request, medium): @@ -321,7 +310,7 @@ class RegistrationSubmitTokenServlet(RestServlet): # Show a failure page with a reason template_vars = {"failure_reason": e.msg} - html = self.failure_email_template.render(**template_vars) + html = self._failure_email_template.render(**template_vars) respond_with_html(request, status_code, html) @@ -334,7 +323,7 @@ class UsernameAvailabilityRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(UsernameAvailabilityRestServlet, self).__init__() + super().__init__() self.hs = hs self.registration_handler = hs.get_registration_handler() self.ratelimiter = FederationRateLimiter( @@ -372,7 +361,7 @@ class RegisterRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(RegisterRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() @@ -385,6 +374,7 @@ class RegisterRestServlet(RestServlet): self.ratelimiter = hs.get_registration_ratelimiter() self.password_policy_handler = hs.get_password_policy_handler() self.clock = hs.get_clock() + self._registration_enabled = self.hs.config.enable_registration self._registration_flows = _calculate_registration_flows( hs.config, self.auth_handler @@ -410,22 +400,6 @@ class RegisterRestServlet(RestServlet): "Do not understand membership kind: %s" % (kind.decode("utf8"),) ) - # we do basic sanity checks here because the auth layer will store these - # in sessions. Pull out the username/password provided to us. - desired_password_hash = None - if "password" in body: - password = body.pop("password") - if not isinstance(password, str) or len(password) > 512: - raise SynapseError(400, "Invalid password") - self.password_policy_handler.validate_password(password) - - # If the password is valid, hash it and store it back on the body. - # This ensures that only the hashed password is handled everywhere. - if "password_hash" in body: - raise SynapseError(400, "Unexpected property: password_hash") - body["password_hash"] = await self.auth_handler.hash(password) - desired_password_hash = body["password_hash"] - # We don't care about usernames for this deployment. In fact, the act # of checking whether they exist already can leak metadata about # which users are already registered. @@ -440,7 +414,12 @@ class RegisterRestServlet(RestServlet): appservice = None if self.auth.has_access_token(request): - appservice = await self.auth.get_appservice_by_req(request) + appservice = self.auth.get_appservice_by_req(request) + + # We need to retrieve the password early in order to pass it to + # application service registration + # This is specific to shadow server registration of users via an AS + password = body.pop("password", None) # fork off as soon as possible for ASes which have completely # different registration flows to normal users @@ -459,23 +438,33 @@ class RegisterRestServlet(RestServlet): access_token = self.auth.get_access_token_from_request(request) - if isinstance(desired_username, str): - result = await self._do_appservice_registration( - desired_username, - desired_password_hash, - desired_display_name, - access_token, - body, - ) - return 200, result # we throw for non 200 responses + if not isinstance(desired_username, str): + raise SynapseError(400, "Desired Username is missing or not a string") + + result = await self._do_appservice_registration( + desired_username, password, desired_display_name, access_token, body + ) + + return 200, result # == Normal User Registration == (everyone else) - if not self.hs.config.enable_registration: + if not self._registration_enabled: raise SynapseError(403, "Registration has been disabled") + # Check if this account is upgrading from a guest account. guest_access_token = body.get("guest_access_token", None) - if "initial_device_display_name" in body and "password_hash" not in body: + # Pull out the provided password and do basic sanity checks early. + # + # Note that we remove the password from the body since the auth layer + # will store the body in the session and we don't want a plaintext + # password store there. + if password is not None: + if not isinstance(password, str) or len(password) > 512: + raise SynapseError(400, "Invalid password") + self.password_policy_handler.validate_password(password) + + if "initial_device_display_name" in body and password is None: # ignore 'initial_device_display_name' if sent without # a password to work around a client bug where it sent # the 'initial_device_display_name' param alone, wiping out @@ -485,6 +474,7 @@ class RegisterRestServlet(RestServlet): session_id = self.auth_handler.get_session_id(body) registered_user_id = None + password_hash = None if session_id: # if we get a registered user id out of here, it means we previously # registered a user for this session, so we could just return the @@ -493,21 +483,43 @@ class RegisterRestServlet(RestServlet): registered_user_id = await self.auth_handler.get_session_data( session_id, "registered_user_id", None ) + # Extract the previously-hashed password from the session. + password_hash = await self.auth_handler.get_session_data( + session_id, "password_hash", None + ) - auth_result, params, session_id = await self.auth_handler.check_auth( - self._registration_flows, - request, - body, - self.hs.get_ip_from_request(request), - "register a new account", - ) + # Check if the user-interactive authentication flows are complete, if + # not this will raise a user-interactive auth error. + try: + auth_result, params, session_id = await self.auth_handler.check_ui_auth( + self._registration_flows, + request, + body, + self.hs.get_ip_from_request(request), + "register a new account", + ) + except InteractiveAuthIncompleteError as e: + # The user needs to provide more steps to complete auth. + # + # Hash the password and store it with the session since the client + # is not required to provide the password again. + # + # If a password hash was previously stored we will not attempt to + # re-hash and store it for efficiency. This assumes the password + # does not change throughout the authentication flow, but this + # should be fine since the data is meant to be consistent. + if not password_hash and password: + password_hash = await self.auth_handler.hash(password) + await self.auth_handler.set_session_data( + e.session_id, "password_hash", password_hash + ) + raise # Check that we're not trying to register a denied 3pid. # # the user-facing checks will probably already have happened in # /register/email/requestToken when we requested a 3pid, but that's not # guaranteed. - if auth_result: for login_type in [LoginType.EMAIL_IDENTITY, LoginType.MSISDN]: if login_type in auth_result: @@ -603,8 +615,12 @@ class RegisterRestServlet(RestServlet): # don't re-register the threepids registered = False else: - # NB: This may be from the auth handler and NOT from the POST - assert_params_in_dict(params, ["password_hash"]) + # If we have a password in this request, prefer it. Otherwise, there + # might be a password hash from an earlier request. + if password: + password_hash = await self.auth_handler.hash(password) + if not password_hash: + raise SynapseError(400, "Missing params: password", Codes.MISSING_PARAM) if not self.hs.config.register_mxid_from_3pid: desired_username = params.get("username", None) @@ -613,7 +629,6 @@ class RegisterRestServlet(RestServlet): pass guest_access_token = params.get("guest_access_token", None) - new_password_hash = params.get("password_hash", None) if desired_username is not None: desired_username = desired_username.lower() @@ -653,13 +668,18 @@ class RegisterRestServlet(RestServlet): Codes.THREEPID_IN_USE, ) + entries = await self.store.get_user_agents_ips_to_ui_auth_session( + session_id + ) + registered_user_id = await self.registration_handler.register_user( localpart=desired_username, - password_hash=new_password_hash, + password_hash=password_hash, guest_access_token=guest_access_token, default_display_name=desired_display_name, threepid=threepid, address=client_addr, + user_agent_ips=entries, ) # Necessary due to auth checks prior to the threepid being # written to the db @@ -677,8 +697,8 @@ class RegisterRestServlet(RestServlet): params=params, ) - # remember that we've now registered that user account, and with - # what user ID (since the user may not have specified) + # Remember that the user account has been registered (and the user + # ID it was registered with, since it might not have been specified). await self.auth_handler.set_session_data( session_id, "registered_user_id", registered_user_id ) @@ -702,12 +722,20 @@ class RegisterRestServlet(RestServlet): return 200, {} async def _do_appservice_registration( - self, username, password_hash, display_name, as_token, body + self, username, password, display_name, as_token, body ): # FIXME: appservice_register() is horribly duplicated with register() # and they should probably just be combined together with a config flag. + + if password: + # Hash the password + # + # In mainline hashing of the password was moved further on in the registration + # flow, but we need it here for the AS use-case of shadow servers + password = await self.auth_handler.hash(password) + user_id = await self.registration_handler.appservice_register( - username, as_token, password_hash, display_name + username, as_token, password, display_name ) result = await self._create_registration_details(user_id, body) @@ -736,7 +764,7 @@ class RegisterRestServlet(RestServlet): (object) params: registration parameters, from which we pull device_id, initial_device_name and inhibit_login Returns: - defer.Deferred: (object) dictionary for response from /register + dictionary for response from /register """ result = {"user_id": user_id, "home_server": self.hs.hostname} if not params.get("inhibit_login", False): diff --git a/synapse/rest/client/v2_alpha/relations.py b/synapse/rest/client/v2_alpha/relations.py
index 89002ffbff..18c75738f8 100644 --- a/synapse/rest/client/v2_alpha/relations.py +++ b/synapse/rest/client/v2_alpha/relations.py
@@ -22,7 +22,7 @@ any time to reflect changes in the MSC. import logging from synapse.api.constants import EventTypes, RelationTypes -from synapse.api.errors import SynapseError +from synapse.api.errors import ShadowBanError, SynapseError from synapse.http.servlet import ( RestServlet, parse_integer, @@ -35,6 +35,7 @@ from synapse.storage.relations import ( PaginationChunk, RelationPaginationToken, ) +from synapse.util.stringutils import random_string from ._base import client_patterns @@ -60,7 +61,7 @@ class RelationSendServlet(RestServlet): ) def __init__(self, hs): - super(RelationSendServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.event_creation_handler = hs.get_event_creation_handler() self.txns = HttpTransactionCache(hs) @@ -111,11 +112,18 @@ class RelationSendServlet(RestServlet): "sender": requester.user.to_string(), } - event, _ = await self.event_creation_handler.create_and_send_nonmember_event( - requester, event_dict=event_dict, txn_id=txn_id - ) + try: + ( + event, + _, + ) = await self.event_creation_handler.create_and_send_nonmember_event( + requester, event_dict=event_dict, txn_id=txn_id + ) + event_id = event.event_id + except ShadowBanError: + event_id = "$" + random_string(43) - return 200, {"event_id": event.event_id} + return 200, {"event_id": event_id} class RelationPaginationServlet(RestServlet): @@ -130,7 +138,7 @@ class RelationPaginationServlet(RestServlet): ) def __init__(self, hs): - super(RelationPaginationServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.clock = hs.get_clock() @@ -225,7 +233,7 @@ class RelationAggregationPaginationServlet(RestServlet): ) def __init__(self, hs): - super(RelationAggregationPaginationServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.event_handler = hs.get_event_handler() @@ -303,7 +311,7 @@ class RelationAggregationGroupPaginationServlet(RestServlet): ) def __init__(self, hs): - super(RelationAggregationGroupPaginationServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.clock = hs.get_clock() diff --git a/synapse/rest/client/v2_alpha/report_event.py b/synapse/rest/client/v2_alpha/report_event.py
index e15927c4ea..215d619ca1 100644 --- a/synapse/rest/client/v2_alpha/report_event.py +++ b/synapse/rest/client/v2_alpha/report_event.py
@@ -32,7 +32,7 @@ class ReportEventRestServlet(RestServlet): PATTERNS = client_patterns("/rooms/(?P<room_id>[^/]*)/report/(?P<event_id>[^/]*)$") def __init__(self, hs): - super(ReportEventRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.clock = hs.get_clock() diff --git a/synapse/rest/client/v2_alpha/room_keys.py b/synapse/rest/client/v2_alpha/room_keys.py
index 59529707df..53de97923f 100644 --- a/synapse/rest/client/v2_alpha/room_keys.py +++ b/synapse/rest/client/v2_alpha/room_keys.py
@@ -37,7 +37,7 @@ class RoomKeysServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(RoomKeysServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_room_keys_handler = hs.get_e2e_room_keys_handler() @@ -248,7 +248,7 @@ class RoomKeysNewVersionServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(RoomKeysNewVersionServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_room_keys_handler = hs.get_e2e_room_keys_handler() @@ -301,7 +301,7 @@ class RoomKeysVersionServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(RoomKeysVersionServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.e2e_room_keys_handler = hs.get_e2e_room_keys_handler() diff --git a/synapse/rest/client/v2_alpha/room_upgrade_rest_servlet.py b/synapse/rest/client/v2_alpha/room_upgrade_rest_servlet.py
index f357015a70..bf030e0ff4 100644 --- a/synapse/rest/client/v2_alpha/room_upgrade_rest_servlet.py +++ b/synapse/rest/client/v2_alpha/room_upgrade_rest_servlet.py
@@ -15,13 +15,14 @@ import logging -from synapse.api.errors import Codes, SynapseError +from synapse.api.errors import Codes, ShadowBanError, SynapseError from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.http.servlet import ( RestServlet, assert_params_in_dict, parse_json_object_from_request, ) +from synapse.util import stringutils from ._base import client_patterns @@ -52,7 +53,7 @@ class RoomUpgradeRestServlet(RestServlet): ) def __init__(self, hs): - super(RoomUpgradeRestServlet, self).__init__() + super().__init__() self._hs = hs self._room_creation_handler = hs.get_room_creation_handler() self._auth = hs.get_auth() @@ -62,7 +63,6 @@ class RoomUpgradeRestServlet(RestServlet): content = parse_json_object_from_request(request) assert_params_in_dict(content, ("new_version",)) - new_version = content["new_version"] new_version = KNOWN_ROOM_VERSIONS.get(content["new_version"]) if new_version is None: @@ -72,9 +72,13 @@ class RoomUpgradeRestServlet(RestServlet): Codes.UNSUPPORTED_ROOM_VERSION, ) - new_room_id = await self._room_creation_handler.upgrade_room( - requester, room_id, new_version - ) + try: + new_room_id = await self._room_creation_handler.upgrade_room( + requester, room_id, new_version + ) + except ShadowBanError: + # Generate a random room ID. + new_room_id = stringutils.random_string(18) ret = {"replacement_room": new_room_id} diff --git a/synapse/rest/client/v2_alpha/sendtodevice.py b/synapse/rest/client/v2_alpha/sendtodevice.py
index db829f3098..bc4f43639a 100644 --- a/synapse/rest/client/v2_alpha/sendtodevice.py +++ b/synapse/rest/client/v2_alpha/sendtodevice.py
@@ -36,7 +36,7 @@ class SendToDeviceRestServlet(servlet.RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(SendToDeviceRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.txns = HttpTransactionCache(hs) diff --git a/synapse/rest/client/v2_alpha/shared_rooms.py b/synapse/rest/client/v2_alpha/shared_rooms.py new file mode 100644
index 0000000000..c866d5151c --- /dev/null +++ b/synapse/rest/client/v2_alpha/shared_rooms.py
@@ -0,0 +1,68 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 Half-Shot +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import logging + +from synapse.api.errors import Codes, SynapseError +from synapse.http.servlet import RestServlet +from synapse.types import UserID + +from ._base import client_patterns + +logger = logging.getLogger(__name__) + + +class UserSharedRoomsServlet(RestServlet): + """ + GET /uk.half-shot.msc2666/user/shared_rooms/{user_id} HTTP/1.1 + """ + + PATTERNS = client_patterns( + "/uk.half-shot.msc2666/user/shared_rooms/(?P<user_id>[^/]*)", + releases=(), # This is an unstable feature + ) + + def __init__(self, hs): + super().__init__() + self.auth = hs.get_auth() + self.store = hs.get_datastore() + self.user_directory_active = hs.config.update_user_directory + + async def on_GET(self, request, user_id): + + if not self.user_directory_active: + raise SynapseError( + code=400, + msg="The user directory is disabled on this server. Cannot determine shared rooms.", + errcode=Codes.FORBIDDEN, + ) + + UserID.from_string(user_id) + + requester = await self.auth.get_user_by_req(request) + if user_id == requester.user.to_string(): + raise SynapseError( + code=400, + msg="You cannot request a list of shared rooms with yourself", + errcode=Codes.FORBIDDEN, + ) + rooms = await self.store.get_shared_rooms_for_users( + requester.user.to_string(), user_id + ) + + return 200, {"joined": list(rooms)} + + +def register_servlets(hs, http_server): + UserSharedRoomsServlet(hs).register(http_server) diff --git a/synapse/rest/client/v2_alpha/sync.py b/synapse/rest/client/v2_alpha/sync.py
index a5c24fbd63..6779df952f 100644 --- a/synapse/rest/client/v2_alpha/sync.py +++ b/synapse/rest/client/v2_alpha/sync.py
@@ -16,8 +16,6 @@ import itertools import logging -from canonicaljson import json - from synapse.api.constants import PresenceState from synapse.api.errors import Codes, StoreError, SynapseError from synapse.api.filtering import DEFAULT_FILTER_COLLECTION, FilterCollection @@ -29,6 +27,7 @@ from synapse.handlers.presence import format_user_presence_state from synapse.handlers.sync import SyncConfig from synapse.http.servlet import RestServlet, parse_boolean, parse_integer, parse_string from synapse.types import StreamToken +from synapse.util import json_decoder from ._base import client_patterns, set_timeline_upper_limit @@ -75,9 +74,10 @@ class SyncRestServlet(RestServlet): ALLOWED_PRESENCE = {"online", "offline", "unavailable"} def __init__(self, hs): - super(SyncRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() + self.store = hs.get_datastore() self.sync_handler = hs.get_sync_handler() self.clock = hs.get_clock() self.filtering = hs.get_filtering() @@ -125,7 +125,7 @@ class SyncRestServlet(RestServlet): filter_collection = DEFAULT_FILTER_COLLECTION elif filter_id.startswith("{"): try: - filter_object = json.loads(filter_id) + filter_object = json_decoder.decode(filter_id) set_timeline_upper_limit( filter_object, self.hs.config.filter_timeline_limit ) @@ -152,10 +152,9 @@ class SyncRestServlet(RestServlet): device_id=device_id, ) + since_token = None if since is not None: - since_token = StreamToken.from_string(since) - else: - since_token = None + since_token = await StreamToken.from_string(self.store, since) # send any outstanding server notices to the user. await self._server_notices_sender.on_user_syncing(user.to_string()) @@ -237,7 +236,7 @@ class SyncRestServlet(RestServlet): "leave": sync_result.groups.leave, }, "device_one_time_keys_count": sync_result.device_one_time_keys_count, - "next_batch": sync_result.next_batch.to_string(), + "next_batch": await sync_result.next_batch.to_string(self.store), } @staticmethod @@ -414,7 +413,7 @@ class SyncRestServlet(RestServlet): result = { "timeline": { "events": serialized_timeline, - "prev_batch": room.timeline.prev_batch.to_string(), + "prev_batch": await room.timeline.prev_batch.to_string(self.store), "limited": room.timeline.limited, }, "state": {"events": serialized_state}, @@ -426,6 +425,7 @@ class SyncRestServlet(RestServlet): result["ephemeral"] = {"events": ephemeral_events} result["unread_notifications"] = room.unread_notifications result["summary"] = room.summary + result["org.matrix.msc2654.unread_count"] = room.unread_count return result diff --git a/synapse/rest/client/v2_alpha/tags.py b/synapse/rest/client/v2_alpha/tags.py
index a3f12e8a77..bf3a79db44 100644 --- a/synapse/rest/client/v2_alpha/tags.py +++ b/synapse/rest/client/v2_alpha/tags.py
@@ -31,7 +31,7 @@ class TagListServlet(RestServlet): PATTERNS = client_patterns("/user/(?P<user_id>[^/]*)/rooms/(?P<room_id>[^/]*)/tags") def __init__(self, hs): - super(TagListServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() @@ -56,7 +56,7 @@ class TagServlet(RestServlet): ) def __init__(self, hs): - super(TagServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.store = hs.get_datastore() self.notifier = hs.get_notifier() diff --git a/synapse/rest/client/v2_alpha/thirdparty.py b/synapse/rest/client/v2_alpha/thirdparty.py
index 23709960ad..0c127a1b5f 100644 --- a/synapse/rest/client/v2_alpha/thirdparty.py +++ b/synapse/rest/client/v2_alpha/thirdparty.py
@@ -28,7 +28,7 @@ class ThirdPartyProtocolsServlet(RestServlet): PATTERNS = client_patterns("/thirdparty/protocols") def __init__(self, hs): - super(ThirdPartyProtocolsServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.appservice_handler = hs.get_application_service_handler() @@ -44,7 +44,7 @@ class ThirdPartyProtocolServlet(RestServlet): PATTERNS = client_patterns("/thirdparty/protocol/(?P<protocol>[^/]+)$") def __init__(self, hs): - super(ThirdPartyProtocolServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.appservice_handler = hs.get_application_service_handler() @@ -65,7 +65,7 @@ class ThirdPartyUserServlet(RestServlet): PATTERNS = client_patterns("/thirdparty/user(/(?P<protocol>[^/]+))?$") def __init__(self, hs): - super(ThirdPartyUserServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.appservice_handler = hs.get_application_service_handler() @@ -87,7 +87,7 @@ class ThirdPartyLocationServlet(RestServlet): PATTERNS = client_patterns("/thirdparty/location(/(?P<protocol>[^/]+))?$") def __init__(self, hs): - super(ThirdPartyLocationServlet, self).__init__() + super().__init__() self.auth = hs.get_auth() self.appservice_handler = hs.get_application_service_handler() diff --git a/synapse/rest/client/v2_alpha/tokenrefresh.py b/synapse/rest/client/v2_alpha/tokenrefresh.py
index 83f3b6b70a..79317c74ba 100644 --- a/synapse/rest/client/v2_alpha/tokenrefresh.py +++ b/synapse/rest/client/v2_alpha/tokenrefresh.py
@@ -28,7 +28,7 @@ class TokenRefreshRestServlet(RestServlet): PATTERNS = client_patterns("/tokenrefresh") def __init__(self, hs): - super(TokenRefreshRestServlet, self).__init__() + super().__init__() async def on_POST(self, request): raise AuthError(403, "tokenrefresh is no longer supported.") diff --git a/synapse/rest/client/v2_alpha/user_directory.py b/synapse/rest/client/v2_alpha/user_directory.py
index 6e8300d6a5..5d4be8adaf 100644 --- a/synapse/rest/client/v2_alpha/user_directory.py +++ b/synapse/rest/client/v2_alpha/user_directory.py
@@ -39,7 +39,7 @@ class UserDirectorySearchRestServlet(RestServlet): Args: hs (synapse.server.HomeServer): server """ - super(UserDirectorySearchRestServlet, self).__init__() + super().__init__() self.hs = hs self.auth = hs.get_auth() self.user_directory_handler = hs.get_user_directory_handler() diff --git a/synapse/rest/client/versions.py b/synapse/rest/client/versions.py
index b1999d051b..c9b9e7f5ff 100644 --- a/synapse/rest/client/versions.py +++ b/synapse/rest/client/versions.py
@@ -19,6 +19,7 @@ import logging import re +from synapse.api.constants import RoomCreationPreset from synapse.http.servlet import RestServlet logger = logging.getLogger(__name__) @@ -28,9 +29,23 @@ class VersionsRestServlet(RestServlet): PATTERNS = [re.compile("^/_matrix/client/versions$")] def __init__(self, hs): - super(VersionsRestServlet, self).__init__() + super().__init__() self.config = hs.config + # Calculate these once since they shouldn't change after start-up. + self.e2ee_forced_public = ( + RoomCreationPreset.PUBLIC_CHAT + in self.config.encryption_enabled_by_default_for_room_presets + ) + self.e2ee_forced_private = ( + RoomCreationPreset.PRIVATE_CHAT + in self.config.encryption_enabled_by_default_for_room_presets + ) + self.e2ee_forced_trusted_private = ( + RoomCreationPreset.TRUSTED_PRIVATE_CHAT + in self.config.encryption_enabled_by_default_for_room_presets + ) + def on_GET(self, request): return ( 200, @@ -63,6 +78,12 @@ class VersionsRestServlet(RestServlet): # Tchap does not currently assume this rule for r0.5.0 # XXX: Remove this when it does "m.lazy_load_members": True, + # Implements additional endpoints as described in MSC2666 + "uk.half-shot.msc2666": True, + # Whether new rooms will be set to encrypted or not (based on presets). + "io.element.e2ee_forced.public": self.e2ee_forced_public, + "io.element.e2ee_forced.private": self.e2ee_forced_private, + "io.element.e2ee_forced.trusted_private": self.e2ee_forced_trusted_private, }, }, ) diff --git a/synapse/rest/consent/consent_resource.py b/synapse/rest/consent/consent_resource.py
index 4386eb4e72..b3e4d5612e 100644 --- a/synapse/rest/consent/consent_resource.py +++ b/synapse/rest/consent/consent_resource.py
@@ -22,8 +22,6 @@ from os import path import jinja2 from jinja2 import TemplateNotFound -from twisted.internet import defer - from synapse.api.errors import NotFoundError, StoreError, SynapseError from synapse.config import ConfigError from synapse.http.server import DirectServeHtmlResource, respond_with_html @@ -135,7 +133,7 @@ class ConsentResource(DirectServeHtmlResource): else: qualified_user_id = UserID(username, self.hs.hostname).to_string() - u = await defer.maybeDeferred(self.store.get_user_by_id, qualified_user_id) + u = await self.store.get_user_by_id(qualified_user_id) if u is None: raise NotFoundError("Unknown user") diff --git a/synapse/rest/health.py b/synapse/rest/health.py new file mode 100644
index 0000000000..0170950bf3 --- /dev/null +++ b/synapse/rest/health.py
@@ -0,0 +1,31 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from twisted.web.resource import Resource + + +class HealthResource(Resource): + """A resource that does nothing except return a 200 with a body of `OK`, + which can be used as a health check. + + Note: `SynapseRequest._should_log_request` ensures that requests to + `/health` do not get logged at INFO. + """ + + isLeaf = 1 + + def render_GET(self, request): + request.setHeader(b"Content-Type", b"text/plain") + return b"OK" diff --git a/synapse/rest/key/v2/remote_key_resource.py b/synapse/rest/key/v2/remote_key_resource.py
index 9b3f85b306..f843f02454 100644 --- a/synapse/rest/key/v2/remote_key_resource.py +++ b/synapse/rest/key/v2/remote_key_resource.py
@@ -15,19 +15,19 @@ import logging from typing import Dict, Set -from canonicaljson import encode_canonical_json, json from signedjson.sign import sign_json from synapse.api.errors import Codes, SynapseError from synapse.crypto.keyring import ServerKeyFetcher -from synapse.http.server import DirectServeJsonResource, respond_with_json_bytes +from synapse.http.server import DirectServeJsonResource, respond_with_json from synapse.http.servlet import parse_integer, parse_json_object_from_request +from synapse.util import json_decoder logger = logging.getLogger(__name__) class RemoteKey(DirectServeJsonResource): - """HTTP resource for retreiving the TLS certificate and NACL signature + """HTTP resource for retrieving the TLS certificate and NACL signature verification keys for a collection of servers. Checks that the reported X.509 TLS certificate matches the one used in the HTTPS connection. Checks that the NACL signature for the remote server is valid. Returns a dict of @@ -35,7 +35,7 @@ class RemoteKey(DirectServeJsonResource): Supports individual GET APIs and a bulk query POST API. - Requsts: + Requests: GET /_matrix/key/v2/query/remote.server.example.com HTTP/1.1 @@ -209,13 +209,15 @@ class RemoteKey(DirectServeJsonResource): # Cast to bytes since postgresql returns a memoryview. json_results.add(bytes(result["key_json"])) + # If there is a cache miss, request the missing keys, then recurse (and + # ensure the result is sent). if cache_misses and query_remote_on_cache_miss: await self.fetcher.get_keys(cache_misses) await self.query_keys(request, query, query_remote_on_cache_miss=False) else: signed_keys = [] for key_json in json_results: - key_json = json.loads(key_json.decode("utf-8")) + key_json = json_decoder.decode(key_json.decode("utf-8")) for signing_key in self.config.key_server_signing_keys: key_json = sign_json(key_json, self.config.server_name, signing_key) @@ -223,4 +225,4 @@ class RemoteKey(DirectServeJsonResource): results = {"server_keys": signed_keys} - respond_with_json_bytes(request, 200, encode_canonical_json(results)) + respond_with_json(request, 200, results, canonical_json=True) diff --git a/synapse/rest/media/v1/_base.py b/synapse/rest/media/v1/_base.py
index 9a847130c0..6568e61829 100644 --- a/synapse/rest/media/v1/_base.py +++ b/synapse/rest/media/v1/_base.py
@@ -17,7 +17,9 @@ import logging import os import urllib +from typing import Awaitable +from twisted.internet.interfaces import IConsumer from twisted.protocols.basic import FileSender from synapse.api.errors import Codes, SynapseError, cs_error @@ -233,21 +235,21 @@ async def respond_with_responder( finish_request(request) -class Responder(object): +class Responder: """Represents a response that can be streamed to the requester. Responder is a context manager which *must* be used, so that any resources held can be cleaned up. """ - def write_to_consumer(self, consumer): + def write_to_consumer(self, consumer: IConsumer) -> Awaitable: """Stream response into consumer Args: - consumer (IConsumer) + consumer: The consumer to stream into. Returns: - Deferred: Resolves once the response has finished being written + Resolves once the response has finished being written """ pass @@ -258,7 +260,7 @@ class Responder(object): pass -class FileInfo(object): +class FileInfo: """Details about a requested/uploaded file. Attributes: diff --git a/synapse/rest/media/v1/filepath.py b/synapse/rest/media/v1/filepath.py
index e25c382c9c..7447eeaebe 100644 --- a/synapse/rest/media/v1/filepath.py +++ b/synapse/rest/media/v1/filepath.py
@@ -33,7 +33,7 @@ def _wrap_in_base_path(func): return _wrapped -class MediaFilePaths(object): +class MediaFilePaths: """Describes where files are stored on disk. Most of the functions have a `*_rel` variant which returns a file path that @@ -80,7 +80,7 @@ class MediaFilePaths(object): self, server_name, file_id, width, height, content_type, method ): top_level_type, sub_type = content_type.split("/") - file_name = "%i-%i-%s-%s" % (width, height, top_level_type, sub_type) + file_name = "%i-%i-%s-%s-%s" % (width, height, top_level_type, sub_type, method) return os.path.join( "remote_thumbnail", server_name, @@ -92,6 +92,23 @@ class MediaFilePaths(object): remote_media_thumbnail = _wrap_in_base_path(remote_media_thumbnail_rel) + # Legacy path that was used to store thumbnails previously. + # Should be removed after some time, when most of the thumbnails are stored + # using the new path. + def remote_media_thumbnail_rel_legacy( + self, server_name, file_id, width, height, content_type + ): + top_level_type, sub_type = content_type.split("/") + file_name = "%i-%i-%s-%s" % (width, height, top_level_type, sub_type) + return os.path.join( + "remote_thumbnail", + server_name, + file_id[0:2], + file_id[2:4], + file_id[4:], + file_name, + ) + def remote_media_thumbnail_dir(self, server_name, file_id): return os.path.join( self.base_path, diff --git a/synapse/rest/media/v1/media_repository.py b/synapse/rest/media/v1/media_repository.py
index 45628c07b4..e1192b47cd 100644 --- a/synapse/rest/media/v1/media_repository.py +++ b/synapse/rest/media/v1/media_repository.py
@@ -18,10 +18,11 @@ import errno import logging import os import shutil -from typing import Dict, Tuple +from typing import IO, Dict, Optional, Tuple import twisted.internet.error import twisted.web.http +from twisted.web.http import Request from twisted.web.resource import Resource from synapse.api.errors import ( @@ -40,6 +41,7 @@ from synapse.util.stringutils import random_string from ._base import ( FileInfo, + Responder, get_filename_from_headers, respond_404, respond_with_responder, @@ -51,7 +53,7 @@ from .media_storage import MediaStorage from .preview_url_resource import PreviewUrlResource from .storage_provider import StorageProviderWrapper from .thumbnail_resource import ThumbnailResource -from .thumbnailer import Thumbnailer +from .thumbnailer import Thumbnailer, ThumbnailError from .upload_resource import UploadResource logger = logging.getLogger(__name__) @@ -60,7 +62,7 @@ logger = logging.getLogger(__name__) UPDATE_RECENTLY_ACCESSED_TS = 60 * 1000 -class MediaRepository(object): +class MediaRepository: def __init__(self, hs): self.hs = hs self.auth = hs.get_auth() @@ -135,20 +137,26 @@ class MediaRepository(object): self.recently_accessed_locals.add(media_id) async def create_content( - self, media_type, upload_name, content, content_length, auth_user - ): + self, + media_type: str, + upload_name: Optional[str], + content: IO, + content_length: int, + auth_user: str, + ) -> str: """Store uploaded content for a local user and return the mxc URL Args: - media_type(str): The content type of the file - upload_name(str): The name of the file + media_type: The content type of the file. + upload_name: The name of the file, if provided. content: A file like object that is the content to store - content_length(int): The length of the content - auth_user(str): The user_id of the uploader + content_length: The length of the content + auth_user: The user_id of the uploader Returns: - Deferred[str]: The mxc url of the stored content + The mxc url of the stored content """ + media_id = random_string(24) file_info = FileInfo(server_name=None, file_id=media_id) @@ -170,19 +178,20 @@ class MediaRepository(object): return "mxc://%s/%s" % (self.server_name, media_id) - async def get_local_media(self, request, media_id, name): + async def get_local_media( + self, request: Request, media_id: str, name: Optional[str] + ) -> None: """Responds to reqests for local media, if exists, or returns 404. Args: - request(twisted.web.http.Request) - media_id (str): The media ID of the content. (This is the same as + request: The incoming request. + media_id: The media ID of the content. (This is the same as the file_id for local content.) - name (str|None): Optional name that, if specified, will be used as + name: Optional name that, if specified, will be used as the filename in the Content-Disposition header of the response. Returns: - Deferred: Resolves once a response has successfully been written - to request + Resolves once a response has successfully been written to request """ media_info = await self.store.get_local_media(media_id) if not media_info or media_info["quarantined_by"]: @@ -203,20 +212,20 @@ class MediaRepository(object): request, responder, media_type, media_length, upload_name ) - async def get_remote_media(self, request, server_name, media_id, name): + async def get_remote_media( + self, request: Request, server_name: str, media_id: str, name: Optional[str] + ) -> None: """Respond to requests for remote media. Args: - request(twisted.web.http.Request) - server_name (str): Remote server_name where the media originated. - media_id (str): The media ID of the content (as defined by the - remote server). - name (str|None): Optional name that, if specified, will be used as + request: The incoming request. + server_name: Remote server_name where the media originated. + media_id: The media ID of the content (as defined by the remote server). + name: Optional name that, if specified, will be used as the filename in the Content-Disposition header of the response. Returns: - Deferred: Resolves once a response has successfully been written - to request + Resolves once a response has successfully been written to request """ if ( self.federation_domain_whitelist is not None @@ -245,17 +254,16 @@ class MediaRepository(object): else: respond_404(request) - async def get_remote_media_info(self, server_name, media_id): + async def get_remote_media_info(self, server_name: str, media_id: str) -> dict: """Gets the media info associated with the remote file, downloading if necessary. Args: - server_name (str): Remote server_name where the media originated. - media_id (str): The media ID of the content (as defined by the - remote server). + server_name: Remote server_name where the media originated. + media_id: The media ID of the content (as defined by the remote server). Returns: - Deferred[dict]: The media_info of the file + The media info of the file """ if ( self.federation_domain_whitelist is not None @@ -278,7 +286,9 @@ class MediaRepository(object): return media_info - async def _get_remote_media_impl(self, server_name, media_id): + async def _get_remote_media_impl( + self, server_name: str, media_id: str + ) -> Tuple[Optional[Responder], dict]: """Looks for media in local cache, if not there then attempt to download from remote server. @@ -288,7 +298,7 @@ class MediaRepository(object): remote server). Returns: - Deferred[(Responder, media_info)] + A tuple of responder and the media info of the file. """ media_info = await self.store.get_cached_remote_media(server_name, media_id) @@ -319,19 +329,21 @@ class MediaRepository(object): responder = await self.media_storage.fetch_media(file_info) return responder, media_info - async def _download_remote_file(self, server_name, media_id, file_id): + async def _download_remote_file( + self, server_name: str, media_id: str, file_id: str + ) -> dict: """Attempt to download the remote file from the given server name, using the given file_id as the local id. Args: - server_name (str): Originating server - media_id (str): The media ID of the content (as defined by the + server_name: Originating server + media_id: The media ID of the content (as defined by the remote server). This is different than the file_id, which is locally generated. - file_id (str): Local file ID + file_id: Local file ID Returns: - Deferred[MediaInfo] + The media info of the file. """ file_info = FileInfo(server_name=server_name, file_id=file_id) @@ -449,13 +461,30 @@ class MediaRepository(object): return t_byte_source async def generate_local_exact_thumbnail( - self, media_id, t_width, t_height, t_method, t_type, url_cache - ): + self, + media_id: str, + t_width: int, + t_height: int, + t_method: str, + t_type: str, + url_cache: str, + ) -> Optional[str]: input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(None, media_id, url_cache=url_cache) ) - thumbnailer = Thumbnailer(input_path) + try: + thumbnailer = Thumbnailer(input_path) + except ThumbnailError as e: + logger.warning( + "Unable to generate a thumbnail for local media %s using a method of %s and type of %s: %s", + media_id, + t_method, + t_type, + e, + ) + return None + t_byte_source = await defer_to_thread( self.hs.get_reactor(), self._generate_thumbnail, @@ -495,14 +524,36 @@ class MediaRepository(object): return output_path + # Could not generate thumbnail. + return None + async def generate_remote_exact_thumbnail( - self, server_name, file_id, media_id, t_width, t_height, t_method, t_type - ): + self, + server_name: str, + file_id: str, + media_id: str, + t_width: int, + t_height: int, + t_method: str, + t_type: str, + ) -> Optional[str]: input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(server_name, file_id, url_cache=False) ) - thumbnailer = Thumbnailer(input_path) + try: + thumbnailer = Thumbnailer(input_path) + except ThumbnailError as e: + logger.warning( + "Unable to generate a thumbnail for remote media %s from %s using a method of %s and type of %s: %s", + media_id, + server_name, + t_method, + t_type, + e, + ) + return None + t_byte_source = await defer_to_thread( self.hs.get_reactor(), self._generate_thumbnail, @@ -548,32 +599,52 @@ class MediaRepository(object): return output_path + # Could not generate thumbnail. + return None + async def _generate_thumbnails( - self, server_name, media_id, file_id, media_type, url_cache=False - ): + self, + server_name: Optional[str], + media_id: str, + file_id: str, + media_type: str, + url_cache: bool = False, + ) -> Optional[dict]: """Generate and store thumbnails for an image. Args: - server_name (str|None): The server name if remote media, else None if local - media_id (str): The media ID of the content. (This is the same as + server_name: The server name if remote media, else None if local + media_id: The media ID of the content. (This is the same as the file_id for local content) - file_id (str): Local file ID - media_type (str): The content type of the file - url_cache (bool): If we are thumbnailing images downloaded for the URL cache, + file_id: Local file ID + media_type: The content type of the file + url_cache: If we are thumbnailing images downloaded for the URL cache, used exclusively by the url previewer Returns: - Deferred[dict]: Dict with "width" and "height" keys of original image + Dict with "width" and "height" keys of original image or None if the + media cannot be thumbnailed. """ requirements = self._get_thumbnail_requirements(media_type) if not requirements: - return + return None input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(server_name, file_id, url_cache=url_cache) ) - thumbnailer = Thumbnailer(input_path) + try: + thumbnailer = Thumbnailer(input_path) + except ThumbnailError as e: + logger.warning( + "Unable to generate thumbnails for remote media %s from %s of type %s: %s", + media_id, + server_name, + media_type, + e, + ) + return None + m_width = thumbnailer.width m_height = thumbnailer.height @@ -584,7 +655,7 @@ class MediaRepository(object): m_height, self.max_image_pixels, ) - return + return None if thumbnailer.transpose_method is not None: m_width, m_height = await defer_to_thread( diff --git a/synapse/rest/media/v1/media_storage.py b/synapse/rest/media/v1/media_storage.py
index 66bc1c3360..a9586fb0b7 100644 --- a/synapse/rest/media/v1/media_storage.py +++ b/synapse/rest/media/v1/media_storage.py
@@ -12,13 +12,11 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import contextlib -import inspect import logging import os import shutil -from typing import Optional +from typing import IO, TYPE_CHECKING, Any, Optional, Sequence from twisted.protocols.basic import FileSender @@ -26,28 +24,39 @@ from synapse.logging.context import defer_to_thread, make_deferred_yieldable from synapse.util.file_consumer import BackgroundFileConsumer from ._base import FileInfo, Responder +from .filepath import MediaFilePaths + +if TYPE_CHECKING: + from synapse.server import HomeServer + + from .storage_provider import StorageProviderWrapper logger = logging.getLogger(__name__) -class MediaStorage(object): +class MediaStorage: """Responsible for storing/fetching files from local sources. Args: - hs (synapse.server.Homeserver) - local_media_directory (str): Base path where we store media on disk - filepaths (MediaFilePaths) - storage_providers ([StorageProvider]): List of StorageProvider that are - used to fetch and store files. + hs + local_media_directory: Base path where we store media on disk + filepaths + storage_providers: List of StorageProvider that are used to fetch and store files. """ - def __init__(self, hs, local_media_directory, filepaths, storage_providers): + def __init__( + self, + hs: "HomeServer", + local_media_directory: str, + filepaths: MediaFilePaths, + storage_providers: Sequence["StorageProviderWrapper"], + ): self.hs = hs self.local_media_directory = local_media_directory self.filepaths = filepaths self.storage_providers = storage_providers - async def store_file(self, source, file_info: FileInfo) -> str: + async def store_file(self, source: IO, file_info: FileInfo) -> str: """Write `source` to the on disk media store, and also any other configured storage providers @@ -69,7 +78,7 @@ class MediaStorage(object): return fname @contextlib.contextmanager - def store_into_file(self, file_info): + def store_into_file(self, file_info: FileInfo): """Context manager used to get a file like object to write into, as described by file_info. @@ -85,7 +94,7 @@ class MediaStorage(object): error. Args: - file_info (FileInfo): Info about the file to store + file_info: Info about the file to store Example: @@ -105,11 +114,7 @@ class MediaStorage(object): async def finish(): for provider in self.storage_providers: - # store_file is supposed to return an Awaitable, but guard - # against improper implementations. - result = provider.store_file(path, file_info) - if inspect.isawaitable(result): - await result + await provider.store_file(path, file_info) finished_called[0] = True @@ -136,21 +141,34 @@ class MediaStorage(object): Returns: Returns a Responder if the file was found, otherwise None. """ + paths = [self._file_info_to_path(file_info)] - path = self._file_info_to_path(file_info) - local_path = os.path.join(self.local_media_directory, path) - if os.path.exists(local_path): - return FileResponder(open(local_path, "rb")) + # fallback for remote thumbnails with no method in the filename + if file_info.thumbnail and file_info.server_name: + paths.append( + self.filepaths.remote_media_thumbnail_rel_legacy( + server_name=file_info.server_name, + file_id=file_info.file_id, + width=file_info.thumbnail_width, + height=file_info.thumbnail_height, + content_type=file_info.thumbnail_type, + ) + ) + + for path in paths: + local_path = os.path.join(self.local_media_directory, path) + if os.path.exists(local_path): + logger.debug("responding with local file %s", local_path) + return FileResponder(open(local_path, "rb")) + logger.debug("local file %s did not exist", local_path) for provider in self.storage_providers: - res = provider.fetch(path, file_info) - # Fetch is supposed to return an Awaitable, but guard against - # improper implementations. - if inspect.isawaitable(res): - res = await res - if res: - logger.debug("Streaming %s from %s", path, provider) - return res + for path in paths: + res = await provider.fetch(path, file_info) # type: Any + if res: + logger.debug("Streaming %s from %s", path, provider) + return res + logger.debug("%s not found on %s", path, provider) return None @@ -169,16 +187,26 @@ class MediaStorage(object): if os.path.exists(local_path): return local_path + # Fallback for paths without method names + # Should be removed in the future + if file_info.thumbnail and file_info.server_name: + legacy_path = self.filepaths.remote_media_thumbnail_rel_legacy( + server_name=file_info.server_name, + file_id=file_info.file_id, + width=file_info.thumbnail_width, + height=file_info.thumbnail_height, + content_type=file_info.thumbnail_type, + ) + legacy_local_path = os.path.join(self.local_media_directory, legacy_path) + if os.path.exists(legacy_local_path): + return legacy_local_path + dirname = os.path.dirname(local_path) if not os.path.exists(dirname): os.makedirs(dirname) for provider in self.storage_providers: - res = provider.fetch(path, file_info) - # Fetch is supposed to return an Awaitable, but guard against - # improper implementations. - if inspect.isawaitable(res): - res = await res + res = await provider.fetch(path, file_info) # type: Any if res: with res: consumer = BackgroundFileConsumer( @@ -190,17 +218,11 @@ class MediaStorage(object): raise Exception("file could not be found") - def _file_info_to_path(self, file_info): + def _file_info_to_path(self, file_info: FileInfo) -> str: """Converts file_info into a relative path. The path is suitable for storing files under a directory, e.g. used to store files on local FS under the base media repository directory. - - Args: - file_info (FileInfo) - - Returns: - str """ if file_info.url_cache: if file_info.thumbnail: diff --git a/synapse/rest/media/v1/preview_url_resource.py b/synapse/rest/media/v1/preview_url_resource.py
index 13d1a6d2ed..dce6c4d168 100644 --- a/synapse/rest/media/v1/preview_url_resource.py +++ b/synapse/rest/media/v1/preview_url_resource.py
@@ -27,9 +27,7 @@ from typing import Dict, Optional from urllib import parse as urlparse import attr -from canonicaljson import json -from twisted.internet import defer from twisted.internet.error import DNSLookupError from synapse.api.errors import Codes, SynapseError @@ -43,6 +41,7 @@ from synapse.http.servlet import parse_integer, parse_string from synapse.logging.context import make_deferred_yieldable, run_in_background from synapse.metrics.background_process_metrics import run_as_background_process from synapse.rest.media.v1._base import get_filename_from_headers +from synapse.util import json_encoder from synapse.util.async_helpers import ObservableDeferred from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.stringutils import random_string @@ -103,7 +102,7 @@ for endpoint, globs in _oembed_globs.items(): _oembed_patterns[re.compile(pattern)] = endpoint -@attr.s +@attr.s(slots=True) class OEmbedResult: # Either HTML content or URL must be provided. html = attr.ib(type=Optional[str]) @@ -228,19 +227,19 @@ class PreviewUrlResource(DirectServeJsonResource): else: logger.info("Returning cached response") - og = await make_deferred_yieldable(defer.maybeDeferred(observable.observe)) + og = await make_deferred_yieldable(observable.observe()) respond_with_json_bytes(request, 200, og, send_cors=True) - async def _do_preview(self, url, user, ts): + async def _do_preview(self, url: str, user: str, ts: int) -> bytes: """Check the db, and download the URL and build a preview Args: - url (str): - user (str): - ts (int): + url: The URL to preview. + user: The user requesting the preview. + ts: The timestamp requested for the preview. Returns: - Deferred[bytes]: json-encoded og data + json-encoded og data """ # check the URL cache in the DB (which will also provide us with # historical previews, if we have any) @@ -355,7 +354,7 @@ class PreviewUrlResource(DirectServeJsonResource): logger.debug("Calculated OG for %s as %s", url, og) - jsonog = json.dumps(og) + jsonog = json_encoder.encode(og) # store OG in history-aware DB cache await self.store.store_url_cache( @@ -451,7 +450,7 @@ class PreviewUrlResource(DirectServeJsonResource): logger.warning("Error downloading oEmbed metadata from %s: %r", url, e) raise OEmbedError() from e - async def _download_url(self, url, user): + async def _download_url(self, url: str, user): # TODO: we should probably honour robots.txt... except in practice # we're most likely being explicitly triggered by a human rather than a # bot, so are we really a robot? @@ -461,7 +460,7 @@ class PreviewUrlResource(DirectServeJsonResource): file_info = FileInfo(server_name=None, file_id=file_id, url_cache=True) # If this URL can be accessed via oEmbed, use that instead. - url_to_download = url + url_to_download = url # type: Optional[str] oembed_url = self._get_oembed_url(url) if oembed_url: # The result might be a new URL to download, or it might be HTML content. @@ -521,9 +520,15 @@ class PreviewUrlResource(DirectServeJsonResource): # FIXME: we should calculate a proper expiration based on the # Cache-Control and Expire headers. But for now, assume 1 hour. expires = ONE_HOUR - etag = headers["ETag"][0] if "ETag" in headers else None + etag = ( + headers[b"ETag"][0].decode("ascii") if b"ETag" in headers else None + ) else: - html_bytes = oembed_result.html.encode("utf-8") # type: ignore + # we can only get here if we did an oembed request and have an oembed_result.html + assert oembed_result.html is not None + assert oembed_url is not None + + html_bytes = oembed_result.html.encode("utf-8") with self.media_storage.store_into_file(file_info) as (f, fname, finish): f.write(html_bytes) await finish() @@ -586,7 +591,7 @@ class PreviewUrlResource(DirectServeJsonResource): logger.debug("Running url preview cache expiry") - if not (await self.store.db.updates.has_completed_background_updates()): + if not (await self.store.db_pool.updates.has_completed_background_updates()): logger.info("Still running DB updates; skipping expiry") return diff --git a/synapse/rest/media/v1/storage_provider.py b/synapse/rest/media/v1/storage_provider.py
index 858680be26..18c9ed48d6 100644 --- a/synapse/rest/media/v1/storage_provider.py +++ b/synapse/rest/media/v1/storage_provider.py
@@ -13,65 +13,66 @@ # See the License for the specific language governing permissions and # limitations under the License. +import inspect import logging import os import shutil - -from twisted.internet import defer +from typing import Optional from synapse.config._base import Config from synapse.logging.context import defer_to_thread, run_in_background +from ._base import FileInfo, Responder from .media_storage import FileResponder logger = logging.getLogger(__name__) -class StorageProvider(object): +class StorageProvider: """A storage provider is a service that can store uploaded media and retrieve them. """ - def store_file(self, path, file_info): + async def store_file(self, path: str, file_info: FileInfo): """Store the file described by file_info. The actual contents can be retrieved by reading the file in file_info.upload_path. Args: - path (str): Relative path of file in local cache - file_info (FileInfo) - - Returns: - Deferred + path: Relative path of file in local cache + file_info: The metadata of the file. """ - pass - def fetch(self, path, file_info): + async def fetch(self, path: str, file_info: FileInfo) -> Optional[Responder]: """Attempt to fetch the file described by file_info and stream it into writer. Args: - path (str): Relative path of file in local cache - file_info (FileInfo) + path: Relative path of file in local cache + file_info: The metadata of the file. Returns: - Deferred(Responder): Returns a Responder if the provider has the file, - otherwise returns None. + Returns a Responder if the provider has the file, otherwise returns None. """ - pass class StorageProviderWrapper(StorageProvider): """Wraps a storage provider and provides various config options Args: - backend (StorageProvider) - store_local (bool): Whether to store new local files or not. - store_synchronous (bool): Whether to wait for file to be successfully + backend: The storage provider to wrap. + store_local: Whether to store new local files or not. + store_synchronous: Whether to wait for file to be successfully uploaded, or todo the upload in the background. - store_remote (bool): Whether remote media should be uploaded + store_remote: Whether remote media should be uploaded """ - def __init__(self, backend, store_local, store_synchronous, store_remote): + def __init__( + self, + backend: StorageProvider, + store_local: bool, + store_synchronous: bool, + store_remote: bool, + ): self.backend = backend self.store_local = store_local self.store_synchronous = store_synchronous @@ -80,28 +81,38 @@ class StorageProviderWrapper(StorageProvider): def __str__(self): return "StorageProviderWrapper[%s]" % (self.backend,) - def store_file(self, path, file_info): + async def store_file(self, path, file_info): if not file_info.server_name and not self.store_local: - return defer.succeed(None) + return None if file_info.server_name and not self.store_remote: - return defer.succeed(None) + return None if self.store_synchronous: - return self.backend.store_file(path, file_info) + # store_file is supposed to return an Awaitable, but guard + # against improper implementations. + result = self.backend.store_file(path, file_info) + if inspect.isawaitable(result): + return await result else: # TODO: Handle errors. - def store(): + async def store(): try: - return self.backend.store_file(path, file_info) + result = self.backend.store_file(path, file_info) + if inspect.isawaitable(result): + return await result except Exception: logger.exception("Error storing file") run_in_background(store) - return defer.succeed(None) + return None - def fetch(self, path, file_info): - return self.backend.fetch(path, file_info) + async def fetch(self, path, file_info): + # store_file is supposed to return an Awaitable, but guard + # against improper implementations. + result = self.backend.fetch(path, file_info) + if inspect.isawaitable(result): + return await result class FileStorageProviderBackend(StorageProvider): @@ -120,7 +131,7 @@ class FileStorageProviderBackend(StorageProvider): def __str__(self): return "FileStorageProviderBackend[%s]" % (self.base_directory,) - def store_file(self, path, file_info): + async def store_file(self, path, file_info): """See StorageProvider.store_file""" primary_fname = os.path.join(self.cache_directory, path) @@ -130,11 +141,11 @@ class FileStorageProviderBackend(StorageProvider): if not os.path.exists(dirname): os.makedirs(dirname) - return defer_to_thread( + return await defer_to_thread( self.hs.get_reactor(), shutil.copyfile, primary_fname, backup_fname ) - def fetch(self, path, file_info): + async def fetch(self, path, file_info): """See StorageProvider.fetch""" backup_fname = os.path.join(self.base_directory, path) diff --git a/synapse/rest/media/v1/thumbnail_resource.py b/synapse/rest/media/v1/thumbnail_resource.py
index a83535b97b..30421b663a 100644 --- a/synapse/rest/media/v1/thumbnail_resource.py +++ b/synapse/rest/media/v1/thumbnail_resource.py
@@ -16,6 +16,7 @@ import logging +from synapse.api.errors import SynapseError from synapse.http.server import DirectServeJsonResource, set_cors_headers from synapse.http.servlet import parse_integer, parse_string @@ -173,7 +174,7 @@ class ThumbnailResource(DirectServeJsonResource): await respond_with_file(request, desired_type, file_path) else: logger.warning("Failed to generate thumbnail") - respond_404(request) + raise SynapseError(400, "Failed to generate thumbnail.") async def _select_or_generate_remote_thumbnail( self, @@ -235,7 +236,7 @@ class ThumbnailResource(DirectServeJsonResource): await respond_with_file(request, desired_type, file_path) else: logger.warning("Failed to generate thumbnail") - respond_404(request) + raise SynapseError(400, "Failed to generate thumbnail.") async def _respond_remote_thumbnail( self, request, server_name, media_id, width, height, method, m_type diff --git a/synapse/rest/media/v1/thumbnailer.py b/synapse/rest/media/v1/thumbnailer.py
index 7126997134..32a8e4f960 100644 --- a/synapse/rest/media/v1/thumbnailer.py +++ b/synapse/rest/media/v1/thumbnailer.py
@@ -15,7 +15,7 @@ import logging from io import BytesIO -from PIL import Image as Image +from PIL import Image logger = logging.getLogger(__name__) @@ -31,12 +31,22 @@ EXIF_TRANSPOSE_MAPPINGS = { } -class Thumbnailer(object): +class ThumbnailError(Exception): + """An error occurred generating a thumbnail.""" + + +class Thumbnailer: FORMATS = {"image/jpeg": "JPEG", "image/png": "PNG"} def __init__(self, input_path): - self.image = Image.open(input_path) + try: + self.image = Image.open(input_path) + except OSError as e: + # If an error occurs opening the image, a thumbnail won't be able to + # be generated. + raise ThumbnailError from e + self.width, self.height = self.image.size self.transpose_method = None try: @@ -73,7 +83,7 @@ class Thumbnailer(object): Args: max_width: The largest possible width. - max_height: The larget possible height. + max_height: The largest possible height. """ if max_width * self.height < max_height * self.width: @@ -107,7 +117,7 @@ class Thumbnailer(object): Args: max_width: The largest possible width. - max_height: The larget possible height. + max_height: The largest possible height. Returns: BytesIO: the bytes of the encoded image ready to be written to disk diff --git a/synapse/rest/media/v1/upload_resource.py b/synapse/rest/media/v1/upload_resource.py
index 3ebf7a68e6..d76f7389e1 100644 --- a/synapse/rest/media/v1/upload_resource.py +++ b/synapse/rest/media/v1/upload_resource.py
@@ -63,6 +63,10 @@ class UploadResource(DirectServeJsonResource): msg="Invalid UTF-8 filename parameter: %r" % (upload_name), code=400 ) + # If the name is falsey (e.g. an empty byte string) ensure it is None. + else: + upload_name = None + headers = request.requestHeaders if headers.hasHeader(b"Content-Type"): diff --git a/synapse/rest/saml2/response_resource.py b/synapse/rest/saml2/response_resource.py
index c10188a5d7..f6668fb5e3 100644 --- a/synapse/rest/saml2/response_resource.py +++ b/synapse/rest/saml2/response_resource.py
@@ -13,10 +13,8 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from twisted.python import failure -from synapse.api.errors import SynapseError -from synapse.http.server import DirectServeHtmlResource, return_html_error +from synapse.http.server import DirectServeHtmlResource class SAML2ResponseResource(DirectServeHtmlResource): @@ -27,21 +25,15 @@ class SAML2ResponseResource(DirectServeHtmlResource): def __init__(self, hs): super().__init__() self._saml_handler = hs.get_saml_handler() - self._error_html_template = hs.config.saml2.saml2_error_html_template async def _async_render_GET(self, request): # We're not expecting any GET request on that resource if everything goes right, # but some IdPs sometimes end up responding with a 302 redirect on this endpoint. # In this case, just tell the user that something went wrong and they should # try to authenticate again. - f = failure.Failure( - SynapseError(400, "Unexpected GET request on /saml2/authn_response") + self._saml_handler._render_error( + request, "unexpected_get", "Unexpected GET request on /saml2/authn_response" ) - return_html_error(f, request, self._error_html_template) async def _async_render_POST(self, request): - try: - await self._saml_handler.handle_saml_response(request) - except Exception: - f = failure.Failure() - return_html_error(f, request, self._error_html_template) + await self._saml_handler.handle_saml_response(request) diff --git a/synapse/rest/synapse/__init__.py b/synapse/rest/synapse/__init__.py new file mode 100644
index 0000000000..c0b733488b --- /dev/null +++ b/synapse/rest/synapse/__init__.py
@@ -0,0 +1,14 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. diff --git a/synapse/rest/synapse/client/__init__.py b/synapse/rest/synapse/client/__init__.py new file mode 100644
index 0000000000..c0b733488b --- /dev/null +++ b/synapse/rest/synapse/client/__init__.py
@@ -0,0 +1,14 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. diff --git a/synapse/rest/synapse/client/password_reset.py b/synapse/rest/synapse/client/password_reset.py new file mode 100644
index 0000000000..9e4fbc0cbd --- /dev/null +++ b/synapse/rest/synapse/client/password_reset.py
@@ -0,0 +1,127 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import logging +from typing import TYPE_CHECKING, Tuple + +from twisted.web.http import Request + +from synapse.api.errors import ThreepidValidationError +from synapse.config.emailconfig import ThreepidBehaviour +from synapse.http.server import DirectServeHtmlResource +from synapse.http.servlet import parse_string +from synapse.util.stringutils import assert_valid_client_secret + +if TYPE_CHECKING: + from synapse.server import HomeServer + +logger = logging.getLogger(__name__) + + +class PasswordResetSubmitTokenResource(DirectServeHtmlResource): + """Handles 3PID validation token submission + + This resource gets mounted under /_synapse/client/password_reset/email/submit_token + """ + + isLeaf = 1 + + def __init__(self, hs: "HomeServer"): + """ + Args: + hs: server + """ + super().__init__() + + self.clock = hs.get_clock() + self.store = hs.get_datastore() + + self._local_threepid_handling_disabled_due_to_email_config = ( + hs.config.local_threepid_handling_disabled_due_to_email_config + ) + self._confirmation_email_template = ( + hs.config.email_password_reset_template_confirmation_html + ) + self._email_password_reset_template_success_html = ( + hs.config.email_password_reset_template_success_html_content + ) + self._failure_email_template = ( + hs.config.email_password_reset_template_failure_html + ) + + # This resource should not be mounted if threepid behaviour is not LOCAL + assert hs.config.threepid_behaviour_email == ThreepidBehaviour.LOCAL + + async def _async_render_GET(self, request: Request) -> Tuple[int, bytes]: + sid = parse_string(request, "sid", required=True) + token = parse_string(request, "token", required=True) + client_secret = parse_string(request, "client_secret", required=True) + assert_valid_client_secret(client_secret) + + # Show a confirmation page, just in case someone accidentally clicked this link when + # they didn't mean to + template_vars = { + "sid": sid, + "token": token, + "client_secret": client_secret, + } + return ( + 200, + self._confirmation_email_template.render(**template_vars).encode("utf-8"), + ) + + async def _async_render_POST(self, request: Request) -> Tuple[int, bytes]: + sid = parse_string(request, "sid", required=True) + token = parse_string(request, "token", required=True) + client_secret = parse_string(request, "client_secret", required=True) + + # Attempt to validate a 3PID session + try: + # Mark the session as valid + next_link = await self.store.validate_threepid_session( + sid, client_secret, token, self.clock.time_msec() + ) + + # Perform a 302 redirect if next_link is set + if next_link: + if next_link.startswith("file:///"): + logger.warning( + "Not redirecting to next_link as it is a local file: address" + ) + else: + next_link_bytes = next_link.encode("utf-8") + request.setHeader("Location", next_link_bytes) + return ( + 302, + ( + b'You are being redirected to <a src="%s">%s</a>.' + % (next_link_bytes, next_link_bytes) + ), + ) + + # Otherwise show the success template + html_bytes = self._email_password_reset_template_success_html.encode( + "utf-8" + ) + status_code = 200 + except ThreepidValidationError as e: + status_code = e.code + + # Show a failure page with a reason + template_vars = {"failure_reason": e.msg} + html_bytes = self._failure_email_template.render(**template_vars).encode( + "utf-8" + ) + + return status_code, html_bytes diff --git a/synapse/rest/well_known.py b/synapse/rest/well_known.py
index 20177b44e7..f591cc6c5c 100644 --- a/synapse/rest/well_known.py +++ b/synapse/rest/well_known.py
@@ -13,17 +13,17 @@ # See the License for the specific language governing permissions and # limitations under the License. -import json import logging from twisted.web.resource import Resource from synapse.http.server import set_cors_headers +from synapse.util import json_encoder logger = logging.getLogger(__name__) -class WellKnownBuilder(object): +class WellKnownBuilder: """Utility to construct the well-known response Args: @@ -67,4 +67,4 @@ class WellKnownResource(Resource): logger.debug("returning: %s", r) request.setHeader(b"Content-Type", b"application/json") - return json.dumps(r).encode("utf-8") + return json_encoder.encode(r).encode("utf-8") diff --git a/synapse/secrets.py b/synapse/secrets.py
index 5f43f81eb0..fb6d90a3b7 100644 --- a/synapse/secrets.py +++ b/synapse/secrets.py
@@ -25,15 +25,19 @@ import sys if sys.version_info[0:2] >= (3, 6): import secrets - def Secrets(): - return secrets + class Secrets: + def token_bytes(self, nbytes=32): + return secrets.token_bytes(nbytes) + + def token_hex(self, nbytes=32): + return secrets.token_hex(nbytes) else: import binascii import os - class Secrets(object): + class Secrets: def token_bytes(self, nbytes=32): return os.urandom(nbytes) diff --git a/synapse/server.py b/synapse/server.py
index a34d8149ff..5e3752c333 100644 --- a/synapse/server.py +++ b/synapse/server.py
@@ -22,10 +22,14 @@ # Imports required for the default HomeServer() implementation import abc +import functools import logging import os +from typing import TYPE_CHECKING, Any, Callable, Dict, List, Optional, TypeVar, cast +import twisted from twisted.mail.smtp import sendmail +from twisted.web.iweb import IPolicyForHTTPS from synapse.api.auth import Auth from synapse.api.filtering import Filtering @@ -93,7 +97,7 @@ from synapse.push.pusherpool import PusherPool from synapse.replication.tcp.client import ReplicationDataHandler from synapse.replication.tcp.handler import ReplicationCommandHandler from synapse.replication.tcp.resource import ReplicationStreamer -from synapse.replication.tcp.streams import STREAMS_MAP +from synapse.replication.tcp.streams import STREAMS_MAP, Stream from synapse.rest.media.v1.media_repository import ( MediaRepository, MediaRepositoryResource, @@ -105,32 +109,75 @@ from synapse.server_notices.worker_server_notices_sender import ( WorkerServerNoticesSender, ) from synapse.state import StateHandler, StateResolutionHandler -from synapse.storage import DataStore, DataStores, Storage +from synapse.storage import Databases, DataStore, Storage from synapse.streams.events import EventSources +from synapse.types import DomainSpecificString from synapse.util import Clock from synapse.util.distributor import Distributor +from synapse.util.ratelimitutils import FederationRateLimiter from synapse.util.stringutils import random_string logger = logging.getLogger(__name__) +if TYPE_CHECKING: + from synapse.handlers.oidc_handler import OidcHandler + from synapse.handlers.saml_handler import SamlHandler -class HomeServer(object): + +T = TypeVar("T", bound=Callable[..., Any]) + + +def cache_in_self(builder: T) -> T: + """Wraps a function called e.g. `get_foo`, checking if `self.foo` exists and + returning if so. If not, calls the given function and sets `self.foo` to it. + + Also ensures that dependency cycles throw an exception correctly, rather + than overflowing the stack. + """ + + if not builder.__name__.startswith("get_"): + raise Exception( + "@cache_in_self can only be used on functions starting with `get_`" + ) + + depname = builder.__name__[len("get_") :] + + building = [False] + + @functools.wraps(builder) + def _get(self): + try: + return getattr(self, depname) + except AttributeError: + pass + + # Prevent cyclic dependencies from deadlocking + if building[0]: + raise ValueError("Cyclic dependency while building %s" % (depname,)) + + building[0] = True + try: + dep = builder(self) + setattr(self, depname, dep) + finally: + building[0] = False + + return dep + + # We cast here as we need to tell mypy that `_get` has the same signature as + # `builder`. + return cast(T, _get) + + +class HomeServer(metaclass=abc.ABCMeta): """A basic homeserver object without lazy component builders. This will need all of the components it requires to either be passed as constructor arguments, or the relevant methods overriding to create them. Typically this would only be used for unit tests. - For every dependency in the DEPENDENCIES list below, this class creates one - method, - def get_DEPENDENCY(self) - which returns the value of that dependency. If no value has yet been set - nor was provided to the constructor, it will attempt to call a lazy builder - method called - def build_DEPENDENCY(self) - which must be implemented by the subclass. This code may call any of the - required "get" methods on the instance to obtain the sub-dependencies that - one requires. + Dependencies should be added by creating a `def get_<depname>(self)` + function, wrapping it in `@cache_in_self`. Attributes: config (synapse.config.homeserver.HomeserverConfig): @@ -138,88 +185,6 @@ class HomeServer(object): we are listening on to provide HTTP services. """ - __metaclass__ = abc.ABCMeta - - DEPENDENCIES = [ - "http_client", - "federation_client", - "federation_server", - "handlers", - "auth", - "room_creation_handler", - "room_shutdown_handler", - "state_handler", - "state_resolution_handler", - "presence_handler", - "sync_handler", - "typing_handler", - "room_list_handler", - "acme_handler", - "auth_handler", - "device_handler", - "stats_handler", - "e2e_keys_handler", - "e2e_room_keys_handler", - "event_handler", - "event_stream_handler", - "initial_sync_handler", - "application_service_api", - "application_service_scheduler", - "application_service_handler", - "device_message_handler", - "profile_handler", - "event_creation_handler", - "deactivate_account_handler", - "set_password_handler", - "notifier", - "event_sources", - "keyring", - "pusherpool", - "event_builder_factory", - "filtering", - "http_client_context_factory", - "proxied_http_client", - "simple_http_client", - "proxied_http_client", - "media_repository", - "media_repository_resource", - "federation_transport_client", - "federation_sender", - "receipts_handler", - "macaroon_generator", - "tcp_replication", - "read_marker_handler", - "action_generator", - "user_directory_handler", - "groups_local_handler", - "groups_server_handler", - "groups_attestation_signing", - "groups_attestation_renewer", - "secrets", - "spam_checker", - "third_party_event_rules", - "room_member_handler", - "federation_registry", - "server_notices_manager", - "server_notices_sender", - "message_handler", - "pagination_handler", - "room_context_handler", - "sendmail", - "registration_handler", - "account_validity_handler", - "cas_handler", - "saml_handler", - "oidc_handler", - "event_client_serializer", - "password_policy_handler", - "storage", - "replication_streamer", - "replication_data_handler", - "replication_streams", - "password_policy_handler", - ] - REQUIRED_ON_MASTER_STARTUP = ["user_directory_handler", "stats_handler"] # This is overridden in derived application classes @@ -234,16 +199,17 @@ class HomeServer(object): config: The full config for the homeserver. """ if not reactor: - from twisted.internet import reactor + from twisted.internet import reactor as _reactor + + reactor = _reactor self._reactor = reactor self.hostname = hostname # the key we use to sign events and requests self.signing_key = config.key.signing_key[0] self.config = config - self._building = {} - self._listening_services = [] - self.start_time = None + self._listening_services = [] # type: List[twisted.internet.tcp.Port] + self.start_time = None # type: Optional[int] self._instance_id = random_string(5) self._instance_name = config.worker_name or "master" @@ -257,13 +223,13 @@ class HomeServer(object): burst_count=config.rc_registration.burst_count, ) - self.datastores = None + self.datastores = None # type: Optional[Databases] # Other kwargs are explicit dependencies for depname in kwargs: setattr(self, depname, kwargs[depname]) - def get_instance_id(self): + def get_instance_id(self) -> str: """A unique ID for this synapse process instance. This is used to distinguish running instances in worker-based @@ -279,13 +245,13 @@ class HomeServer(object): """ return self._instance_name - def setup(self): + def setup(self) -> None: logger.info("Setting up.") self.start_time = int(self.get_clock().time()) - self.datastores = DataStores(self.DATASTORE_CLASS, self) + self.datastores = Databases(self.DATASTORE_CLASS, self) logger.info("Finished setting up.") - def setup_master(self): + def setup_master(self) -> None: """ Some handlers have side effects on instantiation (like registering background updates). This function causes them to be fetched, and @@ -294,192 +260,242 @@ class HomeServer(object): for i in self.REQUIRED_ON_MASTER_STARTUP: getattr(self, "get_" + i)() - def get_reactor(self): + def get_reactor(self) -> twisted.internet.base.ReactorBase: """ Fetch the Twisted reactor in use by this HomeServer. """ return self._reactor - def get_ip_from_request(self, request): + def get_ip_from_request(self, request) -> str: # X-Forwarded-For is handled by our custom request type. return request.getClientIP() - def is_mine(self, domain_specific_string): + def is_mine(self, domain_specific_string: DomainSpecificString) -> bool: return domain_specific_string.domain == self.hostname - def is_mine_id(self, string): + def is_mine_id(self, string: str) -> bool: return string.split(":", 1)[1] == self.hostname - def get_clock(self): + def get_clock(self) -> Clock: return self.clock def get_datastore(self) -> DataStore: + if not self.datastores: + raise Exception("HomeServer.setup must be called before getting datastores") + return self.datastores.main - def get_datastores(self): + def get_datastores(self) -> Databases: + if not self.datastores: + raise Exception("HomeServer.setup must be called before getting datastores") + return self.datastores - def get_config(self): + def get_config(self) -> HomeServerConfig: return self.config - def get_distributor(self): + def get_distributor(self) -> Distributor: return self.distributor def get_registration_ratelimiter(self) -> Ratelimiter: return self.registration_ratelimiter - def build_federation_client(self): + @cache_in_self + def get_federation_client(self) -> FederationClient: return FederationClient(self) - def build_federation_server(self): + @cache_in_self + def get_federation_server(self) -> FederationServer: return FederationServer(self) - def build_handlers(self): + @cache_in_self + def get_handlers(self) -> Handlers: return Handlers(self) - def build_notifier(self): + @cache_in_self + def get_notifier(self) -> Notifier: return Notifier(self) - def build_auth(self): + @cache_in_self + def get_auth(self) -> Auth: return Auth(self) - def build_http_client_context_factory(self): + @cache_in_self + def get_http_client_context_factory(self) -> IPolicyForHTTPS: return ( InsecureInterceptableContextFactory() if self.config.use_insecure_ssl_client_just_for_testing_do_not_use else RegularPolicyForHTTPS() ) - def build_simple_http_client(self): + @cache_in_self + def get_simple_http_client(self) -> SimpleHttpClient: return SimpleHttpClient(self) - def build_proxied_http_client(self): + @cache_in_self + def get_proxied_http_client(self) -> SimpleHttpClient: return SimpleHttpClient( self, http_proxy=os.getenvb(b"http_proxy"), https_proxy=os.getenvb(b"HTTPS_PROXY"), ) - def build_room_creation_handler(self): + @cache_in_self + def get_room_creation_handler(self) -> RoomCreationHandler: return RoomCreationHandler(self) - def build_room_shutdown_handler(self): + @cache_in_self + def get_room_shutdown_handler(self) -> RoomShutdownHandler: return RoomShutdownHandler(self) - def build_sendmail(self): + @cache_in_self + def get_sendmail(self) -> sendmail: return sendmail - def build_state_handler(self): + @cache_in_self + def get_state_handler(self) -> StateHandler: return StateHandler(self) - def build_state_resolution_handler(self): + @cache_in_self + def get_state_resolution_handler(self) -> StateResolutionHandler: return StateResolutionHandler(self) - def build_presence_handler(self): + @cache_in_self + def get_presence_handler(self) -> PresenceHandler: return PresenceHandler(self) - def build_typing_handler(self): + @cache_in_self + def get_typing_handler(self): if self.config.worker.writers.typing == self.get_instance_name(): return TypingWriterHandler(self) else: return FollowerTypingHandler(self) - def build_sync_handler(self): + @cache_in_self + def get_sync_handler(self) -> SyncHandler: return SyncHandler(self) - def build_room_list_handler(self): + @cache_in_self + def get_room_list_handler(self) -> RoomListHandler: return RoomListHandler(self) - def build_auth_handler(self): + @cache_in_self + def get_auth_handler(self) -> AuthHandler: return AuthHandler(self) - def build_macaroon_generator(self): + @cache_in_self + def get_macaroon_generator(self) -> MacaroonGenerator: return MacaroonGenerator(self) - def build_device_handler(self): + @cache_in_self + def get_device_handler(self): if self.config.worker_app: return DeviceWorkerHandler(self) else: return DeviceHandler(self) - def build_device_message_handler(self): + @cache_in_self + def get_device_message_handler(self) -> DeviceMessageHandler: return DeviceMessageHandler(self) - def build_e2e_keys_handler(self): + @cache_in_self + def get_e2e_keys_handler(self) -> E2eKeysHandler: return E2eKeysHandler(self) - def build_e2e_room_keys_handler(self): + @cache_in_self + def get_e2e_room_keys_handler(self) -> E2eRoomKeysHandler: return E2eRoomKeysHandler(self) - def build_acme_handler(self): + @cache_in_self + def get_acme_handler(self) -> AcmeHandler: return AcmeHandler(self) - def build_application_service_api(self): + @cache_in_self + def get_application_service_api(self) -> ApplicationServiceApi: return ApplicationServiceApi(self) - def build_application_service_scheduler(self): + @cache_in_self + def get_application_service_scheduler(self) -> ApplicationServiceScheduler: return ApplicationServiceScheduler(self) - def build_application_service_handler(self): + @cache_in_self + def get_application_service_handler(self) -> ApplicationServicesHandler: return ApplicationServicesHandler(self) - def build_event_handler(self): + @cache_in_self + def get_event_handler(self) -> EventHandler: return EventHandler(self) - def build_event_stream_handler(self): + @cache_in_self + def get_event_stream_handler(self) -> EventStreamHandler: return EventStreamHandler(self) - def build_initial_sync_handler(self): + @cache_in_self + def get_initial_sync_handler(self) -> InitialSyncHandler: return InitialSyncHandler(self) - def build_profile_handler(self): + @cache_in_self + def get_profile_handler(self): if self.config.worker_app: return BaseProfileHandler(self) else: return MasterProfileHandler(self) - def build_event_creation_handler(self): + @cache_in_self + def get_event_creation_handler(self) -> EventCreationHandler: return EventCreationHandler(self) - def build_deactivate_account_handler(self): + @cache_in_self + def get_deactivate_account_handler(self) -> DeactivateAccountHandler: return DeactivateAccountHandler(self) - def build_set_password_handler(self): + @cache_in_self + def get_set_password_handler(self) -> SetPasswordHandler: return SetPasswordHandler(self) - def build_event_sources(self): + @cache_in_self + def get_event_sources(self) -> EventSources: return EventSources(self) - def build_keyring(self): + @cache_in_self + def get_keyring(self) -> Keyring: return Keyring(self) - def build_event_builder_factory(self): + @cache_in_self + def get_event_builder_factory(self) -> EventBuilderFactory: return EventBuilderFactory(self) - def build_filtering(self): + @cache_in_self + def get_filtering(self) -> Filtering: return Filtering(self) - def build_pusherpool(self): + @cache_in_self + def get_pusherpool(self) -> PusherPool: return PusherPool(self) - def build_http_client(self): + @cache_in_self + def get_http_client(self) -> MatrixFederationHttpClient: tls_client_options_factory = context_factory.FederationPolicyForHTTPS( self.config ) return MatrixFederationHttpClient(self, tls_client_options_factory) - def build_media_repository_resource(self): + @cache_in_self + def get_media_repository_resource(self) -> MediaRepositoryResource: # build the media repo resource. This indirects through the HomeServer # to ensure that we only have a single instance of return MediaRepositoryResource(self) - def build_media_repository(self): + @cache_in_self + def get_media_repository(self) -> MediaRepository: return MediaRepository(self) - def build_federation_transport_client(self): + @cache_in_self + def get_federation_transport_client(self) -> TransportLayerClient: return TransportLayerClient(self) - def build_federation_sender(self): + @cache_in_self + def get_federation_sender(self): if self.should_send_federation(): return FederationSender(self) elif not self.config.worker_app: @@ -487,156 +503,156 @@ class HomeServer(object): else: raise Exception("Workers cannot send federation traffic") - def build_receipts_handler(self): + @cache_in_self + def get_receipts_handler(self) -> ReceiptsHandler: return ReceiptsHandler(self) - def build_read_marker_handler(self): + @cache_in_self + def get_read_marker_handler(self) -> ReadMarkerHandler: return ReadMarkerHandler(self) - def build_tcp_replication(self): + @cache_in_self + def get_tcp_replication(self) -> ReplicationCommandHandler: return ReplicationCommandHandler(self) - def build_action_generator(self): + @cache_in_self + def get_action_generator(self) -> ActionGenerator: return ActionGenerator(self) - def build_user_directory_handler(self): + @cache_in_self + def get_user_directory_handler(self) -> UserDirectoryHandler: return UserDirectoryHandler(self) - def build_groups_local_handler(self): + @cache_in_self + def get_groups_local_handler(self): if self.config.worker_app: return GroupsLocalWorkerHandler(self) else: return GroupsLocalHandler(self) - def build_groups_server_handler(self): + @cache_in_self + def get_groups_server_handler(self): if self.config.worker_app: return GroupsServerWorkerHandler(self) else: return GroupsServerHandler(self) - def build_groups_attestation_signing(self): + @cache_in_self + def get_groups_attestation_signing(self) -> GroupAttestationSigning: return GroupAttestationSigning(self) - def build_groups_attestation_renewer(self): + @cache_in_self + def get_groups_attestation_renewer(self) -> GroupAttestionRenewer: return GroupAttestionRenewer(self) - def build_secrets(self): + @cache_in_self + def get_secrets(self) -> Secrets: return Secrets() - def build_stats_handler(self): + @cache_in_self + def get_stats_handler(self) -> StatsHandler: return StatsHandler(self) - def build_spam_checker(self): + @cache_in_self + def get_spam_checker(self): return SpamChecker(self) - def build_third_party_event_rules(self): + @cache_in_self + def get_third_party_event_rules(self) -> ThirdPartyEventRules: return ThirdPartyEventRules(self) - def build_room_member_handler(self): + @cache_in_self + def get_room_member_handler(self): if self.config.worker_app: return RoomMemberWorkerHandler(self) return RoomMemberMasterHandler(self) - def build_federation_registry(self): + @cache_in_self + def get_federation_registry(self) -> FederationHandlerRegistry: return FederationHandlerRegistry(self) - def build_server_notices_manager(self): + @cache_in_self + def get_server_notices_manager(self): if self.config.worker_app: raise Exception("Workers cannot send server notices") return ServerNoticesManager(self) - def build_server_notices_sender(self): + @cache_in_self + def get_server_notices_sender(self): if self.config.worker_app: return WorkerServerNoticesSender(self) return ServerNoticesSender(self) - def build_message_handler(self): + @cache_in_self + def get_message_handler(self) -> MessageHandler: return MessageHandler(self) - def build_pagination_handler(self): + @cache_in_self + def get_pagination_handler(self) -> PaginationHandler: return PaginationHandler(self) - def build_room_context_handler(self): + @cache_in_self + def get_room_context_handler(self) -> RoomContextHandler: return RoomContextHandler(self) - def build_registration_handler(self): + @cache_in_self + def get_registration_handler(self) -> RegistrationHandler: return RegistrationHandler(self) - def build_account_validity_handler(self): + @cache_in_self + def get_account_validity_handler(self) -> AccountValidityHandler: return AccountValidityHandler(self) - def build_cas_handler(self): + @cache_in_self + def get_cas_handler(self) -> CasHandler: return CasHandler(self) - def build_saml_handler(self): + @cache_in_self + def get_saml_handler(self) -> "SamlHandler": from synapse.handlers.saml_handler import SamlHandler return SamlHandler(self) - def build_oidc_handler(self): + @cache_in_self + def get_oidc_handler(self) -> "OidcHandler": from synapse.handlers.oidc_handler import OidcHandler return OidcHandler(self) - def build_event_client_serializer(self): + @cache_in_self + def get_event_client_serializer(self) -> EventClientSerializer: return EventClientSerializer(self) - def build_storage(self) -> Storage: - return Storage(self, self.datastores) + @cache_in_self + def get_password_policy_handler(self) -> PasswordPolicyHandler: + return PasswordPolicyHandler(self) - def build_replication_streamer(self) -> ReplicationStreamer: + @cache_in_self + def get_storage(self) -> Storage: + return Storage(self, self.get_datastores()) + + @cache_in_self + def get_replication_streamer(self) -> ReplicationStreamer: return ReplicationStreamer(self) - def build_replication_data_handler(self): + @cache_in_self + def get_replication_data_handler(self) -> ReplicationDataHandler: return ReplicationDataHandler(self) - def build_replication_streams(self): + @cache_in_self + def get_replication_streams(self) -> Dict[str, Stream]: return {stream.NAME: stream(self) for stream in STREAMS_MAP.values()} - def build_password_policy_handler(self): - return PasswordPolicyHandler(self) + @cache_in_self + def get_federation_ratelimiter(self) -> FederationRateLimiter: + return FederationRateLimiter(self.clock, config=self.config.rc_federation) - def remove_pusher(self, app_id, push_key, user_id): - return self.get_pusherpool().remove_pusher(app_id, push_key, user_id) + async def remove_pusher(self, app_id: str, push_key: str, user_id: str): + return await self.get_pusherpool().remove_pusher(app_id, push_key, user_id) - def should_send_federation(self): + def should_send_federation(self) -> bool: "Should this server be sending federation traffic directly?" return self.config.send_federation and ( not self.config.worker_app or self.config.worker_app == "synapse.app.federation_sender" ) - - -def _make_dependency_method(depname): - def _get(hs): - try: - return getattr(hs, depname) - except AttributeError: - pass - - try: - builder = getattr(hs, "build_%s" % (depname)) - except AttributeError: - raise NotImplementedError( - "%s has no %s nor a builder for it" % (type(hs).__name__, depname) - ) - - # Prevent cyclic dependencies from deadlocking - if depname in hs._building: - raise ValueError("Cyclic dependency while building %s" % (depname,)) - - hs._building[depname] = 1 - try: - dep = builder() - setattr(hs, depname, dep) - finally: - del hs._building[depname] - - return dep - - setattr(HomeServer, "get_%s" % (depname), _get) - - -# Build magic accessors for every dependency -for depname in HomeServer.DEPENDENCIES: - _make_dependency_method(depname) diff --git a/synapse/server.pyi b/synapse/server.pyi deleted file mode 100644
index 1aba408c21..0000000000 --- a/synapse/server.pyi +++ /dev/null
@@ -1,155 +0,0 @@ -from typing import Dict - -import twisted.internet - -import synapse.api.auth -import synapse.config.homeserver -import synapse.crypto.keyring -import synapse.federation.federation_server -import synapse.federation.sender -import synapse.federation.transport.client -import synapse.handlers -import synapse.handlers.auth -import synapse.handlers.deactivate_account -import synapse.handlers.device -import synapse.handlers.e2e_keys -import synapse.handlers.message -import synapse.handlers.presence -import synapse.handlers.register -import synapse.handlers.room -import synapse.handlers.room_member -import synapse.handlers.set_password -import synapse.http.client -import synapse.http.matrixfederationclient -import synapse.notifier -import synapse.push.pusherpool -import synapse.replication.tcp.client -import synapse.replication.tcp.handler -import synapse.rest.media.v1.media_repository -import synapse.server_notices.server_notices_manager -import synapse.server_notices.server_notices_sender -import synapse.state -import synapse.storage -from synapse.events.builder import EventBuilderFactory -from synapse.handlers.typing import FollowerTypingHandler -from synapse.replication.tcp.streams import Stream - -class HomeServer(object): - @property - def config(self) -> synapse.config.homeserver.HomeServerConfig: - pass - @property - def hostname(self) -> str: - pass - def get_auth(self) -> synapse.api.auth.Auth: - pass - def get_auth_handler(self) -> synapse.handlers.auth.AuthHandler: - pass - def get_datastore(self) -> synapse.storage.DataStore: - pass - def get_device_handler(self) -> synapse.handlers.device.DeviceHandler: - pass - def get_e2e_keys_handler(self) -> synapse.handlers.e2e_keys.E2eKeysHandler: - pass - def get_handlers(self) -> synapse.handlers.Handlers: - pass - def get_state_handler(self) -> synapse.state.StateHandler: - pass - def get_state_resolution_handler(self) -> synapse.state.StateResolutionHandler: - pass - def get_simple_http_client(self) -> synapse.http.client.SimpleHttpClient: - """Fetch an HTTP client implementation which doesn't do any blacklisting - or support any HTTP_PROXY settings""" - pass - def get_proxied_http_client(self) -> synapse.http.client.SimpleHttpClient: - """Fetch an HTTP client implementation which doesn't do any blacklisting - but does support HTTP_PROXY settings""" - pass - def get_deactivate_account_handler( - self, - ) -> synapse.handlers.deactivate_account.DeactivateAccountHandler: - pass - def get_room_creation_handler(self) -> synapse.handlers.room.RoomCreationHandler: - pass - def get_room_member_handler(self) -> synapse.handlers.room_member.RoomMemberHandler: - pass - def get_room_shutdown_handler(self) -> synapse.handlers.room.RoomShutdownHandler: - pass - def get_event_creation_handler( - self, - ) -> synapse.handlers.message.EventCreationHandler: - pass - def get_set_password_handler( - self, - ) -> synapse.handlers.set_password.SetPasswordHandler: - pass - def get_federation_sender(self) -> synapse.federation.sender.FederationSender: - pass - def get_federation_transport_client( - self, - ) -> synapse.federation.transport.client.TransportLayerClient: - pass - def get_media_repository_resource( - self, - ) -> synapse.rest.media.v1.media_repository.MediaRepositoryResource: - pass - def get_media_repository( - self, - ) -> synapse.rest.media.v1.media_repository.MediaRepository: - pass - def get_server_notices_manager( - self, - ) -> synapse.server_notices.server_notices_manager.ServerNoticesManager: - pass - def get_server_notices_sender( - self, - ) -> synapse.server_notices.server_notices_sender.ServerNoticesSender: - pass - def get_notifier(self) -> synapse.notifier.Notifier: - pass - def get_presence_handler(self) -> synapse.handlers.presence.BasePresenceHandler: - pass - def get_clock(self) -> synapse.util.Clock: - pass - def get_reactor(self) -> twisted.internet.base.ReactorBase: - pass - def get_keyring(self) -> synapse.crypto.keyring.Keyring: - pass - def get_tcp_replication( - self, - ) -> synapse.replication.tcp.handler.ReplicationCommandHandler: - pass - def get_replication_data_handler( - self, - ) -> synapse.replication.tcp.client.ReplicationDataHandler: - pass - def get_federation_registry( - self, - ) -> synapse.federation.federation_server.FederationHandlerRegistry: - pass - def is_mine_id(self, domain_id: str) -> bool: - pass - def get_instance_id(self) -> str: - pass - def get_instance_name(self) -> str: - pass - def get_event_builder_factory(self) -> EventBuilderFactory: - pass - def get_storage(self) -> synapse.storage.Storage: - pass - def get_registration_handler(self) -> synapse.handlers.register.RegistrationHandler: - pass - def get_macaroon_generator(self) -> synapse.handlers.auth.MacaroonGenerator: - pass - def get_pusherpool(self) -> synapse.push.pusherpool.PusherPool: - pass - def get_replication_streams(self) -> Dict[str, Stream]: - pass - def get_http_client( - self, - ) -> synapse.http.matrixfederationclient.MatrixFederationHttpClient: - pass - def should_send_federation(self) -> bool: - pass - def get_typing_handler(self) -> FollowerTypingHandler: - pass diff --git a/synapse/server_notices/consent_server_notices.py b/synapse/server_notices/consent_server_notices.py
index 3bfc8d7278..3673e7f47e 100644 --- a/synapse/server_notices/consent_server_notices.py +++ b/synapse/server_notices/consent_server_notices.py
@@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from typing import Any from synapse.api.errors import SynapseError from synapse.api.urls import ConsentURIBuilder @@ -22,7 +23,7 @@ from synapse.types import get_localpart_from_id logger = logging.getLogger(__name__) -class ConsentServerNotices(object): +class ConsentServerNotices: """Keeps track of whether we need to send users server_notices about privacy policy consent, and sends one if we do. """ @@ -55,14 +56,11 @@ class ConsentServerNotices(object): self._consent_uri_builder = ConsentURIBuilder(hs.config) - async def maybe_send_server_notice_to_user(self, user_id): + async def maybe_send_server_notice_to_user(self, user_id: str) -> None: """Check if we need to send a notice to this user, and does so if so Args: - user_id (str): user to check - - Returns: - Deferred + user_id: user to check """ if self._server_notice_content is None: # not enabled @@ -105,7 +103,7 @@ class ConsentServerNotices(object): self._users_in_progress.remove(user_id) -def copy_with_str_subst(x, substitutions): +def copy_with_str_subst(x: Any, substitutions: Any) -> Any: """Deep-copy a structure, carrying out string substitions on any strings Args: @@ -121,7 +119,7 @@ def copy_with_str_subst(x, substitutions): if isinstance(x, dict): return {k: copy_with_str_subst(v, substitutions) for (k, v) in x.items()} if isinstance(x, (list, tuple)): - return [copy_with_str_subst(y) for y in x] + return [copy_with_str_subst(y, substitutions) for y in x] # assume it's uninterested and can be shallow-copied. return x diff --git a/synapse/server_notices/resource_limits_server_notices.py b/synapse/server_notices/resource_limits_server_notices.py
index 4404ceff93..2258d306d9 100644 --- a/synapse/server_notices/resource_limits_server_notices.py +++ b/synapse/server_notices/resource_limits_server_notices.py
@@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from typing import List, Tuple from synapse.api.constants import ( EventTypes, @@ -26,7 +27,7 @@ from synapse.server_notices.server_notices_manager import SERVER_NOTICE_ROOM_TAG logger = logging.getLogger(__name__) -class ResourceLimitsServerNotices(object): +class ResourceLimitsServerNotices: """ Keeps track of whether the server has reached it's resource limit and ensures that the client is kept up to date. """ @@ -52,7 +53,7 @@ class ResourceLimitsServerNotices(object): and not hs.config.hs_disabled ) - async def maybe_send_server_notice_to_user(self, user_id): + async def maybe_send_server_notice_to_user(self, user_id: str) -> None: """Check if we need to send a notice to this user, this will be true in two cases. 1. The server has reached its limit does not reflect this @@ -60,10 +61,7 @@ class ResourceLimitsServerNotices(object): actually the server is fine Args: - user_id (str): user to check - - Returns: - Deferred + user_id: user to check """ if not self._enabled: return @@ -115,19 +113,21 @@ class ResourceLimitsServerNotices(object): elif not currently_blocked and limit_msg: # Room is not notifying of a block, when it ought to be. await self._apply_limit_block_notification( - user_id, limit_msg, limit_type + user_id, limit_msg, limit_type # type: ignore ) except SynapseError as e: logger.error("Error sending resource limits server notice: %s", e) - async def _remove_limit_block_notification(self, user_id, ref_events): + async def _remove_limit_block_notification( + self, user_id: str, ref_events: List[str] + ) -> None: """Utility method to remove limit block notifications from the server notices room. Args: - user_id (str): user to notify - ref_events (list[str]): The event_ids of pinned events that are unrelated to - limit blocking and need to be preserved. + user_id: user to notify + ref_events: The event_ids of pinned events that are unrelated to + limit blocking and need to be preserved. """ content = {"pinned": ref_events} await self._server_notices_manager.send_notice( @@ -135,16 +135,16 @@ class ResourceLimitsServerNotices(object): ) async def _apply_limit_block_notification( - self, user_id, event_body, event_limit_type - ): + self, user_id: str, event_body: str, event_limit_type: str + ) -> None: """Utility method to apply limit block notifications in the server notices room. Args: - user_id (str): user to notify - event_body(str): The human readable text that describes the block. - event_limit_type(str): Specifies the type of block e.g. monthly active user - limit has been exceeded. + user_id: user to notify + event_body: The human readable text that describes the block. + event_limit_type: Specifies the type of block e.g. monthly active user + limit has been exceeded. """ content = { "body": event_body, @@ -162,7 +162,7 @@ class ResourceLimitsServerNotices(object): user_id, content, EventTypes.Pinned, "" ) - async def _check_and_set_tags(self, user_id, room_id): + async def _check_and_set_tags(self, user_id: str, room_id: str) -> None: """ Since server notices rooms were originally not with tags, important to check that tags have been set correctly @@ -182,17 +182,16 @@ class ResourceLimitsServerNotices(object): ) self._notifier.on_new_event("account_data_key", max_id, users=[user_id]) - async def _is_room_currently_blocked(self, room_id): + async def _is_room_currently_blocked(self, room_id: str) -> Tuple[bool, List[str]]: """ Determines if the room is currently blocked Args: - room_id(str): The room id of the server notices room + room_id: The room id of the server notices room Returns: - Deferred[Tuple[bool, List]]: bool: Is the room currently blocked - list: The list of pinned events that are unrelated to limit blocking + list: The list of pinned event IDs that are unrelated to limit blocking This list can be used as a convenience in the case where the block is to be lifted and the remaining pinned event references need to be preserved @@ -207,7 +206,7 @@ class ResourceLimitsServerNotices(object): # The user has yet to join the server notices room pass - referenced_events = [] + referenced_events = [] # type: List[str] if pinned_state_event is not None: referenced_events = list(pinned_state_event.content.get("pinned", [])) diff --git a/synapse/server_notices/server_notices_manager.py b/synapse/server_notices/server_notices_manager.py
index bf2454c01c..0422d4c7ce 100644 --- a/synapse/server_notices/server_notices_manager.py +++ b/synapse/server_notices/server_notices_manager.py
@@ -13,8 +13,10 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from typing import Optional from synapse.api.constants import EventTypes, Membership, RoomCreationPreset +from synapse.events import EventBase from synapse.types import UserID, create_requester from synapse.util.caches.descriptors import cached @@ -23,7 +25,7 @@ logger = logging.getLogger(__name__) SERVER_NOTICE_ROOM_TAG = "m.server_notice" -class ServerNoticesManager(object): +class ServerNoticesManager: def __init__(self, hs): """ @@ -50,20 +52,21 @@ class ServerNoticesManager(object): return self._config.server_notices_mxid is not None async def send_notice( - self, user_id, event_content, type=EventTypes.Message, state_key=None - ): + self, + user_id: str, + event_content: dict, + type: str = EventTypes.Message, + state_key: Optional[bool] = None, + ) -> EventBase: """Send a notice to the given user Creates the server notices room, if none exists. Args: - user_id (str): mxid of user to send event to. - event_content (dict): content of event to send - type(EventTypes): type of event - is_state_event(bool): Is the event a state event - - Returns: - Deferred[FrozenEvent] + user_id: mxid of user to send event to. + event_content: content of event to send + type: type of event + is_state_event: Is the event a state event """ room_id = await self.get_or_create_notice_room_for_user(user_id) await self.maybe_invite_user_to_room(user_id, room_id) @@ -89,17 +92,17 @@ class ServerNoticesManager(object): return event @cached() - async def get_or_create_notice_room_for_user(self, user_id): + async def get_or_create_notice_room_for_user(self, user_id: str) -> str: """Get the room for notices for a given user If we have not yet created a notice room for this user, create it, but don't invite the user to it. Args: - user_id (str): complete user id for the user we want a room for + user_id: complete user id for the user we want a room for Returns: - str: room id of notice room. + room id of notice room. """ if not self.is_enabled(): raise Exception("Server notices not enabled") @@ -163,7 +166,7 @@ class ServerNoticesManager(object): logger.info("Created server notices room %s for %s", room_id, user_id) return room_id - async def maybe_invite_user_to_room(self, user_id: str, room_id: str): + async def maybe_invite_user_to_room(self, user_id: str, room_id: str) -> None: """Invite the given user to the given server room, unless the user has already joined or been invited to it. diff --git a/synapse/server_notices/server_notices_sender.py b/synapse/server_notices/server_notices_sender.py
index be74e86641..6870b67ca0 100644 --- a/synapse/server_notices/server_notices_sender.py +++ b/synapse/server_notices/server_notices_sender.py
@@ -12,13 +12,15 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. +from typing import Iterable, Union + from synapse.server_notices.consent_server_notices import ConsentServerNotices from synapse.server_notices.resource_limits_server_notices import ( ResourceLimitsServerNotices, ) -class ServerNoticesSender(object): +class ServerNoticesSender: """A centralised place which sends server notices automatically when Certain Events take place """ @@ -32,22 +34,22 @@ class ServerNoticesSender(object): self._server_notices = ( ConsentServerNotices(hs), ResourceLimitsServerNotices(hs), - ) + ) # type: Iterable[Union[ConsentServerNotices, ResourceLimitsServerNotices]] - async def on_user_syncing(self, user_id): + async def on_user_syncing(self, user_id: str) -> None: """Called when the user performs a sync operation. Args: - user_id (str): mxid of user who synced + user_id: mxid of user who synced """ for sn in self._server_notices: await sn.maybe_send_server_notice_to_user(user_id) - async def on_user_ip(self, user_id): + async def on_user_ip(self, user_id: str) -> None: """Called on the master when a worker process saw a client request. Args: - user_id (str): mxid + user_id: mxid """ # The synchrotrons use a stubbed version of ServerNoticesSender, so # we check for notices to send to the user in on_user_ip as well as diff --git a/synapse/server_notices/worker_server_notices_sender.py b/synapse/server_notices/worker_server_notices_sender.py
index 245ec7c64f..9273e61895 100644 --- a/synapse/server_notices/worker_server_notices_sender.py +++ b/synapse/server_notices/worker_server_notices_sender.py
@@ -12,10 +12,9 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from twisted.internet import defer -class WorkerServerNoticesSender(object): +class WorkerServerNoticesSender: """Stub impl of ServerNoticesSender which does nothing""" def __init__(self, hs): @@ -24,24 +23,18 @@ class WorkerServerNoticesSender(object): hs (synapse.server.HomeServer): """ - def on_user_syncing(self, user_id): + async def on_user_syncing(self, user_id: str) -> None: """Called when the user performs a sync operation. Args: - user_id (str): mxid of user who synced - - Returns: - Deferred + user_id: mxid of user who synced """ - return defer.succeed(None) + return None - def on_user_ip(self, user_id): + async def on_user_ip(self, user_id: str) -> None: """Called on the master when a worker process saw a client request. Args: - user_id (str): mxid - - Returns: - Deferred + user_id: mxid """ raise AssertionError("on_user_ip unexpectedly called on worker") diff --git a/synapse/spam_checker_api/__init__.py b/synapse/spam_checker_api/__init__.py
index 9b78924d96..395ac5ab02 100644 --- a/synapse/spam_checker_api/__init__.py +++ b/synapse/spam_checker_api/__init__.py
@@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from enum import Enum from twisted.internet import defer @@ -25,7 +26,17 @@ if MYPY: logger = logging.getLogger(__name__) -class SpamCheckerApi(object): +class RegistrationBehaviour(Enum): + """ + Enum to define whether a registration request should allowed, denied, or shadow-banned. + """ + + ALLOW = "allow" + SHADOW_BAN = "shadow_ban" + DENY = "deny" + + +class SpamCheckerApi: """A proxy object that gets passed to spam checkers so they can get access to rooms and other relevant information. """ @@ -48,8 +59,10 @@ class SpamCheckerApi(object): twisted.internet.defer.Deferred[list(synapse.events.FrozenEvent)]: The filtered state events in the room. """ - state_ids = yield self._store.get_filtered_current_state_ids( - room_id=room_id, state_filter=StateFilter.from_types(types) + state_ids = yield defer.ensureDeferred( + self._store.get_filtered_current_state_ids( + room_id=room_id, state_filter=StateFilter.from_types(types) + ) ) - state = yield self._store.get_events(state_ids.values()) + state = yield defer.ensureDeferred(self._store.get_events(state_ids.values())) return state.values() diff --git a/synapse/state/__init__.py b/synapse/state/__init__.py
index 25ccef5aa5..31082bb16a 100644 --- a/synapse/state/__init__.py +++ b/synapse/state/__init__.py
@@ -13,31 +13,46 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - +import heapq import logging -from collections import namedtuple -from typing import Awaitable, Dict, Iterable, List, Optional, Set +from collections import defaultdict, namedtuple +from typing import ( + Any, + Awaitable, + Callable, + DefaultDict, + Dict, + Iterable, + List, + Optional, + Sequence, + Set, + Tuple, + Union, + overload, +) import attr from frozendict import frozendict -from prometheus_client import Histogram +from prometheus_client import Counter, Histogram +from typing_extensions import Literal from synapse.api.constants import EventTypes from synapse.api.room_versions import KNOWN_ROOM_VERSIONS, StateResolutionVersions from synapse.events import EventBase from synapse.events.snapshot import EventContext +from synapse.logging.context import ContextResourceUsage from synapse.logging.utils import log_function from synapse.state import v1, v2 -from synapse.storage.data_stores.main.events_worker import EventRedactBehaviour +from synapse.storage.databases.main.events_worker import EventRedactBehaviour from synapse.storage.roommember import ProfileInfo -from synapse.types import StateMap -from synapse.util import Clock +from synapse.types import Collection, StateMap from synapse.util.async_helpers import Linearizer from synapse.util.caches.expiringcache import ExpiringCache from synapse.util.metrics import Measure, measure_func logger = logging.getLogger(__name__) - +metrics_logger = logging.getLogger("synapse.state.metrics") # Metrics for number of state groups involved in a resolution. state_groups_histogram = Histogram( @@ -65,11 +80,17 @@ def _gen_state_id(): return s -class _StateCacheEntry(object): +class _StateCacheEntry: __slots__ = ["state", "state_group", "state_id", "prev_group", "delta_ids"] - def __init__(self, state, state_group, prev_group=None, delta_ids=None): - # dict[(str, str), str] map from (type, state_key) to event_id + def __init__( + self, + state: StateMap[str], + state_group: Optional[int], + prev_group: Optional[int] = None, + delta_ids: Optional[StateMap[str]] = None, + ): + # A map from (type, state_key) to event_id. self.state = frozendict(state) # the ID of a state group if one and only one is involved. @@ -95,7 +116,7 @@ class _StateCacheEntry(object): return len(self.state) -class StateHandler(object): +class StateHandler: """Fetches bits of state from the stores, and does state resolution where necessary """ @@ -107,24 +128,49 @@ class StateHandler(object): self.hs = hs self._state_resolution_handler = hs.get_state_resolution_handler() + @overload async def get_current_state( - self, room_id, event_type=None, state_key="", latest_event_ids=None - ): - """ Retrieves the current state for the room. This is done by + self, + room_id: str, + event_type: Literal[None] = None, + state_key: str = "", + latest_event_ids: Optional[List[str]] = None, + ) -> StateMap[EventBase]: + ... + + @overload + async def get_current_state( + self, + room_id: str, + event_type: str, + state_key: str = "", + latest_event_ids: Optional[List[str]] = None, + ) -> Optional[EventBase]: + ... + + async def get_current_state( + self, + room_id: str, + event_type: Optional[str] = None, + state_key: str = "", + latest_event_ids: Optional[List[str]] = None, + ) -> Union[Optional[EventBase], StateMap[EventBase]]: + """Retrieves the current state for the room. This is done by calling `get_latest_events_in_room` to get the leading edges of the event graph and then resolving any of the state conflicts. This is equivalent to getting the state of an event that were to send next before receiving any new events. - If `event_type` is specified, then the method returns only the one - event (or None) with that `event_type` and `state_key`. - Returns: - map from (type, state_key) to event + If `event_type` is specified, then the method returns only the one + event (or None) with that `event_type` and `state_key`. + + Otherwise, a map from (type, state_key) to event. """ if not latest_event_ids: latest_event_ids = await self.store.get_latest_event_ids_in_room(room_id) + assert latest_event_ids is not None logger.debug("calling resolve_state_groups from get_current_state") ret = await self.resolve_state_groups_for_events(room_id, latest_event_ids) @@ -140,34 +186,30 @@ class StateHandler(object): state_map = await self.store.get_events( list(state.values()), get_prev_content=False ) - state = { + return { key: state_map[e_id] for key, e_id in state.items() if e_id in state_map } - return state - - async def get_current_state_ids(self, room_id, latest_event_ids=None): + async def get_current_state_ids( + self, room_id: str, latest_event_ids: Optional[Iterable[str]] = None + ) -> StateMap[str]: """Get the current state, or the state at a set of events, for a room Args: - room_id (str): - - latest_event_ids (iterable[str]|None): if given, the forward - extremities to resolve. If None, we look them up from the - database (via a cache) + room_id: + latest_event_ids: if given, the forward extremities to resolve. If + None, we look them up from the database (via a cache). Returns: - Deferred[dict[(str, str), str)]]: the state dict, mapping from - (event_type, state_key) -> event_id + the state dict, mapping from (event_type, state_key) -> event_id """ if not latest_event_ids: latest_event_ids = await self.store.get_latest_event_ids_in_room(room_id) + assert latest_event_ids is not None logger.debug("calling resolve_state_groups from get_current_state_ids") ret = await self.resolve_state_groups_for_events(room_id, latest_event_ids) - state = ret.state - - return state + return ret.state async def get_current_users_in_room( self, room_id: str, latest_event_ids: Optional[List[str]] = None @@ -183,32 +225,34 @@ class StateHandler(object): """ if not latest_event_ids: latest_event_ids = await self.store.get_latest_event_ids_in_room(room_id) + assert latest_event_ids is not None + logger.debug("calling resolve_state_groups from get_current_users_in_room") entry = await self.resolve_state_groups_for_events(room_id, latest_event_ids) - joined_users = await self.store.get_joined_users_from_state(room_id, entry) - return joined_users + return await self.store.get_joined_users_from_state(room_id, entry) - async def get_current_hosts_in_room(self, room_id): + async def get_current_hosts_in_room(self, room_id: str) -> Set[str]: event_ids = await self.store.get_latest_event_ids_in_room(room_id) return await self.get_hosts_in_room_at_events(room_id, event_ids) - async def get_hosts_in_room_at_events(self, room_id, event_ids): + async def get_hosts_in_room_at_events( + self, room_id: str, event_ids: List[str] + ) -> Set[str]: """Get the hosts that were in a room at the given event ids Args: - room_id (str): - event_ids (list[str]): + room_id: + event_ids: Returns: - Deferred[list[str]]: the hosts in the room at the given events + The hosts in the room at the given events """ entry = await self.resolve_state_groups_for_events(room_id, event_ids) - joined_hosts = await self.store.get_joined_hosts(room_id, entry) - return joined_hosts + return await self.store.get_joined_hosts(room_id, entry) async def compute_event_context( self, event: EventBase, old_state: Optional[Iterable[EventBase]] = None - ): + ) -> EventContext: """Build an EventContext structure for the event. This works out what the current state should be for the event, and @@ -221,7 +265,7 @@ class StateHandler(object): when receiving an event from federation where we don't have the prev events for, e.g. when backfilling. Returns: - synapse.events.snapshot.EventContext: + The event context. """ if event.internal_metadata.is_outlier(): @@ -262,7 +306,7 @@ class StateHandler(object): # if we're given the state before the event, then we use that state_ids_before_event = { (s.type, s.state_key): s.event_id for s in old_state - } + } # type: StateMap[str] state_group_before_event = None state_group_before_event_prev_group = None deltas_to_state_group_before_event = None @@ -346,19 +390,18 @@ class StateHandler(object): ) @measure_func() - async def resolve_state_groups_for_events(self, room_id, event_ids): + async def resolve_state_groups_for_events( + self, room_id: str, event_ids: Iterable[str] + ) -> _StateCacheEntry: """ Given a list of event_ids this method fetches the state at each event, resolves conflicts between them and returns them. Args: - room_id (str) - event_ids (list[str]) - explicit_room_version (str|None): If set uses the the given room - version to choose the resolution algorithm. If None, then - checks the database for room version. + room_id + event_ids Returns: - Deferred[_StateCacheEntry]: resolved state + The resolved state """ logger.debug("resolve_state_groups event_ids %s", event_ids) @@ -394,7 +437,12 @@ class StateHandler(object): ) return result - async def resolve_events(self, room_version, state_sets, event): + async def resolve_events( + self, + room_version: str, + state_sets: Collection[Iterable[EventBase]], + event: EventBase, + ) -> StateMap[EventBase]: logger.info( "Resolving state for %s with %d groups", event.room_id, len(state_sets) ) @@ -404,22 +452,45 @@ class StateHandler(object): state_map = {ev.event_id: ev for st in state_sets for ev in st} - with Measure(self.clock, "state._resolve_events"): - new_state = await resolve_events_with_store( - self.clock, - event.room_id, - room_version, - state_set_ids, - event_map=state_map, - state_res_store=StateResolutionStore(self.store), - ) + new_state = await self._state_resolution_handler.resolve_events_with_store( + event.room_id, + room_version, + state_set_ids, + event_map=state_map, + state_res_store=StateResolutionStore(self.store), + ) - new_state = {key: state_map[ev_id] for key, ev_id in new_state.items()} + return {key: state_map[ev_id] for key, ev_id in new_state.items()} - return new_state +@attr.s(slots=True) +class _StateResMetrics: + """Keeps track of some usage metrics about state res.""" -class StateResolutionHandler(object): + # System and User CPU time, in seconds + cpu_time = attr.ib(type=float, default=0.0) + + # time spent on database transactions (excluding scheduling time). This roughly + # corresponds to the amount of work done on the db server, excluding event fetches. + db_time = attr.ib(type=float, default=0.0) + + # number of events fetched from the db. + db_events = attr.ib(type=int, default=0) + + +_biggest_room_by_cpu_counter = Counter( + "synapse_state_res_cpu_for_biggest_room_seconds", + "CPU time spent performing state resolution for the single most expensive " + "room for state resolution", +) +_biggest_room_by_db_counter = Counter( + "synapse_state_res_db_for_biggest_room_seconds", + "Database time spent performing state resolution for the single most " + "expensive room for state resolution", +) + + +class StateResolutionHandler: """Responsible for doing state conflict resolution. Note that the storage layer depends on this handler, so all functions must @@ -429,10 +500,9 @@ class StateResolutionHandler(object): def __init__(self, hs): self.clock = hs.get_clock() - # dict of set of event_ids -> _StateCacheEntry. - self._state_cache = None self.resolve_linearizer = Linearizer(name="state_resolve_lock") + # dict of set of event_ids -> _StateCacheEntry. self._state_cache = ExpiringCache( cache_name="state_cache", clock=self.clock, @@ -442,9 +512,25 @@ class StateResolutionHandler(object): reset_expiry_on_get=True, ) + # + # stuff for tracking time spent on state-res by room + # + + # tracks the amount of work done on state res per room + self._state_res_metrics = defaultdict( + _StateResMetrics + ) # type: DefaultDict[str, _StateResMetrics] + + self.clock.looping_call(self._report_metrics, 120 * 1000) + @log_function async def resolve_state_groups( - self, room_id, room_version, state_groups_ids, event_map, state_res_store + self, + room_id: str, + room_version: str, + state_groups_ids: Dict[int, StateMap[str]], + event_map: Optional[Dict[str, EventBase]], + state_res_store: "StateResolutionStore", ): """Resolves conflicts between a set of state groups @@ -452,13 +538,13 @@ class StateResolutionHandler(object): not be called for a single state group Args: - room_id (str): room we are resolving for (used for logging and sanity checks) - room_version (str): version of the room - state_groups_ids (dict[int, dict[(str, str), str]]): - map from state group id to the state in that state group + room_id: room we are resolving for (used for logging and sanity checks) + room_version: version of the room + state_groups_ids: + A map from state group id to the state in that state group (where 'state' is a map from state key to event id) - event_map(dict[str,FrozenEvent]|None): + event_map: a dict from event_id to event, for any events that we happen to have in flight (eg, those currently being persisted). This will be used as a starting point fof finding the state we need; any missing @@ -466,55 +552,31 @@ class StateResolutionHandler(object): If None, all events will be fetched via state_res_store. - state_res_store (StateResolutionStore) + state_res_store Returns: - _StateCacheEntry: resolved state + The resolved state """ - logger.debug("resolve_state_groups state_groups %s", state_groups_ids.keys()) - group_names = frozenset(state_groups_ids.keys()) with (await self.resolve_linearizer.queue(group_names)): - if self._state_cache is not None: - cache = self._state_cache.get(group_names, None) - if cache: - return cache + cache = self._state_cache.get(group_names, None) + if cache: + return cache logger.info( - "Resolving state for %s with %d groups", room_id, len(state_groups_ids) + "Resolving state for %s with groups %s", room_id, list(group_names), ) state_groups_histogram.observe(len(state_groups_ids)) - # start by assuming we won't have any conflicted state, and build up the new - # state map by iterating through the state groups. If we discover a conflict, - # we give up and instead use `resolve_events_with_store`. - # - # XXX: is this actually worthwhile, or should we just let - # resolve_events_with_store do it? - new_state = {} - conflicted_state = False - for st in state_groups_ids.values(): - for key, e_id in st.items(): - if key in new_state: - conflicted_state = True - break - new_state[key] = e_id - if conflicted_state: - break - - if conflicted_state: - logger.info("Resolving conflicted state for %r", room_id) - with Measure(self.clock, "state._resolve_events"): - new_state = await resolve_events_with_store( - self.clock, - room_id, - room_version, - list(state_groups_ids.values()), - event_map=event_map, - state_res_store=state_res_store, - ) + new_state = await self.resolve_events_with_store( + room_id, + room_version, + list(state_groups_ids.values()), + event_map=event_map, + state_res_store=state_res_store, + ) # if the new state matches any of the input state groups, we can # use that state group again. Otherwise we will generate a state_id @@ -524,27 +586,135 @@ class StateResolutionHandler(object): with Measure(self.clock, "state.create_group_ids"): cache = _make_state_cache_entry(new_state, state_groups_ids) - if self._state_cache is not None: - self._state_cache[group_names] = cache + self._state_cache[group_names] = cache return cache + async def resolve_events_with_store( + self, + room_id: str, + room_version: str, + state_sets: Sequence[StateMap[str]], + event_map: Optional[Dict[str, EventBase]], + state_res_store: "StateResolutionStore", + ) -> StateMap[str]: + """ + Args: + room_id: the room we are working in + + room_version: Version of the room + + state_sets: List of dicts of (type, state_key) -> event_id, + which are the different state groups to resolve. + + event_map: + a dict from event_id to event, for any events that we happen to + have in flight (eg, those currently being persisted). This will be + used as a starting point fof finding the state we need; any missing + events will be requested via state_map_factory. + + If None, all events will be fetched via state_res_store. + + state_res_store: a place to fetch events from + + Returns: + a map from (type, state_key) to event_id. + """ + try: + with Measure(self.clock, "state._resolve_events") as m: + v = KNOWN_ROOM_VERSIONS[room_version] + if v.state_res == StateResolutionVersions.V1: + return await v1.resolve_events_with_store( + room_id, state_sets, event_map, state_res_store.get_events + ) + else: + return await v2.resolve_events_with_store( + self.clock, + room_id, + room_version, + state_sets, + event_map, + state_res_store, + ) + finally: + self._record_state_res_metrics(room_id, m.get_resource_usage()) + + def _record_state_res_metrics(self, room_id: str, rusage: ContextResourceUsage): + room_metrics = self._state_res_metrics[room_id] + room_metrics.cpu_time += rusage.ru_utime + rusage.ru_stime + room_metrics.db_time += rusage.db_txn_duration_sec + room_metrics.db_events += rusage.evt_db_fetch_count + + def _report_metrics(self): + if not self._state_res_metrics: + # no state res has happened since the last iteration: don't bother logging. + return + + self._report_biggest( + lambda i: i.cpu_time, "CPU time", _biggest_room_by_cpu_counter, + ) + + self._report_biggest( + lambda i: i.db_time, "DB time", _biggest_room_by_db_counter, + ) + + self._state_res_metrics.clear() -def _make_state_cache_entry(new_state, state_groups_ids): + def _report_biggest( + self, + extract_key: Callable[[_StateResMetrics], Any], + metric_name: str, + prometheus_counter_metric: Counter, + ) -> None: + """Report metrics on the biggest rooms for state res + + Args: + extract_key: a callable which, given a _StateResMetrics, extracts a single + metric to sort by. + metric_name: the name of the metric we have extracted, for the log line + prometheus_counter_metric: a prometheus metric recording the sum of the + the extracted metric + """ + n_to_log = 10 + if not metrics_logger.isEnabledFor(logging.DEBUG): + # only need the most expensive if we don't have debug logging, which + # allows nlargest() to degrade to max() + n_to_log = 1 + + items = self._state_res_metrics.items() + + # log the N biggest rooms + biggest = heapq.nlargest( + n_to_log, items, key=lambda i: extract_key(i[1]) + ) # type: List[Tuple[str, _StateResMetrics]] + metrics_logger.debug( + "%i biggest rooms for state-res by %s: %s", + len(biggest), + metric_name, + ["%s (%gs)" % (r, extract_key(m)) for (r, m) in biggest], + ) + + # report info on the single biggest to prometheus + _, biggest_metrics = biggest[0] + prometheus_counter_metric.inc(extract_key(biggest_metrics)) + + +def _make_state_cache_entry( + new_state: StateMap[str], state_groups_ids: Dict[int, StateMap[str]] +) -> _StateCacheEntry: """Given a resolved state, and a set of input state groups, pick one to base a new state group on (if any), and return an appropriately-constructed _StateCacheEntry. Args: - new_state (dict[(str, str), str]): resolved state map (mapping from - (type, state_key) to event_id) + new_state: resolved state map (mapping from (type, state_key) to event_id) - state_groups_ids (dict[int, dict[(str, str), str]]): - map from state group id to the state in that state group - (where 'state' is a map from state key to event id) + state_groups_ids: + map from state group id to the state in that state group (where + 'state' is a map from state key to event id) Returns: - _StateCacheEntry + The cache entry. """ # if the new state matches any of the input state groups, we can # use that state group again. Otherwise we will generate a state_id @@ -581,49 +751,8 @@ def _make_state_cache_entry(new_state, state_groups_ids): ) -def resolve_events_with_store( - clock: Clock, - room_id: str, - room_version: str, - state_sets: List[StateMap[str]], - event_map: Optional[Dict[str, EventBase]], - state_res_store: "StateResolutionStore", -) -> Awaitable[StateMap[str]]: - """ - Args: - room_id: the room we are working in - - room_version: Version of the room - - state_sets: List of dicts of (type, state_key) -> event_id, - which are the different state groups to resolve. - - event_map: - a dict from event_id to event, for any events that we happen to - have in flight (eg, those currently being persisted). This will be - used as a starting point fof finding the state we need; any missing - events will be requested via state_map_factory. - - If None, all events will be fetched via state_res_store. - - state_res_store: a place to fetch events from - - Returns: - a map from (type, state_key) to event_id. - """ - v = KNOWN_ROOM_VERSIONS[room_version] - if v.state_res == StateResolutionVersions.V1: - return v1.resolve_events_with_store( - room_id, state_sets, event_map, state_res_store.get_events - ) - else: - return v2.resolve_events_with_store( - clock, room_id, room_version, state_sets, event_map, state_res_store - ) - - -@attr.s -class StateResolutionStore(object): +@attr.s(slots=True) +class StateResolutionStore: """Interface that allows state resolution algorithms to access the database in well defined way. @@ -633,15 +762,17 @@ class StateResolutionStore(object): store = attr.ib() - def get_events(self, event_ids, allow_rejected=False): + def get_events( + self, event_ids: Iterable[str], allow_rejected: bool = False + ) -> Awaitable[Dict[str, EventBase]]: """Get events from the database Args: - event_ids (list): The event_ids of the events to fetch - allow_rejected (bool): If True return rejected events. + event_ids: The event_ids of the events to fetch + allow_rejected: If True return rejected events. Returns: - Deferred[dict[str, FrozenEvent]]: Dict from event_id to event. + An awaitable which resolves to a dict from event_id to event. """ return self.store.get_events( @@ -651,7 +782,9 @@ class StateResolutionStore(object): allow_rejected=allow_rejected, ) - def get_auth_chain_difference(self, state_sets: List[Set[str]]): + def get_auth_chain_difference( + self, state_sets: List[Set[str]] + ) -> Awaitable[Set[str]]: """Given sets of state events figure out the auth chain difference (as per state res v2 algorithm). @@ -660,7 +793,7 @@ class StateResolutionStore(object): chain. Returns: - Deferred[Set[str]]: Set of event IDs. + An awaitable that resolves to a set of event IDs. """ return self.store.get_auth_chain_difference(state_sets) diff --git a/synapse/state/v1.py b/synapse/state/v1.py
index ab5e24841d..a493279cbd 100644 --- a/synapse/state/v1.py +++ b/synapse/state/v1.py
@@ -15,14 +15,24 @@ import hashlib import logging -from typing import Awaitable, Callable, Dict, List, Optional +from typing import ( + Awaitable, + Callable, + Dict, + Iterable, + List, + Optional, + Sequence, + Set, + Tuple, +) from synapse import event_auth from synapse.api.constants import EventTypes from synapse.api.errors import AuthError from synapse.api.room_versions import RoomVersions from synapse.events import EventBase -from synapse.types import StateMap +from synapse.types import MutableStateMap, StateMap logger = logging.getLogger(__name__) @@ -32,10 +42,10 @@ POWER_KEY = (EventTypes.PowerLevels, "") async def resolve_events_with_store( room_id: str, - state_sets: List[StateMap[str]], + state_sets: Sequence[StateMap[str]], event_map: Optional[Dict[str, EventBase]], - state_map_factory: Callable[[List[str]], Awaitable], -): + state_map_factory: Callable[[Iterable[str]], Awaitable[Dict[str, EventBase]]], +) -> StateMap[str]: """ Args: room_id: the room we are working in @@ -56,8 +66,7 @@ async def resolve_events_with_store( an Awaitable that resolves to a dict of event_id to event. Returns: - Deferred[dict[(str, str), str]]: - a map from (type, state_key) to event_id. + A map from (type, state_key) to event_id. """ if len(state_sets) == 1: return state_sets[0] @@ -75,8 +84,8 @@ async def resolve_events_with_store( "Asking for %d/%d conflicted events", len(needed_events), needed_event_count ) - # dict[str, FrozenEvent]: a map from state event id to event. Only includes - # the state events which are in conflict (and those in event_map) + # A map from state event id to event. Only includes the state events which + # are in conflict (and those in event_map). state_map = await state_map_factory(needed_events) if event_map is not None: state_map.update(event_map) @@ -91,8 +100,6 @@ async def resolve_events_with_store( # get the ids of the auth events which allow us to authenticate the # conflicted state, picking only from the unconflicting state. - # - # dict[(str, str), str]: a map from state key to event id auth_events = _create_auth_events_from_maps( unconflicted_state, conflicted_state, state_map ) @@ -122,29 +129,30 @@ async def resolve_events_with_store( ) -def _seperate(state_sets): +def _seperate( + state_sets: Iterable[StateMap[str]], +) -> Tuple[MutableStateMap[str], MutableStateMap[Set[str]]]: """Takes the state_sets and figures out which keys are conflicted and which aren't. i.e., which have multiple different event_ids associated with them in different state sets. Args: - state_sets(iterable[dict[(str, str), str]]): + state_sets: List of dicts of (type, state_key) -> event_id, which are the different state groups to resolve. Returns: - (dict[(str, str), str], dict[(str, str), set[str]]): - A tuple of (unconflicted_state, conflicted_state), where: + A tuple of (unconflicted_state, conflicted_state), where: - unconflicted_state is a dict mapping (type, state_key)->event_id - for unconflicted state keys. + unconflicted_state is a dict mapping (type, state_key)->event_id + for unconflicted state keys. - conflicted_state is a dict mapping (type, state_key) to a set of - event ids for conflicted state keys. + conflicted_state is a dict mapping (type, state_key) to a set of + event ids for conflicted state keys. """ state_set_iterator = iter(state_sets) unconflicted_state = dict(next(state_set_iterator)) - conflicted_state = {} + conflicted_state = {} # type: MutableStateMap[Set[str]] for state_set in state_set_iterator: for key, value in state_set.items(): @@ -171,7 +179,21 @@ def _seperate(state_sets): return unconflicted_state, conflicted_state -def _create_auth_events_from_maps(unconflicted_state, conflicted_state, state_map): +def _create_auth_events_from_maps( + unconflicted_state: StateMap[str], + conflicted_state: StateMap[Set[str]], + state_map: Dict[str, EventBase], +) -> StateMap[str]: + """ + + Args: + unconflicted_state: The unconflicted state map. + conflicted_state: The conflicted state map. + state_map: + + Returns: + A map from state key to event id. + """ auth_events = {} for event_ids in conflicted_state.values(): for event_id in event_ids: @@ -179,14 +201,17 @@ def _create_auth_events_from_maps(unconflicted_state, conflicted_state, state_ma keys = event_auth.auth_types_for_event(state_map[event_id]) for key in keys: if key not in auth_events: - event_id = unconflicted_state.get(key, None) - if event_id: - auth_events[key] = event_id + auth_event_id = unconflicted_state.get(key, None) + if auth_event_id: + auth_events[key] = auth_event_id return auth_events def _resolve_with_state( - unconflicted_state_ids, conflicted_state_ids, auth_event_ids, state_map + unconflicted_state_ids: MutableStateMap[str], + conflicted_state_ids: StateMap[Set[str]], + auth_event_ids: StateMap[str], + state_map: Dict[str, EventBase], ): conflicted_state = {} for key, event_ids in conflicted_state_ids.items(): @@ -215,7 +240,9 @@ def _resolve_with_state( return new_state -def _resolve_state_events(conflicted_state, auth_events): +def _resolve_state_events( + conflicted_state: StateMap[List[EventBase]], auth_events: MutableStateMap[EventBase] +) -> StateMap[EventBase]: """ This is where we actually decide which of the conflicted state to use. @@ -255,7 +282,9 @@ def _resolve_state_events(conflicted_state, auth_events): return resolved_state -def _resolve_auth_events(events, auth_events): +def _resolve_auth_events( + events: List[EventBase], auth_events: StateMap[EventBase] +) -> EventBase: reverse = list(reversed(_ordered_events(events))) auth_keys = { @@ -289,7 +318,9 @@ def _resolve_auth_events(events, auth_events): return event -def _resolve_normal_events(events, auth_events): +def _resolve_normal_events( + events: List[EventBase], auth_events: StateMap[EventBase] +) -> EventBase: for event in _ordered_events(events): try: # The signatures have already been checked at this point @@ -309,7 +340,7 @@ def _resolve_normal_events(events, auth_events): return event -def _ordered_events(events): +def _ordered_events(events: Iterable[EventBase]) -> List[EventBase]: def key_func(e): # we have to use utf-8 rather than ascii here because it turns out we allow # people to send us events with non-ascii event IDs :/ diff --git a/synapse/state/v2.py b/synapse/state/v2.py
index 6634955cdc..edf94e7ad6 100644 --- a/synapse/state/v2.py +++ b/synapse/state/v2.py
@@ -16,7 +16,21 @@ import heapq import itertools import logging -from typing import Dict, List, Optional +from typing import ( + Any, + Callable, + Dict, + Generator, + Iterable, + List, + Optional, + Sequence, + Set, + Tuple, + overload, +) + +from typing_extensions import Literal import synapse.state from synapse import event_auth @@ -24,7 +38,7 @@ from synapse.api.constants import EventTypes from synapse.api.errors import AuthError from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.events import EventBase -from synapse.types import StateMap +from synapse.types import MutableStateMap, StateMap from synapse.util import Clock logger = logging.getLogger(__name__) @@ -40,10 +54,10 @@ async def resolve_events_with_store( clock: Clock, room_id: str, room_version: str, - state_sets: List[StateMap[str]], + state_sets: Sequence[StateMap[str]], event_map: Optional[Dict[str, EventBase]], state_res_store: "synapse.state.StateResolutionStore", -): +) -> StateMap[str]: """Resolves the state using the v2 state resolution algorithm Args: @@ -63,8 +77,7 @@ async def resolve_events_with_store( state_res_store: Returns: - Deferred[dict[(str, str), str]]: - a map from (type, state_key) to event_id. + A map from (type, state_key) to event_id. """ logger.debug("Computing conflicted state") @@ -171,18 +184,23 @@ async def resolve_events_with_store( return resolved_state -async def _get_power_level_for_sender(room_id, event_id, event_map, state_res_store): +async def _get_power_level_for_sender( + room_id: str, + event_id: str, + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", +) -> int: """Return the power level of the sender of the given event according to their auth events. Args: - room_id (str) - event_id (str) - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) + room_id + event_id + event_map + state_res_store Returns: - Deferred[int] + The power level. """ event = await _get_event(room_id, event_id, event_map, state_res_store) @@ -217,17 +235,21 @@ async def _get_power_level_for_sender(room_id, event_id, event_map, state_res_st return int(level) -async def _get_auth_chain_difference(state_sets, event_map, state_res_store): +async def _get_auth_chain_difference( + state_sets: Sequence[StateMap[str]], + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", +) -> Set[str]: """Compare the auth chains of each state set and return the set of events that only appear in some but not all of the auth chains. Args: - state_sets (list) - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) + state_sets + event_map + state_res_store Returns: - Deferred[set[str]]: Set of event IDs + Set of event IDs """ difference = await state_res_store.get_auth_chain_difference( @@ -237,17 +259,19 @@ async def _get_auth_chain_difference(state_sets, event_map, state_res_store): return difference -def _seperate(state_sets): +def _seperate( + state_sets: Iterable[StateMap[str]], +) -> Tuple[StateMap[str], StateMap[Set[str]]]: """Return the unconflicted and conflicted state. This is different than in the original algorithm, as this defines a key to be conflicted if one of the state sets doesn't have that key. Args: - state_sets (list) + state_sets Returns: - tuple[dict, dict]: A tuple of unconflicted and conflicted state. The - conflicted state dict is a map from type/state_key to set of event IDs + A tuple of unconflicted and conflicted state. The conflicted state dict + is a map from type/state_key to set of event IDs """ unconflicted_state = {} conflicted_state = {} @@ -260,18 +284,20 @@ def _seperate(state_sets): event_ids.discard(None) conflicted_state[key] = event_ids - return unconflicted_state, conflicted_state + # mypy doesn't understand that discarding None above means that conflicted + # state is StateMap[Set[str]], not StateMap[Set[Optional[Str]]]. + return unconflicted_state, conflicted_state # type: ignore -def _is_power_event(event): +def _is_power_event(event: EventBase) -> bool: """Return whether or not the event is a "power event", as defined by the v2 state resolution algorithm Args: - event (FrozenEvent) + event Returns: - boolean + True if the event is a power event. """ if (event.type, event.state_key) in ( (EventTypes.PowerLevels, ""), @@ -288,19 +314,23 @@ def _is_power_event(event): async def _add_event_and_auth_chain_to_graph( - graph, room_id, event_id, event_map, state_res_store, auth_diff -): + graph: Dict[str, Set[str]], + room_id: str, + event_id: str, + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", + auth_diff: Set[str], +) -> None: """Helper function for _reverse_topological_power_sort that add the event and its auth chain (that is in the auth diff) to the graph Args: - graph (dict[str, set[str]]): A map from event ID to the events auth - event IDs - room_id (str): the room we are working in - event_id (str): Event to add to the graph - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) - auth_diff (set[str]): Set of event IDs that are in the auth difference. + graph: A map from event ID to the events auth event IDs + room_id: the room we are working in + event_id: Event to add to the graph + event_map + state_res_store + auth_diff: Set of event IDs that are in the auth difference. """ state = [event_id] @@ -318,24 +348,29 @@ async def _add_event_and_auth_chain_to_graph( async def _reverse_topological_power_sort( - clock, room_id, event_ids, event_map, state_res_store, auth_diff -): + clock: Clock, + room_id: str, + event_ids: Iterable[str], + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", + auth_diff: Set[str], +) -> List[str]: """Returns a list of the event_ids sorted by reverse topological ordering, and then by power level and origin_server_ts Args: - clock (Clock) - room_id (str): the room we are working in - event_ids (list[str]): The events to sort - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) - auth_diff (set[str]): Set of event IDs that are in the auth difference. + clock + room_id: the room we are working in + event_ids: The events to sort + event_map + state_res_store + auth_diff: Set of event IDs that are in the auth difference. Returns: - Deferred[list[str]]: The sorted list + The sorted list """ - graph = {} + graph = {} # type: Dict[str, Set[str]] for idx, event_id in enumerate(event_ids, start=1): await _add_event_and_auth_chain_to_graph( graph, room_id, event_id, event_map, state_res_store, auth_diff @@ -372,24 +407,30 @@ async def _reverse_topological_power_sort( async def _iterative_auth_checks( - clock, room_id, room_version, event_ids, base_state, event_map, state_res_store -): + clock: Clock, + room_id: str, + room_version: str, + event_ids: List[str], + base_state: StateMap[str], + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", +) -> MutableStateMap[str]: """Sequentially apply auth checks to each event in given list, updating the state as it goes along. Args: - clock (Clock) - room_id (str) - room_version (str) - event_ids (list[str]): Ordered list of events to apply auth checks to - base_state (StateMap[str]): The set of state to start with - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) + clock + room_id + room_version + event_ids: Ordered list of events to apply auth checks to + base_state: The set of state to start with + event_map + state_res_store Returns: - Deferred[StateMap[str]]: Returns the final updated state + Returns the final updated state """ - resolved_state = base_state.copy() + resolved_state = dict(base_state) room_version_obj = KNOWN_ROOM_VERSIONS[room_version] for idx, event_id in enumerate(event_ids, start=1): @@ -439,21 +480,26 @@ async def _iterative_auth_checks( async def _mainline_sort( - clock, room_id, event_ids, resolved_power_event_id, event_map, state_res_store -): + clock: Clock, + room_id: str, + event_ids: List[str], + resolved_power_event_id: Optional[str], + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", +) -> List[str]: """Returns a sorted list of event_ids sorted by mainline ordering based on the given event resolved_power_event_id Args: - clock (Clock) - room_id (str): room we're working in - event_ids (list[str]): Events to sort - resolved_power_event_id (str): The final resolved power level event ID - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) + clock + room_id: room we're working in + event_ids: Events to sort + resolved_power_event_id: The final resolved power level event ID + event_map + state_res_store Returns: - Deferred[list[str]]: The sorted list + The sorted list """ if not event_ids: # It's possible for there to be no event IDs here to sort, so we can @@ -505,59 +551,90 @@ async def _mainline_sort( async def _get_mainline_depth_for_event( - event, mainline_map, event_map, state_res_store -): + event: EventBase, + mainline_map: Dict[str, int], + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", +) -> int: """Get the mainline depths for the given event based on the mainline map Args: - event (FrozenEvent) - mainline_map (dict[str, int]): Map from event_id to mainline depth for - events in the mainline. - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) + event + mainline_map: Map from event_id to mainline depth for events in the mainline. + event_map + state_res_store Returns: - Deferred[int] + The mainline depth """ room_id = event.room_id + tmp_event = event # type: Optional[EventBase] # We do an iterative search, replacing `event with the power level in its # auth events (if any) - while event: + while tmp_event: depth = mainline_map.get(event.event_id) if depth is not None: return depth - auth_events = event.auth_event_ids() - event = None + auth_events = tmp_event.auth_event_ids() + tmp_event = None for aid in auth_events: aev = await _get_event( room_id, aid, event_map, state_res_store, allow_none=True ) if aev and (aev.type, aev.state_key) == (EventTypes.PowerLevels, ""): - event = aev + tmp_event = aev break # Didn't find a power level auth event, so we just return 0 return 0 -async def _get_event(room_id, event_id, event_map, state_res_store, allow_none=False): +@overload +async def _get_event( + room_id: str, + event_id: str, + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", + allow_none: Literal[False] = False, +) -> EventBase: + ... + + +@overload +async def _get_event( + room_id: str, + event_id: str, + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", + allow_none: Literal[True], +) -> Optional[EventBase]: + ... + + +async def _get_event( + room_id: str, + event_id: str, + event_map: Dict[str, EventBase], + state_res_store: "synapse.state.StateResolutionStore", + allow_none: bool = False, +) -> Optional[EventBase]: """Helper function to look up event in event_map, falling back to looking it up in the store Args: - room_id (str) - event_id (str) - event_map (dict[str,FrozenEvent]) - state_res_store (StateResolutionStore) - allow_none (bool): if the event is not found, return None rather than raising + room_id + event_id + event_map + state_res_store + allow_none: if the event is not found, return None rather than raising an exception Returns: - Deferred[Optional[FrozenEvent]] + The event, or none if the event does not exist (and allow_none is True). """ if event_id not in event_map: events = await state_res_store.get_events([event_id], allow_rejected=True) @@ -577,7 +654,9 @@ async def _get_event(room_id, event_id, event_map, state_res_store, allow_none=F return event -def lexicographical_topological_sort(graph, key): +def lexicographical_topological_sort( + graph: Dict[str, Set[str]], key: Callable[[str], Any] +) -> Generator[str, None, None]: """Performs a lexicographic reverse topological sort on the graph. This returns a reverse topological sort (i.e. if node A references B then B @@ -587,20 +666,20 @@ def lexicographical_topological_sort(graph, key): NOTE: `graph` is modified during the sort. Args: - graph (dict[str, set[str]]): A representation of the graph where each - node is a key in the dict and its value are the nodes edges. - key (func): A function that takes a node and returns a value that is - comparable and used to order nodes + graph: A representation of the graph where each node is a key in the + dict and its value are the nodes edges. + key: A function that takes a node and returns a value that is comparable + and used to order nodes Yields: - str: The next node in the topological sort + The next node in the topological sort """ # Note, this is basically Kahn's algorithm except we look at nodes with no # outgoing edges, c.f. # https://en.wikipedia.org/wiki/Topological_sorting#Kahn's_algorithm outdegree_map = graph - reverse_graph = {} + reverse_graph = {} # type: Dict[str, Set[str]] # Lists of nodes with zero out degree. Is actually a tuple of # `(key(node), node)` so that sorting does the right thing diff --git a/synapse/storage/__init__.py b/synapse/storage/__init__.py
index ec89f645d4..bbff3c8d5b 100644 --- a/synapse/storage/__init__.py +++ b/synapse/storage/__init__.py
@@ -17,18 +17,19 @@ """ The storage layer is split up into multiple parts to allow Synapse to run against different configurations of databases (e.g. single or multiple -databases). The `Database` class represents a single physical database. The -`data_stores` are classes that talk directly to a `Database` instance and have -associated schemas, background updates, etc. On top of those there are classes -that provide high level interfaces that combine calls to multiple `data_stores`. +databases). The `DatabasePool` class represents connections to a single physical +database. The `databases` are classes that talk directly to a `DatabasePool` +instance and have associated schemas, background updates, etc. On top of those +there are classes that provide high level interfaces that combine calls to +multiple `databases`. There are also schemas that get applied to every database, regardless of the data stores associated with them (e.g. the schema version tables), which are stored in `synapse.storage.schema`. """ -from synapse.storage.data_stores import DataStores -from synapse.storage.data_stores.main import DataStore +from synapse.storage.databases import Databases +from synapse.storage.databases.main import DataStore from synapse.storage.persist_events import EventsPersistenceStorage from synapse.storage.purge_events import PurgeEventsStorage from synapse.storage.state import StateGroupStorage @@ -36,16 +37,19 @@ from synapse.storage.state import StateGroupStorage __all__ = ["DataStores", "DataStore"] -class Storage(object): +class Storage: """The high level interfaces for talking to various storage layers. """ - def __init__(self, hs, stores: DataStores): + def __init__(self, hs, stores: Databases): # We include the main data store here mainly so that we don't have to # rewrite all the existing code to split it into high vs low level # interfaces. self.main = stores.main - self.persistence = EventsPersistenceStorage(hs, stores) self.purge_events = PurgeEventsStorage(hs, stores) self.state = StateGroupStorage(hs, stores) + + self.persistence = None + if stores.persist_events: + self.persistence = EventsPersistenceStorage(hs, stores) diff --git a/synapse/storage/_base.py b/synapse/storage/_base.py
index 985a042869..ab49d227de 100644 --- a/synapse/storage/_base.py +++ b/synapse/storage/_base.py
@@ -19,12 +19,11 @@ import random from abc import ABCMeta from typing import Any, Optional -from canonicaljson import json - from synapse.storage.database import LoggingTransaction # noqa: F401 from synapse.storage.database import make_in_list_sql_clause # noqa: F401 -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.types import Collection, get_domain_from_id +from synapse.util import json_decoder logger = logging.getLogger(__name__) @@ -37,11 +36,11 @@ class SQLBaseStore(metaclass=ABCMeta): per data store (and not one per physical database). """ - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self.hs = hs self._clock = hs.get_clock() self.database_engine = database.engine - self.db = database + self.db_pool = database self.rand = random.SystemRandom() def process_replication_rows(self, stream_name, instance_name, token, rows): @@ -58,7 +57,6 @@ class SQLBaseStore(metaclass=ABCMeta): """ for host in {get_domain_from_id(u) for u in members_changed}: self._attempt_to_invalidate_cache("is_host_joined", (room_id, host)) - self._attempt_to_invalidate_cache("was_host_joined", (room_id, host)) self._attempt_to_invalidate_cache("get_users_in_room", (room_id,)) self._attempt_to_invalidate_cache("get_room_summary", (room_id,)) @@ -100,13 +98,13 @@ def db_to_json(db_content): if isinstance(db_content, memoryview): db_content = db_content.tobytes() - # Decode it to a Unicode string before feeding it to json.loads, since + # Decode it to a Unicode string before feeding it to the JSON decoder, since # Python 3.5 does not support deserializing bytes. if isinstance(db_content, (bytes, bytearray)): db_content = db_content.decode("utf8") try: - return json.loads(db_content) + return json_decoder.decode(db_content) except Exception: logging.warning("Tried to decode '%r' as JSON and failed", db_content) raise diff --git a/synapse/storage/background_updates.py b/synapse/storage/background_updates.py
index 018826ef69..810721ebe9 100644 --- a/synapse/storage/background_updates.py +++ b/synapse/storage/background_updates.py
@@ -16,18 +16,15 @@ import logging from typing import Optional -from canonicaljson import json - -from twisted.internet import defer - from synapse.metrics.background_process_metrics import run_as_background_process +from synapse.util import json_encoder from . import engines logger = logging.getLogger(__name__) -class BackgroundUpdatePerformance(object): +class BackgroundUpdatePerformance: """Tracks the how long a background update is taking to update its items""" def __init__(self, name): @@ -74,7 +71,7 @@ class BackgroundUpdatePerformance(object): return float(self.total_item_count) / float(self.total_duration_ms) -class BackgroundUpdater(object): +class BackgroundUpdater: """ Background updates are updates to the database that run in the background. Each update processes a batch of data at once. We attempt to limit the impact of each update by monitoring how long each batch takes to @@ -88,7 +85,7 @@ class BackgroundUpdater(object): def __init__(self, hs, database): self._clock = hs.get_clock() - self.db = database + self.db_pool = database # if a background update is currently running, its name. self._current_background_update = None # type: Optional[str] @@ -139,7 +136,7 @@ class BackgroundUpdater(object): # otherwise, check if there are updates to be run. This is important, # as we may be running on a worker which doesn't perform the bg updates # itself, but still wants to wait for them to happen. - updates = await self.db.simple_select_onecol( + updates = await self.db_pool.simple_select_onecol( "background_updates", keyvalues=None, retcol="1", @@ -160,7 +157,7 @@ class BackgroundUpdater(object): if update_name == self._current_background_update: return False - update_exists = await self.db.simple_select_one_onecol( + update_exists = await self.db_pool.simple_select_one_onecol( "background_updates", keyvalues={"update_name": update_name}, retcol="1", @@ -189,10 +186,10 @@ class BackgroundUpdater(object): ORDER BY ordering, update_name """ ) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) if not self._current_background_update: - all_pending_updates = await self.db.runInteraction( + all_pending_updates = await self.db_pool.runInteraction( "background_updates", get_background_updates_txn, ) if not all_pending_updates: @@ -243,7 +240,7 @@ class BackgroundUpdater(object): else: batch_size = self.DEFAULT_BACKGROUND_BATCH_SIZE - progress_json = await self.db.simple_select_one_onecol( + progress_json = await self.db_pool.simple_select_one_onecol( "background_updates", keyvalues={"update_name": update_name}, retcol="progress_json", @@ -308,9 +305,8 @@ class BackgroundUpdater(object): update_name (str): Name of update """ - @defer.inlineCallbacks - def noop_update(progress, batch_size): - yield self._end_background_update(update_name) + async def noop_update(progress, batch_size): + await self._end_background_update(update_name) return 1 self.register_background_update_handler(update_name, noop_update) @@ -402,30 +398,30 @@ class BackgroundUpdater(object): logger.debug("[SQL] %s", sql) c.execute(sql) - if isinstance(self.db.engine, engines.PostgresEngine): + if isinstance(self.db_pool.engine, engines.PostgresEngine): runner = create_index_psql elif psql_only: runner = None else: runner = create_index_sqlite - @defer.inlineCallbacks - def updater(progress, batch_size): + async def updater(progress, batch_size): if runner is not None: logger.info("Adding index %s to %s", index_name, table) - yield self.db.runWithConnection(runner) - yield self._end_background_update(update_name) + await self.db_pool.runWithConnection(runner) + await self._end_background_update(update_name) return 1 self.register_background_update_handler(update_name, updater) - def _end_background_update(self, update_name): + async def _end_background_update(self, update_name: str) -> None: """Removes a completed background update task from the queue. Args: - update_name(str): The name of the completed task to remove + update_name:: The name of the completed task to remove + Returns: - A deferred that completes once the task is removed. + None, completes once the task is removed. """ if update_name != self._current_background_update: raise Exception( @@ -433,11 +429,11 @@ class BackgroundUpdater(object): % update_name ) self._current_background_update = None - return self.db.simple_delete_one( + await self.db_pool.simple_delete_one( "background_updates", keyvalues={"update_name": update_name} ) - def _background_update_progress(self, update_name: str, progress: dict): + async def _background_update_progress(self, update_name: str, progress: dict): """Update the progress of a background update Args: @@ -445,7 +441,7 @@ class BackgroundUpdater(object): progress: The progress of the update. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "background_update_progress", self._background_update_progress_txn, update_name, @@ -461,9 +457,9 @@ class BackgroundUpdater(object): progress(dict): The progress of the update. """ - progress_json = json.dumps(progress) + progress_json = json_encoder.encode(progress) - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, "background_updates", keyvalues={"update_name": update_name}, diff --git a/synapse/storage/data_stores/__init__.py b/synapse/storage/data_stores/__init__.py deleted file mode 100644
index 599ee470d4..0000000000 --- a/synapse/storage/data_stores/__init__.py +++ /dev/null
@@ -1,97 +0,0 @@ -# -*- coding: utf-8 -*- -# Copyright 2019 The Matrix.org Foundation C.I.C. -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -import logging - -from synapse.storage.data_stores.main.events import PersistEventsStore -from synapse.storage.data_stores.state import StateGroupDataStore -from synapse.storage.database import Database, make_conn -from synapse.storage.engines import create_engine -from synapse.storage.prepare_database import prepare_database - -logger = logging.getLogger(__name__) - - -class DataStores(object): - """The various data stores. - - These are low level interfaces to physical databases. - - Attributes: - main (DataStore) - """ - - def __init__(self, main_store_class, hs): - # Note we pass in the main store class here as workers use a different main - # store. - - self.databases = [] - self.main = None - self.state = None - self.persist_events = None - - for database_config in hs.config.database.databases: - db_name = database_config.name - engine = create_engine(database_config.config) - - with make_conn(database_config, engine) as db_conn: - logger.info("Preparing database %r...", db_name) - - engine.check_database(db_conn) - prepare_database( - db_conn, engine, hs.config, data_stores=database_config.data_stores, - ) - - database = Database(hs, database_config, engine) - - if "main" in database_config.data_stores: - logger.info("Starting 'main' data store") - - # Sanity check we don't try and configure the main store on - # multiple databases. - if self.main: - raise Exception("'main' data store already configured") - - self.main = main_store_class(database, db_conn, hs) - - # If we're on a process that can persist events also - # instantiate a `PersistEventsStore` - if hs.config.worker.writers.events == hs.get_instance_name(): - self.persist_events = PersistEventsStore( - hs, database, self.main - ) - - if "state" in database_config.data_stores: - logger.info("Starting 'state' data store") - - # Sanity check we don't try and configure the state store on - # multiple databases. - if self.state: - raise Exception("'state' data store already configured") - - self.state = StateGroupDataStore(database, db_conn, hs) - - db_conn.commit() - - self.databases.append(database) - - logger.info("Database %r prepared", db_name) - - # Sanity check that we have actually configured all the required stores. - if not self.main: - raise Exception("No 'main' data store configured") - - if not self.state: - raise Exception("No 'main' data store configured") diff --git a/synapse/storage/data_stores/main/transactions.py b/synapse/storage/data_stores/main/transactions.py deleted file mode 100644
index a9bf457939..0000000000 --- a/synapse/storage/data_stores/main/transactions.py +++ /dev/null
@@ -1,269 +0,0 @@ -# -*- coding: utf-8 -*- -# Copyright 2014-2016 OpenMarket Ltd -# -# Licensed under the Apache License, Version 2.0 (the "License"); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an "AS IS" BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. - -import logging -from collections import namedtuple - -from canonicaljson import encode_canonical_json - -from twisted.internet import defer - -from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.storage.database import Database -from synapse.util.caches.expiringcache import ExpiringCache - -db_binary_type = memoryview - -logger = logging.getLogger(__name__) - - -_TransactionRow = namedtuple( - "_TransactionRow", - ("id", "transaction_id", "destination", "ts", "response_code", "response_json"), -) - -_UpdateTransactionRow = namedtuple( - "_TransactionRow", ("response_code", "response_json") -) - -SENTINEL = object() - - -class TransactionStore(SQLBaseStore): - """A collection of queries for handling PDUs. - """ - - def __init__(self, database: Database, db_conn, hs): - super(TransactionStore, self).__init__(database, db_conn, hs) - - self._clock.looping_call(self._start_cleanup_transactions, 30 * 60 * 1000) - - self._destination_retry_cache = ExpiringCache( - cache_name="get_destination_retry_timings", - clock=self._clock, - expiry_ms=5 * 60 * 1000, - ) - - def get_received_txn_response(self, transaction_id, origin): - """For an incoming transaction from a given origin, check if we have - already responded to it. If so, return the response code and response - body (as a dict). - - Args: - transaction_id (str) - origin(str) - - Returns: - tuple: None if we have not previously responded to - this transaction or a 2-tuple of (int, dict) - """ - - return self.db.runInteraction( - "get_received_txn_response", - self._get_received_txn_response, - transaction_id, - origin, - ) - - def _get_received_txn_response(self, txn, transaction_id, origin): - result = self.db.simple_select_one_txn( - txn, - table="received_transactions", - keyvalues={"transaction_id": transaction_id, "origin": origin}, - retcols=( - "transaction_id", - "origin", - "ts", - "response_code", - "response_json", - "has_been_referenced", - ), - allow_none=True, - ) - - if result and result["response_code"]: - return result["response_code"], db_to_json(result["response_json"]) - - else: - return None - - def set_received_txn_response(self, transaction_id, origin, code, response_dict): - """Persist the response we returened for an incoming transaction, and - should return for subsequent transactions with the same transaction_id - and origin. - - Args: - txn - transaction_id (str) - origin (str) - code (int) - response_json (str) - """ - - return self.db.simple_insert( - table="received_transactions", - values={ - "transaction_id": transaction_id, - "origin": origin, - "response_code": code, - "response_json": db_binary_type(encode_canonical_json(response_dict)), - "ts": self._clock.time_msec(), - }, - or_ignore=True, - desc="set_received_txn_response", - ) - - @defer.inlineCallbacks - def get_destination_retry_timings(self, destination): - """Gets the current retry timings (if any) for a given destination. - - Args: - destination (str) - - Returns: - None if not retrying - Otherwise a dict for the retry scheme - """ - - result = self._destination_retry_cache.get(destination, SENTINEL) - if result is not SENTINEL: - return result - - result = yield self.db.runInteraction( - "get_destination_retry_timings", - self._get_destination_retry_timings, - destination, - ) - - # We don't hugely care about race conditions between getting and - # invalidating the cache, since we time out fairly quickly anyway. - self._destination_retry_cache[destination] = result - return result - - def _get_destination_retry_timings(self, txn, destination): - result = self.db.simple_select_one_txn( - txn, - table="destinations", - keyvalues={"destination": destination}, - retcols=("destination", "failure_ts", "retry_last_ts", "retry_interval"), - allow_none=True, - ) - - if result and result["retry_last_ts"] > 0: - return result - else: - return None - - def set_destination_retry_timings( - self, destination, failure_ts, retry_last_ts, retry_interval - ): - """Sets the current retry timings for a given destination. - Both timings should be zero if retrying is no longer occuring. - - Args: - destination (str) - failure_ts (int|None) - when the server started failing (ms since epoch) - retry_last_ts (int) - time of last retry attempt in unix epoch ms - retry_interval (int) - how long until next retry in ms - """ - - self._destination_retry_cache.pop(destination, None) - return self.db.runInteraction( - "set_destination_retry_timings", - self._set_destination_retry_timings, - destination, - failure_ts, - retry_last_ts, - retry_interval, - ) - - def _set_destination_retry_timings( - self, txn, destination, failure_ts, retry_last_ts, retry_interval - ): - - if self.database_engine.can_native_upsert: - # Upsert retry time interval if retry_interval is zero (i.e. we're - # resetting it) or greater than the existing retry interval. - - sql = """ - INSERT INTO destinations ( - destination, failure_ts, retry_last_ts, retry_interval - ) - VALUES (?, ?, ?, ?) - ON CONFLICT (destination) DO UPDATE SET - failure_ts = EXCLUDED.failure_ts, - retry_last_ts = EXCLUDED.retry_last_ts, - retry_interval = EXCLUDED.retry_interval - WHERE - EXCLUDED.retry_interval = 0 - OR destinations.retry_interval < EXCLUDED.retry_interval - """ - - txn.execute(sql, (destination, failure_ts, retry_last_ts, retry_interval)) - - return - - self.database_engine.lock_table(txn, "destinations") - - # We need to be careful here as the data may have changed from under us - # due to a worker setting the timings. - - prev_row = self.db.simple_select_one_txn( - txn, - table="destinations", - keyvalues={"destination": destination}, - retcols=("failure_ts", "retry_last_ts", "retry_interval"), - allow_none=True, - ) - - if not prev_row: - self.db.simple_insert_txn( - txn, - table="destinations", - values={ - "destination": destination, - "failure_ts": failure_ts, - "retry_last_ts": retry_last_ts, - "retry_interval": retry_interval, - }, - ) - elif retry_interval == 0 or prev_row["retry_interval"] < retry_interval: - self.db.simple_update_one_txn( - txn, - "destinations", - keyvalues={"destination": destination}, - updatevalues={ - "failure_ts": failure_ts, - "retry_last_ts": retry_last_ts, - "retry_interval": retry_interval, - }, - ) - - def _start_cleanup_transactions(self): - return run_as_background_process( - "cleanup_transactions", self._cleanup_transactions - ) - - def _cleanup_transactions(self): - now = self._clock.time_msec() - month_ago = now - 30 * 24 * 60 * 60 * 1000 - - def _cleanup_transactions_txn(txn): - txn.execute("DELETE FROM received_transactions WHERE ts < ?", (month_ago,)) - - return self.db.runInteraction( - "_cleanup_transactions", _cleanup_transactions_txn - ) diff --git a/synapse/storage/database.py b/synapse/storage/database.py
index 3be20c866a..6116191b16 100644 --- a/synapse/storage/database.py +++ b/synapse/storage/database.py
@@ -28,12 +28,14 @@ from typing import ( Optional, Tuple, TypeVar, + cast, + overload, ) from prometheus_client import Histogram +from typing_extensions import Literal from twisted.enterprise import adbapi -from twisted.internet import defer from synapse.api.errors import StoreError from synapse.config.database import DatabaseConnectionConfig @@ -49,11 +51,11 @@ from synapse.storage.engines import BaseDatabaseEngine, PostgresEngine, Sqlite3E from synapse.storage.types import Connection, Cursor from synapse.types import Collection -logger = logging.getLogger(__name__) - # python 3 does not have a maximum int value MAX_TXN_ID = 2 ** 63 - 1 +logger = logging.getLogger(__name__) + sql_logger = logging.getLogger("synapse.storage.SQL") transaction_logger = logging.getLogger("synapse.storage.txn") perf_logger = logging.getLogger("synapse.storage.TIME") @@ -125,7 +127,7 @@ class LoggingTransaction: method. Args: - txn: The database transcation object to wrap. + txn: The database transaction object to wrap. name: The name of this transactions for logging. database_engine after_callbacks: A list that callbacks will be appended to @@ -160,7 +162,7 @@ class LoggingTransaction: self.after_callbacks = after_callbacks self.exception_callbacks = exception_callbacks - def call_after(self, callback: "Callable[..., None]", *args, **kwargs): + def call_after(self, callback: "Callable[..., None]", *args: Any, **kwargs: Any): """Call the given callback on the main twisted thread after the transaction has finished. Used to invalidate the caches on the correct thread. @@ -171,7 +173,9 @@ class LoggingTransaction: assert self.after_callbacks is not None self.after_callbacks.append((callback, args, kwargs)) - def call_on_exception(self, callback: "Callable[..., None]", *args, **kwargs): + def call_on_exception( + self, callback: "Callable[..., None]", *args: Any, **kwargs: Any + ): # if self.exception_callbacks is None, that means that whatever constructed the # LoggingTransaction isn't expecting there to be any callbacks; assert that # is not the case. @@ -195,7 +199,7 @@ class LoggingTransaction: def description(self) -> Any: return self.txn.description - def execute_batch(self, sql, args): + def execute_batch(self, sql: str, args: Iterable[Iterable[Any]]) -> None: if isinstance(self.database_engine, PostgresEngine): from psycopg2.extras import execute_batch # type: ignore @@ -204,17 +208,17 @@ class LoggingTransaction: for val in args: self.execute(sql, val) - def execute(self, sql: str, *args: Any): + def execute(self, sql: str, *args: Any) -> None: self._do_execute(self.txn.execute, sql, *args) - def executemany(self, sql: str, *args: Any): + def executemany(self, sql: str, *args: Any) -> None: self._do_execute(self.txn.executemany, sql, *args) def _make_sql_one_line(self, sql: str) -> str: "Strip newlines out of SQL so that the loggers in the DB are on one line" return " ".join(line.strip() for line in sql.splitlines() if line.strip()) - def _do_execute(self, func, sql, *args): + def _do_execute(self, func, sql: str, *args: Any) -> None: sql = self._make_sql_one_line(sql) # TODO(paul): Maybe use 'info' and 'debug' for values? @@ -233,29 +237,29 @@ class LoggingTransaction: try: return func(sql, *args) except Exception as e: - logger.debug("[SQL FAIL] {%s} %s", self.name, e) + sql_logger.debug("[SQL FAIL] {%s} %s", self.name, e) raise finally: secs = time.time() - start sql_logger.debug("[SQL time] {%s} %f sec", self.name, secs) sql_query_timer.labels(sql.split()[0]).observe(secs) - def close(self): + def close(self) -> None: self.txn.close() -class PerformanceCounters(object): +class PerformanceCounters: def __init__(self): self.current_counters = {} self.previous_counters = {} - def update(self, key, duration_secs): + def update(self, key: str, duration_secs: float) -> None: count, cum_time = self.current_counters.get(key, (0, 0)) count += 1 cum_time += duration_secs self.current_counters[key] = (count, cum_time) - def interval(self, interval_duration_secs, limit=3): + def interval(self, interval_duration_secs: float, limit: int = 3) -> str: counters = [] for name, (count, cum_time) in self.current_counters.items(): prev_count, prev_time = self.previous_counters.get(name, (0, 0)) @@ -279,7 +283,10 @@ class PerformanceCounters(object): return top_n_counters -class Database(object): +R = TypeVar("R") + + +class DatabasePool: """Wraps a single physical database and connection pool. A single database may be used by multiple data stores. @@ -327,13 +334,12 @@ class Database(object): self._check_safe_to_upsert, ) - def is_running(self): + def is_running(self) -> bool: """Is the database pool currently running """ return self._db_pool.running - @defer.inlineCallbacks - def _check_safe_to_upsert(self): + async def _check_safe_to_upsert(self) -> None: """ Is it safe to use native UPSERT? @@ -342,7 +348,7 @@ class Database(object): If the background updates have not completed, wait 15 sec and check again. """ - updates = yield self.simple_select_list( + updates = await self.simple_select_list( "background_updates", keyvalues=None, retcols=["update_name"], @@ -364,7 +370,7 @@ class Database(object): self._check_safe_to_upsert, ) - def start_profiling(self): + def start_profiling(self) -> None: self._previous_loop_ts = monotonic_time() def loop(): @@ -388,8 +394,33 @@ class Database(object): self._clock.looping_call(loop, 10000) def new_transaction( - self, conn, desc, after_callbacks, exception_callbacks, func, *args, **kwargs - ): + self, + conn: Connection, + desc: str, + after_callbacks: List[_CallbackListEntry], + exception_callbacks: List[_CallbackListEntry], + func: "Callable[..., R]", + *args: Any, + **kwargs: Any + ) -> R: + """Start a new database transaction with the given connection. + + Note: The given func may be called multiple times under certain + failure modes. This is normally fine when in a standard transaction, + but care must be taken if the connection is in `autocommit` mode that + the function will correctly handle being aborted and retried half way + through its execution. + + Args: + conn + desc + after_callbacks + exception_callbacks + func + *args + **kwargs + """ + start = monotonic_time() txn_id = self._TXN_ID @@ -419,7 +450,7 @@ class Database(object): except self.engine.module.OperationalError as e: # This can happen if the database disappears mid # transaction. - logger.warning( + transaction_logger.warning( "[TXN OPERROR] {%s} %s %d/%d", name, e, i, N, ) if i < N: @@ -427,18 +458,20 @@ class Database(object): try: conn.rollback() except self.engine.module.Error as e1: - logger.warning("[TXN EROLL] {%s} %s", name, e1) + transaction_logger.warning("[TXN EROLL] {%s} %s", name, e1) continue raise except self.engine.module.DatabaseError as e: if self.engine.is_deadlock(e): - logger.warning("[TXN DEADLOCK] {%s} %d/%d", name, i, N) + transaction_logger.warning( + "[TXN DEADLOCK] {%s} %d/%d", name, i, N + ) if i < N: i += 1 try: conn.rollback() except self.engine.module.Error as e1: - logger.warning( + transaction_logger.warning( "[TXN EROLL] {%s} %s", name, e1, ) continue @@ -478,7 +511,7 @@ class Database(object): # [2]: https://github.com/python/cpython/blob/v3.8.0/Modules/_sqlite/cursor.c#L236 cursor.close() except Exception as e: - logger.debug("[TXN FAIL] {%s} %s", name, e) + transaction_logger.debug("[TXN FAIL] {%s} %s", name, e) raise finally: end = monotonic_time() @@ -492,8 +525,14 @@ class Database(object): self._txn_perf_counters.update(desc, duration) sql_txn_timer.labels(desc).observe(duration) - @defer.inlineCallbacks - def runInteraction(self, desc: str, func: Callable, *args: Any, **kwargs: Any): + async def runInteraction( + self, + desc: str, + func: "Callable[..., R]", + *args: Any, + db_autocommit: bool = False, + **kwargs: Any + ) -> R: """Starts a transaction on the database and runs a given function Arguments: @@ -502,11 +541,23 @@ class Database(object): database transaction (twisted.enterprise.adbapi.Transaction) as its first argument, followed by `args` and `kwargs`. + db_autocommit: Whether to run the function in "autocommit" mode, + i.e. outside of a transaction. This is useful for transactions + that are only a single query. + + Currently, this is only implemented for Postgres. SQLite will still + run the function inside a transaction. + + WARNING: This means that if func fails half way through then + the changes will *not* be rolled back. `func` may also get + called multiple times if the transaction is retried, so must + correctly handle that case. + args: positional args to pass to `func` kwargs: named args to pass to `func` Returns: - Deferred: The result of func + The result of func """ after_callbacks = [] # type: List[_CallbackListEntry] exception_callbacks = [] # type: List[_CallbackListEntry] @@ -515,13 +566,14 @@ class Database(object): logger.warning("Starting db txn '%s' from sentinel context", desc) try: - result = yield self.runWithConnection( + result = await self.runWithConnection( self.new_transaction, desc, after_callbacks, exception_callbacks, func, *args, + db_autocommit=db_autocommit, **kwargs ) @@ -532,10 +584,15 @@ class Database(object): after_callback(*after_args, **after_kwargs) raise - return result + return cast(R, result) - @defer.inlineCallbacks - def runWithConnection(self, func: Callable, *args: Any, **kwargs: Any): + async def runWithConnection( + self, + func: "Callable[..., R]", + *args: Any, + db_autocommit: bool = False, + **kwargs: Any + ) -> R: """Wraps the .runWithConnection() method on the underlying db_pool. Arguments: @@ -543,10 +600,13 @@ class Database(object): database connection (twisted.enterprise.adbapi.Connection) as its first argument, followed by `args` and `kwargs`. args: positional args to pass to `func` + db_autocommit: Whether to run the function in "autocommit" mode, + i.e. outside of a transaction. This is useful for transaction + that are only a single query. Currently only affects postgres. kwargs: named args to pass to `func` Returns: - Deferred: The result of func + The result of func """ parent_context = current_context() # type: Optional[LoggingContextOrSentinel] if not parent_context: @@ -558,6 +618,13 @@ class Database(object): start_time = monotonic_time() def inner_func(conn, *args, **kwargs): + # We shouldn't be in a transaction. If we are then something + # somewhere hasn't committed after doing work. (This is likely only + # possible during startup, as `run*` will ensure changes are + # committed/rolled back before putting the connection back in the + # pool). + assert not self.engine.in_transaction(conn) + with LoggingContext("runWithConnection", parent_context) as context: sched_duration_sec = monotonic_time() - start_time sql_scheduling_timer.observe(sched_duration_sec) @@ -567,20 +634,25 @@ class Database(object): logger.debug("Reconnecting closed database connection") conn.reconnect() - return func(conn, *args, **kwargs) + try: + if db_autocommit: + self.engine.attempt_to_set_autocommit(conn, True) + + return func(conn, *args, **kwargs) + finally: + if db_autocommit: + self.engine.attempt_to_set_autocommit(conn, False) - result = yield make_deferred_yieldable( + return await make_deferred_yieldable( self._db_pool.runWithConnection(inner_func, *args, **kwargs) ) - return result - @staticmethod - def cursor_to_dict(cursor): + def cursor_to_dict(cursor: Cursor) -> List[Dict[str, Any]]: """Converts a SQL cursor into an list of dicts. Args: - cursor : The DBAPI cursor which has executed a query. + cursor: The DBAPI cursor which has executed a query. Returns: A list of dicts where the key is the column header. """ @@ -588,10 +660,29 @@ class Database(object): results = [dict(zip(col_headers, row)) for row in cursor] return results - def execute(self, desc, decoder, query, *args): + @overload + async def execute( + self, desc: str, decoder: Literal[None], query: str, *args: Any + ) -> List[Tuple[Any, ...]]: + ... + + @overload + async def execute( + self, desc: str, decoder: Callable[[Cursor], R], query: str, *args: Any + ) -> R: + ... + + async def execute( + self, + desc: str, + decoder: Optional[Callable[[Cursor], R]], + query: str, + *args: Any + ) -> R: """Runs a single query for a result set. Args: + desc: description of the transaction, for logging and metrics decoder - The function which can resolve the cursor results to something meaningful. query - The query string to execute @@ -607,29 +698,33 @@ class Database(object): else: return txn.fetchall() - return self.runInteraction(desc, interaction) + return await self.runInteraction(desc, interaction) # "Simple" SQL API methods that operate on a single table with no JOINs, # no complex WHERE clauses, just a dict of values for columns. - @defer.inlineCallbacks - def simple_insert(self, table, values, or_ignore=False, desc="simple_insert"): + async def simple_insert( + self, + table: str, + values: Dict[str, Any], + or_ignore: bool = False, + desc: str = "simple_insert", + ) -> bool: """Executes an INSERT query on the named table. Args: - table : string giving the table name - values : dict of new column names and values for them - or_ignore : bool stating whether an exception should be raised + table: string giving the table name + values: dict of new column names and values for them + or_ignore: bool stating whether an exception should be raised when a conflicting row already exists. If True, False will be returned by the function instead - desc : string giving a description of the transaction + desc: description of the transaction, for logging and metrics Returns: - bool: Whether the row was inserted or not. Only useful when - `or_ignore` is True + Whether the row was inserted or not. Only useful when `or_ignore` is True """ try: - yield self.runInteraction(desc, self.simple_insert_txn, table, values) + await self.runInteraction(desc, self.simple_insert_txn, table, values) except self.engine.module.IntegrityError: # We have to do or_ignore flag at this layer, since we can't reuse # a cursor after we receive an error from the db. @@ -639,7 +734,9 @@ class Database(object): return True @staticmethod - def simple_insert_txn(txn, table, values): + def simple_insert_txn( + txn: LoggingTransaction, table: str, values: Dict[str, Any] + ) -> None: keys, vals = zip(*values.items()) sql = "INSERT INTO %s (%s) VALUES(%s)" % ( @@ -650,11 +747,29 @@ class Database(object): txn.execute(sql, vals) - def simple_insert_many(self, table, values, desc): - return self.runInteraction(desc, self.simple_insert_many_txn, table, values) + async def simple_insert_many( + self, table: str, values: List[Dict[str, Any]], desc: str + ) -> None: + """Executes an INSERT query on the named table. + + Args: + table: string giving the table name + values: dict of new column names and values for them + desc: description of the transaction, for logging and metrics + """ + await self.runInteraction(desc, self.simple_insert_many_txn, table, values) @staticmethod - def simple_insert_many_txn(txn, table, values): + def simple_insert_many_txn( + txn: LoggingTransaction, table: str, values: List[Dict[str, Any]] + ) -> None: + """Executes an INSERT query on the named table. + + Args: + txn: The transaction to use. + table: string giving the table name + values: dict of new column names and values for them + """ if not values: return @@ -682,16 +797,15 @@ class Database(object): txn.executemany(sql, vals) - @defer.inlineCallbacks - def simple_upsert( + async def simple_upsert( self, - table, - keyvalues, - values, - insertion_values={}, - desc="simple_upsert", - lock=True, - ): + table: str, + keyvalues: Dict[str, Any], + values: Dict[str, Any], + insertion_values: Dict[str, Any] = {}, + desc: str = "simple_upsert", + lock: bool = True, + ) -> Optional[bool]: """ `lock` should generally be set to True (the default), but can be set @@ -705,21 +819,20 @@ class Database(object): this table. Args: - table (str): The table to upsert into - keyvalues (dict): The unique key columns and their new values - values (dict): The nonunique columns and their new values - insertion_values (dict): additional key/values to use only when - inserting - lock (bool): True to lock the table when doing the upsert. + table: The table to upsert into + keyvalues: The unique key columns and their new values + values: The nonunique columns and their new values + insertion_values: additional key/values to use only when inserting + desc: description of the transaction, for logging and metrics + lock: True to lock the table when doing the upsert. Returns: - Deferred(None or bool): Native upserts always return None. Emulated - upserts return True if a new entry was created, False if an existing - one was updated. + Native upserts always return None. Emulated upserts return True if a + new entry was created, False if an existing one was updated. """ attempts = 0 while True: try: - result = yield self.runInteraction( + return await self.runInteraction( desc, self.simple_upsert_txn, table, @@ -728,7 +841,6 @@ class Database(object): insertion_values, lock=lock, ) - return result except self.engine.module.IntegrityError as e: attempts += 1 if attempts >= 5: @@ -742,29 +854,34 @@ class Database(object): ) def simple_upsert_txn( - self, txn, table, keyvalues, values, insertion_values={}, lock=True - ): + self, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + values: Dict[str, Any], + insertion_values: Dict[str, Any] = {}, + lock: bool = True, + ) -> Optional[bool]: """ Pick the UPSERT method which works best on the platform. Either the native one (Pg9.5+, recent SQLites), or fall back to an emulated method. Args: txn: The transaction to use. - table (str): The table to upsert into - keyvalues (dict): The unique key tables and their new values - values (dict): The nonunique columns and their new values - insertion_values (dict): additional key/values to use only when - inserting - lock (bool): True to lock the table when doing the upsert. + table: The table to upsert into + keyvalues: The unique key tables and their new values + values: The nonunique columns and their new values + insertion_values: additional key/values to use only when inserting + lock: True to lock the table when doing the upsert. Returns: - None or bool: Native upserts always return None. Emulated - upserts return True if a new entry was created, False if an existing - one was updated. + Native upserts always return None. Emulated upserts return True if a + new entry was created, False if an existing one was updated. """ if self.engine.can_native_upsert and table not in self._unsafe_to_upsert_tables: - return self.simple_upsert_txn_native_upsert( + self.simple_upsert_txn_native_upsert( txn, table, keyvalues, values, insertion_values=insertion_values ) + return None else: return self.simple_upsert_txn_emulated( txn, @@ -776,18 +893,23 @@ class Database(object): ) def simple_upsert_txn_emulated( - self, txn, table, keyvalues, values, insertion_values={}, lock=True - ): + self, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + values: Dict[str, Any], + insertion_values: Dict[str, Any] = {}, + lock: bool = True, + ) -> bool: """ Args: - table (str): The table to upsert into - keyvalues (dict): The unique key tables and their new values - values (dict): The nonunique columns and their new values - insertion_values (dict): additional key/values to use only when - inserting - lock (bool): True to lock the table when doing the upsert. + table: The table to upsert into + keyvalues: The unique key tables and their new values + values: The nonunique columns and their new values + insertion_values: additional key/values to use only when inserting + lock: True to lock the table when doing the upsert. Returns: - bool: Return True if a new entry was created, False if an existing + Returns True if a new entry was created, False if an existing one was updated. """ # We need to lock the table :(, unless we're *really* careful @@ -845,19 +967,21 @@ class Database(object): return True def simple_upsert_txn_native_upsert( - self, txn, table, keyvalues, values, insertion_values={} - ): + self, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + values: Dict[str, Any], + insertion_values: Dict[str, Any] = {}, + ) -> None: """ Use the native UPSERT functionality in recent PostgreSQL versions. Args: - table (str): The table to upsert into - keyvalues (dict): The unique key tables and their new values - values (dict): The nonunique columns and their new values - insertion_values (dict): additional key/values to use only when - inserting - Returns: - None + table: The table to upsert into + keyvalues: The unique key tables and their new values + values: The nonunique columns and their new values + insertion_values: additional key/values to use only when inserting """ allvalues = {} # type: Dict[str, Any] allvalues.update(keyvalues) @@ -885,7 +1009,7 @@ class Database(object): key_names: Collection[str], key_values: Collection[Iterable[Any]], value_names: Collection[str], - value_values: Iterable[Iterable[str]], + value_values: Iterable[Iterable[Any]], ) -> None: """ Upsert, many times. @@ -914,7 +1038,7 @@ class Database(object): key_names: Iterable[str], key_values: Collection[Iterable[Any]], value_names: Collection[str], - value_values: Iterable[Iterable[str]], + value_values: Iterable[Iterable[Any]], ) -> None: """ Upsert, many times, but without native UPSERT support or batching. @@ -987,41 +1111,93 @@ class Database(object): return txn.execute_batch(sql, args) - def simple_select_one( - self, table, keyvalues, retcols, allow_none=False, desc="simple_select_one" - ): + @overload + async def simple_select_one( + self, + table: str, + keyvalues: Dict[str, Any], + retcols: Iterable[str], + allow_none: Literal[False] = False, + desc: str = "simple_select_one", + ) -> Dict[str, Any]: + ... + + @overload + async def simple_select_one( + self, + table: str, + keyvalues: Dict[str, Any], + retcols: Iterable[str], + allow_none: Literal[True] = True, + desc: str = "simple_select_one", + ) -> Optional[Dict[str, Any]]: + ... + + async def simple_select_one( + self, + table: str, + keyvalues: Dict[str, Any], + retcols: Iterable[str], + allow_none: bool = False, + desc: str = "simple_select_one", + ) -> Optional[Dict[str, Any]]: """Executes a SELECT query on the named table, which is expected to return a single row, returning multiple columns from it. Args: - table : string giving the table name - keyvalues : dict of column names and values to select the row with - retcols : list of strings giving the names of the columns to return - - allow_none : If true, return None instead of failing if the SELECT - statement returns no rows + table: string giving the table name + keyvalues: dict of column names and values to select the row with + retcols: list of strings giving the names of the columns to return + allow_none: If true, return None instead of failing if the SELECT + statement returns no rows + desc: description of the transaction, for logging and metrics """ - return self.runInteraction( + return await self.runInteraction( desc, self.simple_select_one_txn, table, keyvalues, retcols, allow_none ) - def simple_select_one_onecol( + @overload + async def simple_select_one_onecol( self, - table, - keyvalues, - retcol, - allow_none=False, - desc="simple_select_one_onecol", - ): + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: Literal[False] = False, + desc: str = "simple_select_one_onecol", + ) -> Any: + ... + + @overload + async def simple_select_one_onecol( + self, + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: Literal[True] = True, + desc: str = "simple_select_one_onecol", + ) -> Optional[Any]: + ... + + async def simple_select_one_onecol( + self, + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: bool = False, + desc: str = "simple_select_one_onecol", + ) -> Optional[Any]: """Executes a SELECT query on the named table, which is expected to return a single row, returning a single column from it. Args: - table : string giving the table name - keyvalues : dict of column names and values to select the row with - retcol : string giving the name of the column to return + table: string giving the table name + keyvalues: dict of column names and values to select the row with + retcol: string giving the name of the column to return + allow_none: If true, return None instead of failing if the SELECT + statement returns no rows + desc: description of the transaction, for logging and metrics """ - return self.runInteraction( + return await self.runInteraction( desc, self.simple_select_one_onecol_txn, table, @@ -1030,10 +1206,39 @@ class Database(object): allow_none=allow_none, ) + @overload @classmethod def simple_select_one_onecol_txn( - cls, txn, table, keyvalues, retcol, allow_none=False - ): + cls, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: Literal[False] = False, + ) -> Any: + ... + + @overload + @classmethod + def simple_select_one_onecol_txn( + cls, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: Literal[True] = True, + ) -> Optional[Any]: + ... + + @classmethod + def simple_select_one_onecol_txn( + cls, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + retcol: str, + allow_none: bool = False, + ) -> Optional[Any]: ret = cls.simple_select_onecol_txn( txn, table=table, keyvalues=keyvalues, retcol=retcol ) @@ -1047,7 +1252,9 @@ class Database(object): raise StoreError(404, "No row found") @staticmethod - def simple_select_onecol_txn(txn, table, keyvalues, retcol): + def simple_select_onecol_txn( + txn: LoggingTransaction, table: str, keyvalues: Dict[str, Any], retcol: str, + ) -> List[Any]: sql = ("SELECT %(retcol)s FROM %(table)s") % {"retcol": retcol, "table": table} if keyvalues: @@ -1058,53 +1265,72 @@ class Database(object): return [r[0] for r in txn] - def simple_select_onecol( - self, table, keyvalues, retcol, desc="simple_select_onecol" - ): + async def simple_select_onecol( + self, + table: str, + keyvalues: Optional[Dict[str, Any]], + retcol: str, + desc: str = "simple_select_onecol", + ) -> List[Any]: """Executes a SELECT query on the named table, which returns a list comprising of the values of the named column from the selected rows. Args: - table (str): table name - keyvalues (dict|None): column names and values to select the rows with - retcol (str): column whos value we wish to retrieve. + table: table name + keyvalues: column names and values to select the rows with + retcol: column whos value we wish to retrieve. + desc: description of the transaction, for logging and metrics Returns: - Deferred: Results in a list + Results in a list """ - return self.runInteraction( + return await self.runInteraction( desc, self.simple_select_onecol_txn, table, keyvalues, retcol ) - def simple_select_list(self, table, keyvalues, retcols, desc="simple_select_list"): + async def simple_select_list( + self, + table: str, + keyvalues: Optional[Dict[str, Any]], + retcols: Iterable[str], + desc: str = "simple_select_list", + ) -> List[Dict[str, Any]]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. Args: - table (str): the table name - keyvalues (dict[str, Any] | None): + table: the table name + keyvalues: column names and values to select the rows with, or None to not apply a WHERE clause. - retcols (iterable[str]): the names of the columns to return + retcols: the names of the columns to return + desc: description of the transaction, for logging and metrics + Returns: - defer.Deferred: resolves to list[dict[str, Any]] + A list of dictionaries. """ - return self.runInteraction( + return await self.runInteraction( desc, self.simple_select_list_txn, table, keyvalues, retcols ) @classmethod - def simple_select_list_txn(cls, txn, table, keyvalues, retcols): + def simple_select_list_txn( + cls, + txn: LoggingTransaction, + table: str, + keyvalues: Optional[Dict[str, Any]], + retcols: Iterable[str], + ) -> List[Dict[str, Any]]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. Args: - txn : Transaction object - table (str): the table name - keyvalues (dict[str, T] | None): + txn: Transaction object + table: the table name + keyvalues: column names and values to select the rows with, or None to not apply a WHERE clause. - retcols (iterable[str]): the names of the columns to return + retcols: the names of the columns to return """ if keyvalues: sql = "SELECT %s FROM %s WHERE %s" % ( @@ -1119,28 +1345,29 @@ class Database(object): return cls.cursor_to_dict(txn) - @defer.inlineCallbacks - def simple_select_many_batch( + async def simple_select_many_batch( self, - table, - column, - iterable, - retcols, - keyvalues={}, - desc="simple_select_many_batch", - batch_size=100, - ): + table: str, + column: str, + iterable: Iterable[Any], + retcols: Iterable[str], + keyvalues: Dict[str, Any] = {}, + desc: str = "simple_select_many_batch", + batch_size: int = 100, + ) -> List[Any]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. - Filters rows by if value of `column` is in `iterable`. + Filters rows by whether the value of `column` is in `iterable`. Args: - table : string giving the table name - column : column name to test for inclusion against `iterable` - iterable : list - keyvalues : dict of column names and values to select the rows with - retcols : list of strings giving the names of the columns to return + table: string giving the table name + column: column name to test for inclusion against `iterable` + iterable: list + retcols: list of strings giving the names of the columns to return + keyvalues: dict of column names and values to select the rows with + desc: description of the transaction, for logging and metrics + batch_size: the number of rows for each select query """ results = [] # type: List[Dict[str, Any]] @@ -1154,7 +1381,7 @@ class Database(object): it_list[i : i + batch_size] for i in range(0, len(it_list), batch_size) ] for chunk in chunks: - rows = yield self.runInteraction( + rows = await self.runInteraction( desc, self.simple_select_many_txn, table, @@ -1169,19 +1396,27 @@ class Database(object): return results @classmethod - def simple_select_many_txn(cls, txn, table, column, iterable, keyvalues, retcols): + def simple_select_many_txn( + cls, + txn: LoggingTransaction, + table: str, + column: str, + iterable: Iterable[Any], + keyvalues: Dict[str, Any], + retcols: Iterable[str], + ) -> List[Dict[str, Any]]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. - Filters rows by if value of `column` is in `iterable`. + Filters rows by whether the value of `column` is in `iterable`. Args: - txn : Transaction object - table : string giving the table name - column : column name to test for inclusion against `iterable` - iterable : list - keyvalues : dict of column names and values to select the rows with - retcols : list of strings giving the names of the columns to return + txn: Transaction object + table: string giving the table name + column: column name to test for inclusion against `iterable` + iterable: list + keyvalues: dict of column names and values to select the rows with + retcols: list of strings giving the names of the columns to return """ if not iterable: return [] @@ -1202,13 +1437,24 @@ class Database(object): txn.execute(sql, values) return cls.cursor_to_dict(txn) - def simple_update(self, table, keyvalues, updatevalues, desc): - return self.runInteraction( + async def simple_update( + self, + table: str, + keyvalues: Dict[str, Any], + updatevalues: Dict[str, Any], + desc: str, + ) -> int: + return await self.runInteraction( desc, self.simple_update_txn, table, keyvalues, updatevalues ) @staticmethod - def simple_update_txn(txn, table, keyvalues, updatevalues): + def simple_update_txn( + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + updatevalues: Dict[str, Any], + ) -> int: if keyvalues: where = "WHERE %s" % " AND ".join("%s = ?" % k for k in keyvalues.keys()) else: @@ -1224,32 +1470,34 @@ class Database(object): return txn.rowcount - def simple_update_one( - self, table, keyvalues, updatevalues, desc="simple_update_one" - ): + async def simple_update_one( + self, + table: str, + keyvalues: Dict[str, Any], + updatevalues: Dict[str, Any], + desc: str = "simple_update_one", + ) -> None: """Executes an UPDATE query on the named table, setting new values for columns in a row matching the key values. Args: - table : string giving the table name - keyvalues : dict of column names and values to select the row with - updatevalues : dict giving column names and values to update - retcols : optional list of column names to return - - If present, retcols gives a list of column names on which to perform - a SELECT statement *before* performing the UPDATE statement. The values - of these will be returned in a dict. - - These are performed within the same transaction, allowing an atomic - get-and-set. This can be used to implement compare-and-set by putting - the update column in the 'keyvalues' dict as well. + table: string giving the table name + keyvalues: dict of column names and values to select the row with + updatevalues: dict giving column names and values to update + desc: description of the transaction, for logging and metrics """ - return self.runInteraction( + await self.runInteraction( desc, self.simple_update_one_txn, table, keyvalues, updatevalues ) @classmethod - def simple_update_one_txn(cls, txn, table, keyvalues, updatevalues): + def simple_update_one_txn( + cls, + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + updatevalues: Dict[str, Any], + ) -> None: rowcount = cls.simple_update_txn(txn, table, keyvalues, updatevalues) if rowcount == 0: @@ -1257,8 +1505,18 @@ class Database(object): if rowcount > 1: raise StoreError(500, "More than one row matched (%s)" % (table,)) + # Ideally we could use the overload decorator here to specify that the + # return type is only optional if allow_none is True, but this does not work + # when you call a static method from an instance. + # See https://github.com/python/mypy/issues/7781 @staticmethod - def simple_select_one_txn(txn, table, keyvalues, retcols, allow_none=False): + def simple_select_one_txn( + txn: LoggingTransaction, + table: str, + keyvalues: Dict[str, Any], + retcols: Iterable[str], + allow_none: bool = False, + ) -> Optional[Dict[str, Any]]: select_sql = "SELECT %s FROM %s WHERE %s" % ( ", ".join(retcols), table, @@ -1277,24 +1535,29 @@ class Database(object): return dict(zip(retcols, row)) - def simple_delete_one(self, table, keyvalues, desc="simple_delete_one"): + async def simple_delete_one( + self, table: str, keyvalues: Dict[str, Any], desc: str = "simple_delete_one" + ) -> None: """Executes a DELETE query on the named table, expecting to delete a single row. Args: - table : string giving the table name - keyvalues : dict of column names and values to select the row with + table: string giving the table name + keyvalues: dict of column names and values to select the row with + desc: description of the transaction, for logging and metrics """ - return self.runInteraction(desc, self.simple_delete_one_txn, table, keyvalues) + await self.runInteraction(desc, self.simple_delete_one_txn, table, keyvalues) @staticmethod - def simple_delete_one_txn(txn, table, keyvalues): + def simple_delete_one_txn( + txn: LoggingTransaction, table: str, keyvalues: Dict[str, Any] + ) -> None: """Executes a DELETE query on the named table, expecting to delete a single row. Args: - table : string giving the table name - keyvalues : dict of column names and values to select the row with + table: string giving the table name + keyvalues: dict of column names and values to select the row with """ sql = "DELETE FROM %s WHERE %s" % ( table, @@ -1307,11 +1570,38 @@ class Database(object): if txn.rowcount > 1: raise StoreError(500, "More than one row matched (%s)" % (table,)) - def simple_delete(self, table, keyvalues, desc): - return self.runInteraction(desc, self.simple_delete_txn, table, keyvalues) + async def simple_delete( + self, table: str, keyvalues: Dict[str, Any], desc: str + ) -> int: + """Executes a DELETE query on the named table. + + Filters rows by the key-value pairs. + + Args: + table: string giving the table name + keyvalues: dict of column names and values to select the row with + desc: description of the transaction, for logging and metrics + + Returns: + The number of deleted rows. + """ + return await self.runInteraction(desc, self.simple_delete_txn, table, keyvalues) @staticmethod - def simple_delete_txn(txn, table, keyvalues): + def simple_delete_txn( + txn: LoggingTransaction, table: str, keyvalues: Dict[str, Any] + ) -> int: + """Executes a DELETE query on the named table. + + Filters rows by the key-value pairs. + + Args: + table: string giving the table name + keyvalues: dict of column names and values to select the row with + + Returns: + The number of deleted rows. + """ sql = "DELETE FROM %s WHERE %s" % ( table, " AND ".join("%s = ?" % (k,) for k in keyvalues), @@ -1320,26 +1610,53 @@ class Database(object): txn.execute(sql, list(keyvalues.values())) return txn.rowcount - def simple_delete_many(self, table, column, iterable, keyvalues, desc): - return self.runInteraction( + async def simple_delete_many( + self, + table: str, + column: str, + iterable: Iterable[Any], + keyvalues: Dict[str, Any], + desc: str, + ) -> int: + """Executes a DELETE query on the named table. + + Filters rows by if value of `column` is in `iterable`. + + Args: + table: string giving the table name + column: column name to test for inclusion against `iterable` + iterable: list + keyvalues: dict of column names and values to select the rows with + desc: description of the transaction, for logging and metrics + + Returns: + Number rows deleted + """ + return await self.runInteraction( desc, self.simple_delete_many_txn, table, column, iterable, keyvalues ) @staticmethod - def simple_delete_many_txn(txn, table, column, iterable, keyvalues): + def simple_delete_many_txn( + txn: LoggingTransaction, + table: str, + column: str, + iterable: Iterable[Any], + keyvalues: Dict[str, Any], + ) -> int: """Executes a DELETE query on the named table. Filters rows by if value of `column` is in `iterable`. Args: - txn : Transaction object - table : string giving the table name - column : column name to test for inclusion against `iterable` - iterable : list - keyvalues : dict of column names and values to select the rows with + txn: Transaction object + table: string giving the table name + column: column name to test for inclusion against `iterable` + iterable: list + keyvalues: dict of column names and values to select the rows with Returns: - int: Number rows deleted + Number rows deleted """ if not iterable: return 0 @@ -1360,8 +1677,14 @@ class Database(object): return txn.rowcount def get_cache_dict( - self, db_conn, table, entity_column, stream_column, max_value, limit=100000 - ): + self, + db_conn: Connection, + table: str, + entity_column: str, + stream_column: str, + max_value: int, + limit: int = 100000, + ) -> Tuple[Dict[Any, int], int]: # Fetch a mapping of room_id -> max stream position for "recent" rooms. # It doesn't really matter how many we get, the StreamChangeCache will # do the right thing to ensure it respects the max size of cache. @@ -1392,65 +1715,19 @@ class Database(object): return cache, min_val - def simple_select_list_paginate( - self, - table, - orderby, - start, - limit, - retcols, - filters=None, - keyvalues=None, - order_direction="ASC", - desc="simple_select_list_paginate", - ): - """ - Executes a SELECT query on the named table with start and limit, - of row numbers, which may return zero or number of rows from start to limit, - returning the result as a list of dicts. - - Args: - table (str): the table name - filters (dict[str, T] | None): - column names and values to filter the rows with, or None to not - apply a WHERE ? LIKE ? clause. - keyvalues (dict[str, T] | None): - column names and values to select the rows with, or None to not - apply a WHERE clause. - orderby (str): Column to order the results by. - start (int): Index to begin the query at. - limit (int): Number of results to return. - retcols (iterable[str]): the names of the columns to return - order_direction (str): Whether the results should be ordered "ASC" or "DESC". - Returns: - defer.Deferred: resolves to list[dict[str, Any]] - """ - return self.runInteraction( - desc, - self.simple_select_list_paginate_txn, - table, - orderby, - start, - limit, - retcols, - filters=filters, - keyvalues=keyvalues, - order_direction=order_direction, - ) - @classmethod def simple_select_list_paginate_txn( cls, - txn, - table, - orderby, - start, - limit, - retcols, - filters=None, - keyvalues=None, - order_direction="ASC", - ): + txn: LoggingTransaction, + table: str, + orderby: str, + start: int, + limit: int, + retcols: Iterable[str], + filters: Optional[Dict[str, Any]] = None, + keyvalues: Optional[Dict[str, Any]] = None, + order_direction: str = "ASC", + ) -> List[Dict[str, Any]]: """ Executes a SELECT query on the named table with start and limit, of row numbers, which may return zero or number of rows from start to limit, @@ -1461,21 +1738,22 @@ class Database(object): using 'AND'. Args: - txn : Transaction object - table (str): the table name - orderby (str): Column to order the results by. - start (int): Index to begin the query at. - limit (int): Number of results to return. - retcols (iterable[str]): the names of the columns to return - filters (dict[str, T] | None): + txn: Transaction object + table: the table name + orderby: Column to order the results by. + start: Index to begin the query at. + limit: Number of results to return. + retcols: the names of the columns to return + filters: column names and values to filter the rows with, or None to not apply a WHERE ? LIKE ? clause. - keyvalues (dict[str, T] | None): + keyvalues: column names and values to select the rows with, or None to not apply a WHERE clause. - order_direction (str): Whether the results should be ordered "ASC" or "DESC". + order_direction: Whether the results should be ordered "ASC" or "DESC". + Returns: - defer.Deferred: resolves to list[dict[str, Any]] + The result as a list of dictionaries. """ if order_direction not in ["ASC", "DESC"]: raise ValueError("order_direction must be one of 'ASC' or 'DESC'.") @@ -1501,51 +1779,65 @@ class Database(object): return cls.cursor_to_dict(txn) - def simple_search_list(self, table, term, col, retcols, desc="simple_search_list"): + async def simple_search_list( + self, + table: str, + term: Optional[str], + col: str, + retcols: Iterable[str], + desc="simple_search_list", + ) -> Optional[List[Dict[str, Any]]]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. Args: - table (str): the table name - term (str | None): - term for searching the table matched to a column. - col (str): column to query term should be matched to - retcols (iterable[str]): the names of the columns to return + table: the table name + term: term for searching the table matched to a column. + col: column to query term should be matched to + retcols: the names of the columns to return + Returns: - defer.Deferred: resolves to list[dict[str, Any]] or None + A list of dictionaries or None. """ - return self.runInteraction( + return await self.runInteraction( desc, self.simple_search_list_txn, table, term, col, retcols ) @classmethod - def simple_search_list_txn(cls, txn, table, term, col, retcols): + def simple_search_list_txn( + cls, + txn: LoggingTransaction, + table: str, + term: Optional[str], + col: str, + retcols: Iterable[str], + ) -> Optional[List[Dict[str, Any]]]: """Executes a SELECT query on the named table, which may return zero or more rows, returning the result as a list of dicts. Args: - txn : Transaction object - table (str): the table name - term (str | None): - term for searching the table matched to a column. - col (str): column to query term should be matched to - retcols (iterable[str]): the names of the columns to return + txn: Transaction object + table: the table name + term: term for searching the table matched to a column. + col: column to query term should be matched to + retcols: the names of the columns to return + Returns: - defer.Deferred: resolves to list[dict[str, Any]] or None + None if no term is given, otherwise a list of dictionaries. """ if term: sql = "SELECT %s FROM %s WHERE %s LIKE ?" % (", ".join(retcols), table, col) termvalues = ["%%" + term + "%%"] txn.execute(sql, termvalues) else: - return 0 + return None return cls.cursor_to_dict(txn) def make_in_list_sql_clause( - database_engine, column: str, iterable: Iterable + database_engine: BaseDatabaseEngine, column: str, iterable: Iterable ) -> Tuple[str, list]: """Returns an SQL clause that checks the given column is in the iterable. diff --git a/synapse/storage/databases/__init__.py b/synapse/storage/databases/__init__.py new file mode 100644
index 0000000000..aa5d490624 --- /dev/null +++ b/synapse/storage/databases/__init__.py
@@ -0,0 +1,119 @@ +# -*- coding: utf-8 -*- +# Copyright 2019 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import logging + +from synapse.storage.database import DatabasePool, make_conn +from synapse.storage.databases.main.events import PersistEventsStore +from synapse.storage.databases.state import StateGroupDataStore +from synapse.storage.engines import create_engine +from synapse.storage.prepare_database import prepare_database + +logger = logging.getLogger(__name__) + + +class Databases: + """The various databases. + + These are low level interfaces to physical databases. + + Attributes: + main (DataStore) + """ + + def __init__(self, main_store_class, hs): + # Note we pass in the main store class here as workers use a different main + # store. + + self.databases = [] + main = None + state = None + persist_events = None + + for database_config in hs.config.database.databases: + db_name = database_config.name + engine = create_engine(database_config.config) + + with make_conn(database_config, engine) as db_conn: + logger.info("[database config %r]: Checking database server", db_name) + engine.check_database(db_conn) + + logger.info( + "[database config %r]: Preparing for databases %r", + db_name, + database_config.databases, + ) + prepare_database( + db_conn, engine, hs.config, databases=database_config.databases, + ) + + database = DatabasePool(hs, database_config, engine) + + if "main" in database_config.databases: + logger.info( + "[database config %r]: Starting 'main' database", db_name + ) + + # Sanity check we don't try and configure the main store on + # multiple databases. + if main: + raise Exception("'main' data store already configured") + + main = main_store_class(database, db_conn, hs) + + # If we're on a process that can persist events also + # instantiate a `PersistEventsStore` + if hs.get_instance_name() in hs.config.worker.writers.events: + persist_events = PersistEventsStore(hs, database, main) + + if "state" in database_config.databases: + logger.info( + "[database config %r]: Starting 'state' database", db_name + ) + + # Sanity check we don't try and configure the state store on + # multiple databases. + if state: + raise Exception("'state' data store already configured") + + state = StateGroupDataStore(database, db_conn, hs) + + db_conn.commit() + + self.databases.append(database) + + logger.info("[database config %r]: prepared", db_name) + + # Closing the context manager doesn't close the connection. + # psycopg will close the connection when the object gets GCed, but *only* + # if the PID is the same as when the connection was opened [1], and + # it may not be if we fork in the meantime. + # + # [1]: https://github.com/psycopg/psycopg2/blob/2_8_5/psycopg/connection_type.c#L1378 + + db_conn.close() + + # Sanity check that we have actually configured all the required stores. + if not main: + raise Exception("No 'main' database configured") + + if not state: + raise Exception("No 'state' database configured") + + # We use local variables here to ensure that the databases do not have + # optional types. + self.main = main + self.state = state + self.persist_events = persist_events diff --git a/synapse/storage/data_stores/main/__init__.py b/synapse/storage/databases/main/__init__.py
index 932458f651..0cb12f4c61 100644 --- a/synapse/storage/data_stores/main/__init__.py +++ b/synapse/storage/databases/main/__init__.py
@@ -18,16 +18,18 @@ import calendar import logging import time +from typing import Any, Dict, List, Optional, Tuple from synapse.api.constants import PresenceState from synapse.config.homeserver import HomeServerConfig -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.engines import PostgresEngine from synapse.storage.util.id_generators import ( IdGenerator, MultiWriterIdGenerator, StreamIdGenerator, ) +from synapse.types import get_domain_from_id from synapse.util.caches.stream_change_cache import StreamChangeCache from .account_data import AccountDataStore @@ -119,7 +121,7 @@ class DataStore( CacheInvalidationWorkerStore, ServerMetricsStore, ): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self.hs = hs self._clock = hs.get_clock() self.database_engine = database.engine @@ -158,23 +160,29 @@ class DataStore( ) if isinstance(self.database_engine, PostgresEngine): + # We set the `writers` to an empty list here as we don't care about + # missing updates over restarts, as we'll not have anything in our + # caches to invalidate. (This reduces the amount of writes to the DB + # that happen). self._cache_id_gen = MultiWriterIdGenerator( db_conn, database, - instance_name="master", + stream_name="caches", + instance_name=hs.get_instance_name(), table="cache_invalidation_stream_by_instance", instance_column="instance_name", id_column="stream_id", sequence_name="cache_invalidation_stream_seq", + writers=[], ) else: self._cache_id_gen = None - super(DataStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) self._presence_on_startup = self._get_active_presence(db_conn) - presence_cache_prefill, min_presence_val = self.db.get_cache_dict( + presence_cache_prefill, min_presence_val = self.db_pool.get_cache_dict( db_conn, "presence_stream", entity_column="user_id", @@ -188,7 +196,7 @@ class DataStore( ) max_device_inbox_id = self._device_inbox_id_gen.get_current_token() - device_inbox_prefill, min_device_inbox_id = self.db.get_cache_dict( + device_inbox_prefill, min_device_inbox_id = self.db_pool.get_cache_dict( db_conn, "device_inbox", entity_column="user_id", @@ -203,7 +211,7 @@ class DataStore( ) # The federation outbox and the local device inbox uses the same # stream_id generator. - device_outbox_prefill, min_device_outbox_id = self.db.get_cache_dict( + device_outbox_prefill, min_device_outbox_id = self.db_pool.get_cache_dict( db_conn, "device_federation_outbox", entity_column="destination", @@ -229,7 +237,7 @@ class DataStore( ) events_max = self._stream_id_gen.get_current_token() - curr_state_delta_prefill, min_curr_state_delta_id = self.db.get_cache_dict( + curr_state_delta_prefill, min_curr_state_delta_id = self.db_pool.get_cache_dict( db_conn, "current_state_delta_stream", entity_column="room_id", @@ -243,7 +251,7 @@ class DataStore( prefilled_cache=curr_state_delta_prefill, ) - _group_updates_prefill, min_group_updates_id = self.db.get_cache_dict( + _group_updates_prefill, min_group_updates_id = self.db_pool.get_cache_dict( db_conn, "local_group_updates", entity_column="user_id", @@ -263,6 +271,9 @@ class DataStore( # Used in _generate_user_daily_visits to keep track of progress self._last_user_visit_update = self._get_start_of_day() + def get_device_stream_token(self) -> int: + return self._device_list_id_gen.get_current_token() + def take_presence_startup_info(self): active_on_startup = self._presence_on_startup self._presence_on_startup = None @@ -282,7 +293,7 @@ class DataStore( txn = db_conn.cursor() txn.execute(sql, (PresenceState.OFFLINE,)) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) txn.close() for row in rows: @@ -290,14 +301,16 @@ class DataStore( return [UserPresenceState(**row) for row in rows] - def count_daily_users(self): + async def count_daily_users(self) -> int: """ Counts the number of users who used this homeserver in the last 24 hours. """ yesterday = int(self._clock.time_msec()) - (1000 * 60 * 60 * 24) - return self.db.runInteraction("count_daily_users", self._count_users, yesterday) + return await self.db_pool.runInteraction( + "count_daily_users", self._count_users, yesterday + ) - def count_monthly_users(self): + async def count_monthly_users(self) -> int: """ Counts the number of users who used this homeserver in the last 30 days. Note this method is intended for phonehome metrics only and is different @@ -305,7 +318,7 @@ class DataStore( amongst other things, includes a 3 day grace period before a user counts. """ thirty_days_ago = int(self._clock.time_msec()) - (1000 * 60 * 60 * 24 * 30) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "count_monthly_users", self._count_users, thirty_days_ago ) @@ -324,15 +337,15 @@ class DataStore( (count,) = txn.fetchone() return count - def count_r30_users(self): + async def count_r30_users(self) -> Dict[str, int]: """ Counts the number of 30 day retained users, defined as:- * Users who have created their accounts more than 30 days ago * Where last seen at most 30 days ago * Where account creation and last_seen are > 30 days apart - Returns counts globaly for a given user as well as breaking - by platform + Returns: + A mapping of counts globally as well as broken out by platform. """ def _count_r30_users(txn): @@ -405,7 +418,7 @@ class DataStore( return results - return self.db.runInteraction("count_r30_users", _count_r30_users) + return await self.db_pool.runInteraction("count_r30_users", _count_r30_users) def _get_start_of_day(self): """ @@ -415,7 +428,7 @@ class DataStore( today_start = calendar.timegm((now.tm_year, now.tm_mon, now.tm_mday, 0, 0, 0)) return today_start * 1000 - def generate_user_daily_visits(self): + async def generate_user_daily_visits(self) -> None: """ Generates daily visit data for use in cohort/ retention analysis """ @@ -470,18 +483,17 @@ class DataStore( # frequently self._last_user_visit_update = now - return self.db.runInteraction( + await self.db_pool.runInteraction( "generate_user_daily_visits", _generate_user_daily_visits ) - def get_users(self): + async def get_users(self) -> List[Dict[str, Any]]: """Function to retrieve a list of users in users table. - Args: Returns: - defer.Deferred: resolves to list[dict[str, Any]] + A list of dictionaries representing users. """ - return self.db.simple_select_list( + return await self.db_pool.simple_select_list( table="users", keyvalues={}, retcols=[ @@ -495,30 +507,40 @@ class DataStore( desc="get_users", ) - def get_users_paginate( - self, start, limit, name=None, guests=True, deactivated=False - ): + async def get_users_paginate( + self, + start: int, + limit: int, + user_id: Optional[str] = None, + name: Optional[str] = None, + guests: bool = True, + deactivated: bool = False, + ) -> Tuple[List[Dict[str, Any]], int]: """Function to retrieve a paginated list of users from users list. This will return a json list of users and the total number of users matching the filter criteria. Args: - start (int): start number to begin the query from - limit (int): number of rows to retrieve - name (string): filter for user names - guests (bool): whether to in include guest users - deactivated (bool): whether to include deactivated users + start: start number to begin the query from + limit: number of rows to retrieve + user_id: search for user_id. ignored if name is not None + name: search for local part of user_id or display name + guests: whether to in include guest users + deactivated: whether to include deactivated users Returns: - defer.Deferred: resolves to list[dict[str, Any]], int + A tuple of a list of mappings from user to information and a count of total users. """ def get_users_paginate_txn(txn): filters = [] - args = [] + args = [self.hs.config.server_name] if name: + filters.append("(name LIKE ? OR displayname LIKE ?)") + args.extend(["@%" + name + "%:%", "%" + name + "%"]) + elif user_id: filters.append("name LIKE ?") - args.append("%" + name + "%") + args.extend(["%" + user_id + "%"]) if not guests: filters.append("is_guest = 0") @@ -528,37 +550,42 @@ class DataStore( where_clause = "WHERE " + " AND ".join(filters) if len(filters) > 0 else "" - sql = "SELECT COUNT(*) as total_users FROM users %s" % (where_clause) - txn.execute(sql, args) - count = txn.fetchone()[0] - - args = [self.hs.config.server_name] + args + [limit, start] - sql = """ - SELECT name, user_type, is_guest, admin, deactivated, displayname, avatar_url + sql_base = """ FROM users as u LEFT JOIN profiles AS p ON u.name = '@' || p.user_id || ':' || ? {} - ORDER BY u.name LIMIT ? OFFSET ? """.format( where_clause ) + sql = "SELECT COUNT(*) as total_users " + sql_base + txn.execute(sql, args) + count = txn.fetchone()[0] + + sql = ( + "SELECT name, user_type, is_guest, admin, deactivated, displayname, avatar_url " + + sql_base + + " ORDER BY u.name LIMIT ? OFFSET ?" + ) + args += [limit, start] txn.execute(sql, args) - users = self.db.cursor_to_dict(txn) + users = self.db_pool.cursor_to_dict(txn) return users, count - return self.db.runInteraction("get_users_paginate_txn", get_users_paginate_txn) + return await self.db_pool.runInteraction( + "get_users_paginate_txn", get_users_paginate_txn + ) - def search_users(self, term): + async def search_users(self, term: str) -> Optional[List[Dict[str, Any]]]: """Function to search users list for one or more users with the matched term. Args: - term (str): search term - col (str): column to query term should be matched to + term: search term + Returns: - defer.Deferred: resolves to list[dict[str, Any]] + A list of dictionaries or None. """ - return self.db.simple_search_list( + return await self.db_pool.simple_search_list( table="users", term=term, col="name", @@ -571,21 +598,24 @@ def check_database_before_upgrade(cur, database_engine, config: HomeServerConfig """Called before upgrading an existing database to check that it is broadly sane compared with the configuration. """ - domain = config.server_name + logger.info("Checking database for consistency with configuration...") - sql = database_engine.convert_param_style( - "SELECT COUNT(*) FROM users WHERE name NOT LIKE ?" - ) - pat = "%:" + domain - cur.execute(sql, (pat,)) - num_not_matching = cur.fetchall()[0][0] - if num_not_matching == 0: + # if there are any users in the database, check that the username matches our + # configured server name. + + cur.execute("SELECT name FROM users LIMIT 1") + rows = cur.fetchall() + if not rows: + return + + user_domain = get_domain_from_id(rows[0][0]) + if user_domain == config.server_name: return raise Exception( "Found users in database not native to %s!\n" - "You cannot changed a synapse server_name after it's been configured" - % (domain,) + "You cannot change a synapse server_name after it's been configured" + % (config.server_name,) ) diff --git a/synapse/storage/data_stores/main/account_data.py b/synapse/storage/databases/main/account_data.py
index 33cc372dfd..ef81d73573 100644 --- a/synapse/storage/data_stores/main/account_data.py +++ b/synapse/storage/databases/main/account_data.py
@@ -16,37 +16,33 @@ import abc import logging -from typing import List, Tuple - -from canonicaljson import json - -from twisted.internet import defer +from typing import Dict, List, Optional, Tuple from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.util.id_generators import StreamIdGenerator -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks +from synapse.types import JsonDict +from synapse.util import json_encoder +from synapse.util.caches.descriptors import _CacheContext, cached from synapse.util.caches.stream_change_cache import StreamChangeCache logger = logging.getLogger(__name__) -class AccountDataWorkerStore(SQLBaseStore): +# The ABCMeta metaclass ensures that it cannot be instantiated without +# the abstract methods being implemented. +class AccountDataWorkerStore(SQLBaseStore, metaclass=abc.ABCMeta): """This is an abstract base class where subclasses must implement `get_max_account_data_stream_id` which can be called in the initializer. """ - # This ABCMeta metaclass ensures that we cannot be instantiated without - # the abstract methods being implemented. - __metaclass__ = abc.ABCMeta - - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): account_max = self.get_max_account_data_stream_id() self._account_data_stream_cache = StreamChangeCache( "AccountDataAndTagsChangeCache", account_max ) - super(AccountDataWorkerStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) @abc.abstractmethod def get_max_account_data_stream_id(self): @@ -58,18 +54,20 @@ class AccountDataWorkerStore(SQLBaseStore): raise NotImplementedError() @cached() - def get_account_data_for_user(self, user_id): + async def get_account_data_for_user( + self, user_id: str + ) -> Tuple[Dict[str, JsonDict], Dict[str, Dict[str, JsonDict]]]: """Get all the client account_data for a user. Args: - user_id(str): The user to get the account_data for. + user_id: The user to get the account_data for. Returns: - A deferred pair of a dict of global account_data and a dict - mapping from room_id string to per room account_data dicts. + A 2-tuple of a dict of global account_data and a dict mapping from + room_id string to per room account_data dicts. """ def get_account_data_for_user_txn(txn): - rows = self.db.simple_select_list_txn( + rows = self.db_pool.simple_select_list_txn( txn, "account_data", {"user_id": user_id}, @@ -80,7 +78,7 @@ class AccountDataWorkerStore(SQLBaseStore): row["account_data_type"]: db_to_json(row["content"]) for row in rows } - rows = self.db.simple_select_list_txn( + rows = self.db_pool.simple_select_list_txn( txn, "room_account_data", {"user_id": user_id}, @@ -94,17 +92,19 @@ class AccountDataWorkerStore(SQLBaseStore): return global_account_data, by_room - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_account_data_for_user", get_account_data_for_user_txn ) - @cachedInlineCallbacks(num_args=2, max_entries=5000) - def get_global_account_data_by_type_for_user(self, data_type, user_id): + @cached(num_args=2, max_entries=5000) + async def get_global_account_data_by_type_for_user( + self, data_type: str, user_id: str + ) -> Optional[JsonDict]: """ Returns: - Deferred: A dict + The account data. """ - result = yield self.db.simple_select_one_onecol( + result = await self.db_pool.simple_select_one_onecol( table="account_data", keyvalues={"user_id": user_id, "account_data_type": data_type}, retcol="content", @@ -118,18 +118,20 @@ class AccountDataWorkerStore(SQLBaseStore): return None @cached(num_args=2) - def get_account_data_for_room(self, user_id, room_id): + async def get_account_data_for_room( + self, user_id: str, room_id: str + ) -> Dict[str, JsonDict]: """Get all the client account_data for a user for a room. Args: - user_id(str): The user to get the account_data for. - room_id(str): The room to get the account_data for. + user_id: The user to get the account_data for. + room_id: The room to get the account_data for. Returns: - A deferred dict of the room account_data + A dict of the room account_data """ def get_account_data_for_room_txn(txn): - rows = self.db.simple_select_list_txn( + rows = self.db_pool.simple_select_list_txn( txn, "room_account_data", {"user_id": user_id, "room_id": room_id}, @@ -140,25 +142,26 @@ class AccountDataWorkerStore(SQLBaseStore): row["account_data_type"]: db_to_json(row["content"]) for row in rows } - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_account_data_for_room", get_account_data_for_room_txn ) @cached(num_args=3, max_entries=5000) - def get_account_data_for_room_and_type(self, user_id, room_id, account_data_type): + async def get_account_data_for_room_and_type( + self, user_id: str, room_id: str, account_data_type: str + ) -> Optional[JsonDict]: """Get the client account_data of given type for a user for a room. Args: - user_id(str): The user to get the account_data for. - room_id(str): The room to get the account_data for. - account_data_type (str): The account data type to get. + user_id: The user to get the account_data for. + room_id: The room to get the account_data for. + account_data_type: The account data type to get. Returns: - A deferred of the room account_data for that type, or None if - there isn't any set. + The room account_data for that type, or None if there isn't any set. """ def get_account_data_for_room_and_type_txn(txn): - content_json = self.db.simple_select_one_onecol_txn( + content_json = self.db_pool.simple_select_one_onecol_txn( txn, table="room_account_data", keyvalues={ @@ -172,7 +175,7 @@ class AccountDataWorkerStore(SQLBaseStore): return db_to_json(content_json) if content_json else None - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_account_data_for_room_and_type", get_account_data_for_room_and_type_txn ) @@ -202,7 +205,7 @@ class AccountDataWorkerStore(SQLBaseStore): txn.execute(sql, (last_id, current_id, limit)) return txn.fetchall() - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_updated_global_account_data", get_updated_global_account_data_txn ) @@ -232,16 +235,18 @@ class AccountDataWorkerStore(SQLBaseStore): txn.execute(sql, (last_id, current_id, limit)) return txn.fetchall() - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_updated_room_account_data", get_updated_room_account_data_txn ) - def get_updated_account_data_for_user(self, user_id, stream_id): + async def get_updated_account_data_for_user( + self, user_id: str, stream_id: int + ) -> Tuple[Dict[str, JsonDict], Dict[str, Dict[str, JsonDict]]]: """Get all the client account_data for a that's changed for a user Args: - user_id(str): The user to get the account_data for. - stream_id(int): The point in the stream since which to get updates + user_id: The user to get the account_data for. + stream_id: The point in the stream since which to get updates Returns: A deferred pair of a dict of global account_data and a dict mapping from room_id string to per room account_data dicts. @@ -275,15 +280,17 @@ class AccountDataWorkerStore(SQLBaseStore): user_id, int(stream_id) ) if not changed: - return defer.succeed(({}, {})) + return ({}, {}) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_updated_account_data_for_user", get_updated_account_data_for_user_txn ) - @cachedInlineCallbacks(num_args=2, cache_context=True, max_entries=5000) - def is_ignored_by(self, ignored_user_id, ignorer_user_id, cache_context): - ignored_account_data = yield self.get_global_account_data_by_type_for_user( + @cached(num_args=2, cache_context=True, max_entries=5000) + async def is_ignored_by( + self, ignored_user_id: str, ignorer_user_id: str, cache_context: _CacheContext + ) -> bool: + ignored_account_data = await self.get_global_account_data_by_type_for_user( "m.ignored_user_list", ignorer_user_id, on_invalidate=cache_context.invalidate, @@ -295,7 +302,7 @@ class AccountDataWorkerStore(SQLBaseStore): class AccountDataStore(AccountDataWorkerStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self._account_data_id_gen = StreamIdGenerator( db_conn, "account_data_max_stream_id", @@ -306,34 +313,37 @@ class AccountDataStore(AccountDataWorkerStore): ], ) - super(AccountDataStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) - def get_max_account_data_stream_id(self): + def get_max_account_data_stream_id(self) -> int: """Get the current max stream id for the private user data stream Returns: - A deferred int. + The maximum stream ID. """ return self._account_data_id_gen.get_current_token() - @defer.inlineCallbacks - def add_account_data_to_room(self, user_id, room_id, account_data_type, content): + async def add_account_data_to_room( + self, user_id: str, room_id: str, account_data_type: str, content: JsonDict + ) -> int: """Add some account_data to a room for a user. + Args: - user_id(str): The user to add a tag for. - room_id(str): The room to add a tag for. - account_data_type(str): The type of account_data to add. - content(dict): A json object to associate with the tag. + user_id: The user to add a tag for. + room_id: The room to add a tag for. + account_data_type: The type of account_data to add. + content: A json object to associate with the tag. + Returns: - A deferred that completes once the account_data has been added. + The maximum stream ID. """ - content_json = json.dumps(content) + content_json = json_encoder.encode(content) - with self._account_data_id_gen.get_next() as next_id: + async with self._account_data_id_gen.get_next() as next_id: # no need to lock here as room_account_data has a unique constraint # on (user_id, room_id, account_data_type) so simple_upsert will # retry if there is a conflict. - yield self.db.simple_upsert( + await self.db_pool.simple_upsert( desc="add_room_account_data", table="room_account_data", keyvalues={ @@ -351,7 +361,7 @@ class AccountDataStore(AccountDataWorkerStore): # doesn't sound any worse than the whole update getting lost, # which is what would happen if we combined the two into one # transaction. - yield self._update_max_stream_id(next_id) + await self._update_max_stream_id(next_id) self._account_data_stream_cache.entity_has_changed(user_id, next_id) self.get_account_data_for_user.invalidate((user_id,)) @@ -360,26 +370,28 @@ class AccountDataStore(AccountDataWorkerStore): (user_id, room_id, account_data_type), content ) - result = self._account_data_id_gen.get_current_token() - return result + return self._account_data_id_gen.get_current_token() - @defer.inlineCallbacks - def add_account_data_for_user(self, user_id, account_data_type, content): + async def add_account_data_for_user( + self, user_id: str, account_data_type: str, content: JsonDict + ) -> int: """Add some account_data to a room for a user. + Args: - user_id(str): The user to add a tag for. - account_data_type(str): The type of account_data to add. - content(dict): A json object to associate with the tag. + user_id: The user to add a tag for. + account_data_type: The type of account_data to add. + content: A json object to associate with the tag. + Returns: - A deferred that completes once the account_data has been added. + The maximum stream ID. """ - content_json = json.dumps(content) + content_json = json_encoder.encode(content) - with self._account_data_id_gen.get_next() as next_id: + async with self._account_data_id_gen.get_next() as next_id: # no need to lock here as account_data has a unique constraint on # (user_id, account_data_type) so simple_upsert will retry if # there is a conflict. - yield self.db.simple_upsert( + await self.db_pool.simple_upsert( desc="add_user_account_data", table="account_data", keyvalues={"user_id": user_id, "account_data_type": account_data_type}, @@ -397,7 +409,7 @@ class AccountDataStore(AccountDataWorkerStore): # Note: This is only here for backwards compat to allow admins to # roll back to a previous Synapse version. Next time we update the # database version we can remove this table. - yield self._update_max_stream_id(next_id) + await self._update_max_stream_id(next_id) self._account_data_stream_cache.entity_has_changed(user_id, next_id) self.get_account_data_for_user.invalidate((user_id,)) @@ -405,14 +417,13 @@ class AccountDataStore(AccountDataWorkerStore): (account_data_type, user_id) ) - result = self._account_data_id_gen.get_current_token() - return result + return self._account_data_id_gen.get_current_token() - def _update_max_stream_id(self, next_id): + async def _update_max_stream_id(self, next_id: int) -> None: """Update the max stream_id Args: - next_id(int): The the revision to advance to. + next_id: The the revision to advance to. """ # Note: This is only here for backwards compat to allow admins to @@ -427,4 +438,4 @@ class AccountDataStore(AccountDataWorkerStore): ) txn.execute(update_max_id_sql, (next_id, next_id)) - return self.db.runInteraction("update_account_data_max_stream_id", _update) + await self.db_pool.runInteraction("update_account_data_max_stream_id", _update) diff --git a/synapse/storage/data_stores/main/appservice.py b/synapse/storage/databases/main/appservice.py
index 56659fed37..85f6b1e3fd 100644 --- a/synapse/storage/data_stores/main/appservice.py +++ b/synapse/storage/databases/main/appservice.py
@@ -16,15 +16,12 @@ import logging import re -from canonicaljson import json - -from twisted.internet import defer - from synapse.appservice import AppServiceTransaction from synapse.config.appservice import load_appservices from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.util import json_encoder logger = logging.getLogger(__name__) @@ -49,13 +46,13 @@ def _make_exclusive_regex(services_cache): class ApplicationServiceWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self.services_cache = load_appservices( hs.hostname, hs.config.app_service_config_files ) self.exclusive_user_regex = _make_exclusive_regex(self.services_cache) - super(ApplicationServiceWorkerStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) def get_app_services(self): return self.services_cache @@ -124,17 +121,15 @@ class ApplicationServiceStore(ApplicationServiceWorkerStore): class ApplicationServiceTransactionWorkerStore( ApplicationServiceWorkerStore, EventsWorkerStore ): - @defer.inlineCallbacks - def get_appservices_by_state(self, state): + async def get_appservices_by_state(self, state): """Get a list of application services based on their state. Args: state(ApplicationServiceState): The state to filter on. Returns: - A Deferred which resolves to a list of ApplicationServices, which - may be empty. + A list of ApplicationServices, which may be empty. """ - results = yield self.db.simple_select_list( + results = await self.db_pool.simple_select_list( "application_services_state", {"state": state}, ["as_id"] ) # NB: This assumes this class is linked with ApplicationServiceStore @@ -147,16 +142,15 @@ class ApplicationServiceTransactionWorkerStore( services.append(service) return services - @defer.inlineCallbacks - def get_appservice_state(self, service): + async def get_appservice_state(self, service): """Get the application service state. Args: service(ApplicationService): The service whose state to set. Returns: - A Deferred which resolves to ApplicationServiceState. + An ApplicationServiceState. """ - result = yield self.db.simple_select_one( + result = await self.db_pool.simple_select_one( "application_services_state", {"as_id": service.id}, ["state"], @@ -167,20 +161,18 @@ class ApplicationServiceTransactionWorkerStore( return result.get("state") return None - def set_appservice_state(self, service, state): + async def set_appservice_state(self, service, state) -> None: """Set the application service state. Args: service(ApplicationService): The service whose state to set. state(ApplicationServiceState): The connectivity state to apply. - Returns: - A Deferred which resolves when the state was set successfully. """ - return self.db.simple_upsert( + await self.db_pool.simple_upsert( "application_services_state", {"as_id": service.id}, {"state": state} ) - def create_appservice_txn(self, service, events): + async def create_appservice_txn(self, service, events): """Atomically creates a new transaction for this application service with the given list of events. @@ -209,7 +201,7 @@ class ApplicationServiceTransactionWorkerStore( new_txn_id = max(highest_txn_id, last_txn_id) + 1 # Insert new txn into txn table - event_ids = json.dumps([e.event_id for e in events]) + event_ids = json_encoder.encode([e.event_id for e in events]) txn.execute( "INSERT INTO application_services_txns(as_id, txn_id, event_ids) " "VALUES(?,?,?)", @@ -217,18 +209,17 @@ class ApplicationServiceTransactionWorkerStore( ) return AppServiceTransaction(service=service, id=new_txn_id, events=events) - return self.db.runInteraction("create_appservice_txn", _create_appservice_txn) + return await self.db_pool.runInteraction( + "create_appservice_txn", _create_appservice_txn + ) - def complete_appservice_txn(self, txn_id, service): + async def complete_appservice_txn(self, txn_id, service) -> None: """Completes an application service transaction. Args: txn_id(str): The transaction ID being completed. service(ApplicationService): The application service which was sent this transaction. - Returns: - A Deferred which resolves if this transaction was stored - successfully. """ txn_id = int(txn_id) @@ -250,7 +241,7 @@ class ApplicationServiceTransactionWorkerStore( ) # Set current txn_id for AS to 'txn_id' - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, "application_services_state", {"as_id": service.id}, @@ -258,26 +249,24 @@ class ApplicationServiceTransactionWorkerStore( ) # Delete txn - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, "application_services_txns", {"txn_id": txn_id, "as_id": service.id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "complete_appservice_txn", _complete_appservice_txn ) - @defer.inlineCallbacks - def get_oldest_unsent_txn(self, service): + async def get_oldest_unsent_txn(self, service): """Get the oldest transaction which has not been sent for this service. Args: service(ApplicationService): The app service to get the oldest txn. Returns: - A Deferred which resolves to an AppServiceTransaction or - None. + An AppServiceTransaction or None. """ def _get_oldest_unsent_txn(txn): @@ -288,7 +277,7 @@ class ApplicationServiceTransactionWorkerStore( " ORDER BY txn_id ASC LIMIT 1", (service.id,), ) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return None @@ -296,7 +285,7 @@ class ApplicationServiceTransactionWorkerStore( return entry - entry = yield self.db.runInteraction( + entry = await self.db_pool.runInteraction( "get_oldest_unsent_appservice_txn", _get_oldest_unsent_txn ) @@ -305,7 +294,7 @@ class ApplicationServiceTransactionWorkerStore( event_ids = db_to_json(entry["event_ids"]) - events = yield self.get_events_as_list(event_ids) + events = await self.get_events_as_list(event_ids) return AppServiceTransaction(service=service, id=entry["txn_id"], events=events) @@ -320,18 +309,17 @@ class ApplicationServiceTransactionWorkerStore( else: return int(last_txn_id[0]) # select 'last_txn' col - def set_appservice_last_pos(self, pos): + async def set_appservice_last_pos(self, pos) -> None: def set_appservice_last_pos_txn(txn): txn.execute( "UPDATE appservice_stream_position SET stream_ordering = ?", (pos,) ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "set_appservice_last_pos", set_appservice_last_pos_txn ) - @defer.inlineCallbacks - def get_new_events_for_appservice(self, current_id, limit): + async def get_new_events_for_appservice(self, current_id, limit): """Get all new evnets""" def get_new_events_for_appservice_txn(txn): @@ -355,11 +343,11 @@ class ApplicationServiceTransactionWorkerStore( return upper_bound, [row[1] for row in rows] - upper_bound, event_ids = yield self.db.runInteraction( + upper_bound, event_ids = await self.db_pool.runInteraction( "get_new_events_for_appservice", get_new_events_for_appservice_txn ) - events = yield self.get_events_as_list(event_ids) + events = await self.get_events_as_list(event_ids) return upper_bound, events diff --git a/synapse/storage/data_stores/main/cache.py b/synapse/storage/databases/main/cache.py
index f39f556c20..1e7637a6f5 100644 --- a/synapse/storage/data_stores/main/cache.py +++ b/synapse/storage/databases/main/cache.py
@@ -26,7 +26,7 @@ from synapse.replication.tcp.streams.events import ( EventsStreamEventRow, ) from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.engines import PostgresEngine from synapse.util.iterutils import batch_iter @@ -39,7 +39,7 @@ CURRENT_STATE_CACHE_NAME = "cs_cache_fake" class CacheInvalidationWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): super().__init__(database, db_conn, hs) self._instance_name = hs.get_instance_name() @@ -92,7 +92,7 @@ class CacheInvalidationWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_updated_caches", get_all_updated_caches_txn ) @@ -202,7 +202,7 @@ class CacheInvalidationWorkerStore(SQLBaseStore): return cache_func.invalidate(keys) - await self.db.runInteraction( + await self.db_pool.runInteraction( "invalidate_cache_and_stream", self._send_invalidation_to_replication, cache_func.__name__, @@ -287,7 +287,7 @@ class CacheInvalidationWorkerStore(SQLBaseStore): if keys is not None: keys = list(keys) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="cache_invalidation_stream_by_instance", values={ @@ -299,8 +299,8 @@ class CacheInvalidationWorkerStore(SQLBaseStore): }, ) - def get_cache_stream_token(self, instance_name): + def get_cache_stream_token_for_writer(self, instance_name: str) -> int: if self._cache_id_gen: - return self._cache_id_gen.get_current_token(instance_name) + return self._cache_id_gen.get_current_token_for_writer(instance_name) else: return 0 diff --git a/synapse/storage/data_stores/main/censor_events.py b/synapse/storage/databases/main/censor_events.py
index 2d48261724..f211ddbaf8 100644 --- a/synapse/storage/data_stores/main/censor_events.py +++ b/synapse/storage/databases/main/censor_events.py
@@ -16,15 +16,13 @@ import logging from typing import TYPE_CHECKING -from twisted.internet import defer - from synapse.events.utils import prune_event_dict from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.cache import CacheInvalidationWorkerStore -from synapse.storage.data_stores.main.events import encode_json -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.cache import CacheInvalidationWorkerStore +from synapse.storage.databases.main.events import encode_json +from synapse.storage.databases.main.events_worker import EventsWorkerStore if TYPE_CHECKING: from synapse.server import HomeServer @@ -34,7 +32,7 @@ logger = logging.getLogger(__name__) class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBaseStore): - def __init__(self, database: Database, db_conn, hs: "HomeServer"): + def __init__(self, database: DatabasePool, db_conn, hs: "HomeServer"): super().__init__(database, db_conn, hs) def _censor_redactions(): @@ -56,7 +54,7 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase return if not ( - await self.db.updates.has_completed_background_update( + await self.db_pool.updates.has_completed_background_update( "redactions_have_censored_ts_idx" ) ): @@ -85,7 +83,7 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase LIMIT ? """ - rows = await self.db.execute( + rows = await self.db_pool.execute( "_censor_redactions_fetch", None, sql, before_ts, 100 ) @@ -123,14 +121,14 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase if pruned_json: self._censor_event_txn(txn, event_id, pruned_json) - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="redactions", keyvalues={"event_id": redaction_id}, updatevalues={"have_censored": True}, ) - await self.db.runInteraction("_update_censor_txn", _update_censor_txn) + await self.db_pool.runInteraction("_update_censor_txn", _update_censor_txn) def _censor_event_txn(self, txn, event_id, pruned_json): """Censor an event by replacing its JSON in the event_json table with the @@ -141,24 +139,23 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase event_id (str): The ID of the event to censor. pruned_json (str): The pruned JSON """ - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="event_json", keyvalues={"event_id": event_id}, updatevalues={"json": pruned_json}, ) - @defer.inlineCallbacks - def expire_event(self, event_id): + async def expire_event(self, event_id: str) -> None: """Retrieve and expire an event that has expired, and delete its associated expiry timestamp. If the event can't be retrieved, delete its associated timestamp so we don't try to expire it again in the future. Args: - event_id (str): The ID of the event to delete. + event_id: The ID of the event to delete. """ # Try to retrieve the event's content from the database or the event cache. - event = yield self.get_event(event_id) + event = await self.get_event(event_id) def delete_expired_event_txn(txn): # Delete the expiry timestamp associated with this event from the database. @@ -193,7 +190,9 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase txn, "_get_event_cache", (event.event_id,) ) - yield self.db.runInteraction("delete_expired_event", delete_expired_event_txn) + await self.db_pool.runInteraction( + "delete_expired_event", delete_expired_event_txn + ) def _delete_event_expiry_txn(self, txn, event_id): """Delete the expiry timestamp associated with an event ID without deleting the @@ -203,6 +202,6 @@ class CensorEventsStore(EventsWorkerStore, CacheInvalidationWorkerStore, SQLBase txn (LoggingTransaction): The transaction to use to perform the deletion. event_id (str): The event ID to delete the associated expiry timestamp of. """ - return self.db.simple_delete_txn( + return self.db_pool.simple_delete_txn( txn=txn, table="event_expiry", keyvalues={"event_id": event_id} ) diff --git a/synapse/storage/data_stores/main/client_ips.py b/synapse/storage/databases/main/client_ips.py
index 995d4764a9..239c7a949c 100644 --- a/synapse/storage/data_stores/main/client_ips.py +++ b/synapse/storage/databases/main/client_ips.py
@@ -14,12 +14,11 @@ # limitations under the License. import logging - -from twisted.internet import defer +from typing import Dict, Optional, Tuple from synapse.metrics.background_process_metrics import wrap_as_background_process from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database, make_tuple_comparison_clause +from synapse.storage.database import DatabasePool, make_tuple_comparison_clause from synapse.util.caches.descriptors import Cache logger = logging.getLogger(__name__) @@ -31,40 +30,40 @@ LAST_SEEN_GRANULARITY = 120 * 1000 class ClientIpBackgroundUpdateStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(ClientIpBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "user_ips_device_index", index_name="user_ips_device_id", table="user_ips", columns=["user_id", "device_id", "last_seen"], ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "user_ips_last_seen_index", index_name="user_ips_last_seen", table="user_ips", columns=["user_id", "last_seen"], ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "user_ips_last_seen_only_index", index_name="user_ips_last_seen_only", table="user_ips", columns=["last_seen"], ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "user_ips_analyze", self._analyze_user_ip ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "user_ips_remove_dupes", self._remove_user_ip_dupes ) # Register a unique index - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "user_ips_device_unique_index", index_name="user_ips_user_token_ip_unique_index", table="user_ips", @@ -73,28 +72,28 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): ) # Drop the old non-unique index - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "user_ips_drop_nonunique_index", self._remove_user_ip_nonunique ) # Update the last seen info in devices. - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "devices_last_seen", self._devices_last_seen_update ) - @defer.inlineCallbacks - def _remove_user_ip_nonunique(self, progress, batch_size): + async def _remove_user_ip_nonunique(self, progress, batch_size): def f(conn): txn = conn.cursor() txn.execute("DROP INDEX IF EXISTS user_ips_user_ip") txn.close() - yield self.db.runWithConnection(f) - yield self.db.updates._end_background_update("user_ips_drop_nonunique_index") + await self.db_pool.runWithConnection(f) + await self.db_pool.updates._end_background_update( + "user_ips_drop_nonunique_index" + ) return 1 - @defer.inlineCallbacks - def _analyze_user_ip(self, progress, batch_size): + async def _analyze_user_ip(self, progress, batch_size): # Background update to analyze user_ips table before we run the # deduplication background update. The table may not have been analyzed # for ages due to the table locks. @@ -104,14 +103,13 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): def user_ips_analyze(txn): txn.execute("ANALYZE user_ips") - yield self.db.runInteraction("user_ips_analyze", user_ips_analyze) + await self.db_pool.runInteraction("user_ips_analyze", user_ips_analyze) - yield self.db.updates._end_background_update("user_ips_analyze") + await self.db_pool.updates._end_background_update("user_ips_analyze") return 1 - @defer.inlineCallbacks - def _remove_user_ip_dupes(self, progress, batch_size): + async def _remove_user_ip_dupes(self, progress, batch_size): # This works function works by scanning the user_ips table in batches # based on `last_seen`. For each row in a batch it searches the rest of # the table to see if there are any duplicates, if there are then they @@ -138,7 +136,7 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): return None # Get a last seen that has roughly `batch_size` since `begin_last_seen` - end_last_seen = yield self.db.runInteraction( + end_last_seen = await self.db_pool.runInteraction( "user_ips_dups_get_last_seen", get_last_seen ) @@ -269,19 +267,18 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): (user_id, access_token, ip, device_id, user_agent, last_seen), ) - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "user_ips_remove_dupes", {"last_seen": end_last_seen} ) - yield self.db.runInteraction("user_ips_dups_remove", remove) + await self.db_pool.runInteraction("user_ips_dups_remove", remove) if last: - yield self.db.updates._end_background_update("user_ips_remove_dupes") + await self.db_pool.updates._end_background_update("user_ips_remove_dupes") return batch_size - @defer.inlineCallbacks - def _devices_last_seen_update(self, progress, batch_size): + async def _devices_last_seen_update(self, progress, batch_size): """Background update to insert last seen info into devices table """ @@ -336,7 +333,7 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): txn.execute_batch(sql, rows) _, _, _, user_id, device_id = rows[-1] - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "devices_last_seen", {"last_user_id": user_id, "last_device_id": device_id}, @@ -344,24 +341,24 @@ class ClientIpBackgroundUpdateStore(SQLBaseStore): return len(rows) - updated = yield self.db.runInteraction( + updated = await self.db_pool.runInteraction( "_devices_last_seen_update", _devices_last_seen_update_txn ) if not updated: - yield self.db.updates._end_background_update("devices_last_seen") + await self.db_pool.updates._end_background_update("devices_last_seen") return updated class ClientIpStore(ClientIpBackgroundUpdateStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): self.client_ip_last_seen = Cache( name="client_ip_last_seen", keylen=4, max_entries=50000 ) - super(ClientIpStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) self.user_ips_max_age = hs.config.user_ips_max_age @@ -378,8 +375,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): if self.user_ips_max_age: self._clock.looping_call(self._prune_old_user_ips, 5 * 1000) - @defer.inlineCallbacks - def insert_client_ip( + async def insert_client_ip( self, user_id, access_token, ip, user_agent, device_id, now=None ): if not now: @@ -390,7 +386,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): last_seen = self.client_ip_last_seen.get(key) except KeyError: last_seen = None - yield self.populate_monthly_active_users(user_id) + await self.populate_monthly_active_users(user_id) # Rate-limited inserts if last_seen is not None and (now - last_seen) < LAST_SEEN_GRANULARITY: return @@ -400,21 +396,21 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): self._batch_row_update[key] = (user_agent, device_id, now) @wrap_as_background_process("update_client_ips") - def _update_client_ips_batch(self): + async def _update_client_ips_batch(self) -> None: # If the DB pool has already terminated, don't try updating - if not self.db.is_running(): + if not self.db_pool.is_running(): return to_update = self._batch_row_update self._batch_row_update = {} - return self.db.runInteraction( + await self.db_pool.runInteraction( "_update_client_ips_batch", self._update_client_ips_batch_txn, to_update ) def _update_client_ips_batch_txn(self, txn, to_update): - if "user_ips" in self.db._unsafe_to_upsert_tables or ( + if "user_ips" in self.db_pool._unsafe_to_upsert_tables or ( not self.database_engine.can_native_upsert ): self.database_engine.lock_table(txn, "user_ips") @@ -423,7 +419,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): (user_id, access_token, ip), (user_agent, device_id, last_seen) = entry try: - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="user_ips", keyvalues={ @@ -445,7 +441,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): # this is always an update rather than an upsert: the row should # already exist, and if it doesn't, that may be because it has been # deleted, and we don't want to re-create it. - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn, table="devices", keyvalues={"user_id": user_id, "device_id": device_id}, @@ -459,25 +455,25 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): # Failed to upsert, log and continue logger.error("Failed to insert client IP %r: %r", entry, e) - @defer.inlineCallbacks - def get_last_client_ip_by_device(self, user_id, device_id): + async def get_last_client_ip_by_device( + self, user_id: str, device_id: Optional[str] + ) -> Dict[Tuple[str, str], dict]: """For each device_id listed, give the user_ip it was last seen on Args: - user_id (str) - device_id (str): If None fetches all devices for the user + user_id: The user to fetch devices for. + device_id: If None fetches all devices for the user Returns: - defer.Deferred: resolves to a dict, where the keys - are (user_id, device_id) tuples. The values are also dicts, with - keys giving the column names + A dictionary mapping a tuple of (user_id, device_id) to dicts, with + keys giving the column names from the devices table. """ keyvalues = {"user_id": user_id} if device_id is not None: keyvalues["device_id"] = device_id - res = yield self.db.simple_select_list( + res = await self.db_pool.simple_select_list( table="devices", keyvalues=keyvalues, retcols=("user_id", "ip", "user_agent", "device_id", "last_seen"), @@ -499,8 +495,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): } return ret - @defer.inlineCallbacks - def get_user_ip_and_agents(self, user): + async def get_user_ip_and_agents(self, user): user_id = user.to_string() results = {} @@ -510,7 +505,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): user_agent, _, last_seen = self._batch_row_update[key] results[(access_token, ip)] = (user_agent, last_seen) - rows = yield self.db.simple_select_list( + rows = await self.db_pool.simple_select_list( table="user_ips", keyvalues={"user_id": user_id}, retcols=["access_token", "ip", "user_agent", "last_seen"], @@ -540,7 +535,7 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): # Nothing to do return - if not await self.db.updates.has_completed_background_update( + if not await self.db_pool.updates.has_completed_background_update( "devices_last_seen" ): # Only start pruning if we have finished populating the devices @@ -573,4 +568,6 @@ class ClientIpStore(ClientIpBackgroundUpdateStore): def _prune_old_user_ips_txn(txn): txn.execute(sql, (timestamp,)) - await self.db.runInteraction("_prune_old_user_ips", _prune_old_user_ips_txn) + await self.db_pool.runInteraction( + "_prune_old_user_ips", _prune_old_user_ips_txn + ) diff --git a/synapse/storage/data_stores/main/deviceinbox.py b/synapse/storage/databases/main/deviceinbox.py
index da297b31fb..d42faa3f1f 100644 --- a/synapse/storage/data_stores/main/deviceinbox.py +++ b/synapse/storage/databases/main/deviceinbox.py
@@ -16,13 +16,10 @@ import logging from typing import List, Tuple -from canonicaljson import json - -from twisted.internet import defer - from synapse.logging.opentracing import log_kv, set_tag, trace from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.util import json_encoder from synapse.util.caches.expiringcache import ExpiringCache logger = logging.getLogger(__name__) @@ -32,24 +29,31 @@ class DeviceInboxWorkerStore(SQLBaseStore): def get_to_device_stream_token(self): return self._device_inbox_id_gen.get_current_token() - def get_new_messages_for_device( - self, user_id, device_id, last_stream_id, current_stream_id, limit=100 - ): + async def get_new_messages_for_device( + self, + user_id: str, + device_id: str, + last_stream_id: int, + current_stream_id: int, + limit: int = 100, + ) -> Tuple[List[dict], int]: """ Args: - user_id(str): The recipient user_id. - device_id(str): The recipient device_id. - current_stream_id(int): The current position of the to device + user_id: The recipient user_id. + device_id: The recipient device_id. + last_stream_id: The last stream ID checked. + current_stream_id: The current position of the to device message stream. + limit: The maximum number of messages to retrieve. + Returns: - Deferred ([dict], int): List of messages for the device and where - in the stream the messages got to. + A list of messages for the device and where in the stream the messages got to. """ has_changed = self._device_inbox_stream_cache.has_entity_changed( user_id, last_stream_id ) if not has_changed: - return defer.succeed(([], current_stream_id)) + return ([], current_stream_id) def get_new_messages_for_device_txn(txn): sql = ( @@ -70,20 +74,22 @@ class DeviceInboxWorkerStore(SQLBaseStore): stream_pos = current_stream_id return messages, stream_pos - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_new_messages_for_device", get_new_messages_for_device_txn ) @trace - @defer.inlineCallbacks - def delete_messages_for_device(self, user_id, device_id, up_to_stream_id): + async def delete_messages_for_device( + self, user_id: str, device_id: str, up_to_stream_id: int + ) -> int: """ Args: - user_id(str): The recipient user_id. - device_id(str): The recipient device_id. - up_to_stream_id(int): Where to delete messages up to. + user_id: The recipient user_id. + device_id: The recipient device_id. + up_to_stream_id: Where to delete messages up to. + Returns: - A deferred that resolves to the number of messages deleted. + The number of messages deleted. """ # If we have cached the last stream id we've deleted up to, we can # check if there is likely to be anything that needs deleting @@ -110,7 +116,7 @@ class DeviceInboxWorkerStore(SQLBaseStore): txn.execute(sql, (user_id, device_id, up_to_stream_id)) return txn.rowcount - count = yield self.db.runInteraction( + count = await self.db_pool.runInteraction( "delete_messages_for_device", delete_messages_for_device_txn ) @@ -129,9 +135,9 @@ class DeviceInboxWorkerStore(SQLBaseStore): return count @trace - def get_new_device_msgs_for_remote( + async def get_new_device_msgs_for_remote( self, destination, last_stream_id, current_stream_id, limit - ): + ) -> Tuple[List[dict], int]: """ Args: destination(str): The name of the remote server. @@ -140,8 +146,7 @@ class DeviceInboxWorkerStore(SQLBaseStore): current_stream_id(int|long): The current position of the device message stream. Returns: - Deferred ([dict], int|long): List of messages for the device and where - in the stream the messages got to. + A list of messages for the device and where in the stream the messages got to. """ set_tag("destination", destination) @@ -154,11 +159,11 @@ class DeviceInboxWorkerStore(SQLBaseStore): ) if not has_changed or last_stream_id == current_stream_id: log_kv({"message": "No new messages in stream"}) - return defer.succeed(([], current_stream_id)) + return ([], current_stream_id) if limit <= 0: # This can happen if we run out of room for EDUs in the transaction. - return defer.succeed(([], last_stream_id)) + return ([], last_stream_id) @trace def get_new_messages_for_remote_destination_txn(txn): @@ -179,21 +184,21 @@ class DeviceInboxWorkerStore(SQLBaseStore): stream_pos = current_stream_id return messages, stream_pos - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_new_device_msgs_for_remote", get_new_messages_for_remote_destination_txn, ) @trace - def delete_device_msgs_for_remote(self, destination, up_to_stream_id): + async def delete_device_msgs_for_remote( + self, destination: str, up_to_stream_id: int + ) -> None: """Used to delete messages when the remote destination acknowledges their receipt. Args: - destination(str): The destination server_name - up_to_stream_id(int): Where to delete messages up to. - Returns: - A deferred that resolves when the messages have been deleted. + destination: The destination server_name + up_to_stream_id: Where to delete messages up to. """ def delete_messages_for_remote_destination_txn(txn): @@ -204,7 +209,7 @@ class DeviceInboxWorkerStore(SQLBaseStore): ) txn.execute(sql, (destination, up_to_stream_id)) - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_device_msgs_for_remote", delete_messages_for_remote_destination_txn ) @@ -269,7 +274,7 @@ class DeviceInboxWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_new_device_messages", get_all_new_device_messages_txn ) @@ -277,30 +282,29 @@ class DeviceInboxWorkerStore(SQLBaseStore): class DeviceInboxBackgroundUpdateStore(SQLBaseStore): DEVICE_INBOX_STREAM_ID = "device_inbox_stream_drop" - def __init__(self, database: Database, db_conn, hs): - super(DeviceInboxBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "device_inbox_stream_index", index_name="device_inbox_stream_id_user_id", table="device_inbox", columns=["stream_id", "user_id"], ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.DEVICE_INBOX_STREAM_ID, self._background_drop_index_device_inbox ) - @defer.inlineCallbacks - def _background_drop_index_device_inbox(self, progress, batch_size): + async def _background_drop_index_device_inbox(self, progress, batch_size): def reindex_txn(conn): txn = conn.cursor() txn.execute("DROP INDEX IF EXISTS device_inbox_stream_id") txn.close() - yield self.db.runWithConnection(reindex_txn) + await self.db_pool.runWithConnection(reindex_txn) - yield self.db.updates._end_background_update(self.DEVICE_INBOX_STREAM_ID) + await self.db_pool.updates._end_background_update(self.DEVICE_INBOX_STREAM_ID) return 1 @@ -308,8 +312,8 @@ class DeviceInboxBackgroundUpdateStore(SQLBaseStore): class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore): DEVICE_INBOX_STREAM_ID = "device_inbox_stream_drop" - def __init__(self, database: Database, db_conn, hs): - super(DeviceInboxStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # Map of (user_id, device_id) to the last stream_id that has been # deleted up to. This is so that we can no op deletions. @@ -321,21 +325,21 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) ) @trace - @defer.inlineCallbacks - def add_messages_to_device_inbox( - self, local_messages_by_user_then_device, remote_messages_by_destination - ): + async def add_messages_to_device_inbox( + self, + local_messages_by_user_then_device: dict, + remote_messages_by_destination: dict, + ) -> int: """Used to send messages from this server. Args: - sender_user_id(str): The ID of the user sending these messages. - local_messages_by_user_and_device(dict): + local_messages_by_user_and_device: Dictionary of user_id to device_id to message. - remote_messages_by_destination(dict): + remote_messages_by_destination: Dictionary of destination server_name to the EDU JSON to send. + Returns: - A deferred stream_id that resolves when the messages have been - inserted. + The new stream_id. """ def add_messages_txn(txn, now_ms, stream_id): @@ -354,13 +358,13 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) ) rows = [] for destination, edu in remote_messages_by_destination.items(): - edu_json = json.dumps(edu) + edu_json = json_encoder.encode(edu) rows.append((destination, stream_id, now_ms, edu_json)) txn.executemany(sql, rows) - with self._device_inbox_id_gen.get_next() as stream_id: + async with self._device_inbox_id_gen.get_next() as stream_id: now_ms = self.clock.time_msec() - yield self.db.runInteraction( + await self.db_pool.runInteraction( "add_messages_to_device_inbox", add_messages_txn, now_ms, stream_id ) for user_id in local_messages_by_user_then_device.keys(): @@ -372,15 +376,14 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) return self._device_inbox_id_gen.get_current_token() - @defer.inlineCallbacks - def add_messages_from_remote_to_device_inbox( - self, origin, message_id, local_messages_by_user_then_device - ): + async def add_messages_from_remote_to_device_inbox( + self, origin: str, message_id: str, local_messages_by_user_then_device: dict + ) -> int: def add_messages_txn(txn, now_ms, stream_id): # Check if we've already inserted a matching message_id for that # origin. This can happen if the origin doesn't receive our # acknowledgement from the first time we received the message. - already_inserted = self.db.simple_select_one_txn( + already_inserted = self.db_pool.simple_select_one_txn( txn, table="device_federation_inbox", keyvalues={"origin": origin, "message_id": message_id}, @@ -392,7 +395,7 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) # Add an entry for this message_id so that we know we've processed # it. - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="device_federation_inbox", values={ @@ -408,9 +411,9 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) txn, stream_id, local_messages_by_user_then_device ) - with self._device_inbox_id_gen.get_next() as stream_id: + async with self._device_inbox_id_gen.get_next() as stream_id: now_ms = self.clock.time_msec() - yield self.db.runInteraction( + await self.db_pool.runInteraction( "add_messages_from_remote_to_device_inbox", add_messages_txn, now_ms, @@ -432,7 +435,7 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) # Handle wildcard device_ids. sql = "SELECT device_id FROM devices WHERE user_id = ?" txn.execute(sql, (user_id,)) - message_json = json.dumps(messages_by_device["*"]) + message_json = json_encoder.encode(messages_by_device["*"]) for row in txn: # Add the message for all devices for this user on this # server. @@ -454,7 +457,7 @@ class DeviceInboxStore(DeviceInboxWorkerStore, DeviceInboxBackgroundUpdateStore) # Only insert into the local inbox if the device exists on # this server device = row[0] - message_json = json.dumps(messages_by_device[device]) + message_json = json_encoder.encode(messages_by_device[device]) messages_json_for_user[device] = message_json if messages_json_for_user: diff --git a/synapse/storage/data_stores/main/devices.py b/synapse/storage/databases/main/devices.py
index 45581a6500..fdf394c612 100644 --- a/synapse/storage/data_stores/main/devices.py +++ b/synapse/storage/databases/main/devices.py
@@ -14,12 +14,9 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. +import abc import logging -from typing import List, Optional, Set, Tuple - -from canonicaljson import json - -from twisted.internet import defer +from typing import Any, Dict, Iterable, List, Optional, Set, Tuple from synapse.api.errors import Codes, StoreError from synapse.logging.opentracing import ( @@ -31,17 +28,13 @@ from synapse.logging.opentracing import ( from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause from synapse.storage.database import ( - Database, + DatabasePool, LoggingTransaction, make_tuple_comparison_clause, ) -from synapse.types import Collection, get_verify_key_from_cross_signing_key -from synapse.util.caches.descriptors import ( - Cache, - cached, - cachedInlineCallbacks, - cachedList, -) +from synapse.types import Collection, JsonDict, get_verify_key_from_cross_signing_key +from synapse.util import json_encoder +from synapse.util.caches.descriptors import Cache, cached, cachedList from synapse.util.iterutils import batch_iter from synapse.util.stringutils import shortstr @@ -55,38 +48,36 @@ BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES = "remove_dup_outbound_pokes" class DeviceWorkerStore(SQLBaseStore): - def get_device(self, user_id, device_id): + async def get_device(self, user_id: str, device_id: str) -> Dict[str, Any]: """Retrieve a device. Only returns devices that are not marked as hidden. Args: - user_id (str): The ID of the user which owns the device - device_id (str): The ID of the device to retrieve + user_id: The ID of the user which owns the device + device_id: The ID of the device to retrieve Returns: - defer.Deferred for a dict containing the device information + A dict containing the device information Raises: StoreError: if the device is not found """ - return self.db.simple_select_one( + return await self.db_pool.simple_select_one( table="devices", keyvalues={"user_id": user_id, "device_id": device_id, "hidden": False}, retcols=("user_id", "device_id", "display_name"), desc="get_device", ) - @defer.inlineCallbacks - def get_devices_by_user(self, user_id): + async def get_devices_by_user(self, user_id: str) -> Dict[str, Dict[str, str]]: """Retrieve all of a user's registered devices. Only returns devices that are not marked as hidden. Args: - user_id (str): + user_id: Returns: - defer.Deferred: resolves to a dict from device_id to a dict - containing "device_id", "user_id" and "display_name" for each - device. + A mapping from device_id to a dict containing "device_id", "user_id" + and "display_name" for each device. """ - devices = yield self.db.simple_select_list( + devices = await self.db_pool.simple_select_list( table="devices", keyvalues={"user_id": user_id, "hidden": False}, retcols=("user_id", "device_id", "display_name"), @@ -96,21 +87,22 @@ class DeviceWorkerStore(SQLBaseStore): return {d["device_id"]: d for d in devices} @trace - @defer.inlineCallbacks - def get_device_updates_by_remote(self, destination, from_stream_id, limit): + async def get_device_updates_by_remote( + self, destination: str, from_stream_id: int, limit: int + ) -> Tuple[int, List[Tuple[str, dict]]]: """Get a stream of device updates to send to the given remote server. Args: - destination (str): The host the device updates are intended for - from_stream_id (int): The minimum stream_id to filter updates by, exclusive - limit (int): Maximum number of device updates to return + destination: The host the device updates are intended for + from_stream_id: The minimum stream_id to filter updates by, exclusive + limit: Maximum number of device updates to return + Returns: - Deferred[tuple[int, list[tuple[string,dict]]]]: - current stream id (ie, the stream id of the last update included in the - response), and the list of updates, where each update is a pair of EDU - type and EDU contents + A mapping from the current stream id (ie, the stream id of the last + update included in the response), and the list of updates, where + each update is a pair of EDU type and EDU contents. """ - now_stream_id = self._device_list_id_gen.get_current_token() + now_stream_id = self.get_device_stream_token() has_changed = self._device_list_federation_stream_cache.has_entity_changed( destination, int(from_stream_id) @@ -118,7 +110,7 @@ class DeviceWorkerStore(SQLBaseStore): if not has_changed: return now_stream_id, [] - updates = yield self.db.runInteraction( + updates = await self.db_pool.runInteraction( "get_device_updates_by_remote", self._get_device_updates_by_remote_txn, destination, @@ -137,7 +129,7 @@ class DeviceWorkerStore(SQLBaseStore): master_key_by_user = {} self_signing_key_by_user = {} for user in users: - cross_signing_key = yield self.get_e2e_cross_signing_key(user, "master") + cross_signing_key = await self.get_e2e_cross_signing_key(user, "master") if cross_signing_key: key_id, verify_key = get_verify_key_from_cross_signing_key( cross_signing_key @@ -150,7 +142,7 @@ class DeviceWorkerStore(SQLBaseStore): "device_id": verify_key.version, } - cross_signing_key = yield self.get_e2e_cross_signing_key( + cross_signing_key = await self.get_e2e_cross_signing_key( user, "self_signing" ) if cross_signing_key: @@ -201,7 +193,7 @@ class DeviceWorkerStore(SQLBaseStore): if update_stream_id > previous_update_stream_id: query_map[key] = (update_stream_id, update_context) - results = yield self._get_device_update_edus_by_remote( + results = await self._get_device_update_edus_by_remote( destination, from_stream_id, query_map ) @@ -214,16 +206,21 @@ class DeviceWorkerStore(SQLBaseStore): return now_stream_id, results def _get_device_updates_by_remote_txn( - self, txn, destination, from_stream_id, now_stream_id, limit + self, + txn: LoggingTransaction, + destination: str, + from_stream_id: int, + now_stream_id: int, + limit: int, ): """Return device update information for a given remote destination Args: - txn (LoggingTransaction): The transaction to execute - destination (str): The host the device updates are intended for - from_stream_id (int): The minimum stream_id to filter updates by, exclusive - now_stream_id (int): The maximum stream_id to filter updates by, inclusive - limit (int): Maximum number of device updates to return + txn: The transaction to execute + destination: The host the device updates are intended for + from_stream_id: The minimum stream_id to filter updates by, exclusive + now_stream_id: The maximum stream_id to filter updates by, inclusive + limit: Maximum number of device updates to return Returns: List: List of device updates @@ -239,25 +236,26 @@ class DeviceWorkerStore(SQLBaseStore): return list(txn) - @defer.inlineCallbacks - def _get_device_update_edus_by_remote(self, destination, from_stream_id, query_map): + async def _get_device_update_edus_by_remote( + self, + destination: str, + from_stream_id: int, + query_map: Dict[Tuple[str, str], Tuple[int, Optional[str]]], + ) -> List[Tuple[str, dict]]: """Returns a list of device update EDUs as well as E2EE keys Args: - destination (str): The host the device updates are intended for - from_stream_id (int): The minimum stream_id to filter updates by, exclusive + destination: The host the device updates are intended for + from_stream_id: The minimum stream_id to filter updates by, exclusive query_map (Dict[(str, str): (int, str|None)]): Dictionary mapping - user_id/device_id to update stream_id and the relevent json-encoded + user_id/device_id to update stream_id and the relevant json-encoded opentracing context Returns: - List[Dict]: List of objects representing an device update EDU - + List of objects representing an device update EDU """ devices = ( - yield self.db.runInteraction( - "_get_e2e_device_keys_txn", - self._get_e2e_device_keys_txn, + await self.get_e2e_device_keys_and_signatures( query_map.keys(), include_all_devices=True, include_deleted_devices=True, @@ -270,7 +268,7 @@ class DeviceWorkerStore(SQLBaseStore): for user_id, user_devices in devices.items(): # The prev_id for the first row is always the last row before # `from_stream_id` - prev_id = yield self._get_last_device_update_for_remote_user( + prev_id = await self._get_last_device_update_for_remote_user( destination, user_id, from_stream_id ) @@ -293,17 +291,11 @@ class DeviceWorkerStore(SQLBaseStore): prev_id = stream_id if device is not None: - key_json = device.get("key_json", None) - if key_json: - result["keys"] = db_to_json(key_json) + keys = device.keys + if keys: + result["keys"] = keys - if "signatures" in device: - for sig_user_id, sigs in device["signatures"].items(): - result["keys"].setdefault("signatures", {}).setdefault( - sig_user_id, {} - ).update(sigs) - - device_display_name = device.get("device_display_name", None) + device_display_name = device.display_name if device_display_name: result["device_display_name"] = device_display_name else: @@ -313,9 +305,9 @@ class DeviceWorkerStore(SQLBaseStore): return results - def _get_last_device_update_for_remote_user( - self, destination, user_id, from_stream_id - ): + async def _get_last_device_update_for_remote_user( + self, destination: str, user_id: str, from_stream_id: int + ) -> int: def f(txn): prev_sent_id_sql = """ SELECT coalesce(max(stream_id), 0) as stream_id @@ -326,19 +318,25 @@ class DeviceWorkerStore(SQLBaseStore): rows = txn.fetchall() return rows[0][0] - return self.db.runInteraction("get_last_device_update_for_remote_user", f) + return await self.db_pool.runInteraction( + "get_last_device_update_for_remote_user", f + ) - def mark_as_sent_devices_by_remote(self, destination, stream_id): + async def mark_as_sent_devices_by_remote( + self, destination: str, stream_id: int + ) -> None: """Mark that updates have successfully been sent to the destination. """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "mark_as_sent_devices_by_remote", self._mark_as_sent_devices_by_remote_txn, destination, stream_id, ) - def _mark_as_sent_devices_by_remote_txn(self, txn, destination, stream_id): + def _mark_as_sent_devices_by_remote_txn( + self, txn: LoggingTransaction, destination: str, stream_id: int + ) -> None: # We update the device_lists_outbound_last_success with the successfully # poked users. sql = """ @@ -350,7 +348,7 @@ class DeviceWorkerStore(SQLBaseStore): txn.execute(sql, (destination, stream_id)) rows = txn.fetchall() - self.db.simple_upsert_many_txn( + self.db_pool.simple_upsert_many_txn( txn=txn, table="device_lists_outbound_last_success", key_names=("destination", "user_id"), @@ -366,17 +364,21 @@ class DeviceWorkerStore(SQLBaseStore): """ txn.execute(sql, (destination, stream_id)) - @defer.inlineCallbacks - def add_user_signature_change_to_streams(self, from_user_id, user_ids): + async def add_user_signature_change_to_streams( + self, from_user_id: str, user_ids: List[str] + ) -> int: """Persist that a user has made new signatures Args: - from_user_id (str): the user who made the signatures - user_ids (list[str]): the users who were signed + from_user_id: the user who made the signatures + user_ids: the users who were signed + + Returns: + THe new stream ID. """ - with self._device_list_id_gen.get_next() as stream_id: - yield self.db.runInteraction( + async with self._device_list_id_gen.get_next() as stream_id: + await self.db_pool.runInteraction( "add_user_sig_change_to_streams", self._add_user_signature_change_txn, from_user_id, @@ -385,45 +387,54 @@ class DeviceWorkerStore(SQLBaseStore): ) return stream_id - def _add_user_signature_change_txn(self, txn, from_user_id, user_ids, stream_id): + def _add_user_signature_change_txn( + self, + txn: LoggingTransaction, + from_user_id: str, + user_ids: List[str], + stream_id: int, + ) -> None: txn.call_after( self._user_signature_stream_cache.entity_has_changed, from_user_id, stream_id, ) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "user_signature_stream", values={ "stream_id": stream_id, "from_user_id": from_user_id, - "user_ids": json.dumps(user_ids), + "user_ids": json_encoder.encode(user_ids), }, ) - def get_device_stream_token(self): - return self._device_list_id_gen.get_current_token() + @abc.abstractmethod + def get_device_stream_token(self) -> int: + """Get the current stream id from the _device_list_id_gen""" + ... @trace - @defer.inlineCallbacks - def get_user_devices_from_cache(self, query_list): + async def get_user_devices_from_cache( + self, query_list: List[Tuple[str, str]] + ) -> Tuple[Set[str], Dict[str, Dict[str, JsonDict]]]: """Get the devices (and keys if any) for remote users from the cache. Args: - query_list(list): List of (user_id, device_ids), if device_ids is + query_list: List of (user_id, device_ids), if device_ids is falsey then return all device ids for that user. Returns: - (user_ids_not_in_cache, results_map), where user_ids_not_in_cache is - a set of user_ids and results_map is a mapping of - user_id -> device_id -> device_info + A tuple of (user_ids_not_in_cache, results_map), where + user_ids_not_in_cache is a set of user_ids and results_map is a + mapping of user_id -> device_id -> device_info. """ user_ids = {user_id for user_id, _ in query_list} - user_map = yield self.get_device_list_last_stream_id_for_remotes(list(user_ids)) + user_map = await self.get_device_list_last_stream_id_for_remotes(list(user_ids)) # We go and check if any of the users need to have their device lists # resynced. If they do then we remove them from the cached list. - users_needing_resync = yield self.get_user_ids_requiring_device_list_resync( + users_needing_resync = await self.get_user_ids_requiring_device_list_resync( user_ids ) user_ids_in_cache = { @@ -437,19 +448,19 @@ class DeviceWorkerStore(SQLBaseStore): continue if device_id: - device = yield self._get_cached_user_device(user_id, device_id) + device = await self._get_cached_user_device(user_id, device_id) results.setdefault(user_id, {})[device_id] = device else: - results[user_id] = yield self.get_cached_devices_for_user(user_id) + results[user_id] = await self.get_cached_devices_for_user(user_id) set_tag("in_cache", results) set_tag("not_in_cache", user_ids_not_in_cache) return user_ids_not_in_cache, results - @cachedInlineCallbacks(num_args=2, tree=True) - def _get_cached_user_device(self, user_id, device_id): - content = yield self.db.simple_select_one_onecol( + @cached(num_args=2, tree=True) + async def _get_cached_user_device(self, user_id: str, device_id: str) -> JsonDict: + content = await self.db_pool.simple_select_one_onecol( table="device_lists_remote_cache", keyvalues={"user_id": user_id, "device_id": device_id}, retcol="content", @@ -457,9 +468,9 @@ class DeviceWorkerStore(SQLBaseStore): ) return db_to_json(content) - @cachedInlineCallbacks() - def get_cached_devices_for_user(self, user_id): - devices = yield self.db.simple_select_list( + @cached() + async def get_cached_devices_for_user(self, user_id: str) -> Dict[str, JsonDict]: + devices = await self.db_pool.simple_select_list( table="device_lists_remote_cache", keyvalues={"user_id": user_id}, retcols=("device_id", "content"), @@ -469,64 +480,19 @@ class DeviceWorkerStore(SQLBaseStore): device["device_id"]: db_to_json(device["content"]) for device in devices } - def get_devices_with_keys_by_user(self, user_id): - """Get all devices (with any device keys) for a user - - Returns: - (stream_id, devices) - """ - return self.db.runInteraction( - "get_devices_with_keys_by_user", - self._get_devices_with_keys_by_user_txn, - user_id, - ) - - def _get_devices_with_keys_by_user_txn(self, txn, user_id): - now_stream_id = self._device_list_id_gen.get_current_token() - - devices = self._get_e2e_device_keys_txn( - txn, [(user_id, None)], include_all_devices=True - ) - - if devices: - user_devices = devices[user_id] - results = [] - for device_id, device in user_devices.items(): - result = {"device_id": device_id} - - key_json = device.get("key_json", None) - if key_json: - result["keys"] = db_to_json(key_json) - - if "signatures" in device: - for sig_user_id, sigs in device["signatures"].items(): - result["keys"].setdefault("signatures", {}).setdefault( - sig_user_id, {} - ).update(sigs) - - device_display_name = device.get("device_display_name", None) - if device_display_name: - result["device_display_name"] = device_display_name - - results.append(result) - - return now_stream_id, results - - return now_stream_id, [] - - def get_users_whose_devices_changed(self, from_key, user_ids): + async def get_users_whose_devices_changed( + self, from_key: int, user_ids: Iterable[str] + ) -> Set[str]: """Get set of users whose devices have changed since `from_key` that are in the given list of user_ids. Args: - from_key (str): The device lists stream token - user_ids (Iterable[str]) + from_key: The device lists stream token + user_ids: The user IDs to query for devices. Returns: - Deferred[set[str]]: The set of user_ids whose devices have changed - since `from_key` + The set of user_ids whose devices have changed since `from_key` """ - from_key = int(from_key) # Get set of users who *may* have changed. Users not in the returned # list have definitely not changed. @@ -535,7 +501,7 @@ class DeviceWorkerStore(SQLBaseStore): ) if not to_check: - return defer.succeed(set()) + return set() def _get_users_whose_devices_changed_txn(txn): changes = set() @@ -555,26 +521,30 @@ class DeviceWorkerStore(SQLBaseStore): return changes - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_whose_devices_changed", _get_users_whose_devices_changed_txn ) - @defer.inlineCallbacks - def get_users_whose_signatures_changed(self, user_id, from_key): + async def get_users_whose_signatures_changed( + self, user_id: str, from_key: int + ) -> Set[str]: """Get the users who have new cross-signing signatures made by `user_id` since `from_key`. Args: - user_id (str): the user who made the signatures - from_key (str): The device lists stream token + user_id: the user who made the signatures + from_key: The device lists stream token + + Returns: + A set of user IDs with updated signatures. """ - from_key = int(from_key) + if self._user_signature_stream_cache.has_entity_changed(user_id, from_key): sql = """ SELECT DISTINCT user_ids FROM user_signature_stream WHERE from_user_id = ? AND stream_id > ? """ - rows = yield self.db.execute( + rows = await self.db_pool.execute( "get_users_whose_signatures_changed", None, sql, user_id, from_key ) return {user for row in rows for user in db_to_json(row[0])} @@ -600,7 +570,7 @@ class DeviceWorkerStore(SQLBaseStore): between the requested tokens due to the limit. The token returned can be used in a subsequent call to this - function to get further updatees. + function to get further updates. The updates are a list of 2-tuples of stream ID and the row data """ @@ -631,17 +601,19 @@ class DeviceWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_device_list_changes_for_remotes", _get_all_device_list_changes_for_remotes, ) @cached(max_entries=10000) - def get_device_list_last_stream_id_for_remote(self, user_id): + async def get_device_list_last_stream_id_for_remote( + self, user_id: str + ) -> Optional[Any]: """Get the last stream_id we got for a user. May be None if we haven't got any information for them. """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="device_lists_remote_extremeties", keyvalues={"user_id": user_id}, retcol="stream_id", @@ -652,10 +624,9 @@ class DeviceWorkerStore(SQLBaseStore): @cachedList( cached_method_name="get_device_list_last_stream_id_for_remote", list_name="user_ids", - inlineCallbacks=True, ) - def get_device_list_last_stream_id_for_remotes(self, user_ids): - rows = yield self.db.simple_select_many_batch( + async def get_device_list_last_stream_id_for_remotes(self, user_ids: str): + rows = await self.db_pool.simple_select_many_batch( table="device_lists_remote_extremeties", column="user_id", iterable=user_ids, @@ -668,8 +639,7 @@ class DeviceWorkerStore(SQLBaseStore): return results - @defer.inlineCallbacks - def get_user_ids_requiring_device_list_resync( + async def get_user_ids_requiring_device_list_resync( self, user_ids: Optional[Collection[str]] = None, ) -> Set[str]: """Given a list of remote users return the list of users that we @@ -680,7 +650,7 @@ class DeviceWorkerStore(SQLBaseStore): The IDs of users whose device lists need resync. """ if user_ids: - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="device_lists_remote_resync", column="user_id", iterable=user_ids, @@ -688,7 +658,7 @@ class DeviceWorkerStore(SQLBaseStore): desc="get_user_ids_requiring_device_list_resync_with_iterable", ) else: - rows = yield self.db.simple_select_list( + rows = await self.db_pool.simple_select_list( table="device_lists_remote_resync", keyvalues=None, retcols=("user_id",), @@ -697,11 +667,11 @@ class DeviceWorkerStore(SQLBaseStore): return {row["user_id"] for row in rows} - def mark_remote_user_device_cache_as_stale(self, user_id: str): + async def mark_remote_user_device_cache_as_stale(self, user_id: str) -> None: """Records that the server has reason to believe the cache of the devices for the remote users is out of date. """ - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="device_lists_remote_resync", keyvalues={"user_id": user_id}, values={}, @@ -709,12 +679,12 @@ class DeviceWorkerStore(SQLBaseStore): desc="make_remote_user_device_cache_as_stale", ) - def mark_remote_user_device_list_as_unsubscribed(self, user_id): + async def mark_remote_user_device_list_as_unsubscribed(self, user_id: str) -> None: """Mark that we no longer track device lists for remote user. """ def _mark_remote_user_device_list_as_unsubscribed_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="device_lists_remote_extremeties", keyvalues={"user_id": user_id}, @@ -723,17 +693,17 @@ class DeviceWorkerStore(SQLBaseStore): txn, self.get_device_list_last_stream_id_for_remote, (user_id,) ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "mark_remote_user_device_list_as_unsubscribed", _mark_remote_user_device_list_as_unsubscribed_txn, ) class DeviceBackgroundUpdateStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(DeviceBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "device_lists_stream_idx", index_name="device_lists_stream_user_id", table="device_lists_stream", @@ -741,7 +711,7 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): ) # create a unique index on device_lists_remote_cache - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "device_lists_remote_cache_unique_idx", index_name="device_lists_remote_cache_unique_id", table="device_lists_remote_cache", @@ -750,7 +720,7 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): ) # And one on device_lists_remote_extremeties - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "device_lists_remote_extremeties_unique_idx", index_name="device_lists_remote_extremeties_unique_idx", table="device_lists_remote_extremeties", @@ -759,35 +729,34 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): ) # once they complete, we can remove the old non-unique indexes. - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( DROP_DEVICE_LIST_STREAMS_NON_UNIQUE_INDEXES, self._drop_device_list_streams_non_unique_indexes, ) # clear out duplicate device list outbound pokes - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES, self._remove_duplicate_outbound_pokes, ) # a pair of background updates that were added during the 1.14 release cycle, # but replaced with 58/06dlols_unique_idx.py - self.db.updates.register_noop_background_update( + self.db_pool.updates.register_noop_background_update( "device_lists_outbound_last_success_unique_idx", ) - self.db.updates.register_noop_background_update( + self.db_pool.updates.register_noop_background_update( "drop_device_lists_outbound_last_success_non_unique_idx", ) - @defer.inlineCallbacks - def _drop_device_list_streams_non_unique_indexes(self, progress, batch_size): + async def _drop_device_list_streams_non_unique_indexes(self, progress, batch_size): def f(conn): txn = conn.cursor() txn.execute("DROP INDEX IF EXISTS device_lists_remote_cache_id") txn.execute("DROP INDEX IF EXISTS device_lists_remote_extremeties_id") txn.close() - yield self.db.runWithConnection(f) - yield self.db.updates._end_background_update( + await self.db_pool.runWithConnection(f) + await self.db_pool.updates._end_background_update( DROP_DEVICE_LIST_STREAMS_NON_UNIQUE_INDEXES ) return 1 @@ -807,7 +776,7 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): def _txn(txn): clause, args = make_tuple_comparison_clause( - self.db.engine, [(x, last_row[x]) for x in KEY_COLS] + self.db_pool.engine, [(x, last_row[x]) for x in KEY_COLS] ) sql = """ SELECT stream_id, destination, user_id, device_id, MAX(ts) AS ts @@ -823,30 +792,32 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): ",".join(KEY_COLS), # ORDER BY ) txn.execute(sql, args + [batch_size]) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) row = None for row in rows: - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, "device_lists_outbound_pokes", {x: row[x] for x in KEY_COLS}, ) row["sent"] = False - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "device_lists_outbound_pokes", row, ) if row: - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES, {"last_row": row}, ) return len(rows) - rows = await self.db.runInteraction(BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES, _txn) + rows = await self.db_pool.runInteraction( + BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES, _txn + ) if not rows: - await self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( BG_UPDATE_REMOVE_DUP_OUTBOUND_POKES ) @@ -854,8 +825,8 @@ class DeviceBackgroundUpdateStore(SQLBaseStore): class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): - def __init__(self, database: Database, db_conn, hs): - super(DeviceStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # Map of (user_id, device_id) -> bool. If there is an entry that implies # the device exists. @@ -865,18 +836,20 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): self._clock.looping_call(self._prune_old_outbound_device_pokes, 60 * 60 * 1000) - @defer.inlineCallbacks - def store_device(self, user_id, device_id, initial_device_display_name): + async def store_device( + self, user_id: str, device_id: str, initial_device_display_name: str + ) -> bool: """Ensure the given device is known; add it to the store if not Args: - user_id (str): id of user associated with the device - device_id (str): id of device - initial_device_display_name (str): initial displayname of the - device. Ignored if device exists. + user_id: id of user associated with the device + device_id: id of device + initial_device_display_name: initial displayname of the device. + Ignored if device exists. + Returns: - defer.Deferred: boolean whether the device was inserted or an - existing device existed with that ID. + Whether the device was inserted or an existing device existed with that ID. + Raises: StoreError: if the device is already in use """ @@ -885,7 +858,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): return False try: - inserted = yield self.db.simple_insert( + inserted = await self.db_pool.simple_insert( "devices", values={ "user_id": user_id, @@ -899,7 +872,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): if not inserted: # if the device already exists, check if it's a real device, or # if the device ID is reserved by something else - hidden = yield self.db.simple_select_one_onecol( + hidden = await self.db_pool.simple_select_one_onecol( "devices", keyvalues={"user_id": user_id, "device_id": device_id}, retcol="hidden", @@ -924,17 +897,14 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): ) raise StoreError(500, "Problem storing device.") - @defer.inlineCallbacks - def delete_device(self, user_id, device_id): + async def delete_device(self, user_id: str, device_id: str) -> None: """Delete a device. Args: - user_id (str): The ID of the user which owns the device - device_id (str): The ID of the device to delete - Returns: - defer.Deferred + user_id: The ID of the user which owns the device + device_id: The ID of the device to delete """ - yield self.db.simple_delete_one( + await self.db_pool.simple_delete_one( table="devices", keyvalues={"user_id": user_id, "device_id": device_id, "hidden": False}, desc="delete_device", @@ -942,17 +912,14 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): self.device_id_exists_cache.invalidate((user_id, device_id)) - @defer.inlineCallbacks - def delete_devices(self, user_id, device_ids): + async def delete_devices(self, user_id: str, device_ids: List[str]) -> None: """Deletes several devices. Args: - user_id (str): The ID of the user which owns the devices - device_ids (list): The IDs of the devices to delete - Returns: - defer.Deferred + user_id: The ID of the user which owns the devices + device_ids: The IDs of the devices to delete """ - yield self.db.simple_delete_many( + await self.db_pool.simple_delete_many( table="devices", column="device_id", iterable=device_ids, @@ -962,50 +929,46 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): for device_id in device_ids: self.device_id_exists_cache.invalidate((user_id, device_id)) - def update_device(self, user_id, device_id, new_display_name=None): + async def update_device( + self, user_id: str, device_id: str, new_display_name: Optional[str] = None + ) -> None: """Update a device. Only updates the device if it is not marked as hidden. Args: - user_id (str): The ID of the user which owns the device - device_id (str): The ID of the device to update - new_display_name (str|None): new displayname for device; None - to leave unchanged + user_id: The ID of the user which owns the device + device_id: The ID of the device to update + new_display_name: new displayname for device; None to leave unchanged Raises: StoreError: if the device is not found - Returns: - defer.Deferred """ updates = {} if new_display_name is not None: updates["display_name"] = new_display_name if not updates: - return defer.succeed(None) - return self.db.simple_update_one( + return None + await self.db_pool.simple_update_one( table="devices", keyvalues={"user_id": user_id, "device_id": device_id, "hidden": False}, updatevalues=updates, desc="update_device", ) - def update_remote_device_list_cache_entry( - self, user_id, device_id, content, stream_id - ): + async def update_remote_device_list_cache_entry( + self, user_id: str, device_id: str, content: JsonDict, stream_id: int + ) -> None: """Updates a single device in the cache of a remote user's devicelist. Note: assumes that we are the only thread that can be updating this user's device list. Args: - user_id (str): User to update device list for - device_id (str): ID of decivice being updated - content (dict): new data on this device - stream_id (int): the version of the device list - - Returns: - Deferred[None] + user_id: User to update device list for + device_id: ID of decivice being updated + content: new data on this device + stream_id: the version of the device list """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "update_remote_device_list_cache_entry", self._update_remote_device_list_cache_entry_txn, user_id, @@ -1015,10 +978,15 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): ) def _update_remote_device_list_cache_entry_txn( - self, txn, user_id, device_id, content, stream_id - ): + self, + txn: LoggingTransaction, + user_id: str, + device_id: str, + content: JsonDict, + stream_id: int, + ) -> None: if content.get("deleted"): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="device_lists_remote_cache", keyvalues={"user_id": user_id, "device_id": device_id}, @@ -1026,11 +994,11 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): txn.call_after(self.device_id_exists_cache.invalidate, (user_id, device_id)) else: - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="device_lists_remote_cache", keyvalues={"user_id": user_id, "device_id": device_id}, - values={"content": json.dumps(content)}, + values={"content": json_encoder.encode(content)}, # we don't need to lock, because we assume we are the only thread # updating this user's devices. lock=False, @@ -1042,7 +1010,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): self.get_device_list_last_stream_id_for_remote.invalidate, (user_id,) ) - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="device_lists_remote_extremeties", keyvalues={"user_id": user_id}, @@ -1052,21 +1020,20 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): lock=False, ) - def update_remote_device_list_cache(self, user_id, devices, stream_id): + async def update_remote_device_list_cache( + self, user_id: str, devices: List[dict], stream_id: int + ) -> None: """Replace the entire cache of the remote user's devices. Note: assumes that we are the only thread that can be updating this user's device list. Args: - user_id (str): User to update device list for - devices (list[dict]): list of device objects supplied over federation - stream_id (int): the version of the device list - - Returns: - Deferred[None] + user_id: User to update device list for + devices: list of device objects supplied over federation + stream_id: the version of the device list """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "update_remote_device_list_cache", self._update_remote_device_list_cache_txn, user_id, @@ -1074,19 +1041,21 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): stream_id, ) - def _update_remote_device_list_cache_txn(self, txn, user_id, devices, stream_id): - self.db.simple_delete_txn( + def _update_remote_device_list_cache_txn( + self, txn: LoggingTransaction, user_id: str, devices: List[dict], stream_id: int + ) -> None: + self.db_pool.simple_delete_txn( txn, table="device_lists_remote_cache", keyvalues={"user_id": user_id} ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="device_lists_remote_cache", values=[ { "user_id": user_id, "device_id": content["device_id"], - "content": json.dumps(content), + "content": json_encoder.encode(content), } for content in devices ], @@ -1098,7 +1067,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): self.get_device_list_last_stream_id_for_remote.invalidate, (user_id,) ) - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="device_lists_remote_extremeties", keyvalues={"user_id": user_id}, @@ -1111,20 +1080,23 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): # If we're replacing the remote user's device list cache presumably # we've done a full resync, so we remove the entry that says we need # to resync - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="device_lists_remote_resync", keyvalues={"user_id": user_id}, ) - @defer.inlineCallbacks - def add_device_change_to_streams(self, user_id, device_ids, hosts): + async def add_device_change_to_streams( + self, user_id: str, device_ids: Collection[str], hosts: List[str] + ): """Persist that a user's devices have been updated, and which hosts (if any) should be poked. """ if not device_ids: return - with self._device_list_id_gen.get_next_mult(len(device_ids)) as stream_ids: - yield self.db.runInteraction( + async with self._device_list_id_gen.get_next_mult( + len(device_ids) + ) as stream_ids: + await self.db_pool.runInteraction( "add_device_change_to_stream", self._add_device_change_to_stream_txn, user_id, @@ -1136,10 +1108,10 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): return stream_ids[-1] context = get_active_span_text_map() - with self._device_list_id_gen.get_next_mult( + async with self._device_list_id_gen.get_next_mult( len(hosts) * len(device_ids) ) as stream_ids: - yield self.db.runInteraction( + await self.db_pool.runInteraction( "add_device_outbound_poke_to_stream", self._add_device_outbound_poke_to_stream_txn, user_id, @@ -1174,7 +1146,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): [(user_id, device_id, min_stream_id) for device_id in device_ids], ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="device_lists_stream", values=[ @@ -1184,7 +1156,13 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): ) def _add_device_outbound_poke_to_stream_txn( - self, txn, user_id, device_ids, hosts, stream_ids, context, + self, + txn: LoggingTransaction, + user_id: str, + device_ids: Collection[str], + hosts: List[str], + stream_ids: List[str], + context: Dict[str, str], ): for host in hosts: txn.call_after( @@ -1196,7 +1174,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): now = self._clock.time_msec() next_stream_id = iter(stream_ids) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="device_lists_outbound_pokes", values=[ @@ -1207,7 +1185,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): "device_id": device_id, "sent": False, "ts": now, - "opentracing_context": json.dumps(context) + "opentracing_context": json_encoder.encode(context) if whitelisted_homeserver(destination) else "{}", } @@ -1216,7 +1194,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): ], ) - def _prune_old_outbound_device_pokes(self, prune_age=24 * 60 * 60 * 1000): + def _prune_old_outbound_device_pokes(self, prune_age: int = 24 * 60 * 60 * 1000): """Delete old entries out of the device_lists_outbound_pokes to ensure that we don't fill up due to dead servers. @@ -1303,7 +1281,7 @@ class DeviceStore(DeviceWorkerStore, DeviceBackgroundUpdateStore): return run_as_background_process( "prune_old_outbound_device_pokes", - self.db.runInteraction, + self.db_pool.runInteraction, "_prune_old_outbound_device_pokes", _prune_txn, ) diff --git a/synapse/storage/data_stores/main/directory.py b/synapse/storage/databases/main/directory.py
index e1d1bc3e05..e5060d4c46 100644 --- a/synapse/storage/data_stores/main/directory.py +++ b/synapse/storage/databases/main/directory.py
@@ -14,30 +14,29 @@ # limitations under the License. from collections import namedtuple -from typing import Optional - -from twisted.internet import defer +from typing import Iterable, List, Optional from synapse.api.errors import SynapseError from synapse.storage._base import SQLBaseStore +from synapse.types import RoomAlias from synapse.util.caches.descriptors import cached RoomAliasMapping = namedtuple("RoomAliasMapping", ("room_id", "room_alias", "servers")) class DirectoryWorkerStore(SQLBaseStore): - @defer.inlineCallbacks - def get_association_from_room_alias(self, room_alias): - """ Get's the room_id and server list for a given room_alias + async def get_association_from_room_alias( + self, room_alias: RoomAlias + ) -> Optional[RoomAliasMapping]: + """Gets the room_id and server list for a given room_alias Args: - room_alias (RoomAlias) + room_alias: The alias to translate to an ID. Returns: - Deferred: results in namedtuple with keys "room_id" and - "servers" or None if no association can be found + The room alias mapping or None if no association can be found. """ - room_id = yield self.db.simple_select_one_onecol( + room_id = await self.db_pool.simple_select_one_onecol( "room_aliases", {"room_alias": room_alias.to_string()}, "room_id", @@ -48,7 +47,7 @@ class DirectoryWorkerStore(SQLBaseStore): if not room_id: return None - servers = yield self.db.simple_select_onecol( + servers = await self.db_pool.simple_select_onecol( "room_alias_servers", {"room_alias": room_alias.to_string()}, "server", @@ -60,8 +59,8 @@ class DirectoryWorkerStore(SQLBaseStore): return RoomAliasMapping(room_id, room_alias.to_string(), servers) - def get_room_alias_creator(self, room_alias): - return self.db.simple_select_one_onecol( + async def get_room_alias_creator(self, room_alias: str) -> str: + return await self.db_pool.simple_select_one_onecol( table="room_aliases", keyvalues={"room_alias": room_alias}, retcol="creator", @@ -69,8 +68,8 @@ class DirectoryWorkerStore(SQLBaseStore): ) @cached(max_entries=5000) - def get_aliases_for_room(self, room_id): - return self.db.simple_select_onecol( + async def get_aliases_for_room(self, room_id: str) -> List[str]: + return await self.db_pool.simple_select_onecol( "room_aliases", {"room_id": room_id}, "room_alias", @@ -79,22 +78,24 @@ class DirectoryWorkerStore(SQLBaseStore): class DirectoryStore(DirectoryWorkerStore): - @defer.inlineCallbacks - def create_room_alias_association(self, room_alias, room_id, servers, creator=None): + async def create_room_alias_association( + self, + room_alias: RoomAlias, + room_id: str, + servers: Iterable[str], + creator: Optional[str] = None, + ) -> None: """ Creates an association between a room alias and room_id/servers Args: - room_alias (RoomAlias) - room_id (str) - servers (list) - creator (str): Optional user_id of creator. - - Returns: - Deferred + room_alias: The alias to create. + room_id: The target of the alias. + servers: A list of servers through which it may be possible to join the room + creator: Optional user_id of creator. """ def alias_txn(txn): - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "room_aliases", { @@ -104,7 +105,7 @@ class DirectoryStore(DirectoryWorkerStore): }, ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="room_alias_servers", values=[ @@ -118,24 +119,22 @@ class DirectoryStore(DirectoryWorkerStore): ) try: - ret = yield self.db.runInteraction( + await self.db_pool.runInteraction( "create_room_alias_association", alias_txn ) except self.database_engine.module.IntegrityError: raise SynapseError( 409, "Room alias %s already exists" % room_alias.to_string() ) - return ret - @defer.inlineCallbacks - def delete_room_alias(self, room_alias): - room_id = yield self.db.runInteraction( + async def delete_room_alias(self, room_alias: RoomAlias) -> str: + room_id = await self.db_pool.runInteraction( "delete_room_alias", self._delete_room_alias_txn, room_alias ) return room_id - def _delete_room_alias_txn(self, txn, room_alias): + def _delete_room_alias_txn(self, txn, room_alias: RoomAlias) -> str: txn.execute( "SELECT room_id FROM room_aliases WHERE room_alias = ?", (room_alias.to_string(),), @@ -160,9 +159,9 @@ class DirectoryStore(DirectoryWorkerStore): return room_id - def update_aliases_for_room( + async def update_aliases_for_room( self, old_room_id: str, new_room_id: str, creator: Optional[str] = None, - ): + ) -> None: """Repoint all of the aliases for a given room, to a different room. Args: @@ -190,6 +189,6 @@ class DirectoryStore(DirectoryWorkerStore): txn, self.get_aliases_for_room, (new_room_id,) ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "_update_aliases_for_room_txn", _update_aliases_for_room_txn ) diff --git a/synapse/storage/data_stores/main/e2e_room_keys.py b/synapse/storage/databases/main/e2e_room_keys.py
index 615364f018..12cecceec2 100644 --- a/synapse/storage/data_stores/main/e2e_room_keys.py +++ b/synapse/storage/databases/main/e2e_room_keys.py
@@ -14,18 +14,18 @@ # See the License for the specific language governing permissions and # limitations under the License. -from canonicaljson import json - -from twisted.internet import defer +from typing import Optional from synapse.api.errors import StoreError from synapse.logging.opentracing import log_kv, trace from synapse.storage._base import SQLBaseStore, db_to_json +from synapse.util import json_encoder class EndToEndRoomKeyStore(SQLBaseStore): - @defer.inlineCallbacks - def update_e2e_room_key(self, user_id, version, room_id, session_id, room_key): + async def update_e2e_room_key( + self, user_id, version, room_id, session_id, room_key + ): """Replaces the encrypted E2E room key for a given session in a given backup Args: @@ -38,7 +38,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): StoreError """ - yield self.db.simple_update_one( + await self.db_pool.simple_update_one( table="e2e_room_keys", keyvalues={ "user_id": user_id, @@ -50,13 +50,12 @@ class EndToEndRoomKeyStore(SQLBaseStore): "first_message_index": room_key["first_message_index"], "forwarded_count": room_key["forwarded_count"], "is_verified": room_key["is_verified"], - "session_data": json.dumps(room_key["session_data"]), + "session_data": json_encoder.encode(room_key["session_data"]), }, desc="update_e2e_room_key", ) - @defer.inlineCallbacks - def add_e2e_room_keys(self, user_id, version, room_keys): + async def add_e2e_room_keys(self, user_id, version, room_keys): """Bulk add room keys to a given backup. Args: @@ -77,7 +76,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): "first_message_index": room_key["first_message_index"], "forwarded_count": room_key["forwarded_count"], "is_verified": room_key["is_verified"], - "session_data": json.dumps(room_key["session_data"]), + "session_data": json_encoder.encode(room_key["session_data"]), } ) log_kv( @@ -89,13 +88,12 @@ class EndToEndRoomKeyStore(SQLBaseStore): } ) - yield self.db.simple_insert_many( + await self.db_pool.simple_insert_many( table="e2e_room_keys", values=values, desc="add_e2e_room_keys" ) @trace - @defer.inlineCallbacks - def get_e2e_room_keys(self, user_id, version, room_id=None, session_id=None): + async def get_e2e_room_keys(self, user_id, version, room_id=None, session_id=None): """Bulk get the E2E room keys for a given backup, optionally filtered to a given room, or a given session. @@ -110,7 +108,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): the backup (or for the specified room) Returns: - A deferred list of dicts giving the session_data and message metadata for + A list of dicts giving the session_data and message metadata for these room keys. """ @@ -125,7 +123,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): if session_id: keyvalues["session_id"] = session_id - rows = yield self.db.simple_select_list( + rows = await self.db_pool.simple_select_list( table="e2e_room_keys", keyvalues=keyvalues, retcols=( @@ -153,7 +151,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): return sessions - def get_e2e_room_keys_multi(self, user_id, version, room_keys): + async def get_e2e_room_keys_multi(self, user_id, version, room_keys): """Get multiple room keys at a time. The difference between this function and get_e2e_room_keys is that this function can be used to retrieve multiple specific keys at a time, whereas get_e2e_room_keys is used for @@ -168,10 +166,10 @@ class EndToEndRoomKeyStore(SQLBaseStore): that we want to query Returns: - Deferred[dict[str, dict[str, dict]]]: a map of room IDs to session IDs to room key + dict[str, dict[str, dict]]: a map of room IDs to session IDs to room key """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_e2e_room_keys_multi", self._get_e2e_room_keys_multi_txn, user_id, @@ -227,15 +225,15 @@ class EndToEndRoomKeyStore(SQLBaseStore): return ret - def count_e2e_room_keys(self, user_id, version): + async def count_e2e_room_keys(self, user_id: str, version: str) -> int: """Get the number of keys in a backup version. Args: - user_id (str): the user whose backup we're querying - version (str): the version ID of the backup we're querying about + user_id: the user whose backup we're querying + version: the version ID of the backup we're querying about """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="e2e_room_keys", keyvalues={"user_id": user_id, "version": version}, retcol="COUNT(*)", @@ -243,8 +241,9 @@ class EndToEndRoomKeyStore(SQLBaseStore): ) @trace - @defer.inlineCallbacks - def delete_e2e_room_keys(self, user_id, version, room_id=None, session_id=None): + async def delete_e2e_room_keys( + self, user_id, version, room_id=None, session_id=None + ): """Bulk delete the E2E room keys for a given backup, optionally filtered to a given room or a given session. @@ -259,7 +258,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): the backup (or for the specified room) Returns: - A deferred of the deletion transaction + The deletion transaction """ keyvalues = {"user_id": user_id, "version": int(version)} @@ -268,7 +267,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): if session_id: keyvalues["session_id"] = session_id - yield self.db.simple_delete( + await self.db_pool.simple_delete( table="e2e_room_keys", keyvalues=keyvalues, desc="delete_e2e_room_keys" ) @@ -284,7 +283,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): raise StoreError(404, "No current backup version") return row[0] - def get_e2e_room_keys_version_info(self, user_id, version=None): + async def get_e2e_room_keys_version_info(self, user_id, version=None): """Get info metadata about a version of our room_keys backup. Args: @@ -294,7 +293,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): Raises: StoreError: with code 404 if there are no e2e_room_keys_versions present Returns: - A deferred dict giving the info metadata for this backup version, with + A dict giving the info metadata for this backup version, with fields including: version(str) algorithm(str) @@ -313,7 +312,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): # it isn't there. raise StoreError(404, "No row found") - result = self.db.simple_select_one_txn( + result = self.db_pool.simple_select_one_txn( txn, table="e2e_room_keys_versions", keyvalues={"user_id": user_id, "version": this_version, "deleted": 0}, @@ -325,12 +324,12 @@ class EndToEndRoomKeyStore(SQLBaseStore): result["etag"] = 0 return result - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_e2e_room_keys_version_info", _get_e2e_room_keys_version_info_txn ) @trace - def create_e2e_room_keys_version(self, user_id, info): + async def create_e2e_room_keys_version(self, user_id: str, info: dict) -> str: """Atomically creates a new version of this user's e2e_room_keys store with the given version info. @@ -339,7 +338,7 @@ class EndToEndRoomKeyStore(SQLBaseStore): info(dict): the info about the backup version to be created Returns: - A deferred string for the newly created version ID + The newly created version ID """ def _create_e2e_room_keys_version_txn(txn): @@ -353,46 +352,50 @@ class EndToEndRoomKeyStore(SQLBaseStore): new_version = str(int(current_version) + 1) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="e2e_room_keys_versions", values={ "user_id": user_id, "version": new_version, "algorithm": info["algorithm"], - "auth_data": json.dumps(info["auth_data"]), + "auth_data": json_encoder.encode(info["auth_data"]), }, ) return new_version - return self.db.runInteraction( + return await self.db_pool.runInteraction( "create_e2e_room_keys_version_txn", _create_e2e_room_keys_version_txn ) @trace - def update_e2e_room_keys_version( - self, user_id, version, info=None, version_etag=None - ): + async def update_e2e_room_keys_version( + self, + user_id: str, + version: str, + info: Optional[dict] = None, + version_etag: Optional[int] = None, + ) -> None: """Update a given backup version Args: - user_id(str): the user whose backup version we're updating - version(str): the version ID of the backup version we're updating - info (dict): the new backup version info to store. If None, then - the backup version info is not updated - version_etag (Optional[int]): etag of the keys in the backup. If - None, then the etag is not updated + user_id: the user whose backup version we're updating + version: the version ID of the backup version we're updating + info: the new backup version info to store. If None, then the backup + version info is not updated. + version_etag: etag of the keys in the backup. If None, then the etag + is not updated. """ updatevalues = {} if info is not None and "auth_data" in info: - updatevalues["auth_data"] = json.dumps(info["auth_data"]) + updatevalues["auth_data"] = json_encoder.encode(info["auth_data"]) if version_etag is not None: updatevalues["etag"] = version_etag if updatevalues: - return self.db.simple_update( + await self.db_pool.simple_update( table="e2e_room_keys_versions", keyvalues={"user_id": user_id, "version": version}, updatevalues=updatevalues, @@ -400,13 +403,15 @@ class EndToEndRoomKeyStore(SQLBaseStore): ) @trace - def delete_e2e_room_keys_version(self, user_id, version=None): + async def delete_e2e_room_keys_version( + self, user_id: str, version: Optional[str] = None + ) -> None: """Delete a given backup version of the user's room keys. Doesn't delete their actual key data. Args: - user_id(str): the user whose backup version we're deleting - version(str): Optional. the version ID of the backup version we're deleting + user_id: the user whose backup version we're deleting + version: Optional. the version ID of the backup version we're deleting If missing, we delete the current backup version info. Raises: StoreError: with code 404 if there are no e2e_room_keys_versions present, @@ -421,19 +426,19 @@ class EndToEndRoomKeyStore(SQLBaseStore): else: this_version = version - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="e2e_room_keys", keyvalues={"user_id": user_id, "version": this_version}, ) - return self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="e2e_room_keys_versions", keyvalues={"user_id": user_id, "version": this_version}, updatevalues={"deleted": 1}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_e2e_room_keys_version", _delete_e2e_room_keys_version_txn ) diff --git a/synapse/storage/data_stores/main/end_to_end_keys.py b/synapse/storage/databases/main/end_to_end_keys.py
index 317c07a829..22e1ed15d0 100644 --- a/synapse/storage/data_stores/main/end_to_end_keys.py +++ b/synapse/storage/databases/main/end_to_end_keys.py
@@ -14,34 +14,79 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from typing import Dict, List, Tuple +import abc +from typing import TYPE_CHECKING, Dict, Iterable, List, Optional, Tuple -from canonicaljson import encode_canonical_json, json +import attr +from canonicaljson import encode_canonical_json from twisted.enterprise.adbapi import Connection -from twisted.internet import defer from synapse.logging.opentracing import log_kv, set_tag, trace from synapse.storage._base import SQLBaseStore, db_to_json from synapse.storage.database import make_in_list_sql_clause +from synapse.storage.types import Cursor +from synapse.types import JsonDict +from synapse.util import json_encoder from synapse.util.caches.descriptors import cached, cachedList from synapse.util.iterutils import batch_iter +if TYPE_CHECKING: + from synapse.handlers.e2e_keys import SignatureListItem + + +@attr.s(slots=True) +class DeviceKeyLookupResult: + """The type returned by get_e2e_device_keys_and_signatures""" + + display_name = attr.ib(type=Optional[str]) + + # the key data from e2e_device_keys_json. Typically includes fields like + # "algorithm", "keys" (including the curve25519 identity key and the ed25519 signing + # key) and "signatures" (a map from (user id) to (key id/device_id) to signature.) + keys = attr.ib(type=Optional[JsonDict]) + class EndToEndKeyWorkerStore(SQLBaseStore): + async def get_e2e_device_keys_for_federation_query( + self, user_id: str + ) -> Tuple[int, List[JsonDict]]: + """Get all devices (with any device keys) for a user + + Returns: + (stream_id, devices) + """ + now_stream_id = self.get_device_stream_token() + + devices = await self.get_e2e_device_keys_and_signatures([(user_id, None)]) + + if devices: + user_devices = devices[user_id] + results = [] + for device_id, device in user_devices.items(): + result = {"device_id": device_id} + + keys = device.keys + if keys: + result["keys"] = keys + + device_display_name = device.display_name + if device_display_name: + result["device_display_name"] = device_display_name + + results.append(result) + + return now_stream_id, results + + return now_stream_id, [] + @trace - @defer.inlineCallbacks - def get_e2e_device_keys( - self, query_list, include_all_devices=False, include_deleted_devices=False - ): - """Fetch a list of device keys. + async def get_e2e_device_keys_for_cs_api( + self, query_list: List[Tuple[str, Optional[str]]] + ) -> Dict[str, Dict[str, JsonDict]]: + """Fetch a list of device keys, formatted suitably for the C/S API. Args: query_list(list): List of pairs of user_ids and device_ids. - include_all_devices (bool): whether to include entries for devices - that don't have device keys - include_deleted_devices (bool): whether to include null entries for - devices which no longer exist (but were in the query_list). - This option only takes effect if include_all_devices is true. Returns: Dict mapping from user-id to dict mapping from device_id to key data. The key data will be a dict in the same format as the @@ -51,13 +96,7 @@ class EndToEndKeyWorkerStore(SQLBaseStore): if not query_list: return {} - results = yield self.db.runInteraction( - "get_e2e_device_keys", - self._get_e2e_device_keys_txn, - query_list, - include_all_devices, - include_deleted_devices, - ) + results = await self.get_e2e_device_keys_and_signatures(query_list) # Build the result structure, un-jsonify the results, and add the # "unsigned" section @@ -65,31 +104,95 @@ class EndToEndKeyWorkerStore(SQLBaseStore): for user_id, device_keys in results.items(): rv[user_id] = {} for device_id, device_info in device_keys.items(): - r = db_to_json(device_info.pop("key_json")) + r = device_info.keys r["unsigned"] = {} - display_name = device_info["device_display_name"] + display_name = device_info.display_name if display_name is not None: r["unsigned"]["device_display_name"] = display_name - if "signatures" in device_info: - for sig_user_id, sigs in device_info["signatures"].items(): - r.setdefault("signatures", {}).setdefault( - sig_user_id, {} - ).update(sigs) rv[user_id][device_id] = r return rv @trace - def _get_e2e_device_keys_txn( - self, txn, query_list, include_all_devices=False, include_deleted_devices=False - ): + async def get_e2e_device_keys_and_signatures( + self, + query_list: List[Tuple[str, Optional[str]]], + include_all_devices: bool = False, + include_deleted_devices: bool = False, + ) -> Dict[str, Dict[str, Optional[DeviceKeyLookupResult]]]: + """Fetch a list of device keys + + Any cross-signatures made on the keys by the owner of the device are also + included. + + The cross-signatures are added to the `signatures` field within the `keys` + object in the response. + + Args: + query_list: List of pairs of user_ids and device_ids. Device id can be None + to indicate "all devices for this user" + + include_all_devices: whether to return devices without device keys + + include_deleted_devices: whether to include null entries for + devices which no longer exist (but were in the query_list). + This option only takes effect if include_all_devices is true. + + Returns: + Dict mapping from user-id to dict mapping from device_id to + key data. + """ set_tag("include_all_devices", include_all_devices) set_tag("include_deleted_devices", include_deleted_devices) + result = await self.db_pool.runInteraction( + "get_e2e_device_keys", + self._get_e2e_device_keys_txn, + query_list, + include_all_devices, + include_deleted_devices, + ) + + # get the (user_id, device_id) tuples to look up cross-signatures for + signature_query = ( + (user_id, device_id) + for user_id, dev in result.items() + for device_id, d in dev.items() + if d is not None and d.keys is not None + ) + + for batch in batch_iter(signature_query, 50): + cross_sigs_result = await self.db_pool.runInteraction( + "get_e2e_cross_signing_signatures", + self._get_e2e_cross_signing_signatures_for_devices_txn, + batch, + ) + + # add each cross-signing signature to the correct device in the result dict. + for (user_id, key_id, device_id, signature) in cross_sigs_result: + target_device_result = result[user_id][device_id] + target_device_signatures = target_device_result.keys.setdefault( + "signatures", {} + ) + signing_user_signatures = target_device_signatures.setdefault( + user_id, {} + ) + signing_user_signatures[key_id] = signature + + log_kv(result) + return result + + def _get_e2e_device_keys_txn( + self, txn, query_list, include_all_devices=False, include_deleted_devices=False + ) -> Dict[str, Dict[str, Optional[DeviceKeyLookupResult]]]: + """Get information on devices from the database + + The results include the device's keys and self-signatures, but *not* any + cross-signing signatures which have been added subsequently (for which, see + get_e2e_device_keys_and_signatures) + """ query_clauses = [] query_params = [] - signature_query_clauses = [] - signature_query_params = [] if include_all_devices is False: include_deleted_devices = False @@ -100,24 +203,16 @@ class EndToEndKeyWorkerStore(SQLBaseStore): for (user_id, device_id) in query_list: query_clause = "user_id = ?" query_params.append(user_id) - signature_query_clause = "target_user_id = ?" - signature_query_params.append(user_id) if device_id is not None: query_clause += " AND device_id = ?" query_params.append(device_id) - signature_query_clause += " AND target_device_id = ?" - signature_query_params.append(device_id) - - signature_query_clause += " AND user_id = ?" - signature_query_params.append(user_id) query_clauses.append(query_clause) - signature_query_clauses.append(signature_query_clause) sql = ( "SELECT user_id, device_id, " - " d.display_name AS device_display_name, " + " d.display_name, " " k.key_json" " FROM devices d" " %s JOIN e2e_device_keys_json k USING (user_id, device_id)" @@ -128,54 +223,53 @@ class EndToEndKeyWorkerStore(SQLBaseStore): ) txn.execute(sql, query_params) - rows = self.db.cursor_to_dict(txn) - result = {} - for row in rows: + result = {} # type: Dict[str, Dict[str, Optional[DeviceKeyLookupResult]]] + for (user_id, device_id, display_name, key_json) in txn: if include_deleted_devices: - deleted_devices.remove((row["user_id"], row["device_id"])) - result.setdefault(row["user_id"], {})[row["device_id"]] = row + deleted_devices.remove((user_id, device_id)) + result.setdefault(user_id, {})[device_id] = DeviceKeyLookupResult( + display_name, db_to_json(key_json) if key_json else None + ) if include_deleted_devices: for user_id, device_id in deleted_devices: result.setdefault(user_id, {})[device_id] = None - # get signatures on the device - signature_sql = ("SELECT * FROM e2e_cross_signing_signatures WHERE %s") % ( - " OR ".join("(" + q + ")" for q in signature_query_clauses) - ) + return result - txn.execute(signature_sql, signature_query_params) - rows = self.db.cursor_to_dict(txn) - - # add each cross-signing signature to the correct device in the result dict. - for row in rows: - signing_user_id = row["user_id"] - signing_key_id = row["key_id"] - target_user_id = row["target_user_id"] - target_device_id = row["target_device_id"] - signature = row["signature"] - - target_user_result = result.get(target_user_id) - if not target_user_result: - continue + def _get_e2e_cross_signing_signatures_for_devices_txn( + self, txn: Cursor, device_query: Iterable[Tuple[str, str]] + ) -> List[Tuple[str, str, str, str]]: + """Get cross-signing signatures for a given list of devices - target_device_result = target_user_result.get(target_device_id) - if not target_device_result: - # note that target_device_result will be None for deleted devices. - continue + Returns signatures made by the owners of the devices. - target_device_signatures = target_device_result.setdefault("signatures", {}) - signing_user_signatures = target_device_signatures.setdefault( - signing_user_id, {} + Returns: a list of results; each entry in the list is a tuple of + (user_id, key_id, target_device_id, signature). + """ + signature_query_clauses = [] + signature_query_params = [] + + for (user_id, device_id) in device_query: + signature_query_clauses.append( + "target_user_id = ? AND target_device_id = ? AND user_id = ?" ) - signing_user_signatures[signing_key_id] = signature + signature_query_params.extend([user_id, device_id, user_id]) - log_kv(result) - return result + signature_sql = """ + SELECT user_id, key_id, target_device_id, signature + FROM e2e_cross_signing_signatures WHERE %s + """ % ( + " OR ".join("(" + q + ")" for q in signature_query_clauses) + ) + + txn.execute(signature_sql, signature_query_params) + return txn.fetchall() - @defer.inlineCallbacks - def get_e2e_one_time_keys(self, user_id, device_id, key_ids): + async def get_e2e_one_time_keys( + self, user_id: str, device_id: str, key_ids: List[str] + ) -> Dict[Tuple[str, str], str]: """Retrieve a number of one-time keys for a user Args: @@ -185,11 +279,10 @@ class EndToEndKeyWorkerStore(SQLBaseStore): retrieve Returns: - deferred resolving to Dict[(str, str), str]: map from (algorithm, - key_id) to json string for key + A map from (algorithm, key_id) to json string for key """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="e2e_one_time_keys_json", column="key_id", iterable=key_ids, @@ -201,17 +294,21 @@ class EndToEndKeyWorkerStore(SQLBaseStore): log_kv({"message": "Fetched one time keys for user", "one_time_keys": result}) return result - @defer.inlineCallbacks - def add_e2e_one_time_keys(self, user_id, device_id, time_now, new_keys): + async def add_e2e_one_time_keys( + self, + user_id: str, + device_id: str, + time_now: int, + new_keys: Iterable[Tuple[str, str, str]], + ) -> None: """Insert some new one time keys for a device. Errors if any of the keys already exist. Args: - user_id(str): id of user to get keys for - device_id(str): id of device to get keys for - time_now(long): insertion time to record (ms since epoch) - new_keys(iterable[(str, str, str)]: keys to add - each a tuple of - (algorithm, key_id, key json) + user_id: id of user to get keys for + device_id: id of device to get keys for + time_now: insertion time to record (ms since epoch) + new_keys: keys to add - each a tuple of (algorithm, key_id, key json) """ def _add_e2e_one_time_keys(txn): @@ -222,7 +319,7 @@ class EndToEndKeyWorkerStore(SQLBaseStore): # a unique constraint. If there is a race of two calls to # `add_e2e_one_time_keys` then they'll conflict and we will only # insert one set. - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="e2e_one_time_keys_json", values=[ @@ -241,15 +338,17 @@ class EndToEndKeyWorkerStore(SQLBaseStore): txn, self.count_e2e_one_time_keys, (user_id, device_id) ) - yield self.db.runInteraction( + await self.db_pool.runInteraction( "add_e2e_one_time_keys_insert", _add_e2e_one_time_keys ) @cached(max_entries=10000) - def count_e2e_one_time_keys(self, user_id, device_id): + async def count_e2e_one_time_keys( + self, user_id: str, device_id: str + ) -> Dict[str, int]: """ Count the number of one time keys the server has for a device Returns: - Dict mapping from algorithm to number of keys for that algorithm. + A mapping from algorithm to number of keys for that algorithm. """ def _count_e2e_one_time_keys(txn): @@ -264,26 +363,27 @@ class EndToEndKeyWorkerStore(SQLBaseStore): result[algorithm] = key_count return result - return self.db.runInteraction( + return await self.db_pool.runInteraction( "count_e2e_one_time_keys", _count_e2e_one_time_keys ) - @defer.inlineCallbacks - def get_e2e_cross_signing_key(self, user_id, key_type, from_user_id=None): + async def get_e2e_cross_signing_key( + self, user_id: str, key_type: str, from_user_id: Optional[str] = None + ) -> Optional[dict]: """Returns a user's cross-signing key. Args: - user_id (str): the user whose key is being requested - key_type (str): the type of key that is being requested: either 'master' + user_id: the user whose key is being requested + key_type: the type of key that is being requested: either 'master' for a master key, 'self_signing' for a self-signing key, or 'user_signing' for a user-signing key - from_user_id (str): if specified, signatures made by this user on + from_user_id: if specified, signatures made by this user on the self-signing key will be included in the result Returns: dict of the key data or None if not found """ - res = yield self.get_e2e_cross_signing_keys_bulk([user_id], from_user_id) + res = await self.get_e2e_cross_signing_keys_bulk([user_id], from_user_id) user_keys = res.get(user_id) if not user_keys: return None @@ -301,7 +401,7 @@ class EndToEndKeyWorkerStore(SQLBaseStore): list_name="user_ids", num_args=1, ) - def _get_bare_e2e_cross_signing_keys_bulk( + async def _get_bare_e2e_cross_signing_keys_bulk( self, user_ids: List[str] ) -> Dict[str, Dict[str, dict]]: """Returns the cross-signing keys for a set of users. The output of this @@ -309,16 +409,15 @@ class EndToEndKeyWorkerStore(SQLBaseStore): the signatures for the calling user need to be fetched. Args: - user_ids (list[str]): the users whose keys are being requested + user_ids: the users whose keys are being requested Returns: - dict[str, dict[str, dict]]: mapping from user ID to key type to key - data. If a user's cross-signing keys were not found, either - their user ID will not be in the dict, or their user ID will map - to None. + A mapping from user ID to key type to key data. If a user's cross-signing + keys were not found, either their user ID will not be in the dict, or + their user ID will map to None. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_bare_e2e_cross_signing_keys_bulk", self._get_bare_e2e_cross_signing_keys_bulk_txn, user_ids, @@ -361,7 +460,7 @@ class EndToEndKeyWorkerStore(SQLBaseStore): ) txn.execute(sql, params) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) for row in rows: user_id = row["user_id"] @@ -420,7 +519,7 @@ class EndToEndKeyWorkerStore(SQLBaseStore): query_params.extend(item) txn.execute(sql, query_params) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) # and add the signatures to the appropriate keys for row in rows: @@ -449,28 +548,26 @@ class EndToEndKeyWorkerStore(SQLBaseStore): return keys - @defer.inlineCallbacks - def get_e2e_cross_signing_keys_bulk( - self, user_ids: List[str], from_user_id: str = None - ) -> defer.Deferred: + async def get_e2e_cross_signing_keys_bulk( + self, user_ids: List[str], from_user_id: Optional[str] = None + ) -> Dict[str, Dict[str, dict]]: """Returns the cross-signing keys for a set of users. Args: - user_ids (list[str]): the users whose keys are being requested - from_user_id (str): if specified, signatures made by this user on + user_ids: the users whose keys are being requested + from_user_id: if specified, signatures made by this user on the self-signing keys will be included in the result Returns: - Deferred[dict[str, dict[str, dict]]]: map of user ID to key type to - key data. If a user's cross-signing keys were not found, either - their user ID will not be in the dict, or their user ID will map - to None. + A map of user ID to key type to key data. If a user's cross-signing + keys were not found, either their user ID will not be in the dict, + or their user ID will map to None. """ - result = yield self._get_bare_e2e_cross_signing_keys_bulk(user_ids) + result = await self._get_bare_e2e_cross_signing_keys_bulk(user_ids) if from_user_id: - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( "get_e2e_cross_signing_signatures", self._get_e2e_cross_signing_signatures_txn, result, @@ -531,14 +628,21 @@ class EndToEndKeyWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_user_signature_changes_for_remotes", _get_all_user_signature_changes_for_remotes_txn, ) + @abc.abstractmethod + def get_device_stream_token(self) -> int: + """Get the current stream id from the _device_list_id_gen""" + ... + class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): - def set_e2e_device_keys(self, user_id, device_id, time_now, device_keys): + async def set_e2e_device_keys( + self, user_id: str, device_id: str, time_now: int, device_keys: JsonDict + ) -> bool: """Stores device keys for a device. Returns whether there was a change or the keys were already in the database. """ @@ -549,7 +653,7 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): set_tag("time_now", time_now) set_tag("device_keys", device_keys) - old_key_json = self.db.simple_select_one_onecol_txn( + old_key_json = self.db_pool.simple_select_one_onecol_txn( txn, table="e2e_device_keys_json", keyvalues={"user_id": user_id, "device_id": device_id}, @@ -565,7 +669,7 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): log_kv({"Message": "Device key already stored."}) return False - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="e2e_device_keys_json", keyvalues={"user_id": user_id, "device_id": device_id}, @@ -574,10 +678,21 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): log_kv({"message": "Device keys stored."}) return True - return self.db.runInteraction("set_e2e_device_keys", _set_e2e_device_keys_txn) + return await self.db_pool.runInteraction( + "set_e2e_device_keys", _set_e2e_device_keys_txn + ) + + async def claim_e2e_one_time_keys( + self, query_list: Iterable[Tuple[str, str, str]] + ) -> Dict[str, Dict[str, Dict[str, bytes]]]: + """Take a list of one time keys out of the database. - def claim_e2e_one_time_keys(self, query_list): - """Take a list of one time keys out of the database""" + Args: + query_list: An iterable of tuples of (user ID, device ID, algorithm). + + Returns: + A map of user ID -> a map device ID -> a map of key ID -> JSON bytes. + """ @trace def _claim_e2e_one_time_keys(txn): @@ -613,11 +728,11 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): ) return result - return self.db.runInteraction( + return await self.db_pool.runInteraction( "claim_e2e_one_time_keys", _claim_e2e_one_time_keys ) - def delete_e2e_keys_by_device(self, user_id, device_id): + async def delete_e2e_keys_by_device(self, user_id: str, device_id: str) -> None: def delete_e2e_keys_by_device_txn(txn): log_kv( { @@ -626,12 +741,12 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): "user_id": user_id, } ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="e2e_device_keys_json", keyvalues={"user_id": user_id, "device_id": device_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="e2e_one_time_keys_json", keyvalues={"user_id": user_id, "device_id": device_id}, @@ -640,11 +755,11 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): txn, self.count_e2e_one_time_keys, (user_id, device_id) ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_e2e_keys_by_device", delete_e2e_keys_by_device_txn ) - def _set_e2e_cross_signing_key_txn(self, txn, user_id, key_type, key): + def _set_e2e_cross_signing_key_txn(self, txn, user_id, key_type, key, stream_id): """Set a user's cross-signing key. Args: @@ -654,6 +769,7 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): for a master key, 'self_signing' for a self-signing key, or 'user_signing' for a user-signing key key (dict): the key data + stream_id (int) """ # the 'key' dict will look something like: # { @@ -679,7 +795,7 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): # We only need to do this for local users, since remote servers should be # responsible for checking this for their own users. if self.hs.is_mine_id(user_id): - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "devices", values={ @@ -691,23 +807,22 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): ) # and finally, store the key itself - with self._cross_signing_id_gen.get_next() as stream_id: - self.db.simple_insert_txn( - txn, - "e2e_cross_signing_keys", - values={ - "user_id": user_id, - "keytype": key_type, - "keydata": json.dumps(key), - "stream_id": stream_id, - }, - ) + self.db_pool.simple_insert_txn( + txn, + "e2e_cross_signing_keys", + values={ + "user_id": user_id, + "keytype": key_type, + "keydata": json_encoder.encode(key), + "stream_id": stream_id, + }, + ) self._invalidate_cache_and_stream( txn, self._get_bare_e2e_cross_signing_keys, (user_id,) ) - def set_e2e_cross_signing_key(self, user_id, key_type, key): + async def set_e2e_cross_signing_key(self, user_id, key_type, key): """Set a user's cross-signing key. Args: @@ -715,22 +830,27 @@ class EndToEndKeyStore(EndToEndKeyWorkerStore, SQLBaseStore): key_type (str): the type of cross-signing key to set key (dict): the key data """ - return self.db.runInteraction( - "add_e2e_cross_signing_key", - self._set_e2e_cross_signing_key_txn, - user_id, - key_type, - key, - ) - def store_e2e_cross_signing_signatures(self, user_id, signatures): + async with self._cross_signing_id_gen.get_next() as stream_id: + return await self.db_pool.runInteraction( + "add_e2e_cross_signing_key", + self._set_e2e_cross_signing_key_txn, + user_id, + key_type, + key, + stream_id, + ) + + async def store_e2e_cross_signing_signatures( + self, user_id: str, signatures: "Iterable[SignatureListItem]" + ) -> None: """Stores cross-signing signatures. Args: - user_id (str): the user who made the signatures - signatures (iterable[SignatureListItem]): signatures to add + user_id: the user who made the signatures + signatures: signatures to add """ - return self.db.simple_insert_many( + await self.db_pool.simple_insert_many( "e2e_cross_signing_signatures", [ { diff --git a/synapse/storage/data_stores/main/event_federation.py b/synapse/storage/databases/main/event_federation.py
index a6bb3221ff..6d3689c09e 100644 --- a/synapse/storage/data_stores/main/event_federation.py +++ b/synapse/storage/databases/main/event_federation.py
@@ -15,16 +15,16 @@ import itertools import logging from queue import Empty, PriorityQueue -from typing import Dict, List, Optional, Set, Tuple - -from twisted.internet import defer +from typing import Dict, Iterable, List, Set, Tuple from synapse.api.errors import StoreError +from synapse.events import EventBase from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import SQLBaseStore, make_in_list_sql_clause -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.data_stores.main.signatures import SignatureWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool, LoggingTransaction +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.storage.databases.main.signatures import SignatureWorkerStore +from synapse.types import Collection from synapse.util.caches.descriptors import cached from synapse.util.iterutils import batch_iter @@ -32,57 +32,51 @@ logger = logging.getLogger(__name__) class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBaseStore): - def get_auth_chain(self, event_ids, include_given=False): + async def get_auth_chain( + self, event_ids: Collection[str], include_given: bool = False + ) -> List[EventBase]: """Get auth events for given event_ids. The events *must* be state events. Args: - event_ids (list): state events - include_given (bool): include the given events in result + event_ids: state events + include_given: include the given events in result Returns: list of events """ - return self.get_auth_chain_ids( + event_ids = await self.get_auth_chain_ids( event_ids, include_given=include_given - ).addCallback(self.get_events_as_list) - - def get_auth_chain_ids( - self, - event_ids: List[str], - include_given: bool = False, - ignore_events: Optional[Set[str]] = None, - ): + ) + return await self.get_events_as_list(event_ids) + + async def get_auth_chain_ids( + self, event_ids: Collection[str], include_given: bool = False, + ) -> List[str]: """Get auth events for given event_ids. The events *must* be state events. Args: event_ids: state events include_given: include the given events in result - ignore_events: Set of events to exclude from the returned auth - chain. This is useful if the caller will just discard the - given events anyway, and saves us from figuring out their auth - chains if not required. Returns: - list of event_ids + An awaitable which resolve to a list of event_ids """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_auth_chain_ids", self._get_auth_chain_ids_txn, event_ids, include_given, - ignore_events, ) - def _get_auth_chain_ids_txn(self, txn, event_ids, include_given, ignore_events): - if ignore_events is None: - ignore_events = set() - + def _get_auth_chain_ids_txn( + self, txn: LoggingTransaction, event_ids: Collection[str], include_given: bool + ) -> List[str]: if include_given: results = set(event_ids) else: results = set() - base_sql = "SELECT auth_id FROM event_auth WHERE " + base_sql = "SELECT DISTINCT auth_id FROM event_auth WHERE " front = set(event_ids) while front: @@ -94,7 +88,6 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas txn.execute(base_sql + clause, args) new_front.update(r[0] for r in txn) - new_front -= ignore_events new_front -= results front = new_front @@ -102,7 +95,7 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas return list(results) - def get_auth_chain_difference(self, state_sets: List[Set[str]]): + async def get_auth_chain_difference(self, state_sets: List[Set[str]]) -> Set[str]: """Given sets of state events figure out the auth chain difference (as per state res v2 algorithm). @@ -111,10 +104,10 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas chain. Returns: - Deferred[Set[str]] + The set of the difference in auth chains. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_auth_chain_difference", self._get_auth_chain_difference_txn, state_sets, @@ -259,13 +252,8 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas # Return all events where not all sets can reach them. return {eid for eid, n in event_to_missing_sets.items() if n} - def get_oldest_events_in_room(self, room_id): - return self.db.runInteraction( - "get_oldest_events_in_room", self._get_oldest_events_in_room_txn, room_id - ) - - def get_oldest_events_with_depth_in_room(self, room_id): - return self.db.runInteraction( + async def get_oldest_events_with_depth_in_room(self, room_id): + return await self.db_pool.runInteraction( "get_oldest_events_with_depth_in_room", self.get_oldest_events_with_depth_in_room_txn, room_id, @@ -286,17 +274,13 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas return dict(txn) - @defer.inlineCallbacks - def get_max_depth_of(self, event_ids): + async def get_max_depth_of(self, event_ids: List[str]) -> int: """Returns the max depth of a set of event IDs Args: - event_ids (list[str]) - - Returns - Deferred[int] + event_ids: The event IDs to calculate the max depth of. """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="events", column="event_id", iterable=event_ids, @@ -309,15 +293,7 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas else: return max(row["depth"] for row in rows) - def _get_oldest_events_in_room_txn(self, txn, room_id): - return self.db.simple_select_onecol_txn( - txn, - table="event_backward_extremities", - keyvalues={"room_id": room_id}, - retcol="event_id", - ) - - def get_prev_events_for_room(self, room_id: str): + async def get_prev_events_for_room(self, room_id: str) -> List[str]: """ Gets a subset of the current forward extremities in the given room. @@ -325,14 +301,14 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas events which refer to hundreds of prev_events. Args: - room_id (str): room_id + room_id: room_id Returns: - Deferred[List[str]]: the event ids of the forward extremites + The event ids of the forward extremities. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_prev_events_for_room", self._get_prev_events_for_room_txn, room_id ) @@ -352,17 +328,19 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas return [row[0] for row in txn] - def get_rooms_with_many_extremities(self, min_count, limit, room_id_filter): + async def get_rooms_with_many_extremities( + self, min_count: int, limit: int, room_id_filter: Iterable[str] + ) -> List[str]: """Get the top rooms with at least N extremities. Args: - min_count (int): The minimum number of extremities - limit (int): The maximum number of rooms to return. - room_id_filter (iterable[str]): room_ids to exclude from the results + min_count: The minimum number of extremities + limit: The maximum number of rooms to return. + room_id_filter: room_ids to exclude from the results Returns: - Deferred[list]: At most `limit` room IDs that have at least - `min_count` extremities, sorted by extremity count. + At most `limit` room IDs that have at least `min_count` extremities, + sorted by extremity count. """ def _get_rooms_with_many_extremities_txn(txn): @@ -387,28 +365,28 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas txn.execute(sql, query_args) return [room_id for room_id, in txn] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_rooms_with_many_extremities", _get_rooms_with_many_extremities_txn ) @cached(max_entries=5000, iterable=True) - def get_latest_event_ids_in_room(self, room_id): - return self.db.simple_select_onecol( + async def get_latest_event_ids_in_room(self, room_id: str) -> List[str]: + return await self.db_pool.simple_select_onecol( table="event_forward_extremities", keyvalues={"room_id": room_id}, retcol="event_id", desc="get_latest_event_ids_in_room", ) - def get_min_depth(self, room_id): - """ For hte given room, get the minimum depth we have seen for it. + async def get_min_depth(self, room_id: str) -> int: + """For the given room, get the minimum depth we have seen for it. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_min_depth", self._get_min_depth_interaction, room_id ) def _get_min_depth_interaction(self, txn, room_id): - min_depth = self.db.simple_select_one_onecol_txn( + min_depth = self.db_pool.simple_select_one_onecol_txn( txn, table="room_depth", keyvalues={"room_id": room_id}, @@ -418,7 +396,9 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas return int(min_depth) if min_depth is not None else None - def get_forward_extremeties_for_room(self, room_id, stream_ordering): + async def get_forward_extremeties_for_room( + self, room_id: str, stream_ordering: int + ) -> List[str]: """For a given room_id and stream_ordering, return the forward extremeties of the room at that point in "time". @@ -426,11 +406,11 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas stream_orderings from that point. Args: - room_id (str): - stream_ordering (int): + room_id: + stream_ordering: Returns: - deferred, which resolves to a list of event_ids + A list of event_ids """ # We want to make the cache more effective, so we clamp to the last # change before the given ordering. @@ -446,10 +426,10 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas if last_change > self.stream_ordering_month_ago: stream_ordering = min(last_change, stream_ordering) - return self._get_forward_extremeties_for_room(room_id, stream_ordering) + return await self._get_forward_extremeties_for_room(room_id, stream_ordering) @cached(max_entries=5000, num_args=2) - def _get_forward_extremeties_for_room(self, room_id, stream_ordering): + async def _get_forward_extremeties_for_room(self, room_id, stream_ordering): """For a given room_id and stream_ordering, return the forward extremeties of the room at that point in "time". @@ -458,7 +438,7 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas """ if stream_ordering <= self.stream_ordering_month_ago: - raise StoreError(400, "stream_ordering too old") + raise StoreError(400, "stream_ordering too old %s" % (stream_ordering,)) sql = """ SELECT event_id FROM stream_ordering_to_exterm @@ -474,31 +454,28 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas txn.execute(sql, (stream_ordering, room_id)) return [event_id for event_id, in txn] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_forward_extremeties_for_room", get_forward_extremeties_for_room_txn ) - def get_backfill_events(self, room_id, event_list, limit): + async def get_backfill_events(self, room_id: str, event_list: list, limit: int): """Get a list of Events for a given topic that occurred before (and including) the events in event_list. Return a list of max size `limit` Args: - txn - room_id (str) - event_list (list) - limit (int) + room_id + event_list + limit """ - return ( - self.db.runInteraction( - "get_backfill_events", - self._get_backfill_events, - room_id, - event_list, - limit, - ) - .addCallback(self.get_events_as_list) - .addCallback(lambda l: sorted(l, key=lambda e: -e.depth)) + event_ids = await self.db_pool.runInteraction( + "get_backfill_events", + self._get_backfill_events, + room_id, + event_list, + limit, ) + events = await self.get_events_as_list(event_ids) + return sorted(events, key=lambda e: -e.depth) def _get_backfill_events(self, txn, room_id, event_list, limit): logger.debug("_get_backfill_events: %s, %r, %s", room_id, event_list, limit) @@ -520,7 +497,7 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas queue = PriorityQueue() for event_id in event_list: - depth = self.db.simple_select_one_onecol_txn( + depth = self.db_pool.simple_select_one_onecol_txn( txn, table="events", keyvalues={"event_id": event_id, "room_id": room_id}, @@ -550,9 +527,8 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas return event_results - @defer.inlineCallbacks - def get_missing_events(self, room_id, earliest_events, latest_events, limit): - ids = yield self.db.runInteraction( + async def get_missing_events(self, room_id, earliest_events, latest_events, limit): + ids = await self.db_pool.runInteraction( "get_missing_events", self._get_missing_events, room_id, @@ -560,8 +536,7 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas latest_events, limit, ) - events = yield self.get_events_as_list(ids) - return events + return await self.get_events_as_list(ids) def _get_missing_events(self, txn, room_id, earliest_events, latest_events, limit): @@ -595,17 +570,13 @@ class EventFederationWorkerStore(EventsWorkerStore, SignatureWorkerStore, SQLBas event_results.reverse() return event_results - @defer.inlineCallbacks - def get_successor_events(self, event_ids): + async def get_successor_events(self, event_ids: Iterable[str]) -> List[str]: """Fetch all events that have the given events as a prev event Args: - event_ids (iterable[str]) - - Returns: - Deferred[list[str]] + event_ids: The events to use as the previous events. """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="event_edges", column="prev_event_id", iterable=event_ids, @@ -628,10 +599,10 @@ class EventFederationStore(EventFederationWorkerStore): EVENT_AUTH_STATE_ONLY = "event_auth_state_only" - def __init__(self, database: Database, db_conn, hs): - super(EventFederationStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_AUTH_STATE_ONLY, self._background_delete_non_state_event_auth ) @@ -658,13 +629,13 @@ class EventFederationStore(EventFederationWorkerStore): return run_as_background_process( "delete_old_forward_extrem_cache", - self.db.runInteraction, + self.db_pool.runInteraction, "_delete_old_forward_extrem_cache", _delete_old_forward_extrem_cache_txn, ) - def clean_room_for_join(self, room_id): - return self.db.runInteraction( + async def clean_room_for_join(self, room_id): + return await self.db_pool.runInteraction( "clean_room_for_join", self._clean_room_for_join_txn, room_id ) @@ -674,8 +645,7 @@ class EventFederationStore(EventFederationWorkerStore): txn.execute(query, (room_id,)) txn.call_after(self.get_latest_event_ids_in_room.invalidate, (room_id,)) - @defer.inlineCallbacks - def _background_delete_non_state_event_auth(self, progress, batch_size): + async def _background_delete_non_state_event_auth(self, progress, batch_size): def delete_event_auth(txn): target_min_stream_id = progress.get("target_min_stream_id_inclusive") max_stream_id = progress.get("max_stream_id_exclusive") @@ -708,17 +678,19 @@ class EventFederationStore(EventFederationWorkerStore): "max_stream_id_exclusive": min_stream_id, } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.EVENT_AUTH_STATE_ONLY, new_progress ) return min_stream_id >= target_min_stream_id - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( self.EVENT_AUTH_STATE_ONLY, delete_event_auth ) if not result: - yield self.db.updates._end_background_update(self.EVENT_AUTH_STATE_ONLY) + await self.db_pool.updates._end_background_update( + self.EVENT_AUTH_STATE_ONLY + ) return batch_size diff --git a/synapse/storage/data_stores/main/event_push_actions.py b/synapse/storage/databases/main/event_push_actions.py
index 504babaa7e..62f1738732 100644 --- a/synapse/storage/data_stores/main/event_push_actions.py +++ b/synapse/storage/databases/main/event_push_actions.py
@@ -15,15 +15,15 @@ # limitations under the License. import logging +from typing import Dict, List, Optional, Tuple, Union -from canonicaljson import json - -from twisted.internet import defer +import attr from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import LoggingTransaction, SQLBaseStore, db_to_json -from synapse.storage.database import Database -from synapse.util.caches.descriptors import cachedInlineCallbacks +from synapse.storage.database import DatabasePool +from synapse.util import json_encoder +from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) @@ -51,7 +51,7 @@ def _serialize_action(actions, is_highlight): else: if actions == DEFAULT_NOTIF_ACTION: return "" - return json.dumps(actions) + return json_encoder.encode(actions) def _deserialize_action(actions, is_highlight): @@ -67,8 +67,8 @@ def _deserialize_action(actions, is_highlight): class EventPushActionsWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(EventPushActionsWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # These get correctly set by _find_stream_orderings_for_times_txn self.stream_ordering_month_ago = None @@ -88,114 +88,141 @@ class EventPushActionsWorkerStore(SQLBaseStore): self._rotate_delay = 3 self._rotate_count = 10000 - @cachedInlineCallbacks(num_args=3, tree=True, max_entries=5000) - def get_unread_event_push_actions_by_room_for_user( - self, room_id, user_id, last_read_event_id - ): - ret = yield self.db.runInteraction( + @cached(num_args=3, tree=True, max_entries=5000) + async def get_unread_event_push_actions_by_room_for_user( + self, room_id: str, user_id: str, last_read_event_id: Optional[str], + ) -> Dict[str, int]: + """Get the notification count, the highlight count and the unread message count + for a given user in a given room after the given read receipt. + + Note that this function assumes the user to be a current member of the room, + since it's either called by the sync handler to handle joined room entries, or by + the HTTP pusher to calculate the badge of unread joined rooms. + + Args: + room_id: The room to retrieve the counts in. + user_id: The user to retrieve the counts for. + last_read_event_id: The event associated with the latest read receipt for + this user in this room. None if no receipt for this user in this room. + + Returns + A dict containing the counts mentioned earlier in this docstring, + respectively under the keys "notify_count", "highlight_count" and + "unread_count". + """ + return await self.db_pool.runInteraction( "get_unread_event_push_actions_by_room", self._get_unread_counts_by_receipt_txn, room_id, user_id, last_read_event_id, ) - return ret def _get_unread_counts_by_receipt_txn( - self, txn, room_id, user_id, last_read_event_id + self, txn, room_id, user_id, last_read_event_id, ): - sql = ( - "SELECT stream_ordering" - " FROM events" - " WHERE room_id = ? AND event_id = ?" - ) - txn.execute(sql, (room_id, last_read_event_id)) - results = txn.fetchall() - if len(results) == 0: - return {"notify_count": 0, "highlight_count": 0} + stream_ordering = None + + if last_read_event_id is not None: + stream_ordering = self.get_stream_id_for_event_txn( + txn, last_read_event_id, allow_none=True, + ) + + if stream_ordering is None: + # Either last_read_event_id is None, or it's an event we don't have (e.g. + # because it's been purged), in which case retrieve the stream ordering for + # the latest membership event from this user in this room (which we assume is + # a join). + event_id = self.db_pool.simple_select_one_onecol_txn( + txn=txn, + table="local_current_membership", + keyvalues={"room_id": room_id, "user_id": user_id}, + retcol="event_id", + ) - stream_ordering = results[0][0] + stream_ordering = self.get_stream_id_for_event_txn(txn, event_id) return self._get_unread_counts_by_pos_txn( txn, room_id, user_id, stream_ordering ) def _get_unread_counts_by_pos_txn(self, txn, room_id, user_id, stream_ordering): - - # First get number of notifications. - # We don't need to put a notif=1 clause as all rows always have - # notif=1 sql = ( - "SELECT count(*)" + "SELECT" + " COUNT(CASE WHEN notif = 1 THEN 1 END)," + " COUNT(CASE WHEN highlight = 1 THEN 1 END)," + " COUNT(CASE WHEN unread = 1 THEN 1 END)" " FROM event_push_actions ea" - " WHERE" - " user_id = ?" - " AND room_id = ?" - " AND stream_ordering > ?" + " WHERE user_id = ?" + " AND room_id = ?" + " AND stream_ordering > ?" ) txn.execute(sql, (user_id, room_id, stream_ordering)) row = txn.fetchone() - notify_count = row[0] if row else 0 + + (notif_count, highlight_count, unread_count) = (0, 0, 0) + + if row: + (notif_count, highlight_count, unread_count) = row txn.execute( """ - SELECT notif_count FROM event_push_summary - WHERE room_id = ? AND user_id = ? AND stream_ordering > ? - """, + SELECT notif_count, unread_count FROM event_push_summary + WHERE room_id = ? AND user_id = ? AND stream_ordering > ? + """, (room_id, user_id, stream_ordering), ) - rows = txn.fetchall() - if rows: - notify_count += rows[0][0] + row = txn.fetchone() - # Now get the number of highlights - sql = ( - "SELECT count(*)" - " FROM event_push_actions ea" - " WHERE" - " highlight = 1" - " AND user_id = ?" - " AND room_id = ?" - " AND stream_ordering > ?" - ) + if row: + notif_count += row[0] - txn.execute(sql, (user_id, room_id, stream_ordering)) - row = txn.fetchone() - highlight_count = row[0] if row else 0 + if row[1] is not None: + # The unread_count column of event_push_summary is NULLable, so we need + # to make sure we don't try increasing the unread counts if it's NULL + # for this row. + unread_count += row[1] - return {"notify_count": notify_count, "highlight_count": highlight_count} + return { + "notify_count": notif_count, + "unread_count": unread_count, + "highlight_count": highlight_count, + } - @defer.inlineCallbacks - def get_push_action_users_in_range(self, min_stream_ordering, max_stream_ordering): + async def get_push_action_users_in_range( + self, min_stream_ordering, max_stream_ordering + ): def f(txn): sql = ( "SELECT DISTINCT(user_id) FROM event_push_actions WHERE" - " stream_ordering >= ? AND stream_ordering <= ?" + " stream_ordering >= ? AND stream_ordering <= ? AND notif = 1" ) txn.execute(sql, (min_stream_ordering, max_stream_ordering)) return [r[0] for r in txn] - ret = yield self.db.runInteraction("get_push_action_users_in_range", f) + ret = await self.db_pool.runInteraction("get_push_action_users_in_range", f) return ret - @defer.inlineCallbacks - def get_unread_push_actions_for_user_in_range_for_http( - self, user_id, min_stream_ordering, max_stream_ordering, limit=20 - ): + async def get_unread_push_actions_for_user_in_range_for_http( + self, + user_id: str, + min_stream_ordering: int, + max_stream_ordering: int, + limit: int = 20, + ) -> List[dict]: """Get a list of the most recent unread push actions for a given user, within the given stream ordering range. Called by the httppusher. Args: - user_id (str): The user to fetch push actions for. - min_stream_ordering(int): The exclusive lower bound on the + user_id: The user to fetch push actions for. + min_stream_ordering: The exclusive lower bound on the stream ordering of event push actions to fetch. - max_stream_ordering(int): The inclusive upper bound on the + max_stream_ordering: The inclusive upper bound on the stream ordering of event push actions to fetch. - limit (int): The maximum number of rows to return. + limit: The maximum number of rows to return. Returns: - A promise which resolves to a list of dicts with the keys "event_id", - "room_id", "stream_ordering", "actions". + A list of dicts with the keys "event_id", "room_id", "stream_ordering", "actions". The list will be ordered by ascending stream_ordering. The list will have between 0~limit entries. """ @@ -222,13 +249,14 @@ class EventPushActionsWorkerStore(SQLBaseStore): " AND ep.user_id = ?" " AND ep.stream_ordering > ?" " AND ep.stream_ordering <= ?" + " AND ep.notif = 1" " ORDER BY ep.stream_ordering ASC LIMIT ?" ) args = [user_id, user_id, min_stream_ordering, max_stream_ordering, limit] txn.execute(sql, args) return txn.fetchall() - after_read_receipt = yield self.db.runInteraction( + after_read_receipt = await self.db_pool.runInteraction( "get_unread_push_actions_for_user_in_range_http_arr", get_after_receipt ) @@ -250,13 +278,14 @@ class EventPushActionsWorkerStore(SQLBaseStore): " AND ep.user_id = ?" " AND ep.stream_ordering > ?" " AND ep.stream_ordering <= ?" + " AND ep.notif = 1" " ORDER BY ep.stream_ordering ASC LIMIT ?" ) args = [user_id, user_id, min_stream_ordering, max_stream_ordering, limit] txn.execute(sql, args) return txn.fetchall() - no_read_receipt = yield self.db.runInteraction( + no_read_receipt = await self.db_pool.runInteraction( "get_unread_push_actions_for_user_in_range_http_nrr", get_no_receipt ) @@ -280,23 +309,25 @@ class EventPushActionsWorkerStore(SQLBaseStore): # one of the subqueries may have hit the limit. return notifs[:limit] - @defer.inlineCallbacks - def get_unread_push_actions_for_user_in_range_for_email( - self, user_id, min_stream_ordering, max_stream_ordering, limit=20 - ): + async def get_unread_push_actions_for_user_in_range_for_email( + self, + user_id: str, + min_stream_ordering: int, + max_stream_ordering: int, + limit: int = 20, + ) -> List[dict]: """Get a list of the most recent unread push actions for a given user, within the given stream ordering range. Called by the emailpusher Args: - user_id (str): The user to fetch push actions for. - min_stream_ordering(int): The exclusive lower bound on the + user_id: The user to fetch push actions for. + min_stream_ordering: The exclusive lower bound on the stream ordering of event push actions to fetch. - max_stream_ordering(int): The inclusive upper bound on the + max_stream_ordering: The inclusive upper bound on the stream ordering of event push actions to fetch. - limit (int): The maximum number of rows to return. + limit: The maximum number of rows to return. Returns: - A promise which resolves to a list of dicts with the keys "event_id", - "room_id", "stream_ordering", "actions", "received_ts". + A list of dicts with the keys "event_id", "room_id", "stream_ordering", "actions", "received_ts". The list will be ordered by descending received_ts. The list will have between 0~limit entries. """ @@ -322,13 +353,14 @@ class EventPushActionsWorkerStore(SQLBaseStore): " AND ep.user_id = ?" " AND ep.stream_ordering > ?" " AND ep.stream_ordering <= ?" + " AND ep.notif = 1" " ORDER BY ep.stream_ordering DESC LIMIT ?" ) args = [user_id, user_id, min_stream_ordering, max_stream_ordering, limit] txn.execute(sql, args) return txn.fetchall() - after_read_receipt = yield self.db.runInteraction( + after_read_receipt = await self.db_pool.runInteraction( "get_unread_push_actions_for_user_in_range_email_arr", get_after_receipt ) @@ -350,13 +382,14 @@ class EventPushActionsWorkerStore(SQLBaseStore): " AND ep.user_id = ?" " AND ep.stream_ordering > ?" " AND ep.stream_ordering <= ?" + " AND ep.notif = 1" " ORDER BY ep.stream_ordering DESC LIMIT ?" ) args = [user_id, user_id, min_stream_ordering, max_stream_ordering, limit] txn.execute(sql, args) return txn.fetchall() - no_read_receipt = yield self.db.runInteraction( + no_read_receipt = await self.db_pool.runInteraction( "get_unread_push_actions_for_user_in_range_email_nrr", get_no_receipt ) @@ -381,62 +414,66 @@ class EventPushActionsWorkerStore(SQLBaseStore): # Now return the first `limit` return notifs[:limit] - def get_if_maybe_push_in_range_for_user(self, user_id, min_stream_ordering): + async def get_if_maybe_push_in_range_for_user( + self, user_id: str, min_stream_ordering: int + ) -> bool: """A fast check to see if there might be something to push for the user since the given stream ordering. May return false positives. Useful to know whether to bother starting a pusher on start up or not. Args: - user_id (str) - min_stream_ordering (int) + user_id + min_stream_ordering Returns: - Deferred[bool]: True if there may be push to process, False if - there definitely isn't. + True if there may be push to process, False if there definitely isn't. """ def _get_if_maybe_push_in_range_for_user_txn(txn): sql = """ SELECT 1 FROM event_push_actions - WHERE user_id = ? AND stream_ordering > ? + WHERE user_id = ? AND stream_ordering > ? AND notif = 1 LIMIT 1 """ txn.execute(sql, (user_id, min_stream_ordering)) return bool(txn.fetchone()) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_if_maybe_push_in_range_for_user", _get_if_maybe_push_in_range_for_user_txn, ) - def add_push_actions_to_staging(self, event_id, user_id_actions): + async def add_push_actions_to_staging( + self, + event_id: str, + user_id_actions: Dict[str, List[Union[dict, str]]], + count_as_unread: bool, + ) -> None: """Add the push actions for the event to the push action staging area. Args: - event_id (str) - user_id_actions (dict[str, list[dict|str])]): A dictionary mapping - user_id to list of push actions, where an action can either be - a string or dict. - - Returns: - Deferred + event_id + user_id_actions: A mapping of user_id to list of push actions, where + an action can either be a string or dict. + count_as_unread: Whether this event should increment unread counts. """ - if not user_id_actions: return # This is a helper function for generating the necessary tuple that - # can be used to inert into the `event_push_actions_staging` table. + # can be used to insert into the `event_push_actions_staging` table. def _gen_entry(user_id, actions): is_highlight = 1 if _action_has_highlight(actions) else 0 + notif = 1 if "notify" in actions else 0 return ( event_id, # event_id column user_id, # user_id column _serialize_action(actions, is_highlight), # actions column - 1, # notif column + notif, # notif column is_highlight, # highlight column + int(count_as_unread), # unread column ) def _add_push_actions_to_staging_txn(txn): @@ -445,8 +482,8 @@ class EventPushActionsWorkerStore(SQLBaseStore): sql = """ INSERT INTO event_push_actions_staging - (event_id, user_id, actions, notif, highlight) - VALUES (?, ?, ?, ?, ?) + (event_id, user_id, actions, notif, highlight, unread) + VALUES (?, ?, ?, ?, ?, ?) """ txn.executemany( @@ -457,21 +494,17 @@ class EventPushActionsWorkerStore(SQLBaseStore): ), ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "add_push_actions_to_staging", _add_push_actions_to_staging_txn ) - @defer.inlineCallbacks - def remove_push_actions_from_staging(self, event_id): + async def remove_push_actions_from_staging(self, event_id: str) -> None: """Called if we failed to persist the event to ensure that stale push actions don't build up in the DB - - Args: - event_id (str) """ try: - res = yield self.db.simple_delete( + res = await self.db_pool.simple_delete( table="event_push_actions_staging", keyvalues={"event_id": event_id}, desc="remove_push_actions_from_staging", @@ -488,7 +521,7 @@ class EventPushActionsWorkerStore(SQLBaseStore): def _find_stream_orderings_for_times(self): return run_as_background_process( "event_push_action_stream_orderings", - self.db.runInteraction, + self.db_pool.runInteraction, "_find_stream_orderings_for_times", self._find_stream_orderings_for_times_txn, ) @@ -509,7 +542,7 @@ class EventPushActionsWorkerStore(SQLBaseStore): "Found stream ordering 1 day ago: it's %d", self.stream_ordering_day_ago ) - def find_first_stream_ordering_after_ts(self, ts): + async def find_first_stream_ordering_after_ts(self, ts: int) -> int: """Gets the stream ordering corresponding to a given timestamp. Specifically, finds the stream_ordering of the first event that was @@ -518,13 +551,12 @@ class EventPushActionsWorkerStore(SQLBaseStore): relatively slow. Args: - ts (int): timestamp in millis + ts: timestamp in millis Returns: - Deferred[int]: stream ordering of the first event received on/after - the timestamp + stream ordering of the first event received on/after the timestamp """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "_find_first_stream_ordering_after_ts_txn", self._find_first_stream_ordering_after_ts_txn, ts, @@ -606,38 +638,39 @@ class EventPushActionsWorkerStore(SQLBaseStore): return range_end - @defer.inlineCallbacks - def get_time_of_last_push_action_before(self, stream_ordering): + async def get_time_of_last_push_action_before(self, stream_ordering): def f(txn): sql = ( "SELECT e.received_ts" " FROM event_push_actions AS ep" " JOIN events e ON ep.room_id = e.room_id AND ep.event_id = e.event_id" - " WHERE ep.stream_ordering > ?" + " WHERE ep.stream_ordering > ? AND notif = 1" " ORDER BY ep.stream_ordering ASC" " LIMIT 1" ) txn.execute(sql, (stream_ordering,)) return txn.fetchone() - result = yield self.db.runInteraction("get_time_of_last_push_action_before", f) + result = await self.db_pool.runInteraction( + "get_time_of_last_push_action_before", f + ) return result[0] if result else None class EventPushActionsStore(EventPushActionsWorkerStore): EPA_HIGHLIGHT_INDEX = "epa_highlight_index" - def __init__(self, database: Database, db_conn, hs): - super(EventPushActionsStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( self.EPA_HIGHLIGHT_INDEX, index_name="event_push_actions_u_highlight", table="event_push_actions", columns=["user_id", "stream_ordering"], ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "event_push_actions_highlights_index", index_name="event_push_actions_highlights_index", table="event_push_actions", @@ -650,8 +683,7 @@ class EventPushActionsStore(EventPushActionsWorkerStore): self._start_rotate_notifs, 30 * 60 * 1000 ) - @defer.inlineCallbacks - def get_push_actions_for_user( + async def get_push_actions_for_user( self, user_id, before=None, limit=50, only_highlight=False ): def f(txn): @@ -676,24 +708,24 @@ class EventPushActionsStore(EventPushActionsWorkerStore): " FROM event_push_actions epa, events e" " WHERE epa.event_id = e.event_id" " AND epa.user_id = ? %s" + " AND epa.notif = 1" " ORDER BY epa.stream_ordering DESC" " LIMIT ?" % (before_clause,) ) txn.execute(sql, args) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - push_actions = yield self.db.runInteraction("get_push_actions_for_user", f) + push_actions = await self.db_pool.runInteraction("get_push_actions_for_user", f) for pa in push_actions: pa["actions"] = _deserialize_action(pa["actions"], pa["highlight"]) return push_actions - @defer.inlineCallbacks - def get_latest_push_action_stream_ordering(self): + async def get_latest_push_action_stream_ordering(self): def f(txn): txn.execute("SELECT MAX(stream_ordering) FROM event_push_actions") return txn.fetchone() - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( "get_latest_push_action_stream_ordering", f ) return result[0] or 0 @@ -747,8 +779,7 @@ class EventPushActionsStore(EventPushActionsWorkerStore): def _start_rotate_notifs(self): return run_as_background_process("rotate_notifs", self._rotate_notifs) - @defer.inlineCallbacks - def _rotate_notifs(self): + async def _rotate_notifs(self): if self._doing_notif_rotation or self.stream_ordering_day_ago is None: return self._doing_notif_rotation = True @@ -757,12 +788,12 @@ class EventPushActionsStore(EventPushActionsWorkerStore): while True: logger.info("Rotating notifications") - caught_up = yield self.db.runInteraction( + caught_up = await self.db_pool.runInteraction( "_rotate_notifs", self._rotate_notifs_txn ) if caught_up: break - yield self.hs.get_clock().sleep(self._rotate_delay) + await self.hs.get_clock().sleep(self._rotate_delay) finally: self._doing_notif_rotation = False @@ -771,7 +802,7 @@ class EventPushActionsStore(EventPushActionsWorkerStore): the archiving process has caught up or not. """ - old_rotate_stream_ordering = self.db.simple_select_one_onecol_txn( + old_rotate_stream_ordering = self.db_pool.simple_select_one_onecol_txn( txn, table="event_push_summary_stream_ordering", keyvalues={}, @@ -807,7 +838,7 @@ class EventPushActionsStore(EventPushActionsWorkerStore): return caught_up def _rotate_notifs_before_txn(self, txn, rotate_to_stream_ordering): - old_rotate_stream_ordering = self.db.simple_select_one_onecol_txn( + old_rotate_stream_ordering = self.db_pool.simple_select_one_onecol_txn( txn, table="event_push_summary_stream_ordering", keyvalues={}, @@ -817,49 +848,100 @@ class EventPushActionsStore(EventPushActionsWorkerStore): # Calculate the new counts that should be upserted into event_push_summary sql = """ SELECT user_id, room_id, - coalesce(old.notif_count, 0) + upd.notif_count, + coalesce(old.%s, 0) + upd.cnt, upd.stream_ordering, old.user_id FROM ( - SELECT user_id, room_id, count(*) as notif_count, + SELECT user_id, room_id, count(*) as cnt, max(stream_ordering) as stream_ordering FROM event_push_actions WHERE ? <= stream_ordering AND stream_ordering < ? AND highlight = 0 + AND %s = 1 GROUP BY user_id, room_id ) AS upd LEFT JOIN event_push_summary AS old USING (user_id, room_id) """ - txn.execute(sql, (old_rotate_stream_ordering, rotate_to_stream_ordering)) - rows = txn.fetchall() + # First get the count of unread messages. + txn.execute( + sql % ("unread_count", "unread"), + (old_rotate_stream_ordering, rotate_to_stream_ordering), + ) - logger.info("Rotating notifications, handling %d rows", len(rows)) + # We need to merge results from the two requests (the one that retrieves the + # unread count and the one that retrieves the notifications count) into a single + # object because we might not have the same amount of rows in each of them. To do + # this, we use a dict indexed on the user ID and room ID to make it easier to + # populate. + summaries = {} # type: Dict[Tuple[str, str], _EventPushSummary] + for row in txn: + summaries[(row[0], row[1])] = _EventPushSummary( + unread_count=row[2], + stream_ordering=row[3], + old_user_id=row[4], + notif_count=0, + ) + + # Then get the count of notifications. + txn.execute( + sql % ("notif_count", "notif"), + (old_rotate_stream_ordering, rotate_to_stream_ordering), + ) + + for row in txn: + if (row[0], row[1]) in summaries: + summaries[(row[0], row[1])].notif_count = row[2] + else: + # Because the rules on notifying are different than the rules on marking + # a message unread, we might end up with messages that notify but aren't + # marked unread, so we might not have a summary for this (user, room) + # tuple to complete. + summaries[(row[0], row[1])] = _EventPushSummary( + unread_count=0, + stream_ordering=row[3], + old_user_id=row[4], + notif_count=row[2], + ) + + logger.info("Rotating notifications, handling %d rows", len(summaries)) # If the `old.user_id` above is NULL then we know there isn't already an # entry in the table, so we simply insert it. Otherwise we update the # existing table. - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_push_summary", values=[ { - "user_id": row[0], - "room_id": row[1], - "notif_count": row[2], - "stream_ordering": row[3], + "user_id": user_id, + "room_id": room_id, + "notif_count": summary.notif_count, + "unread_count": summary.unread_count, + "stream_ordering": summary.stream_ordering, } - for row in rows - if row[4] is None + for ((user_id, room_id), summary) in summaries.items() + if summary.old_user_id is None ], ) txn.executemany( """ - UPDATE event_push_summary SET notif_count = ?, stream_ordering = ? + UPDATE event_push_summary + SET notif_count = ?, unread_count = ?, stream_ordering = ? WHERE user_id = ? AND room_id = ? """, - ((row[2], row[3], row[0], row[1]) for row in rows if row[4] is not None), + ( + ( + summary.notif_count, + summary.unread_count, + summary.stream_ordering, + user_id, + room_id, + ) + for ((user_id, room_id), summary) in summaries.items() + if summary.old_user_id is not None + ), ) txn.execute( @@ -885,3 +967,15 @@ def _action_has_highlight(actions): pass return False + + +@attr.s(slots=True) +class _EventPushSummary: + """Summary of pending event push actions for a given user in a given room. + Used in _rotate_notifs_before_txn to manipulate results from event_push_actions. + """ + + unread_count = attr.ib(type=int) + stream_ordering = attr.ib(type=int) + old_user_id = attr.ib(type=str) + notif_count = attr.ib(type=int) diff --git a/synapse/storage/data_stores/main/events.py b/synapse/storage/databases/main/events.py
index 6f2e0d15cc..18def01f50 100644 --- a/synapse/storage/data_stores/main/events.py +++ b/synapse/storage/databases/main/events.py
@@ -17,13 +17,11 @@ import itertools import logging from collections import OrderedDict, namedtuple -from typing import TYPE_CHECKING, Dict, Iterable, List, Tuple +from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Optional, Set, Tuple import attr from prometheus_client import Counter -from twisted.internet import defer - import synapse.metrics from synapse.api.constants import EventContentFields, EventTypes, RelationTypes from synapse.api.room_versions import RoomVersions @@ -32,16 +30,16 @@ from synapse.events import EventBase # noqa: F401 from synapse.events.snapshot import EventContext # noqa: F401 from synapse.logging.utils import log_function from synapse.storage._base import db_to_json, make_in_list_sql_clause -from synapse.storage.data_stores.main.search import SearchEntry -from synapse.storage.database import Database, LoggingTransaction -from synapse.storage.util.id_generators import StreamIdGenerator +from synapse.storage.database import DatabasePool, LoggingTransaction +from synapse.storage.databases.main.search import SearchEntry +from synapse.storage.util.id_generators import MultiWriterIdGenerator from synapse.types import StateMap, get_domain_from_id from synapse.util.frozenutils import frozendict_json_encoder from synapse.util.iterutils import batch_iter if TYPE_CHECKING: from synapse.server import HomeServer - from synapse.storage.data_stores.main import DataStore + from synapse.storage.databases.main import DataStore logger = logging.getLogger(__name__) @@ -91,35 +89,39 @@ class PersistEventsStore: Note: This is not part of the `DataStore` mixin. """ - def __init__(self, hs: "HomeServer", db: Database, main_data_store: "DataStore"): + def __init__( + self, hs: "HomeServer", db: DatabasePool, main_data_store: "DataStore" + ): self.hs = hs - self.db = db + self.db_pool = db self.store = main_data_store self.database_engine = db.engine self._clock = hs.get_clock() + self._instance_name = hs.get_instance_name() self._ephemeral_messages_enabled = hs.config.enable_ephemeral_messages self.is_mine_id = hs.is_mine_id # Ideally we'd move these ID gens here, unfortunately some other ID # generators are chained off them so doing so is a bit of a PITA. - self._backfill_id_gen = self.store._backfill_id_gen # type: StreamIdGenerator - self._stream_id_gen = self.store._stream_id_gen # type: StreamIdGenerator + self._backfill_id_gen = ( + self.store._backfill_id_gen + ) # type: MultiWriterIdGenerator + self._stream_id_gen = self.store._stream_id_gen # type: MultiWriterIdGenerator # This should only exist on instances that are configured to write assert ( - hs.config.worker.writers.events == hs.get_instance_name() + hs.get_instance_name() in hs.config.worker.writers.events ), "Can only instantiate EventsStore on master" - @defer.inlineCallbacks - def _persist_events_and_state_updates( + async def _persist_events_and_state_updates( self, events_and_contexts: List[Tuple[EventBase, EventContext]], current_state_for_room: Dict[str, StateMap[str]], state_delta_for_room: Dict[str, DeltaState], new_forward_extremeties: Dict[str, List[str]], backfilled: bool = False, - ): + ) -> None: """Persist a set of events alongside updates to the current state and forward extremities tables. @@ -134,7 +136,7 @@ class PersistEventsStore: backfilled Returns: - Deferred: resolves when the events have been persisted + Resolves when the events have been persisted """ # We want to calculate the stream orderings as late as possible, as @@ -162,11 +164,11 @@ class PersistEventsStore: len(events_and_contexts) ) - with stream_ordering_manager as stream_orderings: + async with stream_ordering_manager as stream_orderings: for (event, context), stream in zip(events_and_contexts, stream_orderings): event.internal_metadata.stream_ordering = stream - yield self.db.runInteraction( + await self.db_pool.runInteraction( "persist_events", self._persist_events_txn, events_and_contexts=events_and_contexts, @@ -204,18 +206,17 @@ class PersistEventsStore: (room_id,), list(latest_event_ids) ) - @defer.inlineCallbacks - def _get_events_which_are_prevs(self, event_ids): + async def _get_events_which_are_prevs(self, event_ids: Iterable[str]) -> List[str]: """Filter the supplied list of event_ids to get those which are prev_events of existing (non-outlier/rejected) events. Args: - event_ids (Iterable[str]): event ids to filter + event_ids: event ids to filter Returns: - Deferred[List[str]]: filtered event ids + Filtered event ids """ - results = [] + results = [] # type: List[str] def _get_events_which_are_prevs_txn(txn, batch): sql = """ @@ -238,14 +239,13 @@ class PersistEventsStore: results.extend(r[0] for r in txn if not db_to_json(r[1]).get("soft_failed")) for chunk in batch_iter(event_ids, 100): - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_get_events_which_are_prevs", _get_events_which_are_prevs_txn, chunk ) return results - @defer.inlineCallbacks - def _get_prevs_before_rejected(self, event_ids): + async def _get_prevs_before_rejected(self, event_ids: Iterable[str]) -> Set[str]: """Get soft-failed ancestors to remove from the extremities. Given a set of events, find all those that have been soft-failed or @@ -257,11 +257,11 @@ class PersistEventsStore: are separated by soft failed events. Args: - event_ids (Iterable[str]): Events to find prev events for. Note - that these must have already been persisted. + event_ids: Events to find prev events for. Note that these must have + already been persisted. Returns: - Deferred[set[str]] + The previous events. """ # The set of event_ids to return. This includes all soft-failed events @@ -302,7 +302,7 @@ class PersistEventsStore: existing_prevs.add(prev_event_id) for chunk in batch_iter(event_ids, 100): - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_get_prevs_before_rejected", _get_prevs_before_rejected_txn, chunk ) @@ -376,7 +376,7 @@ class PersistEventsStore: # event's auth chain, but its easier for now just to store them (and # it doesn't take much storage compared to storing the entire event # anyway). - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_auth", values=[ @@ -439,7 +439,7 @@ class PersistEventsStore: """ txn.execute(sql, (stream_id, room_id)) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="current_state_events", keyvalues={"room_id": room_id}, ) else: @@ -587,7 +587,7 @@ class PersistEventsStore: creator = content.get("creator") room_version_id = content.get("room_version", RoomVersions.V1.identifier) - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="rooms", keyvalues={"room_id": room_id}, @@ -599,14 +599,14 @@ class PersistEventsStore: self, txn, new_forward_extremities, max_stream_order ): for room_id, new_extrem in new_forward_extremities.items(): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="event_forward_extremities", keyvalues={"room_id": room_id} ) txn.call_after( self.store.get_latest_event_ids_in_room.invalidate, (room_id,) ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_forward_extremities", values=[ @@ -619,7 +619,7 @@ class PersistEventsStore: # new stream_ordering to new forward extremeties in the room. # This allows us to later efficiently look up the forward extremeties # for a room before a given stream_ordering - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="stream_ordering_to_exterm", values=[ @@ -634,7 +634,9 @@ class PersistEventsStore: ) @classmethod - def _filter_events_and_contexts_for_duplicates(cls, events_and_contexts): + def _filter_events_and_contexts_for_duplicates( + cls, events_and_contexts: List[Tuple[EventBase, EventContext]] + ) -> List[Tuple[EventBase, EventContext]]: """Ensure that we don't have the same event twice. Pick the earliest non-outlier if there is one, else the earliest one. @@ -644,7 +646,9 @@ class PersistEventsStore: Returns: list[(EventBase, EventContext)]: filtered list """ - new_events_and_contexts = OrderedDict() + new_events_and_contexts = ( + OrderedDict() + ) # type: OrderedDict[str, Tuple[EventBase, EventContext]] for event, context in events_and_contexts: prev_event_context = new_events_and_contexts.get(event.event_id) if prev_event_context: @@ -658,7 +662,12 @@ class PersistEventsStore: new_events_and_contexts[event.event_id] = (event, context) return list(new_events_and_contexts.values()) - def _update_room_depths_txn(self, txn, events_and_contexts, backfilled): + def _update_room_depths_txn( + self, + txn, + events_and_contexts: List[Tuple[EventBase, EventContext]], + backfilled: bool, + ): """Update min_depth for each room Args: @@ -667,7 +676,7 @@ class PersistEventsStore: we are persisting backfilled (bool): True if the events were backfilled """ - depth_updates = {} + depth_updates = {} # type: Dict[str, int] for event, context in events_and_contexts: # Remove the any existing cache entries for the event_ids txn.call_after(self.store._invalidate_get_event_cache, event.event_id) @@ -743,7 +752,7 @@ class PersistEventsStore: # change in outlier status to our workers. stream_order = event.internal_metadata.stream_ordering state_group_id = context.state_group - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="ex_outlier_stream", values={ @@ -781,7 +790,7 @@ class PersistEventsStore: d.pop("redacted_because", None) return d - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_json", values=[ @@ -798,11 +807,12 @@ class PersistEventsStore: ], ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="events", values=[ { + "instance_name": self._instance_name, "stream_ordering": event.internal_metadata.stream_ordering, "topological_ordering": event.depth, "depth": event.depth, @@ -827,7 +837,7 @@ class PersistEventsStore: # If we're persisting an unredacted event we go and ensure # that we mark any redactions that reference this event as # requiring censoring. - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn, table="redactions", keyvalues={"redacts": event.event_id}, @@ -969,7 +979,9 @@ class PersistEventsStore: state_values.append(vals) - self.db.simple_insert_many_txn(txn, table="state_events", values=state_values) + self.db_pool.simple_insert_many_txn( + txn, table="state_events", values=state_values + ) # Prefill the event cache self._add_to_cache(txn, events_and_contexts) @@ -1000,7 +1012,7 @@ class PersistEventsStore: ) txn.execute(sql + clause, args) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) for row in rows: event = ev_map[row["event_id"]] if not row["rejects"] and not row["redacts"]: @@ -1020,7 +1032,7 @@ class PersistEventsStore: # invalidate the cache for the redacted event txn.call_after(self.store._invalidate_get_event_cache, event.redacts) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="redactions", values={ @@ -1043,7 +1055,7 @@ class PersistEventsStore: room_id (str): The ID of the room the event was sent to. topological_ordering (int): The position of the event in the room's topology. """ - return self.db.simple_insert_many_txn( + return self.db_pool.simple_insert_many_txn( txn=txn, table="event_labels", values=[ @@ -1065,7 +1077,7 @@ class PersistEventsStore: event_id (str): The event ID the expiry timestamp is associated with. expiry_ts (int): The timestamp at which to expire (delete) the event. """ - return self.db.simple_insert_txn( + return self.db_pool.simple_insert_txn( txn=txn, table="event_expiry", values={"event_id": event_id, "expiry_ts": expiry_ts}, @@ -1089,12 +1101,18 @@ class PersistEventsStore: } ) - self.db.simple_insert_many_txn(txn, table="event_reference_hashes", values=vals) + self.db_pool.simple_insert_many_txn( + txn, table="event_reference_hashes", values=vals + ) def _store_room_members_txn(self, txn, events, backfilled): """Store a room member in the database. """ - self.db.simple_insert_many_txn( + + def str_or_none(val: Any) -> Optional[str]: + return val if isinstance(val, str) else None + + self.db_pool.simple_insert_many_txn( txn, table="room_memberships", values=[ @@ -1104,8 +1122,8 @@ class PersistEventsStore: "sender": event.user_id, "room_id": event.room_id, "membership": event.membership, - "display_name": event.content.get("displayname", None), - "avatar_url": event.content.get("avatar_url", None), + "display_name": str_or_none(event.content.get("displayname")), + "avatar_url": str_or_none(event.content.get("avatar_url")), } for event in events ], @@ -1134,7 +1152,7 @@ class PersistEventsStore: and event.internal_metadata.is_outlier() and event.internal_metadata.is_out_of_band_membership() ): - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="local_current_membership", keyvalues={"room_id": event.room_id, "user_id": event.state_key}, @@ -1172,7 +1190,7 @@ class PersistEventsStore: aggregation_key = relation.get("key") - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="event_relations", values={ @@ -1200,7 +1218,7 @@ class PersistEventsStore: redacted_event_id (str): The event that was redacted. """ - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="event_relations", keyvalues={"event_id": redacted_event_id} ) @@ -1236,7 +1254,7 @@ class PersistEventsStore: # Ignore the event if one of the value isn't an integer. return - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn=txn, table="room_retention", values={ @@ -1295,9 +1313,9 @@ class PersistEventsStore: sql = """ INSERT INTO event_push_actions ( room_id, event_id, user_id, actions, stream_ordering, - topological_ordering, notif, highlight + topological_ordering, notif, highlight, unread ) - SELECT ?, event_id, user_id, actions, ?, ?, notif, highlight + SELECT ?, event_id, user_id, actions, ?, ?, notif, highlight, unread FROM event_push_actions_staging WHERE event_id = ? """ @@ -1317,7 +1335,7 @@ class PersistEventsStore: ) for event, _ in events_and_contexts: - user_ids = self.db.simple_select_onecol_txn( + user_ids = self.db_pool.simple_select_onecol_txn( txn, table="event_push_actions_staging", keyvalues={"event_id": event.event_id}, @@ -1349,7 +1367,7 @@ class PersistEventsStore: ) def _store_rejections_txn(self, txn, event_id, reason): - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="rejections", values={ @@ -1375,7 +1393,7 @@ class PersistEventsStore: state_groups[event.event_id] = context.state_group - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_to_state_groups", values=[ @@ -1397,7 +1415,7 @@ class PersistEventsStore: if min_depth is not None and depth >= min_depth: return - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="room_depth", keyvalues={"room_id": room_id}, @@ -1409,7 +1427,7 @@ class PersistEventsStore: For the given event, update the event edges table and forward and backward extremities tables. """ - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="event_edges", values=[ @@ -1435,7 +1453,7 @@ class PersistEventsStore: Forward extremities are handled when we first start persisting the events. """ - events_by_room = {} + events_by_room = {} # type: Dict[str, List[EventBase]] for ev in events: events_by_room.setdefault(ev.room_id, []).append(ev) diff --git a/synapse/storage/data_stores/main/events_bg_updates.py b/synapse/storage/databases/main/events_bg_updates.py
index 663c94b24f..5e4af2eb51 100644 --- a/synapse/storage/data_stores/main/events_bg_updates.py +++ b/synapse/storage/databases/main/events_bg_updates.py
@@ -15,11 +15,9 @@ import logging -from twisted.internet import defer - from synapse.api.constants import EventContentFields from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool logger = logging.getLogger(__name__) @@ -30,18 +28,18 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): EVENT_FIELDS_SENDER_URL_UPDATE_NAME = "event_fields_sender_url" DELETE_SOFT_FAILED_EXTREMITIES = "delete_soft_failed_extremities" - def __init__(self, database: Database, db_conn, hs): - super(EventsBackgroundUpdatesStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_ORIGIN_SERVER_TS_NAME, self._background_reindex_origin_server_ts ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_FIELDS_SENDER_URL_UPDATE_NAME, self._background_reindex_fields_sender, ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "event_contains_url_index", index_name="event_contains_url_index", table="events", @@ -52,7 +50,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): # an event_id index on event_search is useful for the purge_history # api. Plus it means we get to enforce some integrity with a UNIQUE # clause - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "event_search_event_id_idx", index_name="event_search_event_id_idx", table="event_search", @@ -61,16 +59,16 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): psql_only=True, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.DELETE_SOFT_FAILED_EXTREMITIES, self._cleanup_extremities_bg_update ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "redactions_received_ts", self._redactions_received_ts ) # This index gets deleted in `event_fix_redactions_bytes` update - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "event_fix_redactions_bytes_create_index", index_name="redactions_censored_redacts", table="redactions", @@ -78,15 +76,15 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): where_clause="have_censored", ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "event_fix_redactions_bytes", self._event_fix_redactions_bytes ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "event_store_labels", self._event_store_labels ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "redactions_have_censored_ts_idx", index_name="redactions_have_censored_ts", table="redactions", @@ -94,8 +92,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): where_clause="NOT have_censored", ) - @defer.inlineCallbacks - def _background_reindex_fields_sender(self, progress, batch_size): + async def _background_reindex_fields_sender(self, progress, batch_size): target_min_stream_id = progress["target_min_stream_id_inclusive"] max_stream_id = progress["max_stream_id_exclusive"] rows_inserted = progress.get("rows_inserted", 0) @@ -149,25 +146,24 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): "rows_inserted": rows_inserted + len(rows), } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.EVENT_FIELDS_SENDER_URL_UPDATE_NAME, progress ) return len(rows) - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( self.EVENT_FIELDS_SENDER_URL_UPDATE_NAME, reindex_txn ) if not result: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.EVENT_FIELDS_SENDER_URL_UPDATE_NAME ) return result - @defer.inlineCallbacks - def _background_reindex_origin_server_ts(self, progress, batch_size): + async def _background_reindex_origin_server_ts(self, progress, batch_size): target_min_stream_id = progress["target_min_stream_id_inclusive"] max_stream_id = progress["max_stream_id_exclusive"] rows_inserted = progress.get("rows_inserted", 0) @@ -195,7 +191,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): chunks = [event_ids[i : i + 100] for i in range(0, len(event_ids), 100)] for chunk in chunks: - ev_rows = self.db.simple_select_many_txn( + ev_rows = self.db_pool.simple_select_many_txn( txn, table="event_json", column="event_id", @@ -228,25 +224,24 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): "rows_inserted": rows_inserted + len(rows_to_update), } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.EVENT_ORIGIN_SERVER_TS_NAME, progress ) return len(rows_to_update) - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( self.EVENT_ORIGIN_SERVER_TS_NAME, reindex_search_txn ) if not result: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.EVENT_ORIGIN_SERVER_TS_NAME ) return result - @defer.inlineCallbacks - def _cleanup_extremities_bg_update(self, progress, batch_size): + async def _cleanup_extremities_bg_update(self, progress, batch_size): """Background update to clean out extremities that should have been deleted previously. @@ -374,7 +369,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): to_delete.intersection_update(original_set) - deleted = self.db.simple_delete_many_txn( + deleted = self.db_pool.simple_delete_many_txn( txn=txn, table="event_forward_extremities", column="event_id", @@ -390,7 +385,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): if deleted: # We now need to invalidate the caches of these rooms - rows = self.db.simple_select_many_txn( + rows = self.db_pool.simple_select_many_txn( txn, table="events", column="event_id", @@ -404,7 +399,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): self.get_latest_event_ids_in_room.invalidate, (room_id,) ) - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn=txn, table="_extremities_to_check", column="event_id", @@ -414,26 +409,25 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): return len(original_set) - num_handled = yield self.db.runInteraction( + num_handled = await self.db_pool.runInteraction( "_cleanup_extremities_bg_update", _cleanup_extremities_bg_update_txn ) if not num_handled: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.DELETE_SOFT_FAILED_EXTREMITIES ) def _drop_table_txn(txn): txn.execute("DROP TABLE _extremities_to_check") - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_cleanup_extremities_bg_update_drop_table", _drop_table_txn ) return num_handled - @defer.inlineCallbacks - def _redactions_received_ts(self, progress, batch_size): + async def _redactions_received_ts(self, progress, batch_size): """Handles filling out the `received_ts` column in redactions. """ last_event_id = progress.get("last_event_id", "") @@ -474,23 +468,22 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): txn.execute(sql, (self._clock.time_msec(), last_event_id, upper_event_id)) - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "redactions_received_ts", {"last_event_id": upper_event_id} ) return len(rows) - count = yield self.db.runInteraction( + count = await self.db_pool.runInteraction( "_redactions_received_ts", _redactions_received_ts_txn ) if not count: - yield self.db.updates._end_background_update("redactions_received_ts") + await self.db_pool.updates._end_background_update("redactions_received_ts") return count - @defer.inlineCallbacks - def _event_fix_redactions_bytes(self, progress, batch_size): + async def _event_fix_redactions_bytes(self, progress, batch_size): """Undoes hex encoded censored redacted event JSON. """ @@ -511,16 +504,15 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): txn.execute("DROP INDEX redactions_censored_redacts") - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_event_fix_redactions_bytes", _event_fix_redactions_bytes_txn ) - yield self.db.updates._end_background_update("event_fix_redactions_bytes") + await self.db_pool.updates._end_background_update("event_fix_redactions_bytes") return 1 - @defer.inlineCallbacks - def _event_store_labels(self, progress, batch_size): + async def _event_store_labels(self, progress, batch_size): """Background update handler which will store labels for existing events.""" last_event_id = progress.get("last_event_id", "") @@ -543,7 +535,7 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): try: event_json = db_to_json(event_json_raw) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn=txn, table="event_labels", values=[ @@ -569,17 +561,17 @@ class EventsBackgroundUpdatesStore(SQLBaseStore): nbrows += 1 last_row_event_id = event_id - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "event_store_labels", {"last_event_id": last_row_event_id} ) return nbrows - num_rows = yield self.db.runInteraction( + num_rows = await self.db_pool.runInteraction( desc="event_store_labels", func=_event_store_labels_txn ) if not num_rows: - yield self.db.updates._end_background_update("event_store_labels") + await self.db_pool.updates._end_background_update("event_store_labels") return num_rows diff --git a/synapse/storage/data_stores/main/events_worker.py b/synapse/storage/databases/main/events_worker.py
index e812c67078..f95679ebc4 100644 --- a/synapse/storage/data_stores/main/events_worker.py +++ b/synapse/storage/databases/main/events_worker.py
@@ -13,15 +13,14 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import division - import itertools import logging import threading from collections import namedtuple -from typing import List, Optional, Tuple +from typing import Dict, Iterable, List, Optional, Tuple, overload from constantly import NamedConstant, Names +from typing_extensions import Literal from twisted.internet import defer @@ -32,7 +31,7 @@ from synapse.api.room_versions import ( EventFormatVersions, RoomVersions, ) -from synapse.events import make_event_from_dict +from synapse.events import EventBase, make_event_from_dict from synapse.events.utils import prune_event from synapse.logging.context import PreserveLoggingContext, current_context from synapse.metrics.background_process_metrics import run_as_background_process @@ -40,10 +39,11 @@ from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.replication.tcp.streams import BackfillStream from synapse.replication.tcp.streams.events import EventsStream from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause -from synapse.storage.database import Database -from synapse.storage.util.id_generators import StreamIdGenerator -from synapse.types import get_domain_from_id -from synapse.util.caches.descriptors import Cache, cached, cachedInlineCallbacks +from synapse.storage.database import DatabasePool +from synapse.storage.engines import PostgresEngine +from synapse.storage.util.id_generators import MultiWriterIdGenerator, StreamIdGenerator +from synapse.types import Collection, get_domain_from_id +from synapse.util.caches.descriptors import Cache, cached from synapse.util.iterutils import batch_iter from synapse.util.metrics import Measure @@ -74,30 +74,61 @@ class EventRedactBehaviour(Names): class EventsWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(EventsWorkerStore, self).__init__(database, db_conn, hs) - - if hs.config.worker.writers.events == hs.get_instance_name(): - # We are the process in charge of generating stream ids for events, - # so instantiate ID generators based on the database - self._stream_id_gen = StreamIdGenerator( - db_conn, "events", "stream_ordering", + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) + + if isinstance(database.engine, PostgresEngine): + # If we're using Postgres than we can use `MultiWriterIdGenerator` + # regardless of whether this process writes to the streams or not. + self._stream_id_gen = MultiWriterIdGenerator( + db_conn=db_conn, + db=database, + stream_name="events", + instance_name=hs.get_instance_name(), + table="events", + instance_column="instance_name", + id_column="stream_ordering", + sequence_name="events_stream_seq", + writers=hs.config.worker.writers.events, ) - self._backfill_id_gen = StreamIdGenerator( - db_conn, - "events", - "stream_ordering", - step=-1, - extra_tables=[("ex_outlier_stream", "event_stream_ordering")], + self._backfill_id_gen = MultiWriterIdGenerator( + db_conn=db_conn, + db=database, + stream_name="backfill", + instance_name=hs.get_instance_name(), + table="events", + instance_column="instance_name", + id_column="stream_ordering", + sequence_name="events_backfill_stream_seq", + positive=False, + writers=hs.config.worker.writers.events, ) else: - # Another process is in charge of persisting events and generating - # stream IDs: rely on the replication streams to let us know which - # IDs we can process. - self._stream_id_gen = SlavedIdTracker(db_conn, "events", "stream_ordering") - self._backfill_id_gen = SlavedIdTracker( - db_conn, "events", "stream_ordering", step=-1 - ) + # We shouldn't be running in worker mode with SQLite, but its useful + # to support it for unit tests. + # + # If this process is the writer than we need to use + # `StreamIdGenerator`, otherwise we use `SlavedIdTracker` which gets + # updated over replication. (Multiple writers are not supported for + # SQLite). + if hs.get_instance_name() in hs.config.worker.writers.events: + self._stream_id_gen = StreamIdGenerator( + db_conn, "events", "stream_ordering", + ) + self._backfill_id_gen = StreamIdGenerator( + db_conn, + "events", + "stream_ordering", + step=-1, + extra_tables=[("ex_outlier_stream", "event_stream_ordering")], + ) + else: + self._stream_id_gen = SlavedIdTracker( + db_conn, "events", "stream_ordering" + ) + self._backfill_id_gen = SlavedIdTracker( + db_conn, "events", "stream_ordering", step=-1 + ) self._get_event_cache = Cache( "*getEvent*", @@ -112,69 +143,58 @@ class EventsWorkerStore(SQLBaseStore): def process_replication_rows(self, stream_name, instance_name, token, rows): if stream_name == EventsStream.NAME: - self._stream_id_gen.advance(token) + self._stream_id_gen.advance(instance_name, token) elif stream_name == BackfillStream.NAME: - self._backfill_id_gen.advance(-token) + self._backfill_id_gen.advance(instance_name, -token) super().process_replication_rows(stream_name, instance_name, token, rows) - def get_received_ts(self, event_id): + async def get_received_ts(self, event_id: str) -> Optional[int]: """Get received_ts (when it was persisted) for the event. Raises an exception for unknown events. Args: - event_id (str) + event_id: The event ID to query. Returns: - Deferred[int|None]: Timestamp in milliseconds, or None for events - that were persisted before received_ts was implemented. + Timestamp in milliseconds, or None for events that were persisted + before received_ts was implemented. """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="events", keyvalues={"event_id": event_id}, retcol="received_ts", desc="get_received_ts", ) - def get_received_ts_by_stream_pos(self, stream_ordering): - """Given a stream ordering get an approximate timestamp of when it - happened. - - This is done by simply taking the received ts of the first event that - has a stream ordering greater than or equal to the given stream pos. - If none exists returns the current time, on the assumption that it must - have happened recently. - - Args: - stream_ordering (int) - - Returns: - Deferred[int] - """ - - def _get_approximate_received_ts_txn(txn): - sql = """ - SELECT received_ts FROM events - WHERE stream_ordering >= ? - LIMIT 1 - """ - - txn.execute(sql, (stream_ordering,)) - row = txn.fetchone() - if row and row[0]: - ts = row[0] - else: - ts = self.clock.time_msec() - - return ts + # Inform mypy that if allow_none is False (the default) then get_event + # always returns an EventBase. + @overload + async def get_event( + self, + event_id: str, + redact_behaviour: EventRedactBehaviour = EventRedactBehaviour.REDACT, + get_prev_content: bool = False, + allow_rejected: bool = False, + allow_none: Literal[False] = False, + check_room_id: Optional[str] = None, + ) -> EventBase: + ... - return self.db.runInteraction( - "get_approximate_received_ts", _get_approximate_received_ts_txn - ) + @overload + async def get_event( + self, + event_id: str, + redact_behaviour: EventRedactBehaviour = EventRedactBehaviour.REDACT, + get_prev_content: bool = False, + allow_rejected: bool = False, + allow_none: Literal[True] = False, + check_room_id: Optional[str] = None, + ) -> Optional[EventBase]: + ... - @defer.inlineCallbacks - def get_event( + async def get_event( self, event_id: str, redact_behaviour: EventRedactBehaviour = EventRedactBehaviour.REDACT, @@ -182,7 +202,7 @@ class EventsWorkerStore(SQLBaseStore): allow_rejected: bool = False, allow_none: bool = False, check_room_id: Optional[str] = None, - ): + ) -> Optional[EventBase]: """Get an event from the database by event_id. Args: @@ -207,12 +227,12 @@ class EventsWorkerStore(SQLBaseStore): If there is a mismatch, behave as per allow_none. Returns: - Deferred[EventBase|None] + The event, or None if the event was not found. """ if not isinstance(event_id, str): raise TypeError("Invalid event event_id %r" % (event_id,)) - events = yield self.get_events_as_list( + events = await self.get_events_as_list( [event_id], redact_behaviour=redact_behaviour, get_prev_content=get_prev_content, @@ -230,14 +250,13 @@ class EventsWorkerStore(SQLBaseStore): return event - @defer.inlineCallbacks - def get_events( + async def get_events( self, - event_ids: List[str], + event_ids: Iterable[str], redact_behaviour: EventRedactBehaviour = EventRedactBehaviour.REDACT, get_prev_content: bool = False, allow_rejected: bool = False, - ): + ) -> Dict[str, EventBase]: """Get events from the database Args: @@ -256,9 +275,9 @@ class EventsWorkerStore(SQLBaseStore): omits rejeted events from the response. Returns: - Deferred : Dict from event_id to event. + A mapping from event_id to event. """ - events = yield self.get_events_as_list( + events = await self.get_events_as_list( event_ids, redact_behaviour=redact_behaviour, get_prev_content=get_prev_content, @@ -267,14 +286,13 @@ class EventsWorkerStore(SQLBaseStore): return {e.event_id: e for e in events} - @defer.inlineCallbacks - def get_events_as_list( + async def get_events_as_list( self, - event_ids: List[str], + event_ids: Collection[str], redact_behaviour: EventRedactBehaviour = EventRedactBehaviour.REDACT, get_prev_content: bool = False, allow_rejected: bool = False, - ): + ) -> List[EventBase]: """Get events from the database and return in a list in the same order as given by `event_ids` arg. @@ -295,8 +313,8 @@ class EventsWorkerStore(SQLBaseStore): omits rejected events from the response. Returns: - Deferred[list[EventBase]]: List of events fetched from the database. The - events are in the same order as `event_ids` arg. + List of events fetched from the database. The events are in the same + order as `event_ids` arg. Note that the returned list may be smaller than the list of event IDs if not all events could be fetched. @@ -306,7 +324,7 @@ class EventsWorkerStore(SQLBaseStore): return [] # there may be duplicates so we cast the list to a set - event_entry_map = yield self._get_events_from_cache_or_db( + event_entry_map = await self._get_events_from_cache_or_db( set(event_ids), allow_rejected=allow_rejected ) @@ -341,7 +359,7 @@ class EventsWorkerStore(SQLBaseStore): continue redacted_event_id = entry.event.redacts - event_map = yield self._get_events_from_cache_or_db([redacted_event_id]) + event_map = await self._get_events_from_cache_or_db([redacted_event_id]) original_event_entry = event_map.get(redacted_event_id) if not original_event_entry: # we don't have the redacted event (or it was rejected). @@ -407,7 +425,7 @@ class EventsWorkerStore(SQLBaseStore): if get_prev_content: if "replaces_state" in event.unsigned: - prev = yield self.get_event( + prev = await self.get_event( event.unsigned["replaces_state"], get_prev_content=False, allow_none=True, @@ -419,8 +437,7 @@ class EventsWorkerStore(SQLBaseStore): return events - @defer.inlineCallbacks - def _get_events_from_cache_or_db(self, event_ids, allow_rejected=False): + async def _get_events_from_cache_or_db(self, event_ids, allow_rejected=False): """Fetch a bunch of events from the cache or the database. If events are pulled from the database, they will be cached for future lookups. @@ -435,7 +452,7 @@ class EventsWorkerStore(SQLBaseStore): rejected events are omitted from the response. Returns: - Deferred[Dict[str, _EventCacheEntry]]: + Dict[str, _EventCacheEntry]: map from event id to result """ event_entry_map = self._get_events_from_cache( @@ -453,7 +470,7 @@ class EventsWorkerStore(SQLBaseStore): # the events have been redacted, and if so pulling the redaction event out # of the database to check it. # - missing_events = yield self._get_events_from_db( + missing_events = await self._get_events_from_db( missing_events_ids, allow_rejected=allow_rejected ) @@ -537,7 +554,7 @@ class EventsWorkerStore(SQLBaseStore): event_id for events, _ in event_list for event_id in events } - row_dict = self.db.new_transaction( + row_dict = self.db_pool.new_transaction( conn, "do_fetch", [], [], self._fetch_event_rows, events_to_fetch ) @@ -561,8 +578,7 @@ class EventsWorkerStore(SQLBaseStore): with PreserveLoggingContext(): self.hs.get_reactor().callFromThread(fire, event_list, e) - @defer.inlineCallbacks - def _get_events_from_db(self, event_ids, allow_rejected=False): + async def _get_events_from_db(self, event_ids, allow_rejected=False): """Fetch a bunch of events from the database. Returned events will be added to the cache for future lookups. @@ -576,7 +592,7 @@ class EventsWorkerStore(SQLBaseStore): rejected events are omitted from the response. Returns: - Deferred[Dict[str, _EventCacheEntry]]: + Dict[str, _EventCacheEntry]: map from event id to result. May return extra events which weren't asked for. """ @@ -584,7 +600,7 @@ class EventsWorkerStore(SQLBaseStore): events_to_fetch = event_ids while events_to_fetch: - row_map = yield self._enqueue_events(events_to_fetch) + row_map = await self._enqueue_events(events_to_fetch) # we need to recursively fetch any redactions of those events redaction_ids = set() @@ -610,8 +626,20 @@ class EventsWorkerStore(SQLBaseStore): if not allow_rejected and rejected_reason: continue - d = db_to_json(row["json"]) - internal_metadata = db_to_json(row["internal_metadata"]) + # If the event or metadata cannot be parsed, log the error and act + # as if the event is unknown. + try: + d = db_to_json(row["json"]) + except ValueError: + logger.error("Unable to parse json from event: %s", event_id) + continue + try: + internal_metadata = db_to_json(row["internal_metadata"]) + except ValueError: + logger.error( + "Unable to parse internal_metadata from event: %s", event_id + ) + continue format_version = row["format_version"] if format_version is None: @@ -622,19 +650,38 @@ class EventsWorkerStore(SQLBaseStore): room_version_id = row["room_version_id"] if not room_version_id: - # this should only happen for out-of-band membership events - if not internal_metadata.get("out_of_band_membership"): - logger.warning( - "Room %s for event %s is unknown", d["room_id"], event_id + # this should only happen for out-of-band membership events which + # arrived before #6983 landed. For all other events, we should have + # an entry in the 'rooms' table. + # + # However, the 'out_of_band_membership' flag is unreliable for older + # invites, so just accept it for all membership events. + # + if d["type"] != EventTypes.Member: + raise Exception( + "Room %s for event %s is unknown" % (d["room_id"], event_id) ) - continue - # take a wild stab at the room version based on the event format + # so, assuming this is an out-of-band-invite that arrived before #6983 + # landed, we know that the room version must be v5 or earlier (because + # v6 hadn't been invented at that point, so invites from such rooms + # would have been rejected.) + # + # The main reason we need to know the room version here (other than + # choosing the right python Event class) is in case the event later has + # to be redacted - and all the room versions up to v5 used the same + # redaction algorithm. + # + # So, the following approximations should be adequate. + if format_version == EventFormatVersions.V1: + # if it's event format v1 then it must be room v1 or v2 room_version = RoomVersions.V1 elif format_version == EventFormatVersions.V2: + # if it's event format v2 then it must be room v3 room_version = RoomVersions.V3 else: + # if it's event format v3 then it must be room v4 or v5 room_version = RoomVersions.V5 else: room_version = KNOWN_ROOM_VERSIONS.get(room_version_id) @@ -686,8 +733,7 @@ class EventsWorkerStore(SQLBaseStore): return result_map - @defer.inlineCallbacks - def _enqueue_events(self, events): + async def _enqueue_events(self, events): """Fetches events from the database using the _event_fetch_list. This allows batch and bulk fetching of events - it allows us to fetch events without having to create a new transaction for each request for events. @@ -696,7 +742,7 @@ class EventsWorkerStore(SQLBaseStore): events (Iterable[str]): events to be fetched. Returns: - Deferred[Dict[str, Dict]]: map from event id to row data from the database. + Dict[str, Dict]: map from event id to row data from the database. May contain events that weren't requested. """ @@ -714,12 +760,12 @@ class EventsWorkerStore(SQLBaseStore): if should_start: run_as_background_process( - "fetch_events", self.db.runWithConnection, self._do_fetch + "fetch_events", self.db_pool.runWithConnection, self._do_fetch ) logger.debug("Loading %d events: %s", len(events), events) with PreserveLoggingContext(): - row_map = yield events_d + row_map = await events_d logger.debug("Loaded %d events (%d rows)", len(events), len(row_map)) return row_map @@ -807,20 +853,24 @@ class EventsWorkerStore(SQLBaseStore): return event_dict - def _maybe_redact_event_row(self, original_ev, redactions, event_map): + def _maybe_redact_event_row( + self, + original_ev: EventBase, + redactions: Iterable[str], + event_map: Dict[str, EventBase], + ) -> Optional[EventBase]: """Given an event object and a list of possible redacting event ids, determine whether to honour any of those redactions and if so return a redacted event. Args: - original_ev (EventBase): - redactions (iterable[str]): list of event ids of potential redaction events - event_map (dict[str, EventBase]): other events which have been fetched, in - which we can look up the redaaction events. Map from event id to event. + original_ev: The original event. + redactions: list of event ids of potential redaction events + event_map: other events which have been fetched, in which we can + look up the redaaction events. Map from event id to event. Returns: - Deferred[EventBase|None]: if the event should be redacted, a pruned - event object. Otherwise, None. + If the event should be redacted, a pruned event object. Otherwise, None. """ if original_ev.type == "m.room.create": # we choose to ignore redactions of m.room.create events. @@ -878,12 +928,11 @@ class EventsWorkerStore(SQLBaseStore): # no valid redaction found for this event return None - @defer.inlineCallbacks - def have_events_in_timeline(self, event_ids): + async def have_events_in_timeline(self, event_ids): """Given a list of event ids, check if we have already processed and stored them as non outliers. """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="events", retcols=("event_id",), column="event_id", @@ -894,15 +943,14 @@ class EventsWorkerStore(SQLBaseStore): return {r["event_id"] for r in rows} - @defer.inlineCallbacks - def have_seen_events(self, event_ids): + async def have_seen_events(self, event_ids): """Given a list of event ids, check if we have already processed them. Args: event_ids (iterable[str]): Returns: - Deferred[set[str]]: The events we have already seen. + set[str]: The events we have already seen. """ results = set() @@ -918,41 +966,11 @@ class EventsWorkerStore(SQLBaseStore): # break the input up into chunks of 100 input_iterator = iter(event_ids) for chunk in iter(lambda: list(itertools.islice(input_iterator, 100)), []): - yield self.db.runInteraction( + await self.db_pool.runInteraction( "have_seen_events", have_seen_events_txn, chunk ) return results - def _get_total_state_event_counts_txn(self, txn, room_id): - """ - See get_total_state_event_counts. - """ - # We join against the events table as that has an index on room_id - sql = """ - SELECT COUNT(*) FROM state_events - INNER JOIN events USING (room_id, event_id) - WHERE room_id=? - """ - txn.execute(sql, (room_id,)) - row = txn.fetchone() - return row[0] if row else 0 - - def get_total_state_event_counts(self, room_id): - """ - Gets the total number of state events in a room. - - Args: - room_id (str) - - Returns: - Deferred[int] - """ - return self.db.runInteraction( - "get_total_state_event_counts", - self._get_total_state_event_counts_txn, - room_id, - ) - def _get_current_state_event_counts_txn(self, txn, room_id): """ See get_current_state_event_counts. @@ -962,24 +980,23 @@ class EventsWorkerStore(SQLBaseStore): row = txn.fetchone() return row[0] if row else 0 - def get_current_state_event_counts(self, room_id): + async def get_current_state_event_counts(self, room_id: str) -> int: """ Gets the current number of state events in a room. Args: - room_id (str) + room_id: The room ID to query. Returns: - Deferred[int] + The current number of state events. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_current_state_event_counts", self._get_current_state_event_counts_txn, room_id, ) - @defer.inlineCallbacks - def get_room_complexity(self, room_id): + async def get_room_complexity(self, room_id): """ Get a rough approximation of the complexity of the room. This is used by remote servers to decide whether they wish to join the room or not. @@ -990,9 +1007,9 @@ class EventsWorkerStore(SQLBaseStore): room_id (str) Returns: - Deferred[dict[str:int]] of complexity version to complexity. + dict[str:int] of complexity version to complexity. """ - state_events = yield self.get_current_state_event_counts(room_id) + state_events = await self.get_current_state_event_counts(room_id) # Call this one "v1", so we can introduce new ones as we want to develop # it. @@ -1008,7 +1025,9 @@ class EventsWorkerStore(SQLBaseStore): """The current maximum token that events have reached""" return self._stream_id_gen.get_current_token() - def get_all_new_forward_event_rows(self, last_id, current_id, limit): + async def get_all_new_forward_event_rows( + self, last_id: int, current_id: int, limit: int + ) -> List[Tuple]: """Returns new events, for the Events replication stream Args: @@ -1016,7 +1035,7 @@ class EventsWorkerStore(SQLBaseStore): current_id: the maximum stream_id to return up to limit: the maximum number of rows to return - Returns: Deferred[List[Tuple]] + Returns: a list of events stream rows. Each tuple consists of a stream id as the first element, followed by fields suitable for casting into an EventsStreamRow. @@ -1037,18 +1056,20 @@ class EventsWorkerStore(SQLBaseStore): txn.execute(sql, (last_id, current_id, limit)) return txn.fetchall() - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_new_forward_event_rows", get_all_new_forward_event_rows ) - def get_ex_outlier_stream_rows(self, last_id, current_id): + async def get_ex_outlier_stream_rows( + self, last_id: int, current_id: int + ) -> List[Tuple]: """Returns de-outliered events, for the Events replication stream Args: last_id: the last stream_id from the previous batch. current_id: the maximum stream_id to return up to - Returns: Deferred[List[Tuple]] + Returns: a list of events stream rows. Each tuple consists of a stream id as the first element, followed by fields suitable for casting into an EventsStreamRow. @@ -1071,7 +1092,7 @@ class EventsWorkerStore(SQLBaseStore): txn.execute(sql, (last_id, current_id)) return txn.fetchall() - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_ex_outlier_stream_rows", get_ex_outlier_stream_rows_txn ) @@ -1145,7 +1166,7 @@ class EventsWorkerStore(SQLBaseStore): return new_event_updates, upper_bound, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_new_backfill_event_rows", get_all_new_backfill_event_rows ) @@ -1193,7 +1214,7 @@ class EventsWorkerStore(SQLBaseStore): # we need to make sure that, for every stream id in the results, we get *all* # the rows with that stream id. - rows = await self.db.runInteraction( + rows = await self.db_pool.runInteraction( "get_all_updated_current_state_deltas", get_all_updated_current_state_deltas_txn, ) # type: List[Tuple] @@ -1216,103 +1237,12 @@ class EventsWorkerStore(SQLBaseStore): # stream id. let's run the query again, without a row limit, but for # just one stream id. to_token += 1 - rows = await self.db.runInteraction( + rows = await self.db_pool.runInteraction( "get_deltas_for_stream_id", get_deltas_for_stream_id_txn, to_token ) return rows, to_token, True - @cached(num_args=5, max_entries=10) - def get_all_new_events( - self, - last_backfill_id, - last_forward_id, - current_backfill_id, - current_forward_id, - limit, - ): - """Get all the new events that have arrived at the server either as - new events or as backfilled events""" - have_backfill_events = last_backfill_id != current_backfill_id - have_forward_events = last_forward_id != current_forward_id - - if not have_backfill_events and not have_forward_events: - return defer.succeed(AllNewEventsResult([], [], [], [], [])) - - def get_all_new_events_txn(txn): - sql = ( - "SELECT e.stream_ordering, e.event_id, e.room_id, e.type," - " state_key, redacts" - " FROM events AS e" - " LEFT JOIN redactions USING (event_id)" - " LEFT JOIN state_events USING (event_id)" - " WHERE ? < stream_ordering AND stream_ordering <= ?" - " ORDER BY stream_ordering ASC" - " LIMIT ?" - ) - if have_forward_events: - txn.execute(sql, (last_forward_id, current_forward_id, limit)) - new_forward_events = txn.fetchall() - - if len(new_forward_events) == limit: - upper_bound = new_forward_events[-1][0] - else: - upper_bound = current_forward_id - - sql = ( - "SELECT event_stream_ordering, event_id, state_group" - " FROM ex_outlier_stream" - " WHERE ? > event_stream_ordering" - " AND event_stream_ordering >= ?" - " ORDER BY event_stream_ordering DESC" - ) - txn.execute(sql, (last_forward_id, upper_bound)) - forward_ex_outliers = txn.fetchall() - else: - new_forward_events = [] - forward_ex_outliers = [] - - sql = ( - "SELECT -e.stream_ordering, e.event_id, e.room_id, e.type," - " state_key, redacts" - " FROM events AS e" - " LEFT JOIN redactions USING (event_id)" - " LEFT JOIN state_events USING (event_id)" - " WHERE ? > stream_ordering AND stream_ordering >= ?" - " ORDER BY stream_ordering DESC" - " LIMIT ?" - ) - if have_backfill_events: - txn.execute(sql, (-last_backfill_id, -current_backfill_id, limit)) - new_backfill_events = txn.fetchall() - - if len(new_backfill_events) == limit: - upper_bound = new_backfill_events[-1][0] - else: - upper_bound = current_backfill_id - - sql = ( - "SELECT -event_stream_ordering, event_id, state_group" - " FROM ex_outlier_stream" - " WHERE ? > event_stream_ordering" - " AND event_stream_ordering >= ?" - " ORDER BY event_stream_ordering DESC" - ) - txn.execute(sql, (-last_backfill_id, -upper_bound)) - backward_ex_outliers = txn.fetchall() - else: - new_backfill_events = [] - backward_ex_outliers = [] - - return AllNewEventsResult( - new_forward_events, - new_backfill_events, - forward_ex_outliers, - backward_ex_outliers, - ) - - return self.db.runInteraction("get_all_new_events", get_all_new_events_txn) - async def is_event_after(self, event_id1, event_id2): """Returns True if event_id1 is after event_id2 in the stream """ @@ -1320,9 +1250,9 @@ class EventsWorkerStore(SQLBaseStore): to_2, so_2 = await self.get_event_ordering(event_id2) return (to_1, so_1) > (to_2, so_2) - @cachedInlineCallbacks(max_entries=5000) - def get_event_ordering(self, event_id): - res = yield self.db.simple_select_one( + @cached(max_entries=5000) + async def get_event_ordering(self, event_id): + res = await self.db_pool.simple_select_one( table="events", retcols=["topological_ordering", "stream_ordering"], keyvalues={"event_id": event_id}, @@ -1334,11 +1264,11 @@ class EventsWorkerStore(SQLBaseStore): return (int(res["topological_ordering"]), int(res["stream_ordering"])) - def get_next_event_to_expire(self): + async def get_next_event_to_expire(self) -> Optional[Tuple[str, int]]: """Retrieve the entry with the lowest expiry timestamp in the event_expiry table, or None if there's no more event to expire. - Returns: Deferred[Optional[Tuple[str, int]]] + Returns: A tuple containing the event ID as its first element and an expiry timestamp as its second one, if there's at least one row in the event_expiry table. None otherwise. @@ -1354,17 +1284,6 @@ class EventsWorkerStore(SQLBaseStore): return txn.fetchone() - return self.db.runInteraction( + return await self.db_pool.runInteraction( desc="get_next_event_to_expire", func=get_next_event_to_expire_txn ) - - -AllNewEventsResult = namedtuple( - "AllNewEventsResult", - [ - "new_forward_events", - "new_backfill_events", - "forward_ex_outliers", - "backward_ex_outliers", - ], -) diff --git a/synapse/storage/data_stores/main/filtering.py b/synapse/storage/databases/main/filtering.py
index 342d6622a4..d2f5b9a502 100644 --- a/synapse/storage/data_stores/main/filtering.py +++ b/synapse/storage/databases/main/filtering.py
@@ -17,12 +17,13 @@ from canonicaljson import encode_canonical_json from synapse.api.errors import Codes, SynapseError from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.util.caches.descriptors import cachedInlineCallbacks +from synapse.types import JsonDict +from synapse.util.caches.descriptors import cached class FilteringStore(SQLBaseStore): - @cachedInlineCallbacks(num_args=2) - def get_user_filter(self, user_localpart, filter_id): + @cached(num_args=2) + async def get_user_filter(self, user_localpart, filter_id): # filter_id is BIGINT UNSIGNED, so if it isn't a number, fail # with a coherent error message rather than 500 M_UNKNOWN. try: @@ -30,7 +31,7 @@ class FilteringStore(SQLBaseStore): except ValueError: raise SynapseError(400, "Invalid filter ID", Codes.INVALID_PARAM) - def_json = yield self.db.simple_select_one_onecol( + def_json = await self.db_pool.simple_select_one_onecol( table="user_filters", keyvalues={"user_id": user_localpart, "filter_id": filter_id}, retcol="filter_json", @@ -40,7 +41,7 @@ class FilteringStore(SQLBaseStore): return db_to_json(def_json) - def add_user_filter(self, user_localpart, user_filter): + async def add_user_filter(self, user_localpart: str, user_filter: JsonDict) -> str: def_json = encode_canonical_json(user_filter) # Need an atomic transaction to SELECT the maximal ID so far then @@ -71,4 +72,4 @@ class FilteringStore(SQLBaseStore): return filter_id - return self.db.runInteraction("add_user_filter", _do_txn) + return await self.db_pool.runInteraction("add_user_filter", _do_txn) diff --git a/synapse/storage/data_stores/main/group_server.py b/synapse/storage/databases/main/group_server.py
index 01ff561e1a..7218191965 100644 --- a/synapse/storage/data_stores/main/group_server.py +++ b/synapse/storage/databases/main/group_server.py
@@ -14,14 +14,12 @@ # See the License for the specific language governing permissions and # limitations under the License. -from typing import List, Tuple - -from canonicaljson import json - -from twisted.internet import defer +from typing import Any, Dict, List, Optional, Tuple, Union from synapse.api.errors import SynapseError from synapse.storage._base import SQLBaseStore, db_to_json +from synapse.types import JsonDict +from synapse.util import json_encoder # The category ID for the "default" category. We don't store as null in the # database to avoid the fun of null != null @@ -30,8 +28,8 @@ _DEFAULT_ROLE_ID = "" class GroupServerWorkerStore(SQLBaseStore): - def get_group(self, group_id): - return self.db.simple_select_one( + async def get_group(self, group_id: str) -> Optional[Dict[str, Any]]: + return await self.db_pool.simple_select_one( table="groups", keyvalues={"group_id": group_id}, retcols=( @@ -46,31 +44,35 @@ class GroupServerWorkerStore(SQLBaseStore): desc="get_group", ) - def get_users_in_group(self, group_id, include_private=False): + async def get_users_in_group( + self, group_id: str, include_private: bool = False + ) -> List[Dict[str, Any]]: # TODO: Pagination keyvalues = {"group_id": group_id} if not include_private: keyvalues["is_public"] = True - return self.db.simple_select_list( + return await self.db_pool.simple_select_list( table="group_users", keyvalues=keyvalues, retcols=("user_id", "is_public", "is_admin"), desc="get_users_in_group", ) - def get_invited_users_in_group(self, group_id): + async def get_invited_users_in_group(self, group_id: str) -> List[str]: # TODO: Pagination - return self.db.simple_select_onecol( + return await self.db_pool.simple_select_onecol( table="group_invites", keyvalues={"group_id": group_id}, retcol="user_id", desc="get_invited_users_in_group", ) - def get_rooms_in_group(self, group_id: str, include_private: bool = False): + async def get_rooms_in_group( + self, group_id: str, include_private: bool = False + ) -> List[Dict[str, Union[str, bool]]]: """Retrieve the rooms that belong to a given group. Does not return rooms that lack members. @@ -79,8 +81,7 @@ class GroupServerWorkerStore(SQLBaseStore): include_private: Whether to return private rooms in results Returns: - Deferred[List[Dict[str, str|bool]]]: A list of dictionaries, each in the - form of: + A list of dictionaries, each in the form of: { "room_id": "!a_room_id:example.com", # The ID of the room @@ -117,11 +118,13 @@ class GroupServerWorkerStore(SQLBaseStore): for room_id, is_public in txn ] - return self.db.runInteraction("get_rooms_in_group", _get_rooms_in_group_txn) + return await self.db_pool.runInteraction( + "get_rooms_in_group", _get_rooms_in_group_txn + ) - def get_rooms_for_summary_by_category( + async def get_rooms_for_summary_by_category( self, group_id: str, include_private: bool = False, - ): + ) -> Tuple[List[Dict[str, Any]], Dict[str, Any]]: """Get the rooms and categories that should be included in a summary request Args: @@ -129,7 +132,7 @@ class GroupServerWorkerStore(SQLBaseStore): include_private: Whether to return private rooms in results Returns: - Deferred[Tuple[List, Dict]]: A tuple containing: + A tuple containing: * A list of dictionaries with the keys: * "room_id": str, the room ID @@ -205,13 +208,12 @@ class GroupServerWorkerStore(SQLBaseStore): return rooms, categories - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_rooms_for_summary", _get_rooms_for_summary_txn ) - @defer.inlineCallbacks - def get_group_categories(self, group_id): - rows = yield self.db.simple_select_list( + async def get_group_categories(self, group_id): + rows = await self.db_pool.simple_select_list( table="group_room_categories", keyvalues={"group_id": group_id}, retcols=("category_id", "is_public", "profile"), @@ -226,9 +228,8 @@ class GroupServerWorkerStore(SQLBaseStore): for row in rows } - @defer.inlineCallbacks - def get_group_category(self, group_id, category_id): - category = yield self.db.simple_select_one( + async def get_group_category(self, group_id, category_id): + category = await self.db_pool.simple_select_one( table="group_room_categories", keyvalues={"group_id": group_id, "category_id": category_id}, retcols=("is_public", "profile"), @@ -239,9 +240,8 @@ class GroupServerWorkerStore(SQLBaseStore): return category - @defer.inlineCallbacks - def get_group_roles(self, group_id): - rows = yield self.db.simple_select_list( + async def get_group_roles(self, group_id): + rows = await self.db_pool.simple_select_list( table="group_roles", keyvalues={"group_id": group_id}, retcols=("role_id", "is_public", "profile"), @@ -256,9 +256,8 @@ class GroupServerWorkerStore(SQLBaseStore): for row in rows } - @defer.inlineCallbacks - def get_group_role(self, group_id, role_id): - role = yield self.db.simple_select_one( + async def get_group_role(self, group_id, role_id): + role = await self.db_pool.simple_select_one( table="group_roles", keyvalues={"group_id": group_id, "role_id": role_id}, retcols=("is_public", "profile"), @@ -269,25 +268,25 @@ class GroupServerWorkerStore(SQLBaseStore): return role - def get_local_groups_for_room(self, room_id): + async def get_local_groups_for_room(self, room_id: str) -> List[str]: """Get all of the local group that contain a given room Args: - room_id (str): The ID of a room + room_id: The ID of a room Returns: - Deferred[list[str]]: A twisted.Deferred containing a list of group ids - containing this room + A list of group ids containing this room """ - return self.db.simple_select_onecol( + return await self.db_pool.simple_select_onecol( table="group_rooms", keyvalues={"room_id": room_id}, retcol="group_id", desc="get_local_groups_for_room", ) - def get_users_for_summary_by_role(self, group_id, include_private=False): + async def get_users_for_summary_by_role(self, group_id, include_private=False): """Get the users and roles that should be included in a summary request - Returns ([users], [roles]) + Returns: + ([users], [roles]) """ def _get_users_for_summary_txn(txn): @@ -341,21 +340,24 @@ class GroupServerWorkerStore(SQLBaseStore): return users, roles - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_for_summary_by_role", _get_users_for_summary_txn ) - def is_user_in_group(self, user_id, group_id): - return self.db.simple_select_one_onecol( + async def is_user_in_group(self, user_id: str, group_id: str) -> bool: + result = await self.db_pool.simple_select_one_onecol( table="group_users", keyvalues={"group_id": group_id, "user_id": user_id}, retcol="user_id", allow_none=True, desc="is_user_in_group", - ).addCallback(lambda r: bool(r)) + ) + return bool(result) - def is_user_admin_in_group(self, group_id, user_id): - return self.db.simple_select_one_onecol( + async def is_user_admin_in_group( + self, group_id: str, user_id: str + ) -> Optional[bool]: + return await self.db_pool.simple_select_one_onecol( table="group_users", keyvalues={"group_id": group_id, "user_id": user_id}, retcol="is_admin", @@ -363,10 +365,12 @@ class GroupServerWorkerStore(SQLBaseStore): desc="is_user_admin_in_group", ) - def is_user_invited_to_local_group(self, group_id, user_id): + async def is_user_invited_to_local_group( + self, group_id: str, user_id: str + ) -> Optional[bool]: """Has the group server invited a user? """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="group_invites", keyvalues={"group_id": group_id, "user_id": user_id}, retcol="user_id", @@ -374,7 +378,7 @@ class GroupServerWorkerStore(SQLBaseStore): allow_none=True, ) - def get_users_membership_info_in_group(self, group_id, user_id): + async def get_users_membership_info_in_group(self, group_id, user_id): """Get a dict describing the membership of a user in a group. Example if joined: @@ -385,11 +389,12 @@ class GroupServerWorkerStore(SQLBaseStore): "is_privileged": False, } - Returns an empty dict if the user is not join/invite/etc + Returns: + An empty dict if the user is not join/invite/etc """ def _get_users_membership_in_group_txn(txn): - row = self.db.simple_select_one_txn( + row = self.db_pool.simple_select_one_txn( txn, table="group_users", keyvalues={"group_id": group_id, "user_id": user_id}, @@ -404,7 +409,7 @@ class GroupServerWorkerStore(SQLBaseStore): "is_privileged": row["is_admin"], } - row = self.db.simple_select_one_onecol_txn( + row = self.db_pool.simple_select_one_onecol_txn( txn, table="group_invites", keyvalues={"group_id": group_id, "user_id": user_id}, @@ -417,21 +422,21 @@ class GroupServerWorkerStore(SQLBaseStore): return {} - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_membership_info_in_group", _get_users_membership_in_group_txn ) - def get_publicised_groups_for_user(self, user_id): + async def get_publicised_groups_for_user(self, user_id: str) -> List[str]: """Get all groups a user is publicising """ - return self.db.simple_select_onecol( + return await self.db_pool.simple_select_onecol( table="local_group_membership", keyvalues={"user_id": user_id, "membership": "join", "is_publicised": True}, retcol="group_id", desc="get_publicised_groups_for_user", ) - def get_attestations_need_renewals(self, valid_until_ms): + async def get_attestations_need_renewals(self, valid_until_ms): """Get all attestations that need to be renewed until givent time """ @@ -441,18 +446,17 @@ class GroupServerWorkerStore(SQLBaseStore): WHERE valid_until_ms <= ? """ txn.execute(sql, (valid_until_ms,)) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_attestations_need_renewals", _get_attestations_need_renewals_txn ) - @defer.inlineCallbacks - def get_remote_attestation(self, group_id, user_id): + async def get_remote_attestation(self, group_id, user_id): """Get the attestation that proves the remote agrees that the user is in the group. """ - row = yield self.db.simple_select_one( + row = await self.db_pool.simple_select_one( table="group_attestations_remote", keyvalues={"group_id": group_id, "user_id": user_id}, retcols=("valid_until_ms", "attestation_json"), @@ -466,15 +470,15 @@ class GroupServerWorkerStore(SQLBaseStore): return None - def get_joined_groups(self, user_id): - return self.db.simple_select_onecol( + async def get_joined_groups(self, user_id: str) -> List[str]: + return await self.db_pool.simple_select_onecol( table="local_group_membership", keyvalues={"user_id": user_id, "membership": "join"}, retcol="group_id", desc="get_joined_groups", ) - def get_all_groups_for_user(self, user_id, now_token): + async def get_all_groups_for_user(self, user_id, now_token): def _get_all_groups_for_user_txn(txn): sql = """ SELECT group_id, type, membership, u.content @@ -494,17 +498,17 @@ class GroupServerWorkerStore(SQLBaseStore): for row in txn ] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_groups_for_user", _get_all_groups_for_user_txn ) - def get_groups_changes_for_user(self, user_id, from_token, to_token): + async def get_groups_changes_for_user(self, user_id, from_token, to_token): from_token = int(from_token) has_changed = self._group_updates_stream_cache.has_entity_changed( user_id, from_token ) if not has_changed: - return defer.succeed([]) + return [] def _get_groups_changes_for_user_txn(txn): sql = """ @@ -524,7 +528,7 @@ class GroupServerWorkerStore(SQLBaseStore): for group_id, membership, gtype, content_json in txn ] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_groups_changes_for_user", _get_groups_changes_for_user_txn ) @@ -579,28 +583,47 @@ class GroupServerWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_groups_changes", _get_all_groups_changes_txn ) class GroupServerStore(GroupServerWorkerStore): - def set_group_join_policy(self, group_id, join_policy): + async def set_group_join_policy(self, group_id: str, join_policy: str) -> None: """Set the join policy of a group. join_policy can be one of: * "invite" * "open" """ - return self.db.simple_update_one( + await self.db_pool.simple_update_one( table="groups", keyvalues={"group_id": group_id}, updatevalues={"join_policy": join_policy}, desc="set_group_join_policy", ) - def add_room_to_summary(self, group_id, room_id, category_id, order, is_public): - return self.db.runInteraction( + async def add_room_to_summary( + self, + group_id: str, + room_id: str, + category_id: str, + order: int, + is_public: Optional[bool], + ) -> None: + """Add (or update) room's entry in summary. + + Args: + group_id + room_id + category_id: If not None then adds the category to the end of + the summary if its not already there. + order: If not None inserts the room at that position, e.g. an order + of 1 will put the room first. Otherwise, the room gets added to + the end. + is_public + """ + await self.db_pool.runInteraction( "add_room_to_summary", self._add_room_to_summary_txn, group_id, @@ -611,20 +634,28 @@ class GroupServerStore(GroupServerWorkerStore): ) def _add_room_to_summary_txn( - self, txn, group_id, room_id, category_id, order, is_public - ): + self, + txn, + group_id: str, + room_id: str, + category_id: str, + order: int, + is_public: Optional[bool], + ) -> None: """Add (or update) room's entry in summary. Args: - group_id (str) - room_id (str) - category_id (str): If not None then adds the category to the end of - the summary if its not already there. [Optional] - order (int): If not None inserts the room at that position, e.g. - an order of 1 will put the room first. Otherwise, the room gets - added to the end. + txn + group_id + room_id + category_id: If not None then adds the category to the end of + the summary if its not already there. + order: If not None inserts the room at that position, e.g. an order + of 1 will put the room first. Otherwise, the room gets added to + the end. + is_public """ - room_in_group = self.db.simple_select_one_onecol_txn( + room_in_group = self.db_pool.simple_select_one_onecol_txn( txn, table="group_rooms", keyvalues={"group_id": group_id, "room_id": room_id}, @@ -637,7 +668,7 @@ class GroupServerStore(GroupServerWorkerStore): if category_id is None: category_id = _DEFAULT_CATEGORY_ID else: - cat_exists = self.db.simple_select_one_onecol_txn( + cat_exists = self.db_pool.simple_select_one_onecol_txn( txn, table="group_room_categories", keyvalues={"group_id": group_id, "category_id": category_id}, @@ -648,7 +679,7 @@ class GroupServerStore(GroupServerWorkerStore): raise SynapseError(400, "Category doesn't exist") # TODO: Check category is part of summary already - cat_exists = self.db.simple_select_one_onecol_txn( + cat_exists = self.db_pool.simple_select_one_onecol_txn( txn, table="group_summary_room_categories", keyvalues={"group_id": group_id, "category_id": category_id}, @@ -668,7 +699,7 @@ class GroupServerStore(GroupServerWorkerStore): (group_id, category_id, group_id, category_id), ) - existing = self.db.simple_select_one_txn( + existing = self.db_pool.simple_select_one_txn( txn, table="group_summary_rooms", keyvalues={ @@ -701,7 +732,7 @@ class GroupServerStore(GroupServerWorkerStore): to_update["room_order"] = order if is_public is not None: to_update["is_public"] = is_public - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn, table="group_summary_rooms", keyvalues={ @@ -715,7 +746,7 @@ class GroupServerStore(GroupServerWorkerStore): if is_public is None: is_public = True - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_summary_rooms", values={ @@ -727,11 +758,13 @@ class GroupServerStore(GroupServerWorkerStore): }, ) - def remove_room_from_summary(self, group_id, room_id, category_id): + async def remove_room_from_summary( + self, group_id: str, room_id: str, category_id: str + ) -> int: if category_id is None: category_id = _DEFAULT_CATEGORY_ID - return self.db.simple_delete( + return await self.db_pool.simple_delete( table="group_summary_rooms", keyvalues={ "group_id": group_id, @@ -741,7 +774,13 @@ class GroupServerStore(GroupServerWorkerStore): desc="remove_room_from_summary", ) - def upsert_group_category(self, group_id, category_id, profile, is_public): + async def upsert_group_category( + self, + group_id: str, + category_id: str, + profile: Optional[JsonDict], + is_public: Optional[bool], + ) -> None: """Add/update room category for group """ insertion_values = {} @@ -750,14 +789,14 @@ class GroupServerStore(GroupServerWorkerStore): if profile is None: insertion_values["profile"] = "{}" else: - update_values["profile"] = json.dumps(profile) + update_values["profile"] = json_encoder.encode(profile) if is_public is None: insertion_values["is_public"] = True else: update_values["is_public"] = is_public - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="group_room_categories", keyvalues={"group_id": group_id, "category_id": category_id}, values=update_values, @@ -765,14 +804,20 @@ class GroupServerStore(GroupServerWorkerStore): desc="upsert_group_category", ) - def remove_group_category(self, group_id, category_id): - return self.db.simple_delete( + async def remove_group_category(self, group_id: str, category_id: str) -> int: + return await self.db_pool.simple_delete( table="group_room_categories", keyvalues={"group_id": group_id, "category_id": category_id}, desc="remove_group_category", ) - def upsert_group_role(self, group_id, role_id, profile, is_public): + async def upsert_group_role( + self, + group_id: str, + role_id: str, + profile: Optional[JsonDict], + is_public: Optional[bool], + ) -> None: """Add/remove user role """ insertion_values = {} @@ -781,14 +826,14 @@ class GroupServerStore(GroupServerWorkerStore): if profile is None: insertion_values["profile"] = "{}" else: - update_values["profile"] = json.dumps(profile) + update_values["profile"] = json_encoder.encode(profile) if is_public is None: insertion_values["is_public"] = True else: update_values["is_public"] = is_public - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="group_roles", keyvalues={"group_id": group_id, "role_id": role_id}, values=update_values, @@ -796,15 +841,34 @@ class GroupServerStore(GroupServerWorkerStore): desc="upsert_group_role", ) - def remove_group_role(self, group_id, role_id): - return self.db.simple_delete( + async def remove_group_role(self, group_id: str, role_id: str) -> int: + return await self.db_pool.simple_delete( table="group_roles", keyvalues={"group_id": group_id, "role_id": role_id}, desc="remove_group_role", ) - def add_user_to_summary(self, group_id, user_id, role_id, order, is_public): - return self.db.runInteraction( + async def add_user_to_summary( + self, + group_id: str, + user_id: str, + role_id: str, + order: int, + is_public: Optional[bool], + ) -> None: + """Add (or update) user's entry in summary. + + Args: + group_id + user_id + role_id: If not None then adds the role to the end of the summary if + its not already there. + order: If not None inserts the user at that position, e.g. an order + of 1 will put the user first. Otherwise, the user gets added to + the end. + is_public + """ + await self.db_pool.runInteraction( "add_user_to_summary", self._add_user_to_summary_txn, group_id, @@ -815,20 +879,28 @@ class GroupServerStore(GroupServerWorkerStore): ) def _add_user_to_summary_txn( - self, txn, group_id, user_id, role_id, order, is_public + self, + txn, + group_id: str, + user_id: str, + role_id: str, + order: int, + is_public: Optional[bool], ): """Add (or update) user's entry in summary. Args: - group_id (str) - user_id (str) - role_id (str): If not None then adds the role to the end of - the summary if its not already there. [Optional] - order (int): If not None inserts the user at that position, e.g. - an order of 1 will put the user first. Otherwise, the user gets - added to the end. + txn + group_id + user_id + role_id: If not None then adds the role to the end of the summary if + its not already there. + order: If not None inserts the user at that position, e.g. an order + of 1 will put the user first. Otherwise, the user gets added to + the end. + is_public """ - user_in_group = self.db.simple_select_one_onecol_txn( + user_in_group = self.db_pool.simple_select_one_onecol_txn( txn, table="group_users", keyvalues={"group_id": group_id, "user_id": user_id}, @@ -841,7 +913,7 @@ class GroupServerStore(GroupServerWorkerStore): if role_id is None: role_id = _DEFAULT_ROLE_ID else: - role_exists = self.db.simple_select_one_onecol_txn( + role_exists = self.db_pool.simple_select_one_onecol_txn( txn, table="group_roles", keyvalues={"group_id": group_id, "role_id": role_id}, @@ -852,7 +924,7 @@ class GroupServerStore(GroupServerWorkerStore): raise SynapseError(400, "Role doesn't exist") # TODO: Check role is part of the summary already - role_exists = self.db.simple_select_one_onecol_txn( + role_exists = self.db_pool.simple_select_one_onecol_txn( txn, table="group_summary_roles", keyvalues={"group_id": group_id, "role_id": role_id}, @@ -872,7 +944,7 @@ class GroupServerStore(GroupServerWorkerStore): (group_id, role_id, group_id, role_id), ) - existing = self.db.simple_select_one_txn( + existing = self.db_pool.simple_select_one_txn( txn, table="group_summary_users", keyvalues={"group_id": group_id, "user_id": user_id, "role_id": role_id}, @@ -901,7 +973,7 @@ class GroupServerStore(GroupServerWorkerStore): to_update["user_order"] = order if is_public is not None: to_update["is_public"] = is_public - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn, table="group_summary_users", keyvalues={ @@ -915,7 +987,7 @@ class GroupServerStore(GroupServerWorkerStore): if is_public is None: is_public = True - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_summary_users", values={ @@ -927,50 +999,51 @@ class GroupServerStore(GroupServerWorkerStore): }, ) - def remove_user_from_summary(self, group_id, user_id, role_id): + async def remove_user_from_summary( + self, group_id: str, user_id: str, role_id: str + ) -> int: if role_id is None: role_id = _DEFAULT_ROLE_ID - return self.db.simple_delete( + return await self.db_pool.simple_delete( table="group_summary_users", keyvalues={"group_id": group_id, "role_id": role_id, "user_id": user_id}, desc="remove_user_from_summary", ) - def add_group_invite(self, group_id, user_id): + async def add_group_invite(self, group_id: str, user_id: str) -> None: """Record that the group server has invited a user """ - return self.db.simple_insert( + await self.db_pool.simple_insert( table="group_invites", values={"group_id": group_id, "user_id": user_id}, desc="add_group_invite", ) - def add_user_to_group( + async def add_user_to_group( self, - group_id, - user_id, - is_admin=False, - is_public=True, - local_attestation=None, - remote_attestation=None, - ): + group_id: str, + user_id: str, + is_admin: bool = False, + is_public: bool = True, + local_attestation: dict = None, + remote_attestation: dict = None, + ) -> None: """Add a user to the group server. Args: - group_id (str) - user_id (str) - is_admin (bool) - is_public (bool) - local_attestation (dict): The attestation the GS created to give - to the remote server. Optional if the user and group are on the - same server - remote_attestation (dict): The attestation given to GS by remote + group_id + user_id + is_admin + is_public + local_attestation: The attestation the GS created to give to the remote server. Optional if the user and group are on the same server + remote_attestation: The attestation given to GS by remote server. + Optional if the user and group are on the same server """ def _add_user_to_group_txn(txn): - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_users", values={ @@ -981,14 +1054,14 @@ class GroupServerStore(GroupServerWorkerStore): }, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_invites", keyvalues={"group_id": group_id, "user_id": user_id}, ) if local_attestation: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_attestations_renewals", values={ @@ -998,129 +1071,135 @@ class GroupServerStore(GroupServerWorkerStore): }, ) if remote_attestation: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_attestations_remote", values={ "group_id": group_id, "user_id": user_id, "valid_until_ms": remote_attestation["valid_until_ms"], - "attestation_json": json.dumps(remote_attestation), + "attestation_json": json_encoder.encode(remote_attestation), }, ) - return self.db.runInteraction("add_user_to_group", _add_user_to_group_txn) + await self.db_pool.runInteraction("add_user_to_group", _add_user_to_group_txn) - def remove_user_from_group(self, group_id, user_id): + async def remove_user_from_group(self, group_id: str, user_id: str) -> None: def _remove_user_from_group_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_users", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_invites", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_attestations_renewals", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_attestations_remote", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_summary_users", keyvalues={"group_id": group_id, "user_id": user_id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "remove_user_from_group", _remove_user_from_group_txn ) - def add_room_to_group(self, group_id, room_id, is_public): - return self.db.simple_insert( + async def add_room_to_group( + self, group_id: str, room_id: str, is_public: bool + ) -> None: + await self.db_pool.simple_insert( table="group_rooms", values={"group_id": group_id, "room_id": room_id, "is_public": is_public}, desc="add_room_to_group", ) - def update_room_in_group_visibility(self, group_id, room_id, is_public): - return self.db.simple_update( + async def update_room_in_group_visibility( + self, group_id: str, room_id: str, is_public: bool + ) -> int: + return await self.db_pool.simple_update( table="group_rooms", keyvalues={"group_id": group_id, "room_id": room_id}, updatevalues={"is_public": is_public}, desc="update_room_in_group_visibility", ) - def remove_room_from_group(self, group_id, room_id): + async def remove_room_from_group(self, group_id: str, room_id: str) -> None: def _remove_room_from_group_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_rooms", keyvalues={"group_id": group_id, "room_id": room_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_summary_rooms", keyvalues={"group_id": group_id, "room_id": room_id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "remove_room_from_group", _remove_room_from_group_txn ) - def update_group_publicity(self, group_id, user_id, publicise): + async def update_group_publicity( + self, group_id: str, user_id: str, publicise: bool + ) -> None: """Update whether the user is publicising their membership of the group """ - return self.db.simple_update_one( + await self.db_pool.simple_update_one( table="local_group_membership", keyvalues={"group_id": group_id, "user_id": user_id}, updatevalues={"is_publicised": publicise}, desc="update_group_publicity", ) - @defer.inlineCallbacks - def register_user_group_membership( + async def register_user_group_membership( self, - group_id, - user_id, - membership, - is_admin=False, - content={}, - local_attestation=None, - remote_attestation=None, - is_publicised=False, - ): + group_id: str, + user_id: str, + membership: str, + is_admin: bool = False, + content: JsonDict = {}, + local_attestation: Optional[dict] = None, + remote_attestation: Optional[dict] = None, + is_publicised: bool = False, + ) -> int: """Registers that a local user is a member of a (local or remote) group. Args: - group_id (str) - user_id (str) - membership (str) - is_admin (bool) - content (dict): Content of the membership, e.g. includes the inviter + group_id: The group the member is being added to. + user_id: THe user ID to add to the group. + membership: The type of group membership. + is_admin: Whether the user should be added as a group admin. + content: Content of the membership, e.g. includes the inviter if the user has been invited. - local_attestation (dict): If remote group then store the fact that we + local_attestation: If remote group then store the fact that we have given out an attestation, else None. - remote_attestation (dict): If remote group then store the remote + remote_attestation: If remote group then store the remote attestation from the group, else None. + is_publicised: Whether this should be publicised. """ def _register_user_group_membership_txn(txn, next_id): # TODO: Upsert? - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="local_group_membership", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="local_group_membership", values={ @@ -1129,11 +1208,11 @@ class GroupServerStore(GroupServerWorkerStore): "is_admin": is_admin, "membership": membership, "is_publicised": is_publicised, - "content": json.dumps(content), + "content": json_encoder.encode(content), }, ) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="local_group_updates", values={ @@ -1141,7 +1220,7 @@ class GroupServerStore(GroupServerWorkerStore): "group_id": group_id, "user_id": user_id, "type": "membership", - "content": json.dumps( + "content": json_encoder.encode( {"membership": membership, "content": content} ), }, @@ -1152,7 +1231,7 @@ class GroupServerStore(GroupServerWorkerStore): if membership == "join": if local_attestation: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_attestations_renewals", values={ @@ -1162,23 +1241,23 @@ class GroupServerStore(GroupServerWorkerStore): }, ) if remote_attestation: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="group_attestations_remote", values={ "group_id": group_id, "user_id": user_id, "valid_until_ms": remote_attestation["valid_until_ms"], - "attestation_json": json.dumps(remote_attestation), + "attestation_json": json_encoder.encode(remote_attestation), }, ) else: - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_attestations_renewals", keyvalues={"group_id": group_id, "user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="group_attestations_remote", keyvalues={"group_id": group_id, "user_id": user_id}, @@ -1186,19 +1265,18 @@ class GroupServerStore(GroupServerWorkerStore): return next_id - with self._group_updates_id_gen.get_next() as next_id: - res = yield self.db.runInteraction( + async with self._group_updates_id_gen.get_next() as next_id: + res = await self.db_pool.runInteraction( "register_user_group_membership", _register_user_group_membership_txn, next_id, ) return res - @defer.inlineCallbacks - def create_group( + async def create_group( self, group_id, user_id, name, avatar_url, short_description, long_description - ): - yield self.db.simple_insert( + ) -> None: + await self.db_pool.simple_insert( table="groups", values={ "group_id": group_id, @@ -1211,48 +1289,51 @@ class GroupServerStore(GroupServerWorkerStore): desc="create_group", ) - @defer.inlineCallbacks - def update_group_profile(self, group_id, profile): - yield self.db.simple_update_one( + async def update_group_profile(self, group_id, profile): + await self.db_pool.simple_update_one( table="groups", keyvalues={"group_id": group_id}, updatevalues=profile, desc="update_group_profile", ) - def update_attestation_renewal(self, group_id, user_id, attestation): + async def update_attestation_renewal( + self, group_id: str, user_id: str, attestation: dict + ) -> None: """Update an attestation that we have renewed """ - return self.db.simple_update_one( + await self.db_pool.simple_update_one( table="group_attestations_renewals", keyvalues={"group_id": group_id, "user_id": user_id}, updatevalues={"valid_until_ms": attestation["valid_until_ms"]}, desc="update_attestation_renewal", ) - def update_remote_attestion(self, group_id, user_id, attestation): + async def update_remote_attestion( + self, group_id: str, user_id: str, attestation: dict + ) -> None: """Update an attestation that a remote has renewed """ - return self.db.simple_update_one( + await self.db_pool.simple_update_one( table="group_attestations_remote", keyvalues={"group_id": group_id, "user_id": user_id}, updatevalues={ "valid_until_ms": attestation["valid_until_ms"], - "attestation_json": json.dumps(attestation), + "attestation_json": json_encoder.encode(attestation), }, desc="update_remote_attestion", ) - def remove_attestation_renewal(self, group_id, user_id): + async def remove_attestation_renewal(self, group_id: str, user_id: str) -> int: """Remove an attestation that we thought we should renew, but actually shouldn't. Ideally this would never get called as we would never incorrectly try and do attestations for local users on local groups. Args: - group_id (str) - user_id (str) + group_id + user_id """ - return self.db.simple_delete( + return await self.db_pool.simple_delete( table="group_attestations_renewals", keyvalues={"group_id": group_id, "user_id": user_id}, desc="remove_attestation_renewal", @@ -1261,14 +1342,11 @@ class GroupServerStore(GroupServerWorkerStore): def get_group_stream_token(self): return self._group_updates_id_gen.get_current_token() - def delete_group(self, group_id): + async def delete_group(self, group_id: str) -> None: """Deletes a group fully from the database. Args: - group_id (str) - - Returns: - Deferred + group_id: The group ID to delete. """ def _delete_group_txn(txn): @@ -1288,8 +1366,8 @@ class GroupServerStore(GroupServerWorkerStore): ] for table in tables: - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table=table, keyvalues={"group_id": group_id} ) - return self.db.runInteraction("delete_group", _delete_group_txn) + await self.db_pool.runInteraction("delete_group", _delete_group_txn) diff --git a/synapse/storage/data_stores/main/keys.py b/synapse/storage/databases/main/keys.py
index 4e1642a27a..ad43bb05ab 100644 --- a/synapse/storage/data_stores/main/keys.py +++ b/synapse/storage/databases/main/keys.py
@@ -16,6 +16,7 @@ import itertools import logging +from typing import Dict, Iterable, List, Optional, Tuple from signedjson.key import decode_verify_key_bytes @@ -41,16 +42,17 @@ class KeyStore(SQLBaseStore): @cachedList( cached_method_name="_get_server_verify_key", list_name="server_name_and_key_ids" ) - def get_server_verify_keys(self, server_name_and_key_ids): + async def get_server_verify_keys( + self, server_name_and_key_ids: Iterable[Tuple[str, str]] + ) -> Dict[Tuple[str, str], Optional[FetchKeyResult]]: """ Args: - server_name_and_key_ids (iterable[Tuple[str, str]]): + server_name_and_key_ids: iterable of (server_name, key-id) tuples to fetch keys for Returns: - Deferred: resolves to dict[Tuple[str, str], FetchKeyResult|None]: - map from (server_name, key_id) -> FetchKeyResult, or None if the key is - unknown + A map from (server_name, key_id) -> FetchKeyResult, or None if the + key is unknown """ keys = {} @@ -86,14 +88,19 @@ class KeyStore(SQLBaseStore): _get_keys(txn, batch) return keys - return self.db.runInteraction("get_server_verify_keys", _txn) + return await self.db_pool.runInteraction("get_server_verify_keys", _txn) - def store_server_verify_keys(self, from_server, ts_added_ms, verify_keys): + async def store_server_verify_keys( + self, + from_server: str, + ts_added_ms: int, + verify_keys: Iterable[Tuple[str, str, FetchKeyResult]], + ) -> None: """Stores NACL verification keys for remote servers. Args: - from_server (str): Where the verification keys were looked up - ts_added_ms (int): The time to record that the key was added - verify_keys (iterable[tuple[str, str, FetchKeyResult]]): + from_server: Where the verification keys were looked up + ts_added_ms: The time to record that the key was added + verify_keys: keys to be stored. Each entry is a triplet of (server_name, key_id, key). """ @@ -115,15 +122,9 @@ class KeyStore(SQLBaseStore): # param, which is itself the 2-tuple (server_name, key_id). invalidations.append((server_name, key_id)) - def _invalidate(res): - f = self._get_server_verify_key.invalidate - for i in invalidations: - f((i,)) - return res - - return self.db.runInteraction( + await self.db_pool.runInteraction( "store_server_verify_keys", - self.db.simple_upsert_many_txn, + self.db_pool.simple_upsert_many_txn, table="server_signature_keys", key_names=("server_name", "key_id"), key_values=key_values, @@ -134,24 +135,34 @@ class KeyStore(SQLBaseStore): "verify_key", ), value_values=value_values, - ).addCallback(_invalidate) + ) - def store_server_keys_json( - self, server_name, key_id, from_server, ts_now_ms, ts_expires_ms, key_json_bytes - ): + invalidate = self._get_server_verify_key.invalidate + for i in invalidations: + invalidate((i,)) + + async def store_server_keys_json( + self, + server_name: str, + key_id: str, + from_server: str, + ts_now_ms: int, + ts_expires_ms: int, + key_json_bytes: bytes, + ) -> None: """Stores the JSON bytes for a set of keys from a server The JSON should be signed by the originating server, the intermediate server, and by this server. Updates the value for the (server_name, key_id, from_server) triplet if one already existed. Args: - server_name (str): The name of the server. - key_id (str): The identifer of the key this JSON is for. - from_server (str): The server this JSON was fetched from. - ts_now_ms (int): The time now in milliseconds. - ts_valid_until_ms (int): The time when this json stops being valid. - key_json (bytes): The encoded JSON. + server_name: The name of the server. + key_id: The identifer of the key this JSON is for. + from_server: The server this JSON was fetched from. + ts_now_ms: The time now in milliseconds. + ts_valid_until_ms: The time when this json stops being valid. + key_json_bytes: The encoded JSON. """ - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="server_keys_json", keyvalues={ "server_name": server_name, @@ -169,7 +180,9 @@ class KeyStore(SQLBaseStore): desc="store_server_keys_json", ) - def get_server_keys_json(self, server_keys): + async def get_server_keys_json( + self, server_keys: Iterable[Tuple[str, Optional[str], Optional[str]]] + ) -> Dict[Tuple[str, Optional[str], Optional[str]], List[dict]]: """Retrive the key json for a list of server_keys and key ids. If no keys are found for a given server, key_id and source then that server, key_id, and source triplet entry will be an empty list. @@ -178,8 +191,7 @@ class KeyStore(SQLBaseStore): Args: server_keys (list): List of (server_name, key_id, source) triplets. Returns: - Deferred[dict[Tuple[str, str, str|None], list[dict]]]: - Dict mapping (server_name, key_id, source) triplets to lists of dicts + A mapping from (server_name, key_id, source) triplets to a list of dicts """ def _get_server_keys_json_txn(txn): @@ -190,7 +202,7 @@ class KeyStore(SQLBaseStore): keyvalues["key_id"] = key_id if from_server is not None: keyvalues["from_server"] = from_server - rows = self.db.simple_select_list_txn( + rows = self.db_pool.simple_select_list_txn( txn, "server_keys_json", keyvalues=keyvalues, @@ -205,4 +217,6 @@ class KeyStore(SQLBaseStore): results[(server_name, key_id, from_server)] = rows return results - return self.db.runInteraction("get_server_keys_json", _get_server_keys_json_txn) + return await self.db_pool.runInteraction( + "get_server_keys_json", _get_server_keys_json_txn + ) diff --git a/synapse/storage/data_stores/main/media_repository.py b/synapse/storage/databases/main/media_repository.py
index 15bc13cbd0..cc538c5c10 100644 --- a/synapse/storage/data_stores/main/media_repository.py +++ b/synapse/storage/databases/main/media_repository.py
@@ -12,17 +12,21 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. +from typing import Any, Dict, Iterable, List, Optional, Tuple + from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool + +BG_UPDATE_REMOVE_MEDIA_REPO_INDEX_WITHOUT_METHOD = ( + "media_repository_drop_index_wo_method" +) class MediaRepositoryBackgroundUpdateStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(MediaRepositoryBackgroundUpdateStore, self).__init__( - database, db_conn, hs - ) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( update_name="local_media_repository_url_idx", index_name="local_media_repository_url_idx", table="local_media_repository", @@ -30,19 +34,73 @@ class MediaRepositoryBackgroundUpdateStore(SQLBaseStore): where_clause="url_cache IS NOT NULL", ) + # The following the updates add the method to the unique constraint of + # the thumbnail databases. That fixes an issue, where thumbnails of the + # same resolution, but different methods could overwrite one another. + # This can happen with custom thumbnail configs or with dynamic thumbnailing. + self.db_pool.updates.register_background_index_update( + update_name="local_media_repository_thumbnails_method_idx", + index_name="local_media_repository_thumbn_media_id_width_height_method_key", + table="local_media_repository_thumbnails", + columns=[ + "media_id", + "thumbnail_width", + "thumbnail_height", + "thumbnail_type", + "thumbnail_method", + ], + unique=True, + ) + + self.db_pool.updates.register_background_index_update( + update_name="remote_media_repository_thumbnails_method_idx", + index_name="remote_media_repository_thumbn_media_origin_id_width_height_method_key", + table="remote_media_cache_thumbnails", + columns=[ + "media_origin", + "media_id", + "thumbnail_width", + "thumbnail_height", + "thumbnail_type", + "thumbnail_method", + ], + unique=True, + ) + + self.db_pool.updates.register_background_update_handler( + BG_UPDATE_REMOVE_MEDIA_REPO_INDEX_WITHOUT_METHOD, + self._drop_media_index_without_method, + ) + + async def _drop_media_index_without_method(self, progress, batch_size): + def f(txn): + txn.execute( + "ALTER TABLE local_media_repository_thumbnails DROP CONSTRAINT IF EXISTS local_media_repository_thumbn_media_id_thumbnail_width_thum_key" + ) + txn.execute( + "ALTER TABLE remote_media_cache_thumbnails DROP CONSTRAINT IF EXISTS remote_media_repository_thumbn_media_id_thumbnail_width_thum_key" + ) + + await self.db_pool.runInteraction("drop_media_indices_without_method", f) + await self.db_pool.updates._end_background_update( + BG_UPDATE_REMOVE_MEDIA_REPO_INDEX_WITHOUT_METHOD + ) + return 1 + class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): """Persistence for attachments and avatars""" - def __init__(self, database: Database, db_conn, hs): - super(MediaRepositoryStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - def get_local_media(self, media_id): + async def get_local_media(self, media_id: str) -> Optional[Dict[str, Any]]: """Get the metadata for a local piece of media + Returns: None if the media_id doesn't exist. """ - return self.db.simple_select_one( + return await self.db_pool.simple_select_one( "local_media_repository", {"media_id": media_id}, ( @@ -57,7 +115,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="get_local_media", ) - def store_local_media( + async def store_local_media( self, media_id, media_type, @@ -66,8 +124,8 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): media_length, user_id, url_cache=None, - ): - return self.db.simple_insert( + ) -> None: + await self.db_pool.simple_insert( "local_media_repository", { "media_id": media_id, @@ -81,16 +139,16 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="store_local_media", ) - def mark_local_media_as_safe(self, media_id: str): + async def mark_local_media_as_safe(self, media_id: str) -> None: """Mark a local media as safe from quarantining.""" - return self.db.simple_update_one( + await self.db_pool.simple_update_one( table="local_media_repository", keyvalues={"media_id": media_id}, updatevalues={"safe_from_quarantine": True}, desc="mark_local_media_as_safe", ) - def get_url_cache(self, url, ts): + async def get_url_cache(self, url: str, ts: int) -> Optional[Dict[str, Any]]: """Get the media_id and ts for a cached URL as of the given timestamp Returns: None if the URL isn't cached. @@ -136,12 +194,12 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): ) ) - return self.db.runInteraction("get_url_cache", get_url_cache_txn) + return await self.db_pool.runInteraction("get_url_cache", get_url_cache_txn) - def store_url_cache( + async def store_url_cache( self, url, response_code, etag, expires_ts, og, media_id, download_ts ): - return self.db.simple_insert( + await self.db_pool.simple_insert( "local_media_repository_url_cache", { "url": url, @@ -155,8 +213,8 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="store_url_cache", ) - def get_local_media_thumbnails(self, media_id): - return self.db.simple_select_list( + async def get_local_media_thumbnails(self, media_id: str) -> List[Dict[str, Any]]: + return await self.db_pool.simple_select_list( "local_media_repository_thumbnails", {"media_id": media_id}, ( @@ -169,7 +227,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="get_local_media_thumbnails", ) - def store_local_thumbnail( + async def store_local_thumbnail( self, media_id, thumbnail_width, @@ -178,7 +236,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): thumbnail_method, thumbnail_length, ): - return self.db.simple_insert( + await self.db_pool.simple_insert( "local_media_repository_thumbnails", { "media_id": media_id, @@ -191,8 +249,10 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="store_local_thumbnail", ) - def get_cached_remote_media(self, origin, media_id): - return self.db.simple_select_one( + async def get_cached_remote_media( + self, origin, media_id: str + ) -> Optional[Dict[str, Any]]: + return await self.db_pool.simple_select_one( "remote_media_cache", {"media_origin": origin, "media_id": media_id}, ( @@ -207,7 +267,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="get_cached_remote_media", ) - def store_cached_remote_media( + async def store_cached_remote_media( self, origin, media_id, @@ -217,7 +277,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): upload_name, filesystem_id, ): - return self.db.simple_insert( + await self.db_pool.simple_insert( "remote_media_cache", { "media_origin": origin, @@ -232,12 +292,17 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="store_cached_remote_media", ) - def update_cached_last_access_time(self, local_media, remote_media, time_ms): + async def update_cached_last_access_time( + self, + local_media: Iterable[str], + remote_media: Iterable[Tuple[str, str]], + time_ms: int, + ): """Updates the last access time of the given media Args: - local_media (iterable[str]): Set of media_ids - remote_media (iterable[(str, str)]): Set of (server_name, media_id) + local_media: Set of media_ids + remote_media: Set of (server_name, media_id) time_ms: Current time in milliseconds """ @@ -262,12 +327,14 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): txn.executemany(sql, ((time_ms, media_id) for media_id in local_media)) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "update_cached_last_access_time", update_cache_txn ) - def get_remote_media_thumbnails(self, origin, media_id): - return self.db.simple_select_list( + async def get_remote_media_thumbnails( + self, origin: str, media_id: str + ) -> List[Dict[str, Any]]: + return await self.db_pool.simple_select_list( "remote_media_cache_thumbnails", {"media_origin": origin, "media_id": media_id}, ( @@ -281,7 +348,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="get_remote_media_thumbnails", ) - def store_remote_media_thumbnail( + async def store_remote_media_thumbnail( self, origin, media_id, @@ -292,7 +359,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): thumbnail_method, thumbnail_length, ): - return self.db.simple_insert( + await self.db_pool.simple_insert( "remote_media_cache_thumbnails", { "media_origin": origin, @@ -307,33 +374,35 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): desc="store_remote_media_thumbnail", ) - def get_remote_media_before(self, before_ts): + async def get_remote_media_before(self, before_ts): sql = ( "SELECT media_origin, media_id, filesystem_id" " FROM remote_media_cache" " WHERE last_access_ts < ?" ) - return self.db.execute( - "get_remote_media_before", self.db.cursor_to_dict, sql, before_ts + return await self.db_pool.execute( + "get_remote_media_before", self.db_pool.cursor_to_dict, sql, before_ts ) - def delete_remote_media(self, media_origin, media_id): + async def delete_remote_media(self, media_origin: str, media_id: str) -> None: def delete_remote_media_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, "remote_media_cache", keyvalues={"media_origin": media_origin, "media_id": media_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, "remote_media_cache_thumbnails", keyvalues={"media_origin": media_origin, "media_id": media_id}, ) - return self.db.runInteraction("delete_remote_media", delete_remote_media_txn) + await self.db_pool.runInteraction( + "delete_remote_media", delete_remote_media_txn + ) - def get_expired_url_cache(self, now_ts): + async def get_expired_url_cache(self, now_ts: int) -> List[str]: sql = ( "SELECT media_id FROM local_media_repository_url_cache" " WHERE expires_ts < ?" @@ -345,7 +414,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): txn.execute(sql, (now_ts,)) return [row[0] for row in txn] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_expired_url_cache", _get_expired_url_cache_txn ) @@ -358,9 +427,11 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): def _delete_url_cache_txn(txn): txn.executemany(sql, [(media_id,) for media_id in media_ids]) - return await self.db.runInteraction("delete_url_cache", _delete_url_cache_txn) + return await self.db_pool.runInteraction( + "delete_url_cache", _delete_url_cache_txn + ) - def get_url_cache_media_before(self, before_ts): + async def get_url_cache_media_before(self, before_ts: int) -> List[str]: sql = ( "SELECT media_id FROM local_media_repository" " WHERE created_ts < ? AND url_cache IS NOT NULL" @@ -372,7 +443,7 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): txn.execute(sql, (before_ts,)) return [row[0] for row in txn] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_url_cache_media_before", _get_url_cache_media_before_txn ) @@ -389,6 +460,6 @@ class MediaRepositoryStore(MediaRepositoryBackgroundUpdateStore): txn.executemany(sql, [(media_id,) for media_id in media_ids]) - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "delete_url_cache_media", _delete_url_cache_media_txn ) diff --git a/synapse/storage/data_stores/main/metrics.py b/synapse/storage/databases/main/metrics.py
index dad5bbc602..92099f95ce 100644 --- a/synapse/storage/data_stores/main/metrics.py +++ b/synapse/storage/databases/main/metrics.py
@@ -12,18 +12,34 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -import typing -from collections import Counter -from twisted.internet import defer - -from synapse.metrics import BucketCollector +from synapse.metrics import GaugeBucketCollector from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.event_push_actions import ( +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.event_push_actions import ( EventPushActionsWorkerStore, ) -from synapse.storage.database import Database + +# Collect metrics on the number of forward extremities that exist. +_extremities_collecter = GaugeBucketCollector( + "synapse_forward_extremities", + "Number of rooms on the server with the given number of forward extremities" + " or fewer", + buckets=[1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500], +) + +# we also expose metrics on the "number of excess extremity events", which is +# (E-1)*N, where E is the number of extremities and N is the number of state +# events in the room. This is an approximation to the number of state events +# we could remove from state resolution by reducing the graph to a single +# forward extremity. +_excess_state_events_collecter = GaugeBucketCollector( + "synapse_excess_extremity_events", + "Number of rooms on the server with the given number of excess extremity " + "events, or fewer", + buckets=[0] + [1 << n for n in range(12)], +) class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): @@ -31,21 +47,9 @@ class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): stats and prometheus metrics. """ - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): super().__init__(database, db_conn, hs) - # Collect metrics on the number of forward extremities that exist. - # Counter of number of extremities to count - self._current_forward_extremities_amount = ( - Counter() - ) # type: typing.Counter[int] - - BucketCollector( - "synapse_forward_extremities", - lambda: self._current_forward_extremities_amount, - buckets=[1, 2, 3, 5, 7, 10, 15, 20, 50, 100, 200, 500, "+Inf"], - ) - # Read the extrems every 60 minutes def read_forward_extremities(): # run as a background process to make sure that the database transactions @@ -60,17 +64,27 @@ class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): def fetch(txn): txn.execute( """ - select count(*) c from event_forward_extremities - group by room_id + SELECT t1.c, t2.c + FROM ( + SELECT room_id, COUNT(*) c FROM event_forward_extremities + GROUP BY room_id + ) t1 LEFT JOIN ( + SELECT room_id, COUNT(*) c FROM current_state_events + GROUP BY room_id + ) t2 ON t1.room_id = t2.room_id """ ) return txn.fetchall() - res = await self.db.runInteraction("read_forward_extremities", fetch) - self._current_forward_extremities_amount = Counter([x[0] for x in res]) + res = await self.db_pool.runInteraction("read_forward_extremities", fetch) + + _extremities_collecter.update_data(x[0] for x in res) - @defer.inlineCallbacks - def count_daily_messages(self): + _excess_state_events_collecter.update_data( + (x[0] - 1) * x[1] for x in res if x[1] + ) + + async def count_daily_messages(self): """ Returns an estimate of the number of messages sent in the last day. @@ -88,11 +102,9 @@ class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): (count,) = txn.fetchone() return count - ret = yield self.db.runInteraction("count_messages", _count_messages) - return ret + return await self.db_pool.runInteraction("count_messages", _count_messages) - @defer.inlineCallbacks - def count_daily_sent_messages(self): + async def count_daily_sent_messages(self): def _count_messages(txn): # This is good enough as if you have silly characters in your own # hostname then thats your own fault. @@ -109,11 +121,11 @@ class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): (count,) = txn.fetchone() return count - ret = yield self.db.runInteraction("count_daily_sent_messages", _count_messages) - return ret + return await self.db_pool.runInteraction( + "count_daily_sent_messages", _count_messages + ) - @defer.inlineCallbacks - def count_daily_active_rooms(self): + async def count_daily_active_rooms(self): def _count(txn): sql = """ SELECT COALESCE(COUNT(DISTINCT room_id), 0) FROM events @@ -124,5 +136,4 @@ class ServerMetricsStore(EventPushActionsWorkerStore, SQLBaseStore): (count,) = txn.fetchone() return count - ret = yield self.db.runInteraction("count_daily_active_rooms", _count) - return ret + return await self.db_pool.runInteraction("count_daily_active_rooms", _count) diff --git a/synapse/storage/data_stores/main/monthly_active_users.py b/synapse/storage/databases/main/monthly_active_users.py
index e459cf49a0..e93aad33cd 100644 --- a/synapse/storage/data_stores/main/monthly_active_users.py +++ b/synapse/storage/databases/main/monthly_active_users.py
@@ -13,12 +13,10 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging -from typing import List - -from twisted.internet import defer +from typing import Dict, List from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database, make_in_list_sql_clause +from synapse.storage.database import DatabasePool, make_in_list_sql_clause from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) @@ -29,29 +27,36 @@ LAST_SEEN_GRANULARITY = 60 * 60 * 1000 class MonthlyActiveUsersWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(MonthlyActiveUsersWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._clock = hs.get_clock() self.hs = hs @cached(num_args=0) - def get_monthly_active_count(self): + async def get_monthly_active_count(self) -> int: """Generates current count of monthly active users Returns: - Defered[int]: Number of current monthly active users + Number of current monthly active users """ def _count_users(txn): - sql = "SELECT COALESCE(count(*), 0) FROM monthly_active_users" + # Exclude app service users + sql = """ + SELECT COALESCE(count(*), 0) + FROM monthly_active_users + LEFT JOIN users + ON monthly_active_users.user_id=users.name + WHERE (users.appservice_id IS NULL OR users.appservice_id = ''); + """ txn.execute(sql) (count,) = txn.fetchone() return count - return self.db.runInteraction("count_users", _count_users) + return await self.db_pool.runInteraction("count_users", _count_users) @cached(num_args=0) - def get_monthly_active_count_by_service(self): + async def get_monthly_active_count_by_service(self) -> Dict[str, int]: """Generates current count of monthly active users broken down by service. A service is typically an appservice but also includes native matrix users. Since the `monthly_active_users` table is populated from the `user_ips` table @@ -59,8 +64,7 @@ class MonthlyActiveUsersWorkerStore(SQLBaseStore): method to return anything other than native matrix users. Returns: - Deferred[dict]: dict that includes a mapping between app_service_id - and the number of occurrences. + A mapping between app_service_id and the number of occurrences. """ @@ -76,7 +80,9 @@ class MonthlyActiveUsersWorkerStore(SQLBaseStore): result = txn.fetchall() return dict(result) - return self.db.runInteraction("count_users_by_service", _count_users_by_service) + return await self.db_pool.runInteraction( + "count_users_by_service", _count_users_by_service + ) async def get_registered_reserved_users(self) -> List[str]: """Of the reserved threepids defined in config, retrieve those that are associated @@ -99,17 +105,18 @@ class MonthlyActiveUsersWorkerStore(SQLBaseStore): return users @cached(num_args=1) - def user_last_seen_monthly_active(self, user_id): + async def user_last_seen_monthly_active(self, user_id: str) -> int: """ - Checks if a given user is part of the monthly active user group - Arguments: - user_id (str): user to add/update - Return: - Deferred[int] : timestamp since last seen, None if never seen + Checks if a given user is part of the monthly active user group + Arguments: + user_id: user to add/update + + Return: + Timestamp since last seen, None if never seen """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="monthly_active_users", keyvalues={"user_id": user_id}, retcol="timestamp", @@ -119,8 +126,8 @@ class MonthlyActiveUsersWorkerStore(SQLBaseStore): class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): - def __init__(self, database: Database, db_conn, hs): - super(MonthlyActiveUsersStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._limit_usage_by_mau = hs.config.limit_usage_by_mau self._mau_stats_only = hs.config.mau_stats_only @@ -128,7 +135,7 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): # Do not add more reserved users than the total allowable number # cur = LoggingTransaction( - self.db.new_transaction( + self.db_pool.new_transaction( db_conn, "initialise_mau_threepids", [], @@ -162,7 +169,7 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): is_support = self.is_support_user_txn(txn, user_id) if not is_support: # We do this manually here to avoid hitting #6791 - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="monthly_active_users", keyvalues={"user_id": user_id}, @@ -246,20 +253,16 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): self._invalidate_cache_and_stream(txn, self.get_monthly_active_count, ()) reserved_users = await self.get_registered_reserved_users() - await self.db.runInteraction( + await self.db_pool.runInteraction( "reap_monthly_active_users", _reap_users, reserved_users ) - @defer.inlineCallbacks - def upsert_monthly_active_user(self, user_id): + async def upsert_monthly_active_user(self, user_id: str) -> None: """Updates or inserts the user into the monthly active user table, which is used to track the current MAU usage of the server Args: - user_id (str): user to add/update - - Returns: - Deferred + user_id: user to add/update """ # Support user never to be included in MAU stats. Note I can't easily call this # from upsert_monthly_active_user_txn because then I need a _txn form of @@ -269,11 +272,11 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): # _initialise_reserved_users reasoning that it would be very strange to # include a support user in this context. - is_support = yield self.is_support_user(user_id) + is_support = await self.is_support_user(user_id) if is_support: return - yield self.db.runInteraction( + await self.db_pool.runInteraction( "upsert_monthly_active_user", self.upsert_monthly_active_user_txn, user_id ) @@ -303,7 +306,7 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): # never be a big table and alternative approaches (batching multiple # upserts into a single txn) introduced a lot of extra complexity. # See https://github.com/matrix-org/synapse/issues/3854 for more - is_insert = self.db.simple_upsert_txn( + is_insert = self.db_pool.simple_upsert_txn( txn, table="monthly_active_users", keyvalues={"user_id": user_id}, @@ -320,8 +323,7 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): return is_insert - @defer.inlineCallbacks - def populate_monthly_active_users(self, user_id): + async def populate_monthly_active_users(self, user_id): """Checks on the state of monthly active user limits and optionally add the user to the monthly active tables @@ -330,14 +332,14 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): """ if self._limit_usage_by_mau or self._mau_stats_only: # Trial users and guests should not be included as part of MAU group - is_guest = yield self.is_guest(user_id) + is_guest = await self.is_guest(user_id) if is_guest: return - is_trial = yield self.is_trial_user(user_id) + is_trial = await self.is_trial_user(user_id) if is_trial: return - last_seen_timestamp = yield self.user_last_seen_monthly_active(user_id) + last_seen_timestamp = await self.user_last_seen_monthly_active(user_id) now = self.hs.get_clock().time_msec() # We want to reduce to the total number of db writes, and are happy @@ -350,10 +352,10 @@ class MonthlyActiveUsersStore(MonthlyActiveUsersWorkerStore): # False, there is no point in checking get_monthly_active_count - it # adds no value and will break the logic if max_mau_value is exceeded. if not self._limit_usage_by_mau: - yield self.upsert_monthly_active_user(user_id) + await self.upsert_monthly_active_user(user_id) else: - count = yield self.get_monthly_active_count() + count = await self.get_monthly_active_count() if count < self._max_mau_value: - yield self.upsert_monthly_active_user(user_id) + await self.upsert_monthly_active_user(user_id) elif now - last_seen_timestamp > LAST_SEEN_GRANULARITY: - yield self.upsert_monthly_active_user(user_id) + await self.upsert_monthly_active_user(user_id) diff --git a/synapse/storage/data_stores/main/openid.py b/synapse/storage/databases/main/openid.py
index cc21437e92..2aac64901b 100644 --- a/synapse/storage/data_stores/main/openid.py +++ b/synapse/storage/databases/main/openid.py
@@ -1,9 +1,13 @@ +from typing import Optional + from synapse.storage._base import SQLBaseStore class OpenIdStore(SQLBaseStore): - def insert_open_id_token(self, token, ts_valid_until_ms, user_id): - return self.db.simple_insert( + async def insert_open_id_token( + self, token: str, ts_valid_until_ms: int, user_id: str + ) -> None: + await self.db_pool.simple_insert( table="open_id_tokens", values={ "token": token, @@ -13,7 +17,9 @@ class OpenIdStore(SQLBaseStore): desc="insert_open_id_token", ) - def get_user_id_for_open_id_token(self, token, ts_now_ms): + async def get_user_id_for_open_id_token( + self, token: str, ts_now_ms: int + ) -> Optional[str]: def get_user_id_for_token_txn(txn): sql = ( "SELECT user_id FROM open_id_tokens" @@ -28,6 +34,6 @@ class OpenIdStore(SQLBaseStore): else: return rows[0][0] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_user_id_for_token", get_user_id_for_token_txn ) diff --git a/synapse/storage/data_stores/main/presence.py b/synapse/storage/databases/main/presence.py
index 7574612619..dbbb99cb95 100644 --- a/synapse/storage/data_stores/main/presence.py +++ b/synapse/storage/databases/main/presence.py
@@ -15,23 +15,20 @@ from typing import List, Tuple -from twisted.internet import defer - +from synapse.api.presence import UserPresenceState from synapse.storage._base import SQLBaseStore, make_in_list_sql_clause -from synapse.storage.presence import UserPresenceState from synapse.util.caches.descriptors import cached, cachedList from synapse.util.iterutils import batch_iter class PresenceStore(SQLBaseStore): - @defer.inlineCallbacks - def update_presence(self, presence_states): + async def update_presence(self, presence_states): stream_ordering_manager = self._presence_id_gen.get_next_mult( len(presence_states) ) - with stream_ordering_manager as stream_orderings: - yield self.db.runInteraction( + async with stream_ordering_manager as stream_orderings: + await self.db_pool.runInteraction( "update_presence", self._update_presence_txn, stream_orderings, @@ -48,7 +45,7 @@ class PresenceStore(SQLBaseStore): txn.call_after(self._get_presence_for_user.invalidate, (state.user_id,)) # Actually insert new rows - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="presence_stream", values=[ @@ -124,7 +121,7 @@ class PresenceStore(SQLBaseStore): return updates, upper_bound, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_presence_updates", get_all_presence_updates_txn ) @@ -133,13 +130,10 @@ class PresenceStore(SQLBaseStore): raise NotImplementedError() @cachedList( - cached_method_name="_get_presence_for_user", - list_name="user_ids", - num_args=1, - inlineCallbacks=True, + cached_method_name="_get_presence_for_user", list_name="user_ids", num_args=1, ) - def get_presence_for_users(self, user_ids): - rows = yield self.db.simple_select_many_batch( + async def get_presence_for_users(self, user_ids): + rows = await self.db_pool.simple_select_many_batch( table="presence_stream", column="user_id", iterable=user_ids, @@ -163,24 +157,3 @@ class PresenceStore(SQLBaseStore): def get_current_presence_token(self): return self._presence_id_gen.get_current_token() - - def allow_presence_visible(self, observed_localpart, observer_userid): - return self.db.simple_insert( - table="presence_allow_inbound", - values={ - "observed_user_id": observed_localpart, - "observer_user_id": observer_userid, - }, - desc="allow_presence_visible", - or_ignore=True, - ) - - def disallow_presence_visible(self, observed_localpart, observer_userid): - return self.db.simple_delete_one( - table="presence_allow_inbound", - keyvalues={ - "observed_user_id": observed_localpart, - "observer_user_id": observer_userid, - }, - desc="disallow_presence_visible", - ) diff --git a/synapse/storage/data_stores/main/profile.py b/synapse/storage/databases/main/profile.py
index cd2472feb0..de37866d25 100644 --- a/synapse/storage/data_stores/main/profile.py +++ b/synapse/storage/databases/main/profile.py
@@ -13,14 +13,11 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - -from typing import List, Tuple - -from twisted.internet import defer +from typing import Any, Dict, List, Optional, Tuple from synapse.api.errors import StoreError from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.roommember import ProfileInfo +from synapse.storage.databases.main.roommember import ProfileInfo from synapse.types import UserID from synapse.util.caches.descriptors import cached @@ -28,10 +25,9 @@ BATCH_SIZE = 100 class ProfileWorkerStore(SQLBaseStore): - @defer.inlineCallbacks - def get_profileinfo(self, user_localpart): + async def get_profileinfo(self, user_localpart: str) -> ProfileInfo: try: - profile = yield self.db.simple_select_one( + profile = await self.db_pool.simple_select_one( table="profiles", keyvalues={"user_id": user_localpart}, retcols=("displayname", "avatar_url"), @@ -49,8 +45,8 @@ class ProfileWorkerStore(SQLBaseStore): ) @cached(max_entries=5000) - def get_profile_displayname(self, user_localpart): - return self.db.simple_select_one_onecol( + async def get_profile_displayname(self, user_localpart: str) -> str: + return await self.db_pool.simple_select_one_onecol( table="profiles", keyvalues={"user_id": user_localpart}, retcol="displayname", @@ -58,31 +54,33 @@ class ProfileWorkerStore(SQLBaseStore): ) @cached(max_entries=5000) - def get_profile_avatar_url(self, user_localpart): - return self.db.simple_select_one_onecol( + async def get_profile_avatar_url(self, user_localpart: str) -> str: + return await self.db_pool.simple_select_one_onecol( table="profiles", keyvalues={"user_id": user_localpart}, retcol="avatar_url", desc="get_profile_avatar_url", ) - def get_latest_profile_replication_batch_number(self): + async def get_latest_profile_replication_batch_number(self): def f(txn): txn.execute("SELECT MAX(batch) as maxbatch FROM profiles") - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) return rows[0]["maxbatch"] - return self.db.runInteraction("get_latest_profile_replication_batch_number", f) + return await self.db_pool.runInteraction( + "get_latest_profile_replication_batch_number", f + ) - def get_profile_batch(self, batchnum): - return self.db.simple_select_list( + async def get_profile_batch(self, batchnum): + return await self.db_pool.simple_select_list( table="profiles", keyvalues={"batch": batchnum}, retcols=("user_id", "displayname", "avatar_url", "active"), desc="get_profile_batch", ) - def assign_profile_batch(self): + async def assign_profile_batch(self): def f(txn): sql = ( "UPDATE profiles SET batch = " @@ -94,28 +92,32 @@ class ProfileWorkerStore(SQLBaseStore): txn.execute(sql, (BATCH_SIZE,)) return txn.rowcount - return self.db.runInteraction("assign_profile_batch", f) + return await self.db_pool.runInteraction("assign_profile_batch", f) - def get_replication_hosts(self): + async def get_replication_hosts(self): def f(txn): txn.execute( "SELECT host, last_synced_batch FROM profile_replication_status" ) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) return {r["host"]: r["last_synced_batch"] for r in rows} - return self.db.runInteraction("get_replication_hosts", f) + return await self.db_pool.runInteraction("get_replication_hosts", f) - def update_replication_batch_for_host(self, host, last_synced_batch): - return self.db.simple_upsert( + async def update_replication_batch_for_host( + self, host: str, last_synced_batch: int + ): + return await self.db_pool.simple_upsert( table="profile_replication_status", keyvalues={"host": host}, values={"last_synced_batch": last_synced_batch}, desc="update_replication_batch_for_host", ) - def get_from_remote_profile_cache(self, user_id): - return self.db.simple_select_one( + async def get_from_remote_profile_cache( + self, user_id: str + ) -> Optional[Dict[str, Any]]: + return await self.db_pool.simple_select_one( table="remote_profile_cache", keyvalues={"user_id": user_id}, retcols=("displayname", "avatar_url"), @@ -123,16 +125,18 @@ class ProfileWorkerStore(SQLBaseStore): desc="get_from_remote_profile_cache", ) - def create_profile(self, user_localpart): - return self.db.simple_insert( + async def create_profile(self, user_localpart: str) -> None: + await self.db_pool.simple_insert( table="profiles", values={"user_id": user_localpart}, desc="create_profile" ) - def set_profile_displayname(self, user_localpart, new_displayname, batchnum): + async def set_profile_displayname( + self, user_localpart: str, new_displayname: str, batchnum: int + ) -> None: # Invalidate the read cache for this user self.get_profile_displayname.invalidate((user_localpart,)) - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="profiles", keyvalues={"user_id": user_localpart}, values={"displayname": new_displayname, "batch": batchnum}, @@ -140,11 +144,13 @@ class ProfileWorkerStore(SQLBaseStore): lock=False, # we can do this because user_id has a unique index ) - def set_profile_avatar_url(self, user_localpart, new_avatar_url, batchnum): + async def set_profile_avatar_url( + self, user_localpart: str, new_avatar_url: str, batchnum: int + ) -> None: # Invalidate the read cache for this user self.get_profile_avatar_url.invalidate((user_localpart,)) - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="profiles", keyvalues={"user_id": user_localpart}, values={"avatar_url": new_avatar_url, "batch": batchnum}, @@ -152,9 +158,9 @@ class ProfileWorkerStore(SQLBaseStore): lock=False, # we can do this because user_id has a unique index ) - def set_profiles_active( + async def set_profiles_active( self, users: List[UserID], active: bool, hide: bool, batchnum: int, - ): + ) -> None: """Given a set of users, set active and hidden flags on them. Args: @@ -164,9 +170,6 @@ class ProfileWorkerStore(SQLBaseStore): False and active is False, users will have their profiles erased batchnum: The batch number, used for profile replication - - Returns: - Deferred """ # Convert list of localparts to list of tuples containing localparts user_localparts = [(user.localpart,) for user in users] @@ -181,9 +184,9 @@ class ProfileWorkerStore(SQLBaseStore): value_names += ("avatar_url", "displayname") values = [v + (None, None) for v in values] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "set_profiles_active", - self.db.simple_upsert_many_txn, + self.db_pool.simple_upsert_many_txn, table="profiles", key_names=("user_id",), key_values=user_localparts, @@ -194,10 +197,9 @@ class ProfileWorkerStore(SQLBaseStore): class ProfileStore(ProfileWorkerStore): def __init__(self, database, db_conn, hs): + super().__init__(database, db_conn, hs) - super(ProfileStore, self).__init__(database, db_conn, hs) - - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "profile_replication_status_host_index", index_name="profile_replication_status_idx", table="profile_replication_status", @@ -205,13 +207,15 @@ class ProfileStore(ProfileWorkerStore): unique=True, ) - def add_remote_profile_cache(self, user_id, displayname, avatar_url): + async def add_remote_profile_cache( + self, user_id: str, displayname: str, avatar_url: str + ) -> None: """Ensure we are caching the remote user's profiles. This should only be called when `is_subscribed_remote_profile_for_user` would return true for the user. """ - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="remote_profile_cache", keyvalues={"user_id": user_id}, values={ @@ -222,11 +226,13 @@ class ProfileStore(ProfileWorkerStore): desc="add_remote_profile_cache", ) - def update_remote_profile_cache(self, user_id, displayname, avatar_url): - return self.db.simple_upsert( + async def update_remote_profile_cache( + self, user_id: str, displayname: str, avatar_url: str + ) -> int: + return await self.db_pool.simple_upsert( table="remote_profile_cache", keyvalues={"user_id": user_id}, - updatevalues={ + values={ "displayname": displayname, "avatar_url": avatar_url, "last_check": self._clock.time_msec(), @@ -234,20 +240,21 @@ class ProfileStore(ProfileWorkerStore): desc="update_remote_profile_cache", ) - @defer.inlineCallbacks - def maybe_delete_remote_profile_cache(self, user_id): + async def maybe_delete_remote_profile_cache(self, user_id): """Check if we still care about the remote user's profile, and if we don't then remove their profile from the cache """ - subscribed = yield self.is_subscribed_remote_profile_for_user(user_id) + subscribed = await self.is_subscribed_remote_profile_for_user(user_id) if not subscribed: - yield self.db.simple_delete( + await self.db_pool.simple_delete( table="remote_profile_cache", keyvalues={"user_id": user_id}, desc="delete_remote_profile_cache", ) - def get_remote_profile_cache_entries_that_expire(self, last_checked): + async def get_remote_profile_cache_entries_that_expire( + self, last_checked: int + ) -> Dict[str, str]: """Get all users who haven't been checked since `last_checked` """ @@ -260,18 +267,17 @@ class ProfileStore(ProfileWorkerStore): txn.execute(sql, (last_checked,)) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_remote_profile_cache_entries_that_expire", _get_remote_profile_cache_entries_that_expire_txn, ) - @defer.inlineCallbacks - def is_subscribed_remote_profile_for_user(self, user_id): + async def is_subscribed_remote_profile_for_user(self, user_id): """Check whether we are interested in a remote user's profile. """ - res = yield self.db.simple_select_one_onecol( + res = await self.db_pool.simple_select_one_onecol( table="group_users", keyvalues={"user_id": user_id}, retcol="user_id", @@ -282,7 +288,7 @@ class ProfileStore(ProfileWorkerStore): if res: return True - res = yield self.db.simple_select_one_onecol( + res = await self.db_pool.simple_select_one_onecol( table="group_invites", keyvalues={"user_id": user_id}, retcol="user_id", diff --git a/synapse/storage/data_stores/main/purge_events.py b/synapse/storage/databases/main/purge_events.py
index 6546569139..ecfc6717b3 100644 --- a/synapse/storage/data_stores/main/purge_events.py +++ b/synapse/storage/databases/main/purge_events.py
@@ -14,46 +14,45 @@ # limitations under the License. import logging -from typing import Any, Tuple +from typing import Any, List, Set, Tuple from synapse.api.errors import SynapseError from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.state import StateGroupWorkerStore +from synapse.storage.databases.main.state import StateGroupWorkerStore from synapse.types import RoomStreamToken logger = logging.getLogger(__name__) class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): - def purge_history(self, room_id, token, delete_local_events): + async def purge_history( + self, room_id: str, token: str, delete_local_events: bool + ) -> Set[int]: """Deletes room history before a certain point Args: - room_id (str): - - token (str): A topological token to delete events before - - delete_local_events (bool): + room_id: + token: A topological token to delete events before + delete_local_events: if True, we will delete local events as well as remote ones (instead of just marking them as outliers and deleting their state groups). Returns: - Deferred[set[int]]: The set of state groups that are referenced by - deleted events. + The set of state groups that are referenced by deleted events. """ - return self.db.runInteraction( + parsed_token = await RoomStreamToken.parse(self, token) + + return await self.db_pool.runInteraction( "purge_history", self._purge_history_txn, room_id, - token, + parsed_token, delete_local_events, ) - def _purge_history_txn(self, txn, room_id, token_str, delete_local_events): - token = RoomStreamToken.parse(token_str) - + def _purge_history_txn(self, txn, room_id, token, delete_local_events): # Tables that should be pruned: # event_auth # event_backward_extremities @@ -62,6 +61,7 @@ class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): # event_json # event_push_actions # event_reference_hashes + # event_relations # event_search # event_to_state_groups # events @@ -69,6 +69,7 @@ class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): # room_depth # state_groups # state_groups_state + # destination_rooms # we will build a temporary table listing the events so that we don't # have to keep shovelling the list back and forth across the @@ -209,6 +210,7 @@ class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): "event_edges", "event_forward_extremities", "event_reference_hashes", + "event_relations", "event_search", "rejections", ): @@ -281,17 +283,18 @@ class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): return referenced_state_groups - def purge_room(self, room_id): + async def purge_room(self, room_id: str) -> List[int]: """Deletes all record of a room Args: - room_id (str) + room_id Returns: - Deferred[List[int]]: The list of state groups to delete. + The list of state groups to delete. """ - - return self.db.runInteraction("purge_room", self._purge_room_txn, room_id) + return await self.db_pool.runInteraction( + "purge_room", self._purge_room_txn, room_id + ) def _purge_room_txn(self, txn, room_id): # First we fetch all the state groups that should be deleted, before @@ -334,6 +337,7 @@ class PurgeEventsStore(StateGroupWorkerStore, SQLBaseStore): # and finally, the tables with an index on room_id (or no useful index) for table in ( "current_state_events", + "destination_rooms", "event_backward_extremities", "event_forward_extremities", "event_json", diff --git a/synapse/storage/data_stores/main/push_rule.py b/synapse/storage/databases/main/push_rule.py
index c229248101..711d5aa23d 100644 --- a/synapse/storage/data_stores/main/push_rule.py +++ b/synapse/storage/databases/main/push_rule.py
@@ -13,33 +13,31 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import abc import logging from typing import List, Tuple, Union -from canonicaljson import json - -from twisted.internet import defer - +from synapse.api.errors import NotFoundError, StoreError from synapse.push.baserules import list_with_base_rules from synapse.replication.slave.storage._slaved_id_tracker import SlavedIdTracker from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.storage.data_stores.main.appservice import ApplicationServiceWorkerStore -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.data_stores.main.pusher import PusherWorkerStore -from synapse.storage.data_stores.main.receipts import ReceiptsWorkerStore -from synapse.storage.data_stores.main.roommember import RoomMemberWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.appservice import ApplicationServiceWorkerStore +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.storage.databases.main.pusher import PusherWorkerStore +from synapse.storage.databases.main.receipts import ReceiptsWorkerStore +from synapse.storage.databases.main.roommember import RoomMemberWorkerStore +from synapse.storage.engines import PostgresEngine, Sqlite3Engine from synapse.storage.push_rule import InconsistentRuleException, RuleNotFoundException -from synapse.storage.util.id_generators import ChainedIdGenerator -from synapse.util.caches.descriptors import cachedInlineCallbacks, cachedList +from synapse.storage.util.id_generators import StreamIdGenerator +from synapse.util import json_encoder +from synapse.util.caches.descriptors import cached, cachedList from synapse.util.caches.stream_change_cache import StreamChangeCache logger = logging.getLogger(__name__) -def _load_rules(rawrules, enabled_map): +def _load_rules(rawrules, enabled_map, use_new_defaults=False): ruleslist = [] for rawrule in rawrules: rule = dict(rawrule) @@ -49,7 +47,7 @@ def _load_rules(rawrules, enabled_map): ruleslist.append(rule) # We're going to be mutating this a lot, so do a deep copy - rules = list(list_with_base_rules(ruleslist)) + rules = list(list_with_base_rules(ruleslist, use_new_defaults)) for i, rule in enumerate(rules): rule_id = rule["rule_id"] @@ -63,6 +61,8 @@ def _load_rules(rawrules, enabled_map): return rules +# The ABCMeta metaclass ensures that it cannot be instantiated without +# the abstract methods being implemented. class PushRulesWorkerStore( ApplicationServiceWorkerStore, ReceiptsWorkerStore, @@ -70,28 +70,25 @@ class PushRulesWorkerStore( RoomMemberWorkerStore, EventsWorkerStore, SQLBaseStore, + metaclass=abc.ABCMeta, ): """This is an abstract base class where subclasses must implement `get_max_push_rules_stream_id` which can be called in the initializer. """ - # This ABCMeta metaclass ensures that we cannot be instantiated without - # the abstract methods being implemented. - __metaclass__ = abc.ABCMeta - - def __init__(self, database: Database, db_conn, hs): - super(PushRulesWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) if hs.config.worker.worker_app is None: - self._push_rules_stream_id_gen = ChainedIdGenerator( - self._stream_id_gen, db_conn, "push_rules_stream", "stream_id" - ) # type: Union[ChainedIdGenerator, SlavedIdTracker] + self._push_rules_stream_id_gen = StreamIdGenerator( + db_conn, "push_rules_stream", "stream_id" + ) # type: Union[StreamIdGenerator, SlavedIdTracker] else: self._push_rules_stream_id_gen = SlavedIdTracker( db_conn, "push_rules_stream", "stream_id" ) - push_rules_prefill, push_rules_id = self.db.get_cache_dict( + push_rules_prefill, push_rules_id = self.db_pool.get_cache_dict( db_conn, "push_rules_stream", entity_column="user_id", @@ -105,6 +102,8 @@ class PushRulesWorkerStore( prefilled_cache=push_rules_prefill, ) + self._users_new_default_push_rules = hs.config.users_new_default_push_rules + @abc.abstractmethod def get_max_push_rules_stream_id(self): """Get the position of the push rules stream. @@ -114,9 +113,9 @@ class PushRulesWorkerStore( """ raise NotImplementedError() - @cachedInlineCallbacks(max_entries=5000) - def get_push_rules_for_user(self, user_id): - rows = yield self.db.simple_select_list( + @cached(max_entries=5000) + async def get_push_rules_for_user(self, user_id): + rows = await self.db_pool.simple_select_list( table="push_rules", keyvalues={"user_name": user_id}, retcols=( @@ -132,15 +131,15 @@ class PushRulesWorkerStore( rows.sort(key=lambda row: (-int(row["priority_class"]), -int(row["priority"]))) - enabled_map = yield self.get_push_rules_enabled_for_user(user_id) + enabled_map = await self.get_push_rules_enabled_for_user(user_id) - rules = _load_rules(rows, enabled_map) + use_new_defaults = user_id in self._users_new_default_push_rules - return rules + return _load_rules(rows, enabled_map, use_new_defaults) - @cachedInlineCallbacks(max_entries=5000) - def get_push_rules_enabled_for_user(self, user_id): - results = yield self.db.simple_select_list( + @cached(max_entries=5000) + async def get_push_rules_enabled_for_user(self, user_id): + results = await self.db_pool.simple_select_list( table="push_rules_enable", keyvalues={"user_name": user_id}, retcols=("user_name", "rule_id", "enabled"), @@ -148,9 +147,11 @@ class PushRulesWorkerStore( ) return {r["rule_id"]: False if r["enabled"] == 0 else True for r in results} - def have_push_rules_changed_for_user(self, user_id, last_id): + async def have_push_rules_changed_for_user( + self, user_id: str, last_id: int + ) -> bool: if not self.push_rules_stream_cache.has_entity_changed(user_id, last_id): - return defer.succeed(False) + return False else: def have_push_rules_changed_txn(txn): @@ -162,23 +163,20 @@ class PushRulesWorkerStore( (count,) = txn.fetchone() return bool(count) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "have_push_rules_changed", have_push_rules_changed_txn ) @cachedList( - cached_method_name="get_push_rules_for_user", - list_name="user_ids", - num_args=1, - inlineCallbacks=True, + cached_method_name="get_push_rules_for_user", list_name="user_ids", num_args=1, ) - def bulk_get_push_rules(self, user_ids): + async def bulk_get_push_rules(self, user_ids): if not user_ids: return {} results = {user_id: [] for user_id in user_ids} - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="push_rules", column="user_name", iterable=user_ids, @@ -191,21 +189,26 @@ class PushRulesWorkerStore( for row in rows: results.setdefault(row["user_name"], []).append(row) - enabled_map_by_user = yield self.bulk_get_push_rules_enabled(user_ids) + enabled_map_by_user = await self.bulk_get_push_rules_enabled(user_ids) for user_id, rules in results.items(): - results[user_id] = _load_rules(rules, enabled_map_by_user.get(user_id, {})) + use_new_defaults = user_id in self._users_new_default_push_rules + + results[user_id] = _load_rules( + rules, enabled_map_by_user.get(user_id, {}), use_new_defaults, + ) return results - @defer.inlineCallbacks - def copy_push_rule_from_room_to_room(self, new_room_id, user_id, rule): + async def copy_push_rule_from_room_to_room( + self, new_room_id: str, user_id: str, rule: dict + ) -> None: """Copy a single push rule from one room to another for a specific user. Args: - new_room_id (str): ID of the new room. - user_id (str): ID of user the push rule belongs to. - rule (Dict): A push rule. + new_room_id: ID of the new room. + user_id : ID of user the push rule belongs to. + rule: A push rule. """ # Create new rule id rule_id_scope = "/".join(rule["rule_id"].split("/")[:-1]) @@ -217,7 +220,7 @@ class PushRulesWorkerStore( condition["pattern"] = new_room_id # Add the rule for the new room - yield self.add_push_rule( + await self.add_push_rule( user_id=user_id, rule_id=new_rule_id, priority_class=rule["priority_class"], @@ -225,20 +228,19 @@ class PushRulesWorkerStore( actions=rule["actions"], ) - @defer.inlineCallbacks - def copy_push_rules_from_room_to_room_for_user( - self, old_room_id, new_room_id, user_id - ): + async def copy_push_rules_from_room_to_room_for_user( + self, old_room_id: str, new_room_id: str, user_id: str + ) -> None: """Copy all of the push rules from one room to another for a specific user. Args: - old_room_id (str): ID of the old room. - new_room_id (str): ID of the new room. - user_id (str): ID of user to copy push rules for. + old_room_id: ID of the old room. + new_room_id: ID of the new room. + user_id: ID of user to copy push rules for. """ # Retrieve push rules for this user - user_push_rules = yield self.get_push_rules_for_user(user_id) + user_push_rules = await self.get_push_rules_for_user(user_id) # Get rules relating to the old room and copy them to the new room for rule in user_push_rules: @@ -247,96 +249,20 @@ class PushRulesWorkerStore( (c.get("key") == "room_id" and c.get("pattern") == old_room_id) for c in conditions ): - yield self.copy_push_rule_from_room_to_room(new_room_id, user_id, rule) - - @defer.inlineCallbacks - def bulk_get_push_rules_for_room(self, event, context): - state_group = context.state_group - if not state_group: - # If state_group is None it means it has yet to be assigned a - # state group, i.e. we need to make sure that calls with a state_group - # of None don't hit previous cached calls with a None state_group. - # To do this we set the state_group to a new object as object() != object() - state_group = object() - - current_state_ids = yield defer.ensureDeferred(context.get_current_state_ids()) - result = yield self._bulk_get_push_rules_for_room( - event.room_id, state_group, current_state_ids, event=event - ) - return result - - @cachedInlineCallbacks(num_args=2, cache_context=True) - def _bulk_get_push_rules_for_room( - self, room_id, state_group, current_state_ids, cache_context, event=None - ): - # We don't use `state_group`, its there so that we can cache based - # on it. However, its important that its never None, since two current_state's - # with a state_group of None are likely to be different. - # See bulk_get_push_rules_for_room for how we work around this. - assert state_group is not None - - # We also will want to generate notifs for other people in the room so - # their unread countss are correct in the event stream, but to avoid - # generating them for bot / AS users etc, we only do so for people who've - # sent a read receipt into the room. - - users_in_room = yield self._get_joined_users_from_context( - room_id, - state_group, - current_state_ids, - on_invalidate=cache_context.invalidate, - event=event, - ) - - # We ignore app service users for now. This is so that we don't fill - # up the `get_if_users_have_pushers` cache with AS entries that we - # know don't have pushers, nor even read receipts. - local_users_in_room = { - u - for u in users_in_room - if self.hs.is_mine_id(u) - and not self.get_if_app_services_interested_in_user(u) - } - - # users in the room who have pushers need to get push rules run because - # that's how their pushers work - if_users_with_pushers = yield self.get_if_users_have_pushers( - local_users_in_room, on_invalidate=cache_context.invalidate - ) - user_ids = { - uid for uid, have_pusher in if_users_with_pushers.items() if have_pusher - } - - users_with_receipts = yield self.get_users_with_read_receipts_in_room( - room_id, on_invalidate=cache_context.invalidate - ) - - # any users with pushers must be ours: they have pushers - for uid in users_with_receipts: - if uid in local_users_in_room: - user_ids.add(uid) - - rules_by_user = yield self.bulk_get_push_rules( - user_ids, on_invalidate=cache_context.invalidate - ) - - rules_by_user = {k: v for k, v in rules_by_user.items() if v is not None} - - return rules_by_user + await self.copy_push_rule_from_room_to_room(new_room_id, user_id, rule) @cachedList( cached_method_name="get_push_rules_enabled_for_user", list_name="user_ids", num_args=1, - inlineCallbacks=True, ) - def bulk_get_push_rules_enabled(self, user_ids): + async def bulk_get_push_rules_enabled(self, user_ids): if not user_ids: return {} results = {user_id: {} for user_id in user_ids} - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="push_rules_enable", column="user_name", iterable=user_ids, @@ -394,14 +320,13 @@ class PushRulesWorkerStore( return updates, upper_bound, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_push_rule_updates", get_all_push_rule_updates_txn ) class PushRuleStore(PushRulesWorkerStore): - @defer.inlineCallbacks - def add_push_rule( + async def add_push_rule( self, user_id, rule_id, @@ -410,13 +335,14 @@ class PushRuleStore(PushRulesWorkerStore): actions, before=None, after=None, - ): - conditions_json = json.dumps(conditions) - actions_json = json.dumps(actions) - with self._push_rules_stream_id_gen.get_next() as ids: - stream_id, event_stream_ordering = ids + ) -> None: + conditions_json = json_encoder.encode(conditions) + actions_json = json_encoder.encode(actions) + async with self._push_rules_stream_id_gen.get_next() as stream_id: + event_stream_ordering = self._stream_id_gen.get_current_token() + if before or after: - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_add_push_rule_relative_txn", self._add_push_rule_relative_txn, stream_id, @@ -430,7 +356,7 @@ class PushRuleStore(PushRulesWorkerStore): after, ) else: - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_add_push_rule_highest_priority_txn", self._add_push_rule_highest_priority_txn, stream_id, @@ -461,7 +387,7 @@ class PushRuleStore(PushRulesWorkerStore): relative_to_rule = before or after - res = self.db.simple_select_one_txn( + res = self.db_pool.simple_select_one_txn( txn, table="push_rules", keyvalues={"user_name": user_id, "rule_id": relative_to_rule}, @@ -584,7 +510,7 @@ class PushRuleStore(PushRulesWorkerStore): # We didn't update a row with the given rule_id so insert one push_rule_id = self._push_rule_id_gen.get_next() - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="push_rules", values={ @@ -614,20 +540,44 @@ class PushRuleStore(PushRulesWorkerStore): }, ) - @defer.inlineCallbacks - def delete_push_rule(self, user_id, rule_id): + # ensure we have a push_rules_enable row + # enabledness defaults to true + if isinstance(self.database_engine, PostgresEngine): + sql = """ + INSERT INTO push_rules_enable (id, user_name, rule_id, enabled) + VALUES (?, ?, ?, ?) + ON CONFLICT DO NOTHING + """ + elif isinstance(self.database_engine, Sqlite3Engine): + sql = """ + INSERT OR IGNORE INTO push_rules_enable (id, user_name, rule_id, enabled) + VALUES (?, ?, ?, ?) + """ + else: + raise RuntimeError("Unknown database engine") + + new_enable_id = self._push_rules_enable_id_gen.get_next() + txn.execute(sql, (new_enable_id, user_id, rule_id, 1)) + + async def delete_push_rule(self, user_id: str, rule_id: str) -> None: """ Delete a push rule. Args specify the row to be deleted and can be any of the columns in the push_rule table, but below are the standard ones Args: - user_id (str): The matrix ID of the push rule owner - rule_id (str): The rule_id of the rule to be deleted + user_id: The matrix ID of the push rule owner + rule_id: The rule_id of the rule to be deleted """ def delete_push_rule_txn(txn, stream_id, event_stream_ordering): - self.db.simple_delete_one_txn( + # we don't use simple_delete_one_txn because that would fail if the + # user did not have a push_rule_enable row. + self.db_pool.simple_delete_txn( + txn, "push_rules_enable", {"user_name": user_id, "rule_id": rule_id} + ) + + self.db_pool.simple_delete_one_txn( txn, "push_rules", {"user_name": user_id, "rule_id": rule_id} ) @@ -635,20 +585,40 @@ class PushRuleStore(PushRulesWorkerStore): txn, stream_id, event_stream_ordering, user_id, rule_id, op="DELETE" ) - with self._push_rules_stream_id_gen.get_next() as ids: - stream_id, event_stream_ordering = ids - yield self.db.runInteraction( + async with self._push_rules_stream_id_gen.get_next() as stream_id: + event_stream_ordering = self._stream_id_gen.get_current_token() + + await self.db_pool.runInteraction( "delete_push_rule", delete_push_rule_txn, stream_id, event_stream_ordering, ) - @defer.inlineCallbacks - def set_push_rule_enabled(self, user_id, rule_id, enabled): - with self._push_rules_stream_id_gen.get_next() as ids: - stream_id, event_stream_ordering = ids - yield self.db.runInteraction( + async def set_push_rule_enabled( + self, user_id: str, rule_id: str, enabled: bool, is_default_rule: bool + ) -> None: + """ + Sets the `enabled` state of a push rule. + + Args: + user_id: the user ID of the user who wishes to enable/disable the rule + e.g. '@tina:example.org' + rule_id: the full rule ID of the rule to be enabled/disabled + e.g. 'global/override/.m.rule.roomnotif' + or 'global/override/myCustomRule' + enabled: True if the rule is to be enabled, False if it is to be + disabled + is_default_rule: True if and only if this is a server-default rule. + This skips the check for existence (as only user-created rules + are always stored in the database `push_rules` table). + + Raises: + NotFoundError if the rule does not exist. + """ + async with self._push_rules_stream_id_gen.get_next() as stream_id: + event_stream_ordering = self._stream_id_gen.get_current_token() + await self.db_pool.runInteraction( "_set_push_rule_enabled_txn", self._set_push_rule_enabled_txn, stream_id, @@ -656,13 +626,48 @@ class PushRuleStore(PushRulesWorkerStore): user_id, rule_id, enabled, + is_default_rule, ) def _set_push_rule_enabled_txn( - self, txn, stream_id, event_stream_ordering, user_id, rule_id, enabled + self, + txn, + stream_id, + event_stream_ordering, + user_id, + rule_id, + enabled, + is_default_rule, ): new_id = self._push_rules_enable_id_gen.get_next() - self.db.simple_upsert_txn( + + if not is_default_rule: + # first check it exists; we need to lock for key share so that a + # transaction that deletes the push rule will conflict with this one. + # We also need a push_rule_enable row to exist for every push_rules + # row, otherwise it is possible to simultaneously delete a push rule + # (that has no _enable row) and enable it, resulting in a dangling + # _enable row. To solve this: we either need to use SERIALISABLE or + # ensure we always have a push_rule_enable row for every push_rule + # row. We chose the latter. + for_key_share = "FOR KEY SHARE" + if not isinstance(self.database_engine, PostgresEngine): + # For key share is not applicable/available on SQLite + for_key_share = "" + sql = ( + """ + SELECT 1 FROM push_rules + WHERE user_name = ? AND rule_id = ? + %s + """ + % for_key_share + ) + txn.execute(sql, (user_id, rule_id)) + if txn.fetchone() is None: + # needed to set NOT_FOUND code. + raise NotFoundError("Push rule does not exist.") + + self.db_pool.simple_upsert_txn( txn, "push_rules_enable", {"user_name": user_id, "rule_id": rule_id}, @@ -679,9 +684,32 @@ class PushRuleStore(PushRulesWorkerStore): op="ENABLE" if enabled else "DISABLE", ) - @defer.inlineCallbacks - def set_push_rule_actions(self, user_id, rule_id, actions, is_default_rule): - actions_json = json.dumps(actions) + async def set_push_rule_actions( + self, + user_id: str, + rule_id: str, + actions: List[Union[dict, str]], + is_default_rule: bool, + ) -> None: + """ + Sets the `actions` state of a push rule. + + Will throw NotFoundError if the rule does not exist; the Code for this + is NOT_FOUND. + + Args: + user_id: the user ID of the user who wishes to enable/disable the rule + e.g. '@tina:example.org' + rule_id: the full rule ID of the rule to be enabled/disabled + e.g. 'global/override/.m.rule.roomnotif' + or 'global/override/myCustomRule' + actions: A list of actions (each action being a dict or string), + e.g. ["notify", {"set_tweak": "highlight", "value": false}] + is_default_rule: True if and only if this is a server-default rule. + This skips the check for existence (as only user-created rules + are always stored in the database `push_rules` table). + """ + actions_json = json_encoder.encode(actions) def set_push_rule_actions_txn(txn, stream_id, event_stream_ordering): if is_default_rule: @@ -702,12 +730,19 @@ class PushRuleStore(PushRulesWorkerStore): update_stream=False, ) else: - self.db.simple_update_one_txn( - txn, - "push_rules", - {"user_name": user_id, "rule_id": rule_id}, - {"actions": actions_json}, - ) + try: + self.db_pool.simple_update_one_txn( + txn, + "push_rules", + {"user_name": user_id, "rule_id": rule_id}, + {"actions": actions_json}, + ) + except StoreError as serr: + if serr.code == 404: + # this sets the NOT_FOUND error Code + raise NotFoundError("Push rule does not exist") + else: + raise self._insert_push_rules_update_txn( txn, @@ -719,9 +754,10 @@ class PushRuleStore(PushRulesWorkerStore): data={"actions": actions_json}, ) - with self._push_rules_stream_id_gen.get_next() as ids: - stream_id, event_stream_ordering = ids - yield self.db.runInteraction( + async with self._push_rules_stream_id_gen.get_next() as stream_id: + event_stream_ordering = self._stream_id_gen.get_current_token() + + await self.db_pool.runInteraction( "set_push_rule_actions", set_push_rule_actions_txn, stream_id, @@ -741,7 +777,7 @@ class PushRuleStore(PushRulesWorkerStore): if data is not None: values.update(data) - self.db.simple_insert_txn(txn, "push_rules_stream", values=values) + self.db_pool.simple_insert_txn(txn, "push_rules_stream", values=values) txn.call_after(self.get_push_rules_for_user.invalidate, (user_id,)) txn.call_after(self.get_push_rules_enabled_for_user.invalidate, (user_id,)) @@ -749,11 +785,5 @@ class PushRuleStore(PushRulesWorkerStore): self.push_rules_stream_cache.entity_has_changed, user_id, stream_id ) - def get_push_rules_stream_token(self): - """Get the position of the push rules stream. - Returns a pair of a stream id for the push_rules stream and the - room stream ordering it corresponds to.""" - return self._push_rules_stream_id_gen.get_current_token() - def get_max_push_rules_stream_id(self): - return self.get_push_rules_stream_token()[0] + return self._push_rules_stream_id_gen.get_current_token() diff --git a/synapse/storage/data_stores/main/pusher.py b/synapse/storage/databases/main/pusher.py
index e18f1ca87c..df8609b97b 100644 --- a/synapse/storage/data_stores/main/pusher.py +++ b/synapse/storage/databases/main/pusher.py
@@ -19,10 +19,8 @@ from typing import Iterable, Iterator, List, Tuple from canonicaljson import encode_canonical_json -from twisted.internet import defer - from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.util.caches.descriptors import cachedInlineCallbacks, cachedList +from synapse.util.caches.descriptors import cached, cachedList logger = logging.getLogger(__name__) @@ -34,23 +32,22 @@ class PusherWorkerStore(SQLBaseStore): Drops any rows whose data cannot be decoded """ for r in rows: - dataJson = r["data"] + data_json = r["data"] try: - r["data"] = db_to_json(dataJson) + r["data"] = db_to_json(data_json) except Exception as e: logger.warning( "Invalid JSON in data for pusher %d: %s, %s", r["id"], - dataJson, + data_json, e.args[0], ) continue yield r - @defer.inlineCallbacks - def user_has_pusher(self, user_id): - ret = yield self.db.simple_select_one_onecol( + async def user_has_pusher(self, user_id): + ret = await self.db_pool.simple_select_one_onecol( "pushers", {"user_name": user_id}, "id", allow_none=True ) return ret is not None @@ -61,9 +58,8 @@ class PusherWorkerStore(SQLBaseStore): def get_pushers_by_user_id(self, user_id): return self.get_pushers_by({"user_name": user_id}) - @defer.inlineCallbacks - def get_pushers_by(self, keyvalues): - ret = yield self.db.simple_select_list( + async def get_pushers_by(self, keyvalues): + ret = await self.db_pool.simple_select_list( "pushers", keyvalues, [ @@ -87,16 +83,14 @@ class PusherWorkerStore(SQLBaseStore): ) return self._decode_pushers_rows(ret) - @defer.inlineCallbacks - def get_all_pushers(self): + async def get_all_pushers(self): def get_pushers(txn): txn.execute("SELECT * FROM pushers") - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) return self._decode_pushers_rows(rows) - rows = yield self.db.runInteraction("get_all_pushers", get_pushers) - return rows + return await self.db_pool.runInteraction("get_all_pushers", get_pushers) async def get_all_updated_pushers_rows( self, instance_name: str, last_id: int, current_id: int, limit: int @@ -160,23 +154,20 @@ class PusherWorkerStore(SQLBaseStore): return updates, upper_bound, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_updated_pushers_rows", get_all_updated_pushers_rows_txn ) - @cachedInlineCallbacks(num_args=1, max_entries=15000) - def get_if_user_has_pusher(self, user_id): + @cached(num_args=1, max_entries=15000) + async def get_if_user_has_pusher(self, user_id): # This only exists for the cachedList decorator raise NotImplementedError() @cachedList( - cached_method_name="get_if_user_has_pusher", - list_name="user_ids", - num_args=1, - inlineCallbacks=True, + cached_method_name="get_if_user_has_pusher", list_name="user_ids", num_args=1, ) - def get_if_users_have_pushers(self, user_ids): - rows = yield self.db.simple_select_many_batch( + async def get_if_users_have_pushers(self, user_ids): + rows = await self.db_pool.simple_select_many_batch( table="pushers", column="user_name", iterable=user_ids, @@ -189,34 +180,38 @@ class PusherWorkerStore(SQLBaseStore): return result - @defer.inlineCallbacks - def update_pusher_last_stream_ordering( + async def update_pusher_last_stream_ordering( self, app_id, pushkey, user_id, last_stream_ordering - ): - yield self.db.simple_update_one( + ) -> None: + await self.db_pool.simple_update_one( "pushers", {"app_id": app_id, "pushkey": pushkey, "user_name": user_id}, {"last_stream_ordering": last_stream_ordering}, desc="update_pusher_last_stream_ordering", ) - @defer.inlineCallbacks - def update_pusher_last_stream_ordering_and_success( - self, app_id, pushkey, user_id, last_stream_ordering, last_success - ): + async def update_pusher_last_stream_ordering_and_success( + self, + app_id: str, + pushkey: str, + user_id: str, + last_stream_ordering: int, + last_success: int, + ) -> bool: """Update the last stream ordering position we've processed up to for the given pusher. Args: - app_id (str) - pushkey (str) - last_stream_ordering (int) - last_success (int) + app_id + pushkey + user_id + last_stream_ordering + last_success Returns: - Deferred[bool]: True if the pusher still exists; False if it has been deleted. + True if the pusher still exists; False if it has been deleted. """ - updated = yield self.db.simple_update( + updated = await self.db_pool.simple_update( table="pushers", keyvalues={"app_id": app_id, "pushkey": pushkey, "user_name": user_id}, updatevalues={ @@ -228,18 +223,18 @@ class PusherWorkerStore(SQLBaseStore): return bool(updated) - @defer.inlineCallbacks - def update_pusher_failing_since(self, app_id, pushkey, user_id, failing_since): - yield self.db.simple_update( + async def update_pusher_failing_since( + self, app_id, pushkey, user_id, failing_since + ) -> None: + await self.db_pool.simple_update( table="pushers", keyvalues={"app_id": app_id, "pushkey": pushkey, "user_name": user_id}, updatevalues={"failing_since": failing_since}, desc="update_pusher_failing_since", ) - @defer.inlineCallbacks - def get_throttle_params_by_room(self, pusher_id): - res = yield self.db.simple_select_list( + async def get_throttle_params_by_room(self, pusher_id): + res = await self.db_pool.simple_select_list( "pusher_throttle", {"pusher": pusher_id}, ["room_id", "last_sent_ts", "throttle_ms"], @@ -255,11 +250,10 @@ class PusherWorkerStore(SQLBaseStore): return params_by_room - @defer.inlineCallbacks - def set_throttle_params(self, pusher_id, room_id, params): + async def set_throttle_params(self, pusher_id, room_id, params) -> None: # no need to lock because `pusher_throttle` has a primary key on # (pusher, room_id) so simple_upsert will retry - yield self.db.simple_upsert( + await self.db_pool.simple_upsert( "pusher_throttle", {"pusher": pusher_id, "room_id": room_id}, params, @@ -272,8 +266,7 @@ class PusherStore(PusherWorkerStore): def get_pushers_stream_token(self): return self._pushers_id_gen.get_current_token() - @defer.inlineCallbacks - def add_pusher( + async def add_pusher( self, user_id, access_token, @@ -287,11 +280,11 @@ class PusherStore(PusherWorkerStore): data, last_stream_ordering, profile_tag="", - ): - with self._pushers_id_gen.get_next() as stream_id: + ) -> None: + async with self._pushers_id_gen.get_next() as stream_id: # no need to lock because `pushers` has a unique key on # (app_id, pushkey, user_name) so simple_upsert will retry - yield self.db.simple_upsert( + await self.db_pool.simple_upsert( table="pushers", keyvalues={"app_id": app_id, "pushkey": pushkey, "user_name": user_id}, values={ @@ -316,21 +309,22 @@ class PusherStore(PusherWorkerStore): if user_has_pusher is not True: # invalidate, since we the user might not have had a pusher before - yield self.db.runInteraction( + await self.db_pool.runInteraction( "add_pusher", self._invalidate_cache_and_stream, self.get_if_user_has_pusher, (user_id,), ) - @defer.inlineCallbacks - def delete_pusher_by_app_id_pushkey_user_id(self, app_id, pushkey, user_id): + async def delete_pusher_by_app_id_pushkey_user_id( + self, app_id, pushkey, user_id + ) -> None: def delete_pusher_txn(txn, stream_id): self._invalidate_cache_and_stream( txn, self.get_if_user_has_pusher, (user_id,) ) - self.db.simple_delete_one_txn( + self.db_pool.simple_delete_one_txn( txn, "pushers", {"app_id": app_id, "pushkey": pushkey, "user_name": user_id}, @@ -339,7 +333,7 @@ class PusherStore(PusherWorkerStore): # it's possible for us to end up with duplicate rows for # (app_id, pushkey, user_id) at different stream_ids, but that # doesn't really matter. - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="deleted_pushers", values={ @@ -350,5 +344,7 @@ class PusherStore(PusherWorkerStore): }, ) - with self._pushers_id_gen.get_next() as stream_id: - yield self.db.runInteraction("delete_pusher", delete_pusher_txn, stream_id) + async with self._pushers_id_gen.get_next() as stream_id: + await self.db_pool.runInteraction( + "delete_pusher", delete_pusher_txn, stream_id + ) diff --git a/synapse/storage/data_stores/main/receipts.py b/synapse/storage/databases/main/receipts.py
index 1d723f2d34..c79ddff680 100644 --- a/synapse/storage/data_stores/main/receipts.py +++ b/synapse/storage/databases/main/receipts.py
@@ -16,33 +16,30 @@ import abc import logging -from typing import List, Tuple - -from canonicaljson import json +from typing import Any, Dict, List, Optional, Tuple from twisted.internet import defer from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.util.id_generators import StreamIdGenerator +from synapse.util import json_encoder from synapse.util.async_helpers import ObservableDeferred -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks, cachedList +from synapse.util.caches.descriptors import cached, cachedList from synapse.util.caches.stream_change_cache import StreamChangeCache logger = logging.getLogger(__name__) -class ReceiptsWorkerStore(SQLBaseStore): +# The ABCMeta metaclass ensures that it cannot be instantiated without +# the abstract methods being implemented. +class ReceiptsWorkerStore(SQLBaseStore, metaclass=abc.ABCMeta): """This is an abstract base class where subclasses must implement `get_max_receipt_stream_id` which can be called in the initializer. """ - # This ABCMeta metaclass ensures that we cannot be instantiated without - # the abstract methods being implemented. - __metaclass__ = abc.ABCMeta - - def __init__(self, database: Database, db_conn, hs): - super(ReceiptsWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._receipts_stream_cache = StreamChangeCache( "ReceiptsRoomChangeCache", self.get_max_receipt_stream_id() @@ -57,14 +54,16 @@ class ReceiptsWorkerStore(SQLBaseStore): """ raise NotImplementedError() - @cachedInlineCallbacks() - def get_users_with_read_receipts_in_room(self, room_id): - receipts = yield self.get_receipts_for_room(room_id, "m.read") + @cached() + async def get_users_with_read_receipts_in_room(self, room_id): + receipts = await self.get_receipts_for_room(room_id, "m.read") return {r["user_id"] for r in receipts} @cached(num_args=2) - def get_receipts_for_room(self, room_id, receipt_type): - return self.db.simple_select_list( + async def get_receipts_for_room( + self, room_id: str, receipt_type: str + ) -> List[Dict[str, Any]]: + return await self.db_pool.simple_select_list( table="receipts_linearized", keyvalues={"room_id": room_id, "receipt_type": receipt_type}, retcols=("user_id", "event_id"), @@ -72,8 +71,10 @@ class ReceiptsWorkerStore(SQLBaseStore): ) @cached(num_args=3) - def get_last_receipt_event_id_for_user(self, user_id, room_id, receipt_type): - return self.db.simple_select_one_onecol( + async def get_last_receipt_event_id_for_user( + self, user_id: str, room_id: str, receipt_type: str + ) -> Optional[str]: + return await self.db_pool.simple_select_one_onecol( table="receipts_linearized", keyvalues={ "room_id": room_id, @@ -85,9 +86,9 @@ class ReceiptsWorkerStore(SQLBaseStore): allow_none=True, ) - @cachedInlineCallbacks(num_args=2) - def get_receipts_for_user(self, user_id, receipt_type): - rows = yield self.db.simple_select_list( + @cached(num_args=2) + async def get_receipts_for_user(self, user_id, receipt_type): + rows = await self.db_pool.simple_select_list( table="receipts_linearized", keyvalues={"user_id": user_id, "receipt_type": receipt_type}, retcols=("room_id", "event_id"), @@ -96,8 +97,7 @@ class ReceiptsWorkerStore(SQLBaseStore): return {row["room_id"]: row["event_id"] for row in rows} - @defer.inlineCallbacks - def get_receipts_for_user_with_orderings(self, user_id, receipt_type): + async def get_receipts_for_user_with_orderings(self, user_id, receipt_type): def f(txn): sql = ( "SELECT rl.room_id, rl.event_id," @@ -111,7 +111,9 @@ class ReceiptsWorkerStore(SQLBaseStore): txn.execute(sql, (user_id,)) return txn.fetchall() - rows = yield self.db.runInteraction("get_receipts_for_user_with_orderings", f) + rows = await self.db_pool.runInteraction( + "get_receipts_for_user_with_orderings", f + ) return { row[0]: { "event_id": row[1], @@ -121,56 +123,61 @@ class ReceiptsWorkerStore(SQLBaseStore): for row in rows } - @defer.inlineCallbacks - def get_linearized_receipts_for_rooms(self, room_ids, to_key, from_key=None): + async def get_linearized_receipts_for_rooms( + self, room_ids: List[str], to_key: int, from_key: Optional[int] = None + ) -> List[dict]: """Get receipts for multiple rooms for sending to clients. Args: - room_ids (list): List of room_ids. - to_key (int): Max stream id to fetch receipts upto. - from_key (int): Min stream id to fetch receipts from. None fetches + room_id: List of room_ids. + to_key: Max stream id to fetch receipts upto. + from_key: Min stream id to fetch receipts from. None fetches from the start. Returns: - list: A list of receipts. + A list of receipts. """ room_ids = set(room_ids) if from_key is not None: # Only ask the database about rooms where there have been new # receipts added since `from_key` - room_ids = yield self._receipts_stream_cache.get_entities_changed( + room_ids = self._receipts_stream_cache.get_entities_changed( room_ids, from_key ) - results = yield self._get_linearized_receipts_for_rooms( + results = await self._get_linearized_receipts_for_rooms( room_ids, to_key, from_key=from_key ) return [ev for res in results.values() for ev in res] - def get_linearized_receipts_for_room(self, room_id, to_key, from_key=None): + async def get_linearized_receipts_for_room( + self, room_id: str, to_key: int, from_key: Optional[int] = None + ) -> List[dict]: """Get receipts for a single room for sending to clients. Args: - room_ids (str): The room id. - to_key (int): Max stream id to fetch receipts upto. - from_key (int): Min stream id to fetch receipts from. None fetches + room_ids: The room id. + to_key: Max stream id to fetch receipts upto. + from_key: Min stream id to fetch receipts from. None fetches from the start. Returns: - Deferred[list]: A list of receipts. + A list of receipts. """ if from_key is not None: # Check the cache first to see if any new receipts have been added # since`from_key`. If not we can no-op. if not self._receipts_stream_cache.has_entity_changed(room_id, from_key): - defer.succeed([]) + return [] - return self._get_linearized_receipts_for_room(room_id, to_key, from_key) + return await self._get_linearized_receipts_for_room(room_id, to_key, from_key) - @cachedInlineCallbacks(num_args=3, tree=True) - def _get_linearized_receipts_for_room(self, room_id, to_key, from_key=None): + @cached(num_args=3, tree=True) + async def _get_linearized_receipts_for_room( + self, room_id: str, to_key: int, from_key: Optional[int] = None + ) -> List[dict]: """See get_linearized_receipts_for_room """ @@ -190,11 +197,11 @@ class ReceiptsWorkerStore(SQLBaseStore): txn.execute(sql, (room_id, to_key)) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) return rows - rows = yield self.db.runInteraction("get_linearized_receipts_for_room", f) + rows = await self.db_pool.runInteraction("get_linearized_receipts_for_room", f) if not rows: return [] @@ -211,9 +218,8 @@ class ReceiptsWorkerStore(SQLBaseStore): cached_method_name="_get_linearized_receipts_for_room", list_name="room_ids", num_args=3, - inlineCallbacks=True, ) - def _get_linearized_receipts_for_rooms(self, room_ids, to_key, from_key=None): + async def _get_linearized_receipts_for_rooms(self, room_ids, to_key, from_key=None): if not room_ids: return {} @@ -240,9 +246,9 @@ class ReceiptsWorkerStore(SQLBaseStore): txn.execute(sql + clause, [to_key] + list(args)) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - txn_results = yield self.db.runInteraction( + txn_results = await self.db_pool.runInteraction( "_get_linearized_receipts_for_rooms", f ) @@ -268,12 +274,14 @@ class ReceiptsWorkerStore(SQLBaseStore): } return results - def get_users_sent_receipts_between(self, last_id: int, current_id: int): + async def get_users_sent_receipts_between( + self, last_id: int, current_id: int + ) -> List[str]: """Get all users who sent receipts between `last_id` exclusive and `current_id` inclusive. Returns: - Deferred[List[str]] + The list of users. """ if last_id == current_id: @@ -288,7 +296,7 @@ class ReceiptsWorkerStore(SQLBaseStore): return [r[0] for r in txn] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_sent_receipts_between", _get_users_sent_receipts_between_txn ) @@ -340,12 +348,12 @@ class ReceiptsWorkerStore(SQLBaseStore): return updates, upper_bound, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_updated_receipts", get_all_updated_receipts_txn ) def _invalidate_get_users_with_receipts_in_room( - self, room_id, receipt_type, user_id + self, room_id: str, receipt_type: str, user_id: str ): if receipt_type != "m.read": return @@ -371,14 +379,14 @@ class ReceiptsWorkerStore(SQLBaseStore): class ReceiptsStore(ReceiptsWorkerStore): - def __init__(self, database: Database, db_conn, hs): + def __init__(self, database: DatabasePool, db_conn, hs): # We instantiate this first as the ReceiptsWorkerStore constructor # needs to be able to call get_max_receipt_stream_id self._receipts_id_gen = StreamIdGenerator( db_conn, "receipts_linearized", "stream_id" ) - super(ReceiptsStore, self).__init__(database, db_conn, hs) + super().__init__(database, db_conn, hs) def get_max_receipt_stream_id(self): return self._receipts_id_gen.get_current_token() @@ -393,7 +401,7 @@ class ReceiptsStore(ReceiptsWorkerStore): otherwise, the rx timestamp of the event that the RR corresponds to (or 0 if the event is unknown) """ - res = self.db.simple_select_one_txn( + res = self.db_pool.simple_select_one_txn( txn, table="events", retcols=["stream_ordering", "received_ts"], @@ -446,7 +454,7 @@ class ReceiptsStore(ReceiptsWorkerStore): (user_id, room_id, receipt_type), ) - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="receipts_linearized", keyvalues={ @@ -457,7 +465,7 @@ class ReceiptsStore(ReceiptsWorkerStore): values={ "stream_id": stream_id, "event_id": event_id, - "data": json.dumps(data), + "data": json_encoder.encode(data), }, # receipts_linearized has a unique constraint on # (user_id, room_id, receipt_type), so no need to lock @@ -471,15 +479,21 @@ class ReceiptsStore(ReceiptsWorkerStore): return rx_ts - @defer.inlineCallbacks - def insert_receipt(self, room_id, receipt_type, user_id, event_ids, data): + async def insert_receipt( + self, + room_id: str, + receipt_type: str, + user_id: str, + event_ids: List[str], + data: dict, + ) -> Optional[Tuple[int, int]]: """Insert a receipt, either from local client or remote server. Automatically does conversion between linearized and graph representations. """ if not event_ids: - return + return None if len(event_ids) == 1: linearized_event_id = event_ids[0] @@ -506,13 +520,12 @@ class ReceiptsStore(ReceiptsWorkerStore): else: raise RuntimeError("Unrecognized event_ids: %r" % (event_ids,)) - linearized_event_id = yield self.db.runInteraction( + linearized_event_id = await self.db_pool.runInteraction( "insert_receipt_conv", graph_to_linear ) - stream_id_manager = self._receipts_id_gen.get_next() - with stream_id_manager as stream_id: - event_ts = yield self.db.runInteraction( + async with self._receipts_id_gen.get_next() as stream_id: + event_ts = await self.db_pool.runInteraction( "insert_linearized_receipt", self.insert_linearized_receipt_txn, room_id, @@ -534,14 +547,16 @@ class ReceiptsStore(ReceiptsWorkerStore): now - event_ts, ) - yield self.insert_graph_receipt(room_id, receipt_type, user_id, event_ids, data) + await self.insert_graph_receipt(room_id, receipt_type, user_id, event_ids, data) max_persisted_id = self._receipts_id_gen.get_current_token() return stream_id, max_persisted_id - def insert_graph_receipt(self, room_id, receipt_type, user_id, event_ids, data): - return self.db.runInteraction( + async def insert_graph_receipt( + self, room_id, receipt_type, user_id, event_ids, data + ): + return await self.db_pool.runInteraction( "insert_graph_receipt", self.insert_graph_receipt_txn, room_id, @@ -567,7 +582,7 @@ class ReceiptsStore(ReceiptsWorkerStore): self._get_linearized_receipts_for_room.invalidate_many, (room_id,) ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="receipts_graph", keyvalues={ @@ -576,14 +591,14 @@ class ReceiptsStore(ReceiptsWorkerStore): "user_id": user_id, }, ) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="receipts_graph", values={ "room_id": room_id, "receipt_type": receipt_type, "user_id": user_id, - "event_ids": json.dumps(event_ids), - "data": json.dumps(data), + "event_ids": json_encoder.encode(event_ids), + "data": json_encoder.encode(data), }, ) diff --git a/synapse/storage/data_stores/main/registration.py b/synapse/storage/databases/main/registration.py
index 7930041998..f949308994 100644 --- a/synapse/storage/data_stores/main/registration.py +++ b/synapse/storage/databases/main/registration.py
@@ -17,20 +17,17 @@ import logging import re -from typing import List, Optional - -from twisted.internet import defer -from twisted.internet.defer import Deferred +from typing import Any, Dict, List, Optional, Tuple from synapse.api.constants import UserTypes from synapse.api.errors import Codes, StoreError, SynapseError, ThreepidValidationError from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.types import Cursor from synapse.storage.util.sequence import build_sequence_generator from synapse.types import UserID -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks +from synapse.util.caches.descriptors import cached THIRTY_MINUTES_IN_MS = 30 * 60 * 1000 @@ -38,19 +35,22 @@ logger = logging.getLogger(__name__) class RegistrationWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(RegistrationWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.config = hs.config self.clock = hs.get_clock() + # Note: we don't check this sequence for consistency as we'd have to + # call `find_max_generated_user_id_localpart` each time, which is + # expensive if there are many entries. self._user_id_seq = build_sequence_generator( database.engine, find_max_generated_user_id_localpart, "user_id_seq", ) @cached() - def get_user_by_id(self, user_id): - return self.db.simple_select_one( + async def get_user_by_id(self, user_id: str) -> Optional[Dict[str, Any]]: + return await self.db_pool.simple_select_one( table="users", keyvalues={"name": user_id}, retcols=[ @@ -69,19 +69,15 @@ class RegistrationWorkerStore(SQLBaseStore): desc="get_user_by_id", ) - @defer.inlineCallbacks - def is_trial_user(self, user_id): + async def is_trial_user(self, user_id: str) -> bool: """Checks if user is in the "trial" period, i.e. within the first N days of registration defined by `mau_trial_days` config Args: - user_id (str) - - Returns: - Deferred[bool] + user_id: The user to check for trial status. """ - info = yield self.get_user_by_id(user_id) + info = await self.get_user_by_id(user_id) if not info: return False @@ -91,42 +87,39 @@ class RegistrationWorkerStore(SQLBaseStore): return is_trial @cached() - def get_user_by_access_token(self, token): + async def get_user_by_access_token(self, token: str) -> Optional[dict]: """Get a user from the given access token. Args: - token (str): The access token of a user. + token: The access token of a user. Returns: - defer.Deferred: None, if the token did not match, otherwise dict - including the keys `name`, `is_guest`, `device_id`, `token_id`, - `valid_until_ms`. + None, if the token did not match, otherwise dict + including the keys `name`, `is_guest`, `device_id`, `token_id`, + `valid_until_ms`. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_user_by_access_token", self._query_for_auth, token ) - @cachedInlineCallbacks() - def get_expiration_ts_for_user(self, user_id): + @cached() + async def get_expiration_ts_for_user(self, user_id: str) -> Optional[int]: """Get the expiration timestamp for the account bearing a given user ID. Args: - user_id (str): The ID of the user. + user_id: The ID of the user. Returns: - defer.Deferred: None, if the account has no expiration timestamp, - otherwise int representation of the timestamp (as a number of - milliseconds since epoch). + None, if the account has no expiration timestamp, otherwise int + representation of the timestamp (as a number of milliseconds since epoch). """ - res = yield self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="account_validity", keyvalues={"user_id": user_id}, retcol="expiration_ts_ms", allow_none=True, desc="get_expiration_ts_for_user", ) - return res - @defer.inlineCallbacks - def is_account_expired(self, user_id: str, current_ts: int): + async def is_account_expired(self, user_id: str, current_ts: int) -> bool: """ Returns whether an user account is expired. @@ -135,33 +128,34 @@ class RegistrationWorkerStore(SQLBaseStore): current_ts: The current timestamp Returns: - Deferred[bool]: whether the user account has expired + Whether the user account has expired """ - expiration_ts = yield self.get_expiration_ts_for_user(user_id) - if expiration_ts is not None and current_ts >= expiration_ts: - return True - return False - - @defer.inlineCallbacks - def set_account_validity_for_user( - self, user_id, expiration_ts, email_sent, renewal_token=None - ): + expiration_ts = await self.get_expiration_ts_for_user(user_id) + return expiration_ts is not None and current_ts >= expiration_ts + + async def set_account_validity_for_user( + self, + user_id: str, + expiration_ts: int, + email_sent: bool, + renewal_token: Optional[str] = None, + ) -> None: """Updates the account validity properties of the given account, with the given values. Args: - user_id (str): ID of the account to update properties for. - expiration_ts (int): New expiration date, as a timestamp in milliseconds + user_id: ID of the account to update properties for. + expiration_ts: New expiration date, as a timestamp in milliseconds since epoch. - email_sent (bool): True means a renewal email has been sent for this - account and there's no need to send another one for the current validity + email_sent: True means a renewal email has been sent for this account + and there's no need to send another one for the current validity period. - renewal_token (str): Renewal token the user can use to extend the validity + renewal_token: Renewal token the user can use to extend the validity of their account. Defaults to no token. """ def set_account_validity_for_user_txn(txn): - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn=txn, table="account_validity", keyvalues={"user_id": user_id}, @@ -175,12 +169,11 @@ class RegistrationWorkerStore(SQLBaseStore): txn, self.get_expiration_ts_for_user, (user_id,) ) - yield self.db.runInteraction( + await self.db_pool.runInteraction( "set_account_validity_for_user", set_account_validity_for_user_txn ) - @defer.inlineCallbacks - def get_expired_users(self): + async def get_expired_users(self): """Get UserIDs of all expired users. Users who are not active, or do not have profile information, are @@ -205,76 +198,71 @@ class RegistrationWorkerStore(SQLBaseStore): return [UserID.from_string(row[0]) for row in rows] - res = yield self.db.runInteraction( + res = await self.db_pool.runInteraction( "get_expired_users", get_expired_users_txn, self.clock.time_msec() ) + return res - @defer.inlineCallbacks - def set_renewal_token_for_user(self, user_id, renewal_token): + async def set_renewal_token_for_user( + self, user_id: str, renewal_token: str + ) -> None: """Defines a renewal token for a given user. Args: - user_id (str): ID of the user to set the renewal token for. - renewal_token (str): Random unique string that will be used to renew the + user_id: ID of the user to set the renewal token for. + renewal_token: Random unique string that will be used to renew the user's account. Raises: StoreError: The provided token is already set for another user. """ - yield self.db.simple_update_one( + await self.db_pool.simple_update_one( table="account_validity", keyvalues={"user_id": user_id}, updatevalues={"renewal_token": renewal_token}, desc="set_renewal_token_for_user", ) - @defer.inlineCallbacks - def get_user_from_renewal_token(self, renewal_token): + async def get_user_from_renewal_token(self, renewal_token: str) -> str: """Get a user ID from a renewal token. Args: - renewal_token (str): The renewal token to perform the lookup with. + renewal_token: The renewal token to perform the lookup with. Returns: - defer.Deferred[str]: The ID of the user to which the token belongs. + The ID of the user to which the token belongs. """ - res = yield self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="account_validity", keyvalues={"renewal_token": renewal_token}, retcol="user_id", desc="get_user_from_renewal_token", ) - return res - - @defer.inlineCallbacks - def get_renewal_token_for_user(self, user_id): + async def get_renewal_token_for_user(self, user_id: str) -> str: """Get the renewal token associated with a given user ID. Args: - user_id (str): The user ID to lookup a token for. + user_id: The user ID to lookup a token for. Returns: - defer.Deferred[str]: The renewal token associated with this user ID. + The renewal token associated with this user ID. """ - res = yield self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="account_validity", keyvalues={"user_id": user_id}, retcol="renewal_token", desc="get_renewal_token_for_user", ) - return res - - @defer.inlineCallbacks - def get_users_expiring_soon(self): + async def get_users_expiring_soon(self) -> List[Dict[str, int]]: """Selects users whose account will expire in the [now, now + renew_at] time window (see configuration for account_validity for information on what renew_at refers to). Returns: - Deferred: Resolves to a list[dict[user_id (str), expiration_ts_ms (int)]] + A list of dictionaries mapping user ID to expiration time (in milliseconds). """ def select_users_txn(txn, now_ms, renew_at): @@ -284,50 +272,45 @@ class RegistrationWorkerStore(SQLBaseStore): ) values = [False, now_ms, renew_at] txn.execute(sql, values) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - res = yield self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_expiring_soon", select_users_txn, self.clock.time_msec(), self.config.account_validity.renew_at, ) - return res - - @defer.inlineCallbacks - def set_renewal_mail_status(self, user_id, email_sent): + async def set_renewal_mail_status(self, user_id: str, email_sent: bool) -> None: """Sets or unsets the flag that indicates whether a renewal email has been sent to the user (and the user hasn't renewed their account yet). Args: - user_id (str): ID of the user to set/unset the flag for. - email_sent (bool): Flag which indicates whether a renewal email has been sent + user_id: ID of the user to set/unset the flag for. + email_sent: Flag which indicates whether a renewal email has been sent to this user. """ - yield self.db.simple_update_one( + await self.db_pool.simple_update_one( table="account_validity", keyvalues={"user_id": user_id}, updatevalues={"email_sent": email_sent}, desc="set_renewal_mail_status", ) - @defer.inlineCallbacks - def delete_account_validity_for_user(self, user_id): + async def delete_account_validity_for_user(self, user_id: str) -> None: """Deletes the entry for the given user in the account validity table, removing their expiration date and renewal token. Args: - user_id (str): ID of the user to remove from the account validity table. + user_id: ID of the user to remove from the account validity table. """ - yield self.db.simple_delete_one( + await self.db_pool.simple_delete_one( table="account_validity", keyvalues={"user_id": user_id}, desc="delete_account_validity_for_user", ) - @defer.inlineCallbacks - def get_info_for_users( + async def get_info_for_users( self, user_ids: List[str], ): """Return the user info for a given set of users @@ -361,7 +344,7 @@ class RegistrationWorkerStore(SQLBaseStore): return rows - info_rows = yield self.db.runInteraction( + info_rows = await self.db_pool.runInteraction( "get_info_for_users", _get_info_for_users_txn ) @@ -375,16 +358,16 @@ class RegistrationWorkerStore(SQLBaseStore): for user_id, deactivated, expiration in info_rows } - async def is_server_admin(self, user): + async def is_server_admin(self, user: UserID) -> bool: """Determines if a user is an admin of this homeserver. Args: - user (UserID): user ID of the user to test + user: user ID of the user to test - Returns (bool): + Returns: true iff the user is a server admin, false otherwise. """ - res = await self.db.simple_select_one_onecol( + res = await self.db_pool.simple_select_one_onecol( table="users", keyvalues={"name": user.to_string()}, retcol="admin", @@ -394,28 +377,27 @@ class RegistrationWorkerStore(SQLBaseStore): return bool(res) if res else False - def set_server_admin(self, user, admin): + async def set_server_admin(self, user: UserID, admin: bool) -> None: """Sets whether a user is an admin of this homeserver. Args: - user (UserID): user ID of the user to test - admin (bool): true iff the user is to be a server admin, - false otherwise. + user: user ID of the user to test + admin: true iff the user is to be a server admin, false otherwise. """ def set_server_admin_txn(txn): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, "users", {"name": user.to_string()}, {"admin": 1 if admin else 0} ) self._invalidate_cache_and_stream( txn, self.get_user_by_id, (user.to_string(),) ) - return self.db.runInteraction("set_server_admin", set_server_admin_txn) + await self.db_pool.runInteraction("set_server_admin", set_server_admin_txn) def _query_for_auth(self, txn, token): sql = ( - "SELECT users.name, users.is_guest, access_tokens.id as token_id," + "SELECT users.name, users.is_guest, users.shadow_banned, access_tokens.id as token_id," " access_tokens.device_id, access_tokens.valid_until_ms" " FROM users" " INNER JOIN access_tokens on users.name = access_tokens.user_id" @@ -423,43 +405,42 @@ class RegistrationWorkerStore(SQLBaseStore): ) txn.execute(sql, (token,)) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if rows: return rows[0] return None - @cachedInlineCallbacks() - def is_real_user(self, user_id): + @cached() + async def is_real_user(self, user_id: str) -> bool: """Determines if the user is a real user, ie does not have a 'user_type'. Args: - user_id (str): user id to test + user_id: user id to test Returns: - Deferred[bool]: True if user 'user_type' is null or empty string + True if user 'user_type' is null or empty string """ - res = yield self.db.runInteraction( + return await self.db_pool.runInteraction( "is_real_user", self.is_real_user_txn, user_id ) - return res @cached() - def is_support_user(self, user_id): + async def is_support_user(self, user_id: str) -> bool: """Determines if the user is of type UserTypes.SUPPORT Args: - user_id (str): user id to test + user_id: user id to test Returns: - Deferred[bool]: True if user is of type UserTypes.SUPPORT + True if user is of type UserTypes.SUPPORT """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "is_support_user", self.is_support_user_txn, user_id ) def is_real_user_txn(self, txn, user_id): - res = self.db.simple_select_one_onecol_txn( + res = self.db_pool.simple_select_one_onecol_txn( txn=txn, table="users", keyvalues={"name": user_id}, @@ -469,7 +450,7 @@ class RegistrationWorkerStore(SQLBaseStore): return res is None def is_support_user_txn(self, txn, user_id): - res = self.db.simple_select_one_onecol_txn( + res = self.db_pool.simple_select_one_onecol_txn( txn=txn, table="users", keyvalues={"name": user_id}, @@ -478,9 +459,11 @@ class RegistrationWorkerStore(SQLBaseStore): ) return True if res == UserTypes.SUPPORT else False - def get_users_by_id_case_insensitive(self, user_id): + async def get_users_by_id_case_insensitive(self, user_id: str) -> Dict[str, str]: """Gets users that match user_id case insensitively. - Returns a mapping of user_id -> password_hash. + + Returns: + A mapping of user_id -> password_hash. """ def f(txn): @@ -488,11 +471,11 @@ class RegistrationWorkerStore(SQLBaseStore): txn.execute(sql, (user_id,)) return dict(txn) - return self.db.runInteraction("get_users_by_id_case_insensitive", f) + return await self.db_pool.runInteraction("get_users_by_id_case_insensitive", f) async def get_user_by_external_id( self, auth_provider: str, external_id: str - ) -> str: + ) -> Optional[str]: """Look up a user by their external auth id Args: @@ -500,9 +483,9 @@ class RegistrationWorkerStore(SQLBaseStore): external_id: id on that system Returns: - str|None: the mxid of the user, or None if they are not known + the mxid of the user, or None if they are not known """ - return await self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="user_external_ids", keyvalues={"auth_provider": auth_provider, "external_id": external_id}, retcol="user_id", @@ -510,21 +493,19 @@ class RegistrationWorkerStore(SQLBaseStore): desc="get_user_by_external_id", ) - @defer.inlineCallbacks - def count_all_users(self): + async def count_all_users(self): """Counts all users registered on the homeserver.""" def _count_users(txn): txn.execute("SELECT COUNT(*) AS users FROM users") - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if rows: return rows[0]["users"] return 0 - ret = yield self.db.runInteraction("count_users", _count_users) - return ret + return await self.db_pool.runInteraction("count_users", _count_users) - def count_daily_user_type(self): + async def count_daily_user_type(self) -> Dict[str, int]: """ Counts 1) native non guest users 2) native guests users @@ -553,10 +534,11 @@ class RegistrationWorkerStore(SQLBaseStore): results[row[0]] = row[1] return results - return self.db.runInteraction("count_daily_user_type", _count_daily_user_type) + return await self.db_pool.runInteraction( + "count_daily_user_type", _count_daily_user_type + ) - @defer.inlineCallbacks - def count_nonbridged_users(self): + async def count_nonbridged_users(self): def _count_users(txn): txn.execute( """ @@ -567,29 +549,26 @@ class RegistrationWorkerStore(SQLBaseStore): (count,) = txn.fetchone() return count - ret = yield self.db.runInteraction("count_users", _count_users) - return ret + return await self.db_pool.runInteraction("count_users", _count_users) - @defer.inlineCallbacks - def count_real_users(self): + async def count_real_users(self): """Counts all users without a special user_type registered on the homeserver.""" def _count_users(txn): txn.execute("SELECT COUNT(*) AS users FROM users where user_type is null") - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if rows: return rows[0]["users"] return 0 - ret = yield self.db.runInteraction("count_real_users", _count_users) - return ret + return await self.db_pool.runInteraction("count_real_users", _count_users) async def generate_user_id(self) -> str: """Generate a suitable localpart for a guest user Returns: a (hopefully) free localpart """ - next_id = await self.db.runInteraction( + next_id = await self.db_pool.runInteraction( "generate_user_id", self._user_id_seq.get_next_id_txn ) @@ -605,7 +584,7 @@ class RegistrationWorkerStore(SQLBaseStore): Returns: The user ID or None if no user id/threepid mapping exists """ - user_id = await self.db.runInteraction( + user_id = await self.db_pool.runInteraction( "get_user_id_by_threepid", self.get_user_id_by_threepid_txn, medium, address ) return user_id @@ -621,7 +600,7 @@ class RegistrationWorkerStore(SQLBaseStore): Returns: str|None: user id or None if no user id/threepid mapping exists """ - ret = self.db.simple_select_one_txn( + ret = self.db_pool.simple_select_one_txn( txn, "user_threepids", {"medium": medium, "address": address}, @@ -632,61 +611,57 @@ class RegistrationWorkerStore(SQLBaseStore): return ret["user_id"] return None - @defer.inlineCallbacks - def user_add_threepid(self, user_id, medium, address, validated_at, added_at): - yield self.db.simple_upsert( + async def user_add_threepid(self, user_id, medium, address, validated_at, added_at): + await self.db_pool.simple_upsert( "user_threepids", {"medium": medium, "address": address}, {"user_id": user_id, "validated_at": validated_at, "added_at": added_at}, ) - @defer.inlineCallbacks - def user_get_threepids(self, user_id): - ret = yield self.db.simple_select_list( + async def user_get_threepids(self, user_id): + return await self.db_pool.simple_select_list( "user_threepids", {"user_id": user_id}, ["medium", "address", "validated_at", "added_at"], "user_get_threepids", ) - return ret - def user_delete_threepid(self, user_id, medium, address): - return self.db.simple_delete( + async def user_delete_threepid(self, user_id, medium, address) -> None: + await self.db_pool.simple_delete( "user_threepids", keyvalues={"user_id": user_id, "medium": medium, "address": address}, desc="user_delete_threepid", ) - def user_delete_threepids(self, user_id: str): + async def user_delete_threepids(self, user_id: str) -> None: """Delete all threepid this user has bound Args: user_id: The user id to delete all threepids of """ - return self.db.simple_delete( + await self.db_pool.simple_delete( "user_threepids", keyvalues={"user_id": user_id}, desc="user_delete_threepids", ) - def add_user_bound_threepid(self, user_id, medium, address, id_server): + async def add_user_bound_threepid( + self, user_id: str, medium: str, address: str, id_server: str + ): """The server proxied a bind request to the given identity server on behalf of the given user. We need to remember this in case the user asks us to unbind the threepid. Args: - user_id (str) - medium (str) - address (str) - id_server (str) - - Returns: - Deferred + user_id + medium + address + id_server """ # We need to use an upsert, in case they user had already bound the # threepid - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="user_threepid_id_server", keyvalues={ "user_id": user_id, @@ -699,41 +674,40 @@ class RegistrationWorkerStore(SQLBaseStore): desc="add_user_bound_threepid", ) - def user_get_bound_threepids(self, user_id): + async def user_get_bound_threepids(self, user_id: str) -> List[Dict[str, Any]]: """Get the threepids that a user has bound to an identity server through the homeserver The homeserver remembers where binds to an identity server occurred. Using this method can retrieve those threepids. Args: - user_id (str): The ID of the user to retrieve threepids for + user_id: The ID of the user to retrieve threepids for Returns: - Deferred[list[dict]]: List of dictionaries containing the following: + List of dictionaries containing the following keys: medium (str): The medium of the threepid (e.g "email") address (str): The address of the threepid (e.g "bob@example.com") """ - return self.db.simple_select_list( + return await self.db_pool.simple_select_list( table="user_threepid_id_server", keyvalues={"user_id": user_id}, retcols=["medium", "address"], desc="user_get_bound_threepids", ) - def remove_user_bound_threepid(self, user_id, medium, address, id_server): + async def remove_user_bound_threepid( + self, user_id: str, medium: str, address: str, id_server: str + ) -> None: """The server proxied an unbind request to the given identity server on behalf of the given user, so we remove the mapping of threepid to identity server. Args: - user_id (str) - medium (str) - address (str) - id_server (str) - - Returns: - Deferred + user_id + medium + address + id_server """ - return self.db.simple_delete( + await self.db_pool.simple_delete( table="user_threepid_id_server", keyvalues={ "user_id": user_id, @@ -744,37 +718,39 @@ class RegistrationWorkerStore(SQLBaseStore): desc="remove_user_bound_threepid", ) - def get_id_servers_user_bound(self, user_id, medium, address): + async def get_id_servers_user_bound( + self, user_id: str, medium: str, address: str + ) -> List[str]: """Get the list of identity servers that the server proxied bind requests to for given user and threepid Args: - user_id (str) - medium (str) - address (str) + user_id: The user to query for identity servers. + medium: The medium to query for identity servers. + address: The address to query for identity servers. Returns: - Deferred[list[str]]: Resolves to a list of identity servers + A list of identity servers """ - return self.db.simple_select_onecol( + return await self.db_pool.simple_select_onecol( table="user_threepid_id_server", keyvalues={"user_id": user_id, "medium": medium, "address": address}, retcol="id_server", desc="get_id_servers_user_bound", ) - @cachedInlineCallbacks() - def get_user_deactivated_status(self, user_id): + @cached() + async def get_user_deactivated_status(self, user_id: str) -> bool: """Retrieve the value for the `deactivated` property for the provided user. Args: - user_id (str): The ID of the user to retrieve the status for. + user_id: The ID of the user to retrieve the status for. Returns: - defer.Deferred(bool): The requested value. + True if the user was deactivated, false if the user is still active. """ - res = yield self.db.simple_select_one_onecol( + res = await self.db_pool.simple_select_one_onecol( table="users", keyvalues={"name": user_id}, retcol="deactivated", @@ -784,24 +760,29 @@ class RegistrationWorkerStore(SQLBaseStore): # Convert the integer into a boolean. return res == 1 - def get_threepid_validation_session( - self, medium, client_secret, address=None, sid=None, validated=True - ): + async def get_threepid_validation_session( + self, + medium: Optional[str], + client_secret: str, + address: Optional[str] = None, + sid: Optional[str] = None, + validated: Optional[bool] = True, + ) -> Optional[Dict[str, Any]]: """Gets a session_id and last_send_attempt (if available) for a combination of validation metadata Args: - medium (str|None): The medium of the 3PID - address (str|None): The address of the 3PID - sid (str|None): The ID of the validation session - client_secret (str): A unique string provided by the client to help identify this + medium: The medium of the 3PID + client_secret: A unique string provided by the client to help identify this validation attempt - validated (bool|None): Whether sessions should be filtered by + address: The address of the 3PID + sid: The ID of the validation session + validated: Whether sessions should be filtered by whether they have been validated already or not. None to perform no filtering Returns: - Deferred[dict|None]: A dict containing the following: + A dict containing the following: * address - address of the 3pid * medium - medium of the 3pid * client_secret - a secret provided by the client for this validation session @@ -841,57 +822,57 @@ class RegistrationWorkerStore(SQLBaseStore): sql += " LIMIT 1" txn.execute(sql, list(keyvalues.values())) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return None return rows[0] - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_threepid_validation_session", get_threepid_validation_session_txn ) - def delete_threepid_session(self, session_id): + async def delete_threepid_session(self, session_id: str) -> None: """Removes a threepid validation session from the database. This can be done after validation has been performed and whatever action was waiting on it has been carried out Args: - session_id (str): The ID of the session to delete + session_id: The ID of the session to delete """ def delete_threepid_session_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="threepid_validation_token", keyvalues={"session_id": session_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="threepid_validation_session", keyvalues={"session_id": session_id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_threepid_session", delete_threepid_session_txn ) class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): - def __init__(self, database: Database, db_conn, hs): - super(RegistrationBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.clock = hs.get_clock() self.config = hs.config - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "access_tokens_device_index", index_name="access_tokens_device_id", table="access_tokens", columns=["user_id", "device_id"], ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "users_creation_ts", index_name="users_creation_ts", table="users", @@ -901,18 +882,19 @@ class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): # we no longer use refresh tokens, but it's possible that some people # might have a background update queued to build this index. Just # clear the background update. - self.db.updates.register_noop_background_update("refresh_tokens_device_index") + self.db_pool.updates.register_noop_background_update( + "refresh_tokens_device_index" + ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "user_threepids_grandfather", self._bg_user_threepids_grandfather ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "users_set_deactivated_flag", self._background_update_set_deactivated_flag ) - @defer.inlineCallbacks - def _background_update_set_deactivated_flag(self, progress, batch_size): + async def _background_update_set_deactivated_flag(self, progress, batch_size): """Retrieves a list of all deactivated users and sets the 'deactivated' flag to 1 for each of them. """ @@ -940,7 +922,7 @@ class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): (last_user, batch_size), ) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return True, 0 @@ -954,7 +936,7 @@ class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): logger.info("Marked %d rows as deactivated", rows_processed_nb) - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "users_set_deactivated_flag", {"user_id": rows[-1]["name"]} ) @@ -963,17 +945,18 @@ class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): else: return False, len(rows) - end, nb_processed = yield self.db.runInteraction( + end, nb_processed = await self.db_pool.runInteraction( "users_set_deactivated_flag", _background_update_set_deactivated_flag_txn ) if end: - yield self.db.updates._end_background_update("users_set_deactivated_flag") + await self.db_pool.updates._end_background_update( + "users_set_deactivated_flag" + ) return nb_processed - @defer.inlineCallbacks - def _bg_user_threepids_grandfather(self, progress, batch_size): + async def _bg_user_threepids_grandfather(self, progress, batch_size): """We now track which identity servers a user binds their 3PID to, so we need to handle the case of existing bindings where we didn't track this. @@ -994,20 +977,21 @@ class RegistrationBackgroundUpdateStore(RegistrationWorkerStore): txn.executemany(sql, [(id_server,) for id_server in id_servers]) if id_servers: - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_bg_user_threepids_grandfather", _bg_user_threepids_grandfather_txn ) - yield self.db.updates._end_background_update("user_threepids_grandfather") + await self.db_pool.updates._end_background_update("user_threepids_grandfather") return 1 class RegistrationStore(RegistrationBackgroundUpdateStore): - def __init__(self, database: Database, db_conn, hs): - super(RegistrationStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self._account_validity = hs.config.account_validity + self._ignore_unknown_session_error = hs.config.request_token_inhibit_3pid_errors if self._account_validity.enabled: self._clock.call_later( @@ -1028,23 +1012,26 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): hs.get_clock().looping_call(start_cull, THIRTY_MINUTES_IN_MS) - @defer.inlineCallbacks - def add_access_token_to_user(self, user_id, token, device_id, valid_until_ms): + async def add_access_token_to_user( + self, + user_id: str, + token: str, + device_id: Optional[str], + valid_until_ms: Optional[int], + ) -> None: """Adds an access token for the given user. Args: - user_id (str): The user ID. - token (str): The new access token to add. - device_id (str): ID of the device to associate with the access - token - valid_until_ms (int|None): when the token is valid until. None for - no expiry. + user_id: The user ID. + token: The new access token to add. + device_id: ID of the device to associate with the access token + valid_until_ms: when the token is valid until. None for no expiry. Raises: StoreError if there was a problem adding this. """ next_id = self._access_tokens_id_gen.get_next() - yield self.db.simple_insert( + await self.db_pool.simple_insert( "access_tokens", { "id": next_id, @@ -1056,40 +1043,40 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): desc="add_access_token_to_user", ) - def register_user( + async def register_user( self, - user_id, - password_hash=None, - was_guest=False, - make_guest=False, - appservice_id=None, - create_profile_with_displayname=None, - admin=False, - user_type=None, - ): + user_id: str, + password_hash: Optional[str] = None, + was_guest: bool = False, + make_guest: bool = False, + appservice_id: Optional[str] = None, + create_profile_with_displayname: Optional[str] = None, + admin: bool = False, + user_type: Optional[str] = None, + shadow_banned: bool = False, + ) -> None: """Attempts to register an account. Args: - user_id (str): The desired user ID to register. - password_hash (str|None): Optional. The password hash for this user. - was_guest (bool): Optional. Whether this is a guest account being - upgraded to a non-guest account. - make_guest (boolean): True if the the new user should be guest, - false to add a regular user account. - appservice_id (str): The ID of the appservice registering the user. - create_profile_with_displayname (unicode): Optionally create a profile for + user_id: The desired user ID to register. + password_hash: Optional. The password hash for this user. + was_guest: Whether this is a guest account being upgraded to a + non-guest account. + make_guest: True if the the new user should be guest, false to add a + regular user account. + appservice_id: The ID of the appservice registering the user. + create_profile_with_displayname: Optionally create a profile for the user, setting their displayname to the given value - admin (boolean): is an admin user? - user_type (str|None): type of user. One of the values from - api.constants.UserTypes, or None for a normal user. + admin: is an admin user? + user_type: type of user. One of the values from api.constants.UserTypes, + or None for a normal user. + shadow_banned: Whether the user is shadow-banned, i.e. they may be + told their requests succeeded but we ignore them. Raises: StoreError if the user_id could not be registered. - - Returns: - Deferred """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "register_user", self._register_user, user_id, @@ -1100,6 +1087,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): create_profile_with_displayname, admin, user_type, + shadow_banned, ) def _register_user( @@ -1113,6 +1101,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): create_profile_with_displayname, admin, user_type, + shadow_banned, ): user_id_obj = UserID.from_string(user_id) @@ -1123,7 +1112,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): # Ensure that the guest user actually exists # ``allow_none=False`` makes this raise an exception # if the row isn't in the database. - self.db.simple_select_one_txn( + self.db_pool.simple_select_one_txn( txn, "users", keyvalues={"name": user_id, "is_guest": 1}, @@ -1131,7 +1120,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): allow_none=False, ) - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, "users", keyvalues={"name": user_id, "is_guest": 1}, @@ -1142,10 +1131,11 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): "appservice_id": appservice_id, "admin": 1 if admin else 0, "user_type": user_type, + "shadow_banned": shadow_banned, }, ) else: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "users", values={ @@ -1156,6 +1146,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): "appservice_id": appservice_id, "admin": 1 if admin else 0, "user_type": user_type, + "shadow_banned": shadow_banned, }, ) @@ -1188,11 +1179,10 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) self._invalidate_cache_and_stream(txn, self.get_user_by_id, (user_id,)) - txn.call_after(self.is_guest.invalidate, (user_id,)) - def record_user_external_id( + async def record_user_external_id( self, auth_provider: str, external_id: str, user_id: str - ) -> Deferred: + ) -> None: """Record a mapping from an external user id to a mxid Args: @@ -1200,7 +1190,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): external_id: id on that system user_id: complete mxid that it is mapped to """ - return self.db.simple_insert( + await self.db_pool.simple_insert( table="user_external_ids", values={ "auth_provider": auth_provider, @@ -1210,7 +1200,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): desc="record_user_external_id", ) - def user_set_password_hash(self, user_id, password_hash): + async def user_set_password_hash(self, user_id: str, password_hash: str) -> None: """ NB. This does *not* evict any cache because the one use for this removes most of the entries subsequently anyway so it would be @@ -1218,29 +1208,30 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): """ def user_set_password_hash_txn(txn): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, "users", {"name": user_id}, {"password_hash": password_hash} ) self._invalidate_cache_and_stream(txn, self.get_user_by_id, (user_id,)) - return self.db.runInteraction( + await self.db_pool.runInteraction( "user_set_password_hash", user_set_password_hash_txn ) - def user_set_consent_version(self, user_id, consent_version): + async def user_set_consent_version( + self, user_id: str, consent_version: str + ) -> None: """Updates the user table to record privacy policy consent Args: - user_id (str): full mxid of the user to update - consent_version (str): version of the policy the user has consented - to + user_id: full mxid of the user to update + consent_version: version of the policy the user has consented to Raises: StoreError(404) if user not found """ def f(txn): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="users", keyvalues={"name": user_id}, @@ -1248,23 +1239,24 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) self._invalidate_cache_and_stream(txn, self.get_user_by_id, (user_id,)) - return self.db.runInteraction("user_set_consent_version", f) + await self.db_pool.runInteraction("user_set_consent_version", f) - def user_set_consent_server_notice_sent(self, user_id, consent_version): + async def user_set_consent_server_notice_sent( + self, user_id: str, consent_version: str + ) -> None: """Updates the user table to record that we have sent the user a server notice about privacy policy consent Args: - user_id (str): full mxid of the user to update - consent_version (str): version of the policy we have notified the - user about + user_id: full mxid of the user to update + consent_version: version of the policy we have notified the user about Raises: StoreError(404) if user not found """ def f(txn): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="users", keyvalues={"name": user_id}, @@ -1272,22 +1264,25 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) self._invalidate_cache_and_stream(txn, self.get_user_by_id, (user_id,)) - return self.db.runInteraction("user_set_consent_server_notice_sent", f) + await self.db_pool.runInteraction("user_set_consent_server_notice_sent", f) - def user_delete_access_tokens(self, user_id, except_token_id=None, device_id=None): + async def user_delete_access_tokens( + self, + user_id: str, + except_token_id: Optional[str] = None, + device_id: Optional[str] = None, + ) -> List[Tuple[str, int, Optional[str]]]: """ Invalidate access tokens belonging to a user Args: - user_id (str): ID of user the tokens belong to - except_token_id (str): list of access_tokens IDs which should - *not* be deleted - device_id (str|None): ID of device the tokens are associated with. + user_id: ID of user the tokens belong to + except_token_id: access_tokens ID which should *not* be deleted + device_id: ID of device the tokens are associated with. If None, tokens associated with any device (or no device) will be deleted Returns: - defer.Deferred[list[str, int, str|None, int]]: a list of - (token, token id, device id) for each of the deleted tokens + A tuple of (token, token id, device id) for each of the deleted tokens """ def f(txn): @@ -1318,11 +1313,11 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): return tokens_and_devices - return self.db.runInteraction("user_delete_access_tokens", f) + return await self.db_pool.runInteraction("user_delete_access_tokens", f) - def delete_access_token(self, access_token): + async def delete_access_token(self, access_token: str) -> None: def f(txn): - self.db.simple_delete_one_txn( + self.db_pool.simple_delete_one_txn( txn, table="access_tokens", keyvalues={"token": access_token} ) @@ -1330,11 +1325,11 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): txn, self.get_user_by_access_token, (access_token,) ) - return self.db.runInteraction("delete_access_token", f) + await self.db_pool.runInteraction("delete_access_token", f) - @cachedInlineCallbacks() - def is_guest(self, user_id): - res = yield self.db.simple_select_one_onecol( + @cached() + async def is_guest(self, user_id: str) -> bool: + res = await self.db_pool.simple_select_one_onecol( table="users", keyvalues={"name": user_id}, retcol="is_guest", @@ -1344,36 +1339,36 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): return res if res else False - def add_user_pending_deactivation(self, user_id): + async def add_user_pending_deactivation(self, user_id: str) -> None: """ Adds a user to the table of users who need to be parted from all the rooms they're in """ - return self.db.simple_insert( + await self.db_pool.simple_insert( "users_pending_deactivation", values={"user_id": user_id}, desc="add_user_pending_deactivation", ) - def del_user_pending_deactivation(self, user_id): + async def del_user_pending_deactivation(self, user_id: str) -> None: """ Removes the given user to the table of users who need to be parted from all the rooms they're in, effectively marking that user as fully deactivated. """ # XXX: This should be simple_delete_one but we failed to put a unique index on # the table, so somehow duplicate entries have ended up in it. - return self.db.simple_delete( + await self.db_pool.simple_delete( "users_pending_deactivation", keyvalues={"user_id": user_id}, desc="del_user_pending_deactivation", ) - def get_user_pending_deactivation(self): + async def get_user_pending_deactivation(self) -> Optional[str]: """ Gets one user from the table of users waiting to be parted from all the rooms they're in. """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( "users_pending_deactivation", keyvalues={}, retcol="user_id", @@ -1381,29 +1376,30 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): desc="get_users_pending_deactivation", ) - def validate_threepid_session(self, session_id, client_secret, token, current_ts): + async def validate_threepid_session( + self, session_id: str, client_secret: str, token: str, current_ts: int + ) -> Optional[str]: """Attempt to validate a threepid session using a token Args: - session_id (str): The id of a validation session - client_secret (str): A unique string provided by the client to - help identify this validation attempt - token (str): A validation token - current_ts (int): The current unix time in milliseconds. Used for - checking token expiry status + session_id: The id of a validation session + client_secret: A unique string provided by the client to help identify + this validation attempt + token: A validation token + current_ts: The current unix time in milliseconds. Used for checking + token expiry status Raises: ThreepidValidationError: if a matching validation token was not found or has expired Returns: - deferred str|None: A str representing a link to redirect the user - to if there is one. + A str representing a link to redirect the user to if there is one. """ # Insert everything into a transaction in order to run atomically def validate_threepid_session_txn(txn): - row = self.db.simple_select_one_txn( + row = self.db_pool.simple_select_one_txn( txn, table="threepid_validation_session", keyvalues={"session_id": session_id}, @@ -1412,16 +1408,23 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) if not row: - raise ThreepidValidationError(400, "Unknown session_id") + if self._ignore_unknown_session_error: + # If we need to inhibit the error caused by an incorrect session ID, + # use None as placeholder values for the client secret and the + # validation timestamp. + # It shouldn't be an issue because they're both only checked after + # the token check, which should fail. And if it doesn't for some + # reason, the next check is on the client secret, which is NOT NULL, + # so we don't have to worry about the client secret matching by + # accident. + row = {"client_secret": None, "validated_at": None} + else: + raise ThreepidValidationError(400, "Unknown session_id") + retrieved_client_secret = row["client_secret"] validated_at = row["validated_at"] - if retrieved_client_secret != client_secret: - raise ThreepidValidationError( - 400, "This client_secret does not match the provided session_id" - ) - - row = self.db.simple_select_one_txn( + row = self.db_pool.simple_select_one_txn( txn, table="threepid_validation_token", keyvalues={"session_id": session_id, "token": token}, @@ -1436,6 +1439,11 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): expires = row["expires"] next_link = row["next_link"] + if retrieved_client_secret != client_secret: + raise ThreepidValidationError( + 400, "This client_secret does not match the provided session_id" + ) + # If the session is already validated, no need to revalidate if validated_at: return next_link @@ -1446,7 +1454,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) # Looks good. Validate the session - self.db.simple_update_txn( + self.db_pool.simple_update_txn( txn, table="threepid_validation_session", keyvalues={"session_id": session_id}, @@ -1456,78 +1464,40 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): return next_link # Return next_link if it exists - return self.db.runInteraction( + return await self.db_pool.runInteraction( "validate_threepid_session_txn", validate_threepid_session_txn ) - def upsert_threepid_validation_session( + async def start_or_continue_validation_session( self, - medium, - address, - client_secret, - send_attempt, - session_id, - validated_at=None, - ): - """Upsert a threepid validation session - Args: - medium (str): The medium of the 3PID - address (str): The address of the 3PID - client_secret (str): A unique string provided by the client to - help identify this validation attempt - send_attempt (int): The latest send_attempt on this session - session_id (str): The id of this validation session - validated_at (int|None): The unix timestamp in milliseconds of - when the session was marked as valid - """ - insertion_values = { - "medium": medium, - "address": address, - "client_secret": client_secret, - } - - if validated_at: - insertion_values["validated_at"] = validated_at - - return self.db.simple_upsert( - table="threepid_validation_session", - keyvalues={"session_id": session_id}, - values={"last_send_attempt": send_attempt}, - insertion_values=insertion_values, - desc="upsert_threepid_validation_session", - ) - - def start_or_continue_validation_session( - self, - medium, - address, - session_id, - client_secret, - send_attempt, - next_link, - token, - token_expires, - ): + medium: str, + address: str, + session_id: str, + client_secret: str, + send_attempt: int, + next_link: Optional[str], + token: str, + token_expires: int, + ) -> None: """Creates a new threepid validation session if it does not already exist and associates a new validation token with it Args: - medium (str): The medium of the 3PID - address (str): The address of the 3PID - session_id (str): The id of this validation session - client_secret (str): A unique string provided by the client to - help identify this validation attempt - send_attempt (int): The latest send_attempt on this session - next_link (str|None): The link to redirect the user to upon - successful validation - token (str): The validation token - token_expires (int): The timestamp for which after the token - will no longer be valid + medium: The medium of the 3PID + address: The address of the 3PID + session_id: The id of this validation session + client_secret: A unique string provided by the client to help + identify this validation attempt + send_attempt: The latest send_attempt on this session + next_link: The link to redirect the user to upon successful validation + token: The validation token + token_expires: The timestamp for which after the token will no + longer be valid """ def start_or_continue_validation_session_txn(txn): # Create or update a validation session - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="threepid_validation_session", keyvalues={"session_id": session_id}, @@ -1540,7 +1510,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) # Create a new validation token with this session ID - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="threepid_validation_token", values={ @@ -1551,12 +1521,12 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): }, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "start_or_continue_validation_session", start_or_continue_validation_session_txn, ) - def cull_expired_threepid_validation_tokens(self): + async def cull_expired_threepid_validation_tokens(self) -> None: """Remove threepid validation tokens with expiry dates that have passed""" def cull_expired_threepid_validation_tokens_txn(txn, ts): @@ -1564,24 +1534,25 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): DELETE FROM threepid_validation_token WHERE expires < ? """ - return txn.execute(sql, (ts,)) + txn.execute(sql, (ts,)) - return self.db.runInteraction( + await self.db_pool.runInteraction( "cull_expired_threepid_validation_tokens", cull_expired_threepid_validation_tokens_txn, self.clock.time_msec(), ) - @defer.inlineCallbacks - def set_user_deactivated_status(self, user_id, deactivated): + async def set_user_deactivated_status( + self, user_id: str, deactivated: bool + ) -> None: """Set the `deactivated` property for the provided user to the provided value. Args: - user_id (str): The ID of the user to set the status for. - deactivated (bool): The value to set for `deactivated`. + user_id: The ID of the user to set the status for. + deactivated: The value to set for `deactivated`. """ - yield self.db.runInteraction( + await self.db_pool.runInteraction( "set_user_deactivated_status", self.set_user_deactivated_status_txn, user_id, @@ -1589,7 +1560,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) def set_user_deactivated_status_txn(self, txn, user_id, deactivated): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn=txn, table="users", keyvalues={"name": user_id}, @@ -1598,9 +1569,9 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): self._invalidate_cache_and_stream( txn, self.get_user_deactivated_status, (user_id,) ) + txn.call_after(self.is_guest.invalidate, (user_id,)) - @defer.inlineCallbacks - def _set_expiration_date_when_missing(self): + async def _set_expiration_date_when_missing(self): """ Retrieves the list of registered users that don't have an expiration date, and adds an expiration date for each of them. @@ -1617,14 +1588,14 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): ) txn.execute(sql, []) - res = self.db.cursor_to_dict(txn) + res = self.db_pool.cursor_to_dict(txn) if res: for user in res: self.set_expiration_date_for_user_txn( txn, user["name"], use_delta=True ) - yield self.db.runInteraction( + await self.db_pool.runInteraction( "get_users_with_no_expiration_date", select_users_with_no_expiration_date_txn, ) @@ -1648,7 +1619,7 @@ class RegistrationStore(RegistrationBackgroundUpdateStore): expiration_ts, ) - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, "account_validity", keyvalues={"user_id": user_id}, diff --git a/synapse/storage/data_stores/main/rejections.py b/synapse/storage/databases/main/rejections.py
index 27e5a2084a..1e361aaa9a 100644 --- a/synapse/storage/data_stores/main/rejections.py +++ b/synapse/storage/databases/main/rejections.py
@@ -14,6 +14,7 @@ # limitations under the License. import logging +from typing import Optional from synapse.storage._base import SQLBaseStore @@ -21,8 +22,8 @@ logger = logging.getLogger(__name__) class RejectionsStore(SQLBaseStore): - def get_rejection_reason(self, event_id): - return self.db.simple_select_one_onecol( + async def get_rejection_reason(self, event_id: str) -> Optional[str]: + return await self.db_pool.simple_select_one_onecol( table="rejections", retcol="reason", keyvalues={"event_id": event_id}, diff --git a/synapse/storage/data_stores/main/relations.py b/synapse/storage/databases/main/relations.py
index 7d477f8d01..5cd61547f7 100644 --- a/synapse/storage/data_stores/main/relations.py +++ b/synapse/storage/databases/main/relations.py
@@ -14,56 +14,53 @@ # limitations under the License. import logging +from typing import Optional import attr from synapse.api.constants import RelationTypes +from synapse.events import EventBase from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.stream import generate_pagination_where_clause +from synapse.storage.databases.main.stream import generate_pagination_where_clause from synapse.storage.relations import ( AggregationPaginationToken, PaginationChunk, RelationPaginationToken, ) -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks +from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) class RelationsWorkerStore(SQLBaseStore): @cached(tree=True) - def get_relations_for_event( + async def get_relations_for_event( self, - event_id, - relation_type=None, - event_type=None, - aggregation_key=None, - limit=5, - direction="b", - from_token=None, - to_token=None, - ): + event_id: str, + relation_type: Optional[str] = None, + event_type: Optional[str] = None, + aggregation_key: Optional[str] = None, + limit: int = 5, + direction: str = "b", + from_token: Optional[RelationPaginationToken] = None, + to_token: Optional[RelationPaginationToken] = None, + ) -> PaginationChunk: """Get a list of relations for an event, ordered by topological ordering. Args: - event_id (str): Fetch events that relate to this event ID. - relation_type (str|None): Only fetch events with this relation - type, if given. - event_type (str|None): Only fetch events with this event type, if - given. - aggregation_key (str|None): Only fetch events with this aggregation - key, if given. - limit (int): Only fetch the most recent `limit` events. - direction (str): Whether to fetch the most recent first (`"b"`) or - the oldest first (`"f"`). - from_token (RelationPaginationToken|None): Fetch rows from the given - token, or from the start if None. - to_token (RelationPaginationToken|None): Fetch rows up to the given - token, or up to the end if None. + event_id: Fetch events that relate to this event ID. + relation_type: Only fetch events with this relation type, if given. + event_type: Only fetch events with this event type, if given. + aggregation_key: Only fetch events with this aggregation key, if given. + limit: Only fetch the most recent `limit` events. + direction: Whether to fetch the most recent first (`"b"`) or the + oldest first (`"f"`). + from_token: Fetch rows from the given token, or from the start if None. + to_token: Fetch rows up to the given token, or up to the end if None. Returns: - Deferred[PaginationChunk]: List of event IDs that match relations - requested. The rows are of the form `{"event_id": "..."}`. + List of event IDs that match relations requested. The rows are of + the form `{"event_id": "..."}`. """ where_clause = ["relates_to_id = ?"] @@ -129,20 +126,20 @@ class RelationsWorkerStore(SQLBaseStore): chunk=list(events[:limit]), next_batch=next_batch, prev_batch=from_token ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_recent_references_for_event", _get_recent_references_for_event_txn ) @cached(tree=True) - def get_aggregation_groups_for_event( + async def get_aggregation_groups_for_event( self, - event_id, - event_type=None, - limit=5, - direction="b", - from_token=None, - to_token=None, - ): + event_id: str, + event_type: Optional[str] = None, + limit: int = 5, + direction: str = "b", + from_token: Optional[AggregationPaginationToken] = None, + to_token: Optional[AggregationPaginationToken] = None, + ) -> PaginationChunk: """Get a list of annotations on the event, grouped by event type and aggregation key, sorted by count. @@ -150,21 +147,17 @@ class RelationsWorkerStore(SQLBaseStore): on an event. Args: - event_id (str): Fetch events that relate to this event ID. - event_type (str|None): Only fetch events with this event type, if - given. - limit (int): Only fetch the `limit` groups. - direction (str): Whether to fetch the highest count first (`"b"`) or + event_id: Fetch events that relate to this event ID. + event_type: Only fetch events with this event type, if given. + limit: Only fetch the `limit` groups. + direction: Whether to fetch the highest count first (`"b"`) or the lowest count first (`"f"`). - from_token (AggregationPaginationToken|None): Fetch rows from the - given token, or from the start if None. - to_token (AggregationPaginationToken|None): Fetch rows up to the - given token, or up to the end if None. - + from_token: Fetch rows from the given token, or from the start if None. + to_token: Fetch rows up to the given token, or up to the end if None. Returns: - Deferred[PaginationChunk]: List of groups of annotations that - match. Each row is a dict with `type`, `key` and `count` fields. + List of groups of annotations that match. Each row is a dict with + `type`, `key` and `count` fields. """ where_clause = ["relates_to_id = ?", "relation_type = ?"] @@ -223,22 +216,22 @@ class RelationsWorkerStore(SQLBaseStore): chunk=list(events[:limit]), next_batch=next_batch, prev_batch=from_token ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_aggregation_groups_for_event", _get_aggregation_groups_for_event_txn ) - @cachedInlineCallbacks() - def get_applicable_edit(self, event_id): + @cached() + async def get_applicable_edit(self, event_id: str) -> Optional[EventBase]: """Get the most recent edit (if any) that has happened for the given event. Correctly handles checking whether edits were allowed to happen. Args: - event_id (str): The original event ID + event_id: The original event ID Returns: - Deferred[EventBase|None]: Returns the most recent edit, if any. + The most recent edit, if any. """ # We only allow edits for `m.room.message` events that have the same sender @@ -268,28 +261,29 @@ class RelationsWorkerStore(SQLBaseStore): if row: return row[0] - edit_id = yield self.db.runInteraction( + edit_id = await self.db_pool.runInteraction( "get_applicable_edit", _get_applicable_edit_txn ) if not edit_id: - return + return None - edit_event = yield self.get_event(edit_id, allow_none=True) - return edit_event + return await self.get_event(edit_id, allow_none=True) - def has_user_annotated_event(self, parent_id, event_type, aggregation_key, sender): + async def has_user_annotated_event( + self, parent_id: str, event_type: str, aggregation_key: str, sender: str + ) -> bool: """Check if a user has already annotated an event with the same key (e.g. already liked an event). Args: - parent_id (str): The event being annotated - event_type (str): The event type of the annotation - aggregation_key (str): The aggregation key of the annotation - sender (str): The sender of the annotation + parent_id: The event being annotated + event_type: The event type of the annotation + aggregation_key: The aggregation key of the annotation + sender: The sender of the annotation Returns: - Deferred[bool] + True if the event is already annotated. """ sql = """ @@ -318,7 +312,7 @@ class RelationsWorkerStore(SQLBaseStore): return bool(txn.fetchone()) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_if_user_has_annotated_event", _get_if_user_has_annotated_event ) diff --git a/synapse/storage/data_stores/main/room.py b/synapse/storage/databases/main/room.py
index 93b6380f13..8fab8de973 100644 --- a/synapse/storage/data_stores/main/room.py +++ b/synapse/storage/databases/main/room.py
@@ -21,26 +21,19 @@ from abc import abstractmethod from enum import Enum from typing import Any, Dict, List, Optional, Tuple -from canonicaljson import json - -from twisted.internet import defer - from synapse.api.constants import EventTypes from synapse.api.errors import StoreError from synapse.api.room_versions import RoomVersion, RoomVersions from synapse.storage._base import SQLBaseStore, db_to_json -from synapse.storage.data_stores.main.search import SearchStore -from synapse.storage.database import Database, LoggingTransaction -from synapse.types import ThirdPartyInstanceID -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks +from synapse.storage.database import DatabasePool, LoggingTransaction +from synapse.storage.databases.main.search import SearchStore +from synapse.types import JsonDict, ThirdPartyInstanceID +from synapse.util import json_encoder +from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) -OpsLevel = collections.namedtuple( - "OpsLevel", ("ban_level", "kick_level", "redact_level") -) - RatelimitOverride = collections.namedtuple( "RatelimitOverride", ("messages_per_second", "burst_count") ) @@ -75,20 +68,20 @@ class RoomSortOrder(Enum): class RoomWorkerStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.config = hs.config - def get_room(self, room_id): + async def get_room(self, room_id: str) -> dict: """Retrieve a room. Args: - room_id (str): The ID of the room to retrieve. + room_id: The ID of the room to retrieve. Returns: A dict containing the room information, or None if the room is unknown. """ - return self.db.simple_select_one( + return await self.db_pool.simple_select_one( table="rooms", keyvalues={"room_id": room_id}, retcols=("room_id", "is_public", "creator"), @@ -96,7 +89,7 @@ class RoomWorkerStore(SQLBaseStore): allow_none=True, ) - def get_room_with_stats(self, room_id: str): + async def get_room_with_stats(self, room_id: str) -> Optional[Dict[str, Any]]: """Retrieve room with statistics. Args: @@ -111,7 +104,8 @@ class RoomWorkerStore(SQLBaseStore): curr.local_users_in_room AS joined_local_members, rooms.room_version AS version, rooms.creator, state.encryption, state.is_federatable AS federatable, rooms.is_public AS public, state.join_rules, state.guest_access, - state.history_visibility, curr.current_state_events AS state_events + state.history_visibility, curr.current_state_events AS state_events, + state.avatar, state.topic FROM rooms LEFT JOIN room_stats_state state USING (room_id) LEFT JOIN room_stats_current curr USING (room_id) @@ -120,7 +114,7 @@ class RoomWorkerStore(SQLBaseStore): txn.execute(sql, [room_id]) # Catch error if sql returns empty result to return "None" instead of an error try: - res = self.db.cursor_to_dict(txn)[0] + res = self.db_pool.cursor_to_dict(txn)[0] except IndexError: return None @@ -128,25 +122,29 @@ class RoomWorkerStore(SQLBaseStore): res["public"] = bool(res["public"]) return res - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_room_with_stats", get_room_with_stats_txn, room_id ) - def get_public_room_ids(self): - return self.db.simple_select_onecol( + async def get_public_room_ids(self) -> List[str]: + return await self.db_pool.simple_select_onecol( table="rooms", keyvalues={"is_public": True}, retcol="room_id", desc="get_public_room_ids", ) - def count_public_rooms(self, network_tuple, ignore_non_federatable): + async def count_public_rooms( + self, + network_tuple: Optional[ThirdPartyInstanceID], + ignore_non_federatable: bool, + ) -> int: """Counts the number of public rooms as tracked in the room_stats_current and room_stats_state table. Args: - network_tuple (ThirdPartyInstanceID|None) - ignore_non_federatable (bool): If true filters out non-federatable rooms + network_tuple + ignore_non_federatable: If true filters out non-federatable rooms """ def _count_public_rooms_txn(txn): @@ -190,10 +188,11 @@ class RoomWorkerStore(SQLBaseStore): txn.execute(sql, query_args) return txn.fetchone()[0] - return self.db.runInteraction("count_public_rooms", _count_public_rooms_txn) + return await self.db_pool.runInteraction( + "count_public_rooms", _count_public_rooms_txn + ) - @defer.inlineCallbacks - def get_largest_public_rooms( + async def get_largest_public_rooms( self, network_tuple: Optional[ThirdPartyInstanceID], search_filter: Optional[dict], @@ -323,21 +322,21 @@ class RoomWorkerStore(SQLBaseStore): def _get_largest_public_rooms_txn(txn): txn.execute(sql, query_args) - results = self.db.cursor_to_dict(txn) + results = self.db_pool.cursor_to_dict(txn) if not forwards: results.reverse() return results - ret_val = yield self.db.runInteraction( + ret_val = await self.db_pool.runInteraction( "get_largest_public_rooms", _get_largest_public_rooms_txn ) - defer.returnValue(ret_val) + return ret_val @cached(max_entries=10000) - def is_room_blocked(self, room_id): - return self.db.simple_select_one_onecol( + async def is_room_blocked(self, room_id: str) -> Optional[bool]: + return await self.db_pool.simple_select_one_onecol( table="blocked_rooms", keyvalues={"room_id": room_id}, retcol="1", @@ -345,23 +344,22 @@ class RoomWorkerStore(SQLBaseStore): desc="is_room_blocked", ) - @defer.inlineCallbacks - def is_room_published(self, room_id): + async def is_room_published(self, room_id: str) -> bool: """Check whether a room has been published in the local public room directory. Args: - room_id (str) + room_id Returns: - bool: Whether the room is currently published in the room directory + Whether the room is currently published in the room directory """ # Get room information - room_info = yield self.get_room(room_id) + room_info = await self.get_room(room_id) if not room_info: - defer.returnValue(False) + return False # Check the is_public value - defer.returnValue(room_info.get("is_public", False)) + return room_info.get("is_public", False) async def get_rooms_paginate( self, @@ -523,12 +521,12 @@ class RoomWorkerStore(SQLBaseStore): room_count = txn.fetchone() return rooms, room_count[0] - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_rooms_paginate", _get_rooms_paginate_txn, ) - @cachedInlineCallbacks(max_entries=10000) - def get_ratelimit_for_user(self, user_id): + @cached(max_entries=10000) + async def get_ratelimit_for_user(self, user_id): """Check if there are any overrides for ratelimiting for the given user @@ -540,7 +538,7 @@ class RoomWorkerStore(SQLBaseStore): of RatelimitOverride are None or 0 then ratelimitng has been disabled for that user entirely. """ - row = yield self.db.simple_select_one( + row = await self.db_pool.simple_select_one( table="ratelimit_override", keyvalues={"user_id": user_id}, retcols=("messages_per_second", "burst_count"), @@ -556,8 +554,8 @@ class RoomWorkerStore(SQLBaseStore): else: return None - @cachedInlineCallbacks() - def get_retention_policy_for_room(self, room_id): + @cached() + async def get_retention_policy_for_room(self, room_id): """Get the retention policy for a given room. If no retention policy has been found for this room, returns a policy defined @@ -575,7 +573,7 @@ class RoomWorkerStore(SQLBaseStore): # maximum, in order not to filter out events we should filter out when sending to # the client. if not self.config.retention_enabled: - defer.returnValue({"min_lifetime": None, "max_lifetime": None}) + return {"min_lifetime": None, "max_lifetime": None} def get_retention_policy_for_room_txn(txn): txn.execute( @@ -587,21 +585,19 @@ class RoomWorkerStore(SQLBaseStore): (room_id,), ) - return self.db.cursor_to_dict(txn) + return self.db_pool.cursor_to_dict(txn) - ret = yield self.db.runInteraction( + ret = await self.db_pool.runInteraction( "get_retention_policy_for_room", get_retention_policy_for_room_txn, ) # If we don't know this room ID, ret will be None, in this case return the default # policy. if not ret: - defer.returnValue( - { - "min_lifetime": self.config.retention_default_min_lifetime, - "max_lifetime": self.config.retention_default_max_lifetime, - } - ) + return { + "min_lifetime": self.config.retention_default_min_lifetime, + "max_lifetime": self.config.retention_default_max_lifetime, + } row = ret[0] @@ -615,17 +611,16 @@ class RoomWorkerStore(SQLBaseStore): if row["max_lifetime"] is None: row["max_lifetime"] = self.config.retention_default_max_lifetime - defer.returnValue(row) + return row - def get_media_mxcs_in_room(self, room_id): + async def get_media_mxcs_in_room(self, room_id: str) -> Tuple[List[str], List[str]]: """Retrieves all the local and remote media MXC URIs in a given room Args: - room_id (str) + room_id Returns: - The local and remote media as a lists of tuples where the key is - the hostname and the value is the media ID. + The local and remote media as a lists of the media IDs. """ def _get_media_mxcs_in_room_txn(txn): @@ -641,11 +636,13 @@ class RoomWorkerStore(SQLBaseStore): return local_media_mxcs, remote_media_mxcs - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_media_ids_in_room", _get_media_mxcs_in_room_txn ) - def quarantine_media_ids_in_room(self, room_id, quarantined_by): + async def quarantine_media_ids_in_room( + self, room_id: str, quarantined_by: str + ) -> int: """For a room loops through all events with media and quarantines the associated media """ @@ -658,7 +655,7 @@ class RoomWorkerStore(SQLBaseStore): txn, local_mxcs, remote_mxcs, quarantined_by ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "quarantine_media_in_room", _quarantine_media_in_room_txn ) @@ -721,9 +718,9 @@ class RoomWorkerStore(SQLBaseStore): return local_media_mxcs, remote_media_mxcs - def quarantine_media_by_id( + async def quarantine_media_by_id( self, server_name: str, media_id: str, quarantined_by: str, - ): + ) -> int: """quarantines a single local or remote media id Args: @@ -742,11 +739,13 @@ class RoomWorkerStore(SQLBaseStore): txn, local_mxcs, remote_mxcs, quarantined_by ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "quarantine_media_by_user", _quarantine_media_by_id_txn ) - def quarantine_media_ids_by_user(self, user_id: str, quarantined_by: str): + async def quarantine_media_ids_by_user( + self, user_id: str, quarantined_by: str + ) -> int: """quarantines all local media associated with a single user Args: @@ -758,7 +757,7 @@ class RoomWorkerStore(SQLBaseStore): local_media_ids = self._get_media_ids_by_user_txn(txn, user_id) return self._quarantine_media_txn(txn, local_media_ids, [], quarantined_by) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "quarantine_media_by_user", _quarantine_media_by_user_txn ) @@ -876,7 +875,7 @@ class RoomWorkerStore(SQLBaseStore): return updates, upto_token, limited - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_all_new_public_rooms", get_all_new_public_rooms ) @@ -885,27 +884,26 @@ class RoomBackgroundUpdateStore(SQLBaseStore): REMOVE_TOMESTONED_ROOMS_BG_UPDATE = "remove_tombstoned_rooms_from_directory" ADD_ROOMS_ROOM_VERSION_COLUMN = "add_rooms_room_version_column" - def __init__(self, database: Database, db_conn, hs): - super(RoomBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.config = hs.config - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "insert_room_retention", self._background_insert_retention, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.REMOVE_TOMESTONED_ROOMS_BG_UPDATE, self._remove_tombstoned_rooms_from_directory, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.ADD_ROOMS_ROOM_VERSION_COLUMN, self._background_add_rooms_room_version_column, ) - @defer.inlineCallbacks - def _background_insert_retention(self, progress, batch_size): + async def _background_insert_retention(self, progress, batch_size): """Retrieves a list of all rooms within a range and inserts an entry for each of them into the room_retention table. NULLs the property's columns if missing from the retention event in the room's @@ -929,7 +927,7 @@ class RoomBackgroundUpdateStore(SQLBaseStore): (last_room, batch_size), ) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return True @@ -941,7 +939,7 @@ class RoomBackgroundUpdateStore(SQLBaseStore): ev = db_to_json(row["json"]) retention_policy = ev["content"] - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn=txn, table="room_retention", values={ @@ -954,7 +952,7 @@ class RoomBackgroundUpdateStore(SQLBaseStore): logger.info("Inserted %d rows into room_retention", len(rows)) - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, "insert_room_retention", {"room_id": rows[-1]["room_id"]} ) @@ -963,14 +961,14 @@ class RoomBackgroundUpdateStore(SQLBaseStore): else: return False - end = yield self.db.runInteraction( + end = await self.db_pool.runInteraction( "insert_room_retention", _background_insert_retention_txn, ) if end: - yield self.db.updates._end_background_update("insert_room_retention") + await self.db_pool.updates._end_background_update("insert_room_retention") - defer.returnValue(batch_size) + return batch_size async def _background_add_rooms_room_version_column( self, progress: dict, batch_size: int @@ -1012,7 +1010,7 @@ class RoomBackgroundUpdateStore(SQLBaseStore): # mainly for paranoia as much badness would happen if we don't # insert the row and then try and get the room version for the # room. - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="rooms", keyvalues={"room_id": room_id}, @@ -1021,19 +1019,19 @@ class RoomBackgroundUpdateStore(SQLBaseStore): ) new_last_room_id = room_id - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.ADD_ROOMS_ROOM_VERSION_COLUMN, {"room_id": new_last_room_id} ) return False - end = await self.db.runInteraction( + end = await self.db_pool.runInteraction( "_background_add_rooms_room_version_column", _background_add_rooms_room_version_column_txn, ) if end: - await self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.ADD_ROOMS_ROOM_VERSION_COLUMN ) @@ -1067,12 +1065,12 @@ class RoomBackgroundUpdateStore(SQLBaseStore): return [row[0] for row in txn] - rooms = await self.db.runInteraction( + rooms = await self.db_pool.runInteraction( "get_tombstoned_directory_rooms", _get_rooms ) if not rooms: - await self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.REMOVE_TOMESTONED_ROOMS_BG_UPDATE ) return 0 @@ -1081,7 +1079,7 @@ class RoomBackgroundUpdateStore(SQLBaseStore): logger.info("Removing tombstoned room %s from the directory", room_id) await self.set_room_is_public(room_id, False) - await self.db.updates._background_update_progress( + await self.db_pool.updates._background_update_progress( self.REMOVE_TOMESTONED_ROOMS_BG_UPDATE, {"room_id": rooms[-1]} ) @@ -1097,8 +1095,8 @@ class RoomBackgroundUpdateStore(SQLBaseStore): class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.config = hs.config @@ -1108,7 +1106,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): Called when we join a room over federation, and overwrites any room version currently in the table. """ - await self.db.simple_upsert( + await self.db_pool.simple_upsert( desc="upsert_room_on_join", table="rooms", keyvalues={"room_id": room_id}, @@ -1119,8 +1117,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): lock=False, ) - @defer.inlineCallbacks - def store_room( + async def store_room( self, room_id: str, room_creator_user_id: str, @@ -1141,7 +1138,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): try: def store_room_txn(txn, next_id): - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, "rooms", { @@ -1152,7 +1149,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): }, ) if is_public: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="public_room_list_stream", values={ @@ -1162,8 +1159,10 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): }, ) - with self._public_room_id_gen.get_next() as next_id: - yield self.db.runInteraction("store_room_txn", store_room_txn, next_id) + async with self._public_room_id_gen.get_next() as next_id: + await self.db_pool.runInteraction( + "store_room_txn", store_room_txn, next_id + ) except Exception as e: logger.error("store_room with room_id=%s failed: %s", room_id, e) raise StoreError(500, "Problem creating room.") @@ -1173,7 +1172,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): When we receive an invite over federation, store the version of the room if we don't already know the room version. """ - await self.db.simple_upsert( + await self.db_pool.simple_upsert( desc="maybe_store_room_on_invite", table="rooms", keyvalues={"room_id": room_id}, @@ -1188,17 +1187,16 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): lock=False, ) - @defer.inlineCallbacks - def set_room_is_public(self, room_id, is_public): + async def set_room_is_public(self, room_id, is_public): def set_room_is_public_txn(txn, next_id): - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="rooms", keyvalues={"room_id": room_id}, updatevalues={"is_public": is_public}, ) - entries = self.db.simple_select_list_txn( + entries = self.db_pool.simple_select_list_txn( txn, table="public_room_list_stream", keyvalues={ @@ -1216,7 +1214,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): add_to_stream = bool(entries[-1]["visibility"]) != is_public if add_to_stream: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="public_room_list_stream", values={ @@ -1228,14 +1226,13 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): }, ) - with self._public_room_id_gen.get_next() as next_id: - yield self.db.runInteraction( + async with self._public_room_id_gen.get_next() as next_id: + await self.db_pool.runInteraction( "set_room_is_public", set_room_is_public_txn, next_id ) self.hs.get_notifier().on_new_replication_data() - @defer.inlineCallbacks - def set_room_is_public_appservice( + async def set_room_is_public_appservice( self, room_id, appservice_id, network_id, is_public ): """Edit the appservice/network specific public room list. @@ -1256,7 +1253,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): def set_room_is_public_appservice_txn(txn, next_id): if is_public: try: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="appservice_room_list", values={ @@ -1269,7 +1266,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): # We've already inserted, nothing to do. return else: - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="appservice_room_list", keyvalues={ @@ -1279,7 +1276,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): }, ) - entries = self.db.simple_select_list_txn( + entries = self.db_pool.simple_select_list_txn( txn, table="public_room_list_stream", keyvalues={ @@ -1297,7 +1294,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): add_to_stream = bool(entries[-1]["visibility"]) != is_public if add_to_stream: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="public_room_list_stream", values={ @@ -1309,16 +1306,16 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): }, ) - with self._public_room_id_gen.get_next() as next_id: - yield self.db.runInteraction( + async with self._public_room_id_gen.get_next() as next_id: + await self.db_pool.runInteraction( "set_room_is_public_appservice", set_room_is_public_appservice_txn, next_id, ) self.hs.get_notifier().on_new_replication_data() - def get_room_count(self): - """Retrieve a list of all rooms + async def get_room_count(self) -> int: + """Retrieve the total number of rooms. """ def f(txn): @@ -1327,13 +1324,19 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): row = txn.fetchone() return row[0] or 0 - return self.db.runInteraction("get_rooms", f) + return await self.db_pool.runInteraction("get_rooms", f) - def add_event_report( - self, room_id, event_id, user_id, reason, content, received_ts - ): + async def add_event_report( + self, + room_id: str, + event_id: str, + user_id: str, + reason: str, + content: JsonDict, + received_ts: int, + ) -> None: next_id = self._event_reports_id_gen.get_next() - return self.db.simple_insert( + await self.db_pool.simple_insert( table="event_reports", values={ "id": next_id, @@ -1342,60 +1345,150 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): "event_id": event_id, "user_id": user_id, "reason": reason, - "content": json.dumps(content), + "content": json_encoder.encode(content), }, desc="add_event_report", ) + async def get_event_reports_paginate( + self, + start: int, + limit: int, + direction: str = "b", + user_id: Optional[str] = None, + room_id: Optional[str] = None, + ) -> Tuple[List[Dict[str, Any]], int]: + """Retrieve a paginated list of event reports + + Args: + start: event offset to begin the query from + limit: number of rows to retrieve + direction: Whether to fetch the most recent first (`"b"`) or the + oldest first (`"f"`) + user_id: search for user_id. Ignored if user_id is None + room_id: search for room_id. Ignored if room_id is None + Returns: + event_reports: json list of event reports + count: total number of event reports matching the filter criteria + """ + + def _get_event_reports_paginate_txn(txn): + filters = [] + args = [] + + if user_id: + filters.append("er.user_id LIKE ?") + args.extend(["%" + user_id + "%"]) + if room_id: + filters.append("er.room_id LIKE ?") + args.extend(["%" + room_id + "%"]) + + if direction == "b": + order = "DESC" + else: + order = "ASC" + + where_clause = "WHERE " + " AND ".join(filters) if len(filters) > 0 else "" + + sql = """ + SELECT COUNT(*) as total_event_reports + FROM event_reports AS er + {} + """.format( + where_clause + ) + txn.execute(sql, args) + count = txn.fetchone()[0] + + sql = """ + SELECT + er.id, + er.received_ts, + er.room_id, + er.event_id, + er.user_id, + er.reason, + er.content, + events.sender, + room_aliases.room_alias, + event_json.json AS event_json + FROM event_reports AS er + LEFT JOIN room_aliases + ON room_aliases.room_id = er.room_id + JOIN events + ON events.event_id = er.event_id + JOIN event_json + ON event_json.event_id = er.event_id + {where_clause} + ORDER BY er.received_ts {order} + LIMIT ? + OFFSET ? + """.format( + where_clause=where_clause, order=order, + ) + + args += [limit, start] + txn.execute(sql, args) + event_reports = self.db_pool.cursor_to_dict(txn) + + if count > 0: + for row in event_reports: + try: + row["content"] = db_to_json(row["content"]) + row["event_json"] = db_to_json(row["event_json"]) + except Exception: + continue + + return event_reports, count + + return await self.db_pool.runInteraction( + "get_event_reports_paginate", _get_event_reports_paginate_txn + ) + def get_current_public_room_stream_id(self): return self._public_room_id_gen.get_current_token() - @defer.inlineCallbacks - def block_room(self, room_id, user_id): + async def block_room(self, room_id: str, user_id: str) -> None: """Marks the room as blocked. Can be called multiple times. Args: - room_id (str): Room to block - user_id (str): Who blocked it - - Returns: - Deferred + room_id: Room to block + user_id: Who blocked it """ - yield self.db.simple_upsert( + await self.db_pool.simple_upsert( table="blocked_rooms", keyvalues={"room_id": room_id}, values={}, insertion_values={"user_id": user_id}, desc="block_room", ) - yield self.db.runInteraction( + await self.db_pool.runInteraction( "block_room_invalidation", self._invalidate_cache_and_stream, self.is_room_blocked, (room_id,), ) - @defer.inlineCallbacks - def get_rooms_for_retention_period_in_range( - self, min_ms, max_ms, include_null=False - ): + async def get_rooms_for_retention_period_in_range( + self, min_ms: Optional[int], max_ms: Optional[int], include_null: bool = False + ) -> Dict[str, dict]: """Retrieves all of the rooms within the given retention range. Optionally includes the rooms which don't have a retention policy. Args: - min_ms (int|None): Duration in milliseconds that define the lower limit of + min_ms: Duration in milliseconds that define the lower limit of the range to handle (exclusive). If None, doesn't set a lower limit. - max_ms (int|None): Duration in milliseconds that define the upper limit of + max_ms: Duration in milliseconds that define the upper limit of the range to handle (inclusive). If None, doesn't set an upper limit. - include_null (bool): Whether to include rooms which retention policy is NULL + include_null: Whether to include rooms which retention policy is NULL in the returned set. Returns: - dict[str, dict]: The rooms within this range, along with their retention - policy. The key is "room_id", and maps to a dict describing the retention - policy associated with this room ID. The keys for this nested dict are - "min_lifetime" (int|None), and "max_lifetime" (int|None). + The rooms within this range, along with their retention + policy. The key is "room_id", and maps to a dict describing the retention + policy associated with this room ID. The keys for this nested dict are + "min_lifetime" (int|None), and "max_lifetime" (int|None). """ def get_rooms_for_retention_period_in_range_txn(txn): @@ -1425,7 +1518,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): txn.execute(sql, args) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) rooms_dict = {} for row in rows: @@ -1441,7 +1534,7 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): txn.execute(sql) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) # If a room isn't already in the dict (i.e. it doesn't have a retention # policy in its state), add it with a null policy. @@ -1454,9 +1547,9 @@ class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore): return rooms_dict - rooms = yield self.db.runInteraction( + rooms = await self.db_pool.runInteraction( "get_rooms_for_retention_period_in_range", get_rooms_for_retention_period_in_range_txn, ) - defer.returnValue(rooms) + return rooms diff --git a/synapse/storage/data_stores/main/roommember.py b/synapse/storage/databases/main/roommember.py
index a92e401e88..86ffe2479e 100644 --- a/synapse/storage/data_stores/main/roommember.py +++ b/synapse/storage/databases/main/roommember.py
@@ -13,13 +13,12 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import logging -from typing import Iterable, List, Set - -from twisted.internet import defer +from typing import TYPE_CHECKING, Dict, FrozenSet, Iterable, List, Optional, Set from synapse.api.constants import EventTypes, Membership +from synapse.events import EventBase +from synapse.events.snapshot import EventContext from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage._base import ( @@ -28,8 +27,8 @@ from synapse.storage._base import ( db_to_json, make_in_list_sql_clause, ) -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.events_worker import EventsWorkerStore from synapse.storage.engines import Sqlite3Engine from synapse.storage.roommember import ( GetRoomsForUserWithStreamOrdering, @@ -37,12 +36,15 @@ from synapse.storage.roommember import ( ProfileInfo, RoomsForUser, ) -from synapse.types import Collection, get_domain_from_id +from synapse.types import Collection, PersistedEventPosition, get_domain_from_id from synapse.util.async_helpers import Linearizer from synapse.util.caches import intern_string -from synapse.util.caches.descriptors import cached, cachedInlineCallbacks, cachedList +from synapse.util.caches.descriptors import _CacheContext, cached, cachedList from synapse.util.metrics import Measure +if TYPE_CHECKING: + from synapse.state import _StateCacheEntry + logger = logging.getLogger(__name__) @@ -51,8 +53,8 @@ _CURRENT_STATE_MEMBERSHIP_UPDATE_NAME = "current_state_events_membership" class RoomMemberWorkerStore(EventsWorkerStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomMemberWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # Is the current_state_events.membership up to date? Or is the # background update still running? @@ -87,8 +89,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): lambda: self._known_servers_count, ) - @defer.inlineCallbacks - def _count_known_servers(self): + async def _count_known_servers(self): """ Count the servers that this server knows about. @@ -116,7 +117,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): txn.execute(query) return list(txn)[0][0] - count = yield self.db.runInteraction("get_known_servers", _transact) + count = await self.db_pool.runInteraction("get_known_servers", _transact) # We always know about ourselves, even if we have nothing in # room_memberships (for example, the server is new). @@ -128,7 +129,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): membership column is up to date """ - pending_update = self.db.simple_select_one_txn( + pending_update = self.db_pool.simple_select_one_txn( txn, table="background_updates", keyvalues={"update_name": _CURRENT_STATE_MEMBERSHIP_UPDATE_NAME}, @@ -144,18 +145,18 @@ class RoomMemberWorkerStore(EventsWorkerStore): 15.0, run_as_background_process, "_check_safe_current_state_events_membership_updated", - self.db.runInteraction, + self.db_pool.runInteraction, "_check_safe_current_state_events_membership_updated", self._check_safe_current_state_events_membership_updated_txn, ) @cached(max_entries=100000, iterable=True) - def get_users_in_room(self, room_id): - return self.db.runInteraction( + async def get_users_in_room(self, room_id: str) -> List[str]: + return await self.db_pool.runInteraction( "get_users_in_room", self.get_users_in_room_txn, room_id ) - def get_users_in_room_txn(self, txn, room_id): + def get_users_in_room_txn(self, txn, room_id: str) -> List[str]: # If we can assume current_state_events.membership is up to date # then we can avoid a join, which is a Very Good Thing given how # frequently this function gets called. @@ -178,14 +179,13 @@ class RoomMemberWorkerStore(EventsWorkerStore): return [r[0] for r in txn] @cached(max_entries=100000) - def get_room_summary(self, room_id): + async def get_room_summary(self, room_id: str) -> Dict[str, MemberSummary]: """ Get the details of a room roughly suitable for use by the room summary extension to /sync. Useful when lazy loading room members. Args: - room_id (str): The room ID to query + room_id: The room ID to query Returns: - Deferred[dict[str, MemberSummary]: - dict of membership states, pointing to a MemberSummary named tuple. + dict of membership states, pointing to a MemberSummary named tuple. """ def _get_room_summary_txn(txn): @@ -259,80 +259,63 @@ class RoomMemberWorkerStore(EventsWorkerStore): return res - return self.db.runInteraction("get_room_summary", _get_room_summary_txn) - - def _get_user_counts_in_room_txn(self, txn, room_id): - """ - Get the user count in a room by membership. - - Args: - room_id (str) - membership (Membership) - - Returns: - Deferred[int] - """ - sql = """ - SELECT m.membership, count(*) FROM room_memberships as m - INNER JOIN current_state_events as c USING(event_id) - WHERE c.type = 'm.room.member' AND c.room_id = ? - GROUP BY m.membership - """ - - txn.execute(sql, (room_id,)) - return {row[0]: row[1] for row in txn} + return await self.db_pool.runInteraction( + "get_room_summary", _get_room_summary_txn + ) @cached() - def get_invited_rooms_for_local_user(self, user_id): - """ Get all the rooms the *local* user is invited to + async def get_invited_rooms_for_local_user(self, user_id: str) -> RoomsForUser: + """Get all the rooms the *local* user is invited to. Args: - user_id (str): The user ID. + user_id: The user ID. + Returns: - A deferred list of RoomsForUser. + A list of RoomsForUser. """ - return self.get_rooms_for_local_user_where_membership_is( + return await self.get_rooms_for_local_user_where_membership_is( user_id, [Membership.INVITE] ) - @defer.inlineCallbacks - def get_invite_for_local_user_in_room(self, user_id, room_id): - """Gets the invite for the given *local* user and room + async def get_invite_for_local_user_in_room( + self, user_id: str, room_id: str + ) -> Optional[RoomsForUser]: + """Gets the invite for the given *local* user and room. Args: - user_id (str) - room_id (str) + user_id: The user ID to find the invite of. + room_id: The room to user was invited to. Returns: - Deferred: Resolves to either a RoomsForUser or None if no invite was - found. + Either a RoomsForUser or None if no invite was found. """ - invites = yield self.get_invited_rooms_for_local_user(user_id) + invites = await self.get_invited_rooms_for_local_user(user_id) for invite in invites: if invite.room_id == room_id: return invite return None - @defer.inlineCallbacks - def get_rooms_for_local_user_where_membership_is(self, user_id, membership_list): - """ Get all the rooms for this *local* user where the membership for this user + async def get_rooms_for_local_user_where_membership_is( + self, user_id: str, membership_list: Collection[str] + ) -> List[RoomsForUser]: + """Get all the rooms for this *local* user where the membership for this user matches one in the membership list. Filters out forgotten rooms. Args: - user_id (str): The user ID. - membership_list (list): A list of synapse.api.constants.Membership - values which the user must be in. + user_id: The user ID. + membership_list: A list of synapse.api.constants.Membership + values which the user must be in. Returns: - Deferred[list[RoomsForUser]] + The RoomsForUser that the user matches the membership types. """ if not membership_list: - return defer.succeed(None) + return [] - rooms = yield self.db.runInteraction( + rooms = await self.db_pool.runInteraction( "get_rooms_for_local_user_where_membership_is", self._get_rooms_for_local_user_where_membership_is_txn, user_id, @@ -340,12 +323,12 @@ class RoomMemberWorkerStore(EventsWorkerStore): ) # Now we filter out forgotten rooms - forgotten_rooms = yield self.get_forgotten_rooms_for_user(user_id) + forgotten_rooms = await self.get_forgotten_rooms_for_user(user_id) return [room for room in rooms if room.room_id not in forgotten_rooms] def _get_rooms_for_local_user_where_membership_is_txn( - self, txn, user_id, membership_list - ): + self, txn, user_id: str, membership_list: List[str] + ) -> List[RoomsForUser]: # Paranoia check. if not self.hs.is_mine_id(user_id): raise Exception( @@ -369,38 +352,41 @@ class RoomMemberWorkerStore(EventsWorkerStore): ) txn.execute(sql, (user_id, *args)) - results = [RoomsForUser(**r) for r in self.db.cursor_to_dict(txn)] + results = [RoomsForUser(**r) for r in self.db_pool.cursor_to_dict(txn)] return results @cached(max_entries=500000, iterable=True) - def get_rooms_for_user_with_stream_ordering(self, user_id): + async def get_rooms_for_user_with_stream_ordering( + self, user_id: str + ) -> FrozenSet[GetRoomsForUserWithStreamOrdering]: """Returns a set of room_ids the user is currently joined to. If a remote user only returns rooms this server is currently participating in. Args: - user_id (str) + user_id Returns: - Deferred[frozenset[GetRoomsForUserWithStreamOrdering]]: Returns - the rooms the user is in currently, along with the stream ordering - of the most recent join for that user and room. + Returns the rooms the user is in currently, along with the stream + ordering of the most recent join for that user and room. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_rooms_for_user_with_stream_ordering", self._get_rooms_for_user_with_stream_ordering_txn, user_id, ) - def _get_rooms_for_user_with_stream_ordering_txn(self, txn, user_id): + def _get_rooms_for_user_with_stream_ordering_txn( + self, txn, user_id: str + ) -> FrozenSet[GetRoomsForUserWithStreamOrdering]: # We use `current_state_events` here and not `local_current_membership` # as a) this gets called with remote users and b) this only gets called # for rooms the server is participating in. if self._current_state_events_membership_up_to_date: sql = """ - SELECT room_id, e.stream_ordering + SELECT room_id, e.instance_name, e.stream_ordering FROM current_state_events AS c INNER JOIN events AS e USING (room_id, event_id) WHERE @@ -410,7 +396,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): """ else: sql = """ - SELECT room_id, e.stream_ordering + SELECT room_id, e.instance_name, e.stream_ordering FROM current_state_events AS c INNER JOIN room_memberships AS m USING (room_id, event_id) INNER JOIN events AS e USING (room_id, event_id) @@ -421,9 +407,12 @@ class RoomMemberWorkerStore(EventsWorkerStore): """ txn.execute(sql, (user_id, Membership.JOIN)) - results = frozenset(GetRoomsForUserWithStreamOrdering(*row) for row in txn) - - return results + return frozenset( + GetRoomsForUserWithStreamOrdering( + room_id, PersistedEventPosition(instance, stream_id) + ) + for room_id, instance, stream_id in txn + ) async def get_users_server_still_shares_room_with( self, user_ids: Collection[str] @@ -453,42 +442,44 @@ class RoomMemberWorkerStore(EventsWorkerStore): return {row[0] for row in txn} - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "get_users_server_still_shares_room_with", _get_users_server_still_shares_room_with_txn, ) - @defer.inlineCallbacks - def get_rooms_for_user(self, user_id, on_invalidate=None): + async def get_rooms_for_user(self, user_id: str, on_invalidate=None): """Returns a set of room_ids the user is currently joined to. If a remote user only returns rooms this server is currently participating in. """ - rooms = yield self.get_rooms_for_user_with_stream_ordering( + rooms = await self.get_rooms_for_user_with_stream_ordering( user_id, on_invalidate=on_invalidate ) return frozenset(r.room_id for r in rooms) - @cachedInlineCallbacks(max_entries=500000, cache_context=True, iterable=True) - def get_users_who_share_room_with_user(self, user_id, cache_context): + @cached(max_entries=500000, cache_context=True, iterable=True) + async def get_users_who_share_room_with_user( + self, user_id: str, cache_context: _CacheContext + ) -> Set[str]: """Returns the set of users who share a room with `user_id` """ - room_ids = yield self.get_rooms_for_user( + room_ids = await self.get_rooms_for_user( user_id, on_invalidate=cache_context.invalidate ) user_who_share_room = set() for room_id in room_ids: - user_ids = yield self.get_users_in_room( + user_ids = await self.get_users_in_room( room_id, on_invalidate=cache_context.invalidate ) user_who_share_room.update(user_ids) return user_who_share_room - @defer.inlineCallbacks - def get_joined_users_from_context(self, event, context): + async def get_joined_users_from_context( + self, event: EventBase, context: EventContext + ): state_group = context.state_group if not state_group: # If state_group is None it means it has yet to be assigned a @@ -497,14 +488,12 @@ class RoomMemberWorkerStore(EventsWorkerStore): # To do this we set the state_group to a new object as object() != object() state_group = object() - current_state_ids = yield defer.ensureDeferred(context.get_current_state_ids()) - result = yield self._get_joined_users_from_context( + current_state_ids = await context.get_current_state_ids() + return await self._get_joined_users_from_context( event.room_id, state_group, current_state_ids, event=event, context=context ) - return result - @defer.inlineCallbacks - def get_joined_users_from_state(self, room_id, state_entry): + async def get_joined_users_from_state(self, room_id, state_entry): state_group = state_entry.state_group if not state_group: # If state_group is None it means it has yet to be assigned a @@ -514,16 +503,12 @@ class RoomMemberWorkerStore(EventsWorkerStore): state_group = object() with Measure(self._clock, "get_joined_users_from_state"): - return ( - yield self._get_joined_users_from_context( - room_id, state_group, state_entry.state, context=state_entry - ) + return await self._get_joined_users_from_context( + room_id, state_group, state_entry.state, context=state_entry ) - @cachedInlineCallbacks( - num_args=2, cache_context=True, iterable=True, max_entries=100000 - ) - def _get_joined_users_from_context( + @cached(num_args=2, cache_context=True, iterable=True, max_entries=100000) + async def _get_joined_users_from_context( self, room_id, state_group, @@ -535,7 +520,6 @@ class RoomMemberWorkerStore(EventsWorkerStore): # We don't use `state_group`, it's there so that we can cache based # on it. However, it's important that it's never None, since two current_states # with a state_group of None are likely to be different. - # See bulk_get_push_rules_for_room for how we work around this. assert state_group is not None users_in_room = {} @@ -588,7 +572,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): missing_member_event_ids.append(event_id) if missing_member_event_ids: - event_to_memberships = yield self._get_joined_profiles_from_event_ids( + event_to_memberships = await self._get_joined_profiles_from_event_ids( missing_member_event_ids ) users_in_room.update((row for row in event_to_memberships.values() if row)) @@ -608,23 +592,21 @@ class RoomMemberWorkerStore(EventsWorkerStore): raise NotImplementedError() @cachedList( - cached_method_name="_get_joined_profile_from_event_id", - list_name="event_ids", - inlineCallbacks=True, + cached_method_name="_get_joined_profile_from_event_id", list_name="event_ids", ) - def _get_joined_profiles_from_event_ids(self, event_ids): + async def _get_joined_profiles_from_event_ids(self, event_ids: Iterable[str]): """For given set of member event_ids check if they point to a join event and if so return the associated user and profile info. Args: - event_ids (Iterable[str]): The member event IDs to lookup + event_ids: The member event IDs to lookup Returns: - Deferred[dict[str, Tuple[str, ProfileInfo]|None]]: Map from event ID + dict[str, Tuple[str, ProfileInfo]|None]: Map from event ID to `user_id` and ProfileInfo (or None if not join event). """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="room_memberships", column="event_id", iterable=event_ids, @@ -644,8 +626,8 @@ class RoomMemberWorkerStore(EventsWorkerStore): for row in rows } - @cachedInlineCallbacks(max_entries=10000) - def is_host_joined(self, room_id, host): + @cached(max_entries=10000) + async def is_host_joined(self, room_id: str, host: str) -> bool: if "%" in host or "_" in host: raise Exception("Invalid host name") @@ -664,47 +646,9 @@ class RoomMemberWorkerStore(EventsWorkerStore): # the returned user actually has the correct domain. like_clause = "%:" + host - rows = yield self.db.execute("is_host_joined", None, sql, room_id, like_clause) - - if not rows: - return False - - user_id = rows[0][0] - if get_domain_from_id(user_id) != host: - # This can only happen if the host name has something funky in it - raise Exception("Invalid host name") - - return True - - @cachedInlineCallbacks() - def was_host_joined(self, room_id, host): - """Check whether the server is or ever was in the room. - - Args: - room_id (str) - host (str) - - Returns: - Deferred: Resolves to True if the host is/was in the room, otherwise - False. - """ - if "%" in host or "_" in host: - raise Exception("Invalid host name") - - sql = """ - SELECT user_id FROM room_memberships - WHERE room_id = ? - AND user_id LIKE ? - AND membership = 'join' - LIMIT 1 - """ - - # We do need to be careful to ensure that host doesn't have any wild cards - # in it, but we checked above for known ones and we'll check below that - # the returned user actually has the correct domain. - like_clause = "%:" + host - - rows = yield self.db.execute("was_host_joined", None, sql, room_id, like_clause) + rows = await self.db_pool.execute( + "is_host_joined", None, sql, room_id, like_clause + ) if not rows: return False @@ -716,8 +660,7 @@ class RoomMemberWorkerStore(EventsWorkerStore): return True - @defer.inlineCallbacks - def get_joined_hosts(self, room_id, state_entry): + async def get_joined_hosts(self, room_id: str, state_entry): state_group = state_entry.state_group if not state_group: # If state_group is None it means it has yet to be assigned a @@ -727,32 +670,28 @@ class RoomMemberWorkerStore(EventsWorkerStore): state_group = object() with Measure(self._clock, "get_joined_hosts"): - return ( - yield self._get_joined_hosts( - room_id, state_group, state_entry.state, state_entry=state_entry - ) + return await self._get_joined_hosts( + room_id, state_group, state_entry.state, state_entry=state_entry ) - @cachedInlineCallbacks(num_args=2, max_entries=10000, iterable=True) - # @defer.inlineCallbacks - def _get_joined_hosts(self, room_id, state_group, current_state_ids, state_entry): + @cached(num_args=2, max_entries=10000, iterable=True) + async def _get_joined_hosts( + self, room_id, state_group, current_state_ids, state_entry + ): # We don't use `state_group`, its there so that we can cache based # on it. However, its important that its never None, since two current_state's # with a state_group of None are likely to be different. - # See bulk_get_push_rules_for_room for how we work around this. assert state_group is not None - cache = yield self._get_joined_hosts_cache(room_id) - joined_hosts = yield cache.get_destinations(state_entry) - - return joined_hosts + cache = await self._get_joined_hosts_cache(room_id) + return await cache.get_destinations(state_entry) @cached(max_entries=10000) - def _get_joined_hosts_cache(self, room_id): + def _get_joined_hosts_cache(self, room_id: str) -> "_JoinedHostsCache": return _JoinedHostsCache(self, room_id) - @cachedInlineCallbacks(num_args=2) - def did_forget(self, user_id, room_id): + @cached(num_args=2) + async def did_forget(self, user_id: str, room_id: str) -> bool: """Returns whether user_id has elected to discard history for room_id. Returns False if they have since re-joined.""" @@ -774,18 +713,18 @@ class RoomMemberWorkerStore(EventsWorkerStore): rows = txn.fetchall() return rows[0][0] - count = yield self.db.runInteraction("did_forget_membership", f) + count = await self.db_pool.runInteraction("did_forget_membership", f) return count == 0 @cached() - def get_forgotten_rooms_for_user(self, user_id): + async def get_forgotten_rooms_for_user(self, user_id: str) -> Set[str]: """Gets all rooms the user has forgotten. Args: - user_id (str) + user_id: The user ID to query the rooms of. Returns: - Deferred[set[str]] + The forgotten rooms. """ def _get_forgotten_rooms_for_user_txn(txn): @@ -811,22 +750,21 @@ class RoomMemberWorkerStore(EventsWorkerStore): txn.execute(sql, (user_id,)) return {row[0] for row in txn if row[1] == 0} - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_forgotten_rooms_for_user", _get_forgotten_rooms_for_user_txn ) - @defer.inlineCallbacks - def get_rooms_user_has_been_in(self, user_id): + async def get_rooms_user_has_been_in(self, user_id: str) -> Set[str]: """Get all rooms that the user has ever been in. Args: - user_id (str) + user_id: The user ID to get the rooms of. Returns: - Deferred[set[str]]: Set of room IDs. + Set of room IDs. """ - room_ids = yield self.db.simple_select_onecol( + room_ids = await self.db_pool.simple_select_onecol( table="room_memberships", keyvalues={"membership": Membership.JOIN, "user_id": user_id}, retcol="room_id", @@ -835,13 +773,13 @@ class RoomMemberWorkerStore(EventsWorkerStore): return set(room_ids) - def get_membership_from_event_ids( + async def get_membership_from_event_ids( self, member_event_ids: Iterable[str] ) -> List[dict]: """Get user_id and membership of a set of event IDs. """ - return self.db.simple_select_many_batch( + return await self.db_pool.simple_select_many_batch( table="room_memberships", column="event_id", iterable=member_event_ids, @@ -877,23 +815,23 @@ class RoomMemberWorkerStore(EventsWorkerStore): return bool(txn.fetchone()) - return await self.db.runInteraction( + return await self.db_pool.runInteraction( "is_local_host_in_room_ignoring_users", _is_local_host_in_room_ignoring_users_txn, ) class RoomMemberBackgroundUpdateStore(SQLBaseStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomMemberBackgroundUpdateStore, self).__init__(database, db_conn, hs) - self.db.updates.register_background_update_handler( + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) + self.db_pool.updates.register_background_update_handler( _MEMBERSHIP_PROFILE_UPDATE_NAME, self._background_add_membership_profile ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( _CURRENT_STATE_MEMBERSHIP_UPDATE_NAME, self._background_current_state_membership, ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( "room_membership_forgotten_idx", index_name="room_memberships_user_room_forgotten", table="room_memberships", @@ -901,8 +839,7 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): where_clause="forgotten = 1", ) - @defer.inlineCallbacks - def _background_add_membership_profile(self, progress, batch_size): + async def _background_add_membership_profile(self, progress, batch_size): target_min_stream_id = progress.get( "target_min_stream_id_inclusive", self._min_stream_order_on_start ) @@ -926,7 +863,7 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): txn.execute(sql, (target_min_stream_id, max_stream_id, batch_size)) - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return 0 @@ -961,25 +898,24 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): "max_stream_id_exclusive": min_stream_id, } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, _MEMBERSHIP_PROFILE_UPDATE_NAME, progress ) return len(rows) - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( _MEMBERSHIP_PROFILE_UPDATE_NAME, add_membership_profile_txn ) if not result: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( _MEMBERSHIP_PROFILE_UPDATE_NAME ) return result - @defer.inlineCallbacks - def _background_current_state_membership(self, progress, batch_size): + async def _background_current_state_membership(self, progress, batch_size): """Update the new membership column on current_state_events. This works by iterating over all rooms in alphebetical order. @@ -1013,7 +949,7 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): last_processed_room = next_room - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, _CURRENT_STATE_MEMBERSHIP_UPDATE_NAME, {"last_processed_room": last_processed_room}, @@ -1025,14 +961,14 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): # string, which will compare before all room IDs correctly. last_processed_room = progress.get("last_processed_room", "") - row_count, finished = yield self.db.runInteraction( + row_count, finished = await self.db_pool.runInteraction( "_background_current_state_membership_update", _background_current_state_membership_txn, last_processed_room, ) if finished: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( _CURRENT_STATE_MEMBERSHIP_UPDATE_NAME ) @@ -1040,10 +976,10 @@ class RoomMemberBackgroundUpdateStore(SQLBaseStore): class RoomMemberStore(RoomMemberWorkerStore, RoomMemberBackgroundUpdateStore): - def __init__(self, database: Database, db_conn, hs): - super(RoomMemberStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - def forget(self, user_id, room_id): + async def forget(self, user_id: str, room_id: str) -> None: """Indicate that user_id wishes to discard history for room_id.""" def f(txn): @@ -1064,10 +1000,10 @@ class RoomMemberStore(RoomMemberWorkerStore, RoomMemberBackgroundUpdateStore): txn, self.get_forgotten_rooms_for_user, (user_id,) ) - return self.db.runInteraction("forget_membership", f) + await self.db_pool.runInteraction("forget_membership", f) -class _JoinedHostsCache(object): +class _JoinedHostsCache: """Cache for joined hosts in a room that is optimised to handle updates via state deltas. """ @@ -1084,17 +1020,19 @@ class _JoinedHostsCache(object): self._len = 0 - @defer.inlineCallbacks - def get_destinations(self, state_entry): + async def get_destinations(self, state_entry: "_StateCacheEntry") -> Set[str]: """Get set of destinations for a state entry Args: - state_entry(synapse.state._StateCacheEntry) + state_entry + + Returns: + The destinations as a set. """ if state_entry.state_group == self.state_group: return frozenset(self.hosts_to_joined_users) - with (yield self.linearizer.queue(())): + with (await self.linearizer.queue(())): if state_entry.state_group == self.state_group: pass elif state_entry.prev_group == self.state_group: @@ -1106,7 +1044,7 @@ class _JoinedHostsCache(object): user_id = state_key known_joins = self.hosts_to_joined_users.setdefault(host, set()) - event = yield self.store.get_event(event_id) + event = await self.store.get_event(event_id) if event.membership == Membership.JOIN: known_joins.add(user_id) else: @@ -1115,7 +1053,7 @@ class _JoinedHostsCache(object): if not known_joins: self.hosts_to_joined_users.pop(host, None) else: - joined_users = yield self.store.get_joined_users_from_state( + joined_users = await self.store.get_joined_users_from_state( self.room_id, state_entry ) diff --git a/synapse/storage/data_stores/main/schema/delta/12/v12.sql b/synapse/storage/databases/main/schema/delta/12/v12.sql
index 5964c5aaac..5964c5aaac 100644 --- a/synapse/storage/data_stores/main/schema/delta/12/v12.sql +++ b/synapse/storage/databases/main/schema/delta/12/v12.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/13/v13.sql b/synapse/storage/databases/main/schema/delta/13/v13.sql
index f8649e5d99..f8649e5d99 100644 --- a/synapse/storage/data_stores/main/schema/delta/13/v13.sql +++ b/synapse/storage/databases/main/schema/delta/13/v13.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/14/v14.sql b/synapse/storage/databases/main/schema/delta/14/v14.sql
index a831920da6..a831920da6 100644 --- a/synapse/storage/data_stores/main/schema/delta/14/v14.sql +++ b/synapse/storage/databases/main/schema/delta/14/v14.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/15/appservice_txns.sql b/synapse/storage/databases/main/schema/delta/15/appservice_txns.sql
index e4f5e76aec..e4f5e76aec 100644 --- a/synapse/storage/data_stores/main/schema/delta/15/appservice_txns.sql +++ b/synapse/storage/databases/main/schema/delta/15/appservice_txns.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/15/presence_indices.sql b/synapse/storage/databases/main/schema/delta/15/presence_indices.sql
index 6b8d0f1ca7..6b8d0f1ca7 100644 --- a/synapse/storage/data_stores/main/schema/delta/15/presence_indices.sql +++ b/synapse/storage/databases/main/schema/delta/15/presence_indices.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/15/v15.sql b/synapse/storage/databases/main/schema/delta/15/v15.sql
index 9523d2bcc3..9523d2bcc3 100644 --- a/synapse/storage/data_stores/main/schema/delta/15/v15.sql +++ b/synapse/storage/databases/main/schema/delta/15/v15.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/events_order_index.sql b/synapse/storage/databases/main/schema/delta/16/events_order_index.sql
index a48f215170..a48f215170 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/events_order_index.sql +++ b/synapse/storage/databases/main/schema/delta/16/events_order_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/remote_media_cache_index.sql b/synapse/storage/databases/main/schema/delta/16/remote_media_cache_index.sql
index 7a15265cb1..7a15265cb1 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/remote_media_cache_index.sql +++ b/synapse/storage/databases/main/schema/delta/16/remote_media_cache_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/remove_duplicates.sql b/synapse/storage/databases/main/schema/delta/16/remove_duplicates.sql
index 65c97b5e2f..65c97b5e2f 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/remove_duplicates.sql +++ b/synapse/storage/databases/main/schema/delta/16/remove_duplicates.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/room_alias_index.sql b/synapse/storage/databases/main/schema/delta/16/room_alias_index.sql
index f82486132b..f82486132b 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/room_alias_index.sql +++ b/synapse/storage/databases/main/schema/delta/16/room_alias_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/unique_constraints.sql b/synapse/storage/databases/main/schema/delta/16/unique_constraints.sql
index 5b8de52c33..5b8de52c33 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/unique_constraints.sql +++ b/synapse/storage/databases/main/schema/delta/16/unique_constraints.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/16/users.sql b/synapse/storage/databases/main/schema/delta/16/users.sql
index cd0709250d..cd0709250d 100644 --- a/synapse/storage/data_stores/main/schema/delta/16/users.sql +++ b/synapse/storage/databases/main/schema/delta/16/users.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/17/drop_indexes.sql b/synapse/storage/databases/main/schema/delta/17/drop_indexes.sql
index 7c9a90e27f..7c9a90e27f 100644 --- a/synapse/storage/data_stores/main/schema/delta/17/drop_indexes.sql +++ b/synapse/storage/databases/main/schema/delta/17/drop_indexes.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/17/server_keys.sql b/synapse/storage/databases/main/schema/delta/17/server_keys.sql
index 70b247a06b..70b247a06b 100644 --- a/synapse/storage/data_stores/main/schema/delta/17/server_keys.sql +++ b/synapse/storage/databases/main/schema/delta/17/server_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/17/user_threepids.sql b/synapse/storage/databases/main/schema/delta/17/user_threepids.sql
index c17715ac80..c17715ac80 100644 --- a/synapse/storage/data_stores/main/schema/delta/17/user_threepids.sql +++ b/synapse/storage/databases/main/schema/delta/17/user_threepids.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/18/server_keys_bigger_ints.sql b/synapse/storage/databases/main/schema/delta/18/server_keys_bigger_ints.sql
index 6e0871c92b..6e0871c92b 100644 --- a/synapse/storage/data_stores/main/schema/delta/18/server_keys_bigger_ints.sql +++ b/synapse/storage/databases/main/schema/delta/18/server_keys_bigger_ints.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/19/event_index.sql b/synapse/storage/databases/main/schema/delta/19/event_index.sql
index 18b97b4332..18b97b4332 100644 --- a/synapse/storage/data_stores/main/schema/delta/19/event_index.sql +++ b/synapse/storage/databases/main/schema/delta/19/event_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/20/dummy.sql b/synapse/storage/databases/main/schema/delta/20/dummy.sql
index e0ac49d1ec..e0ac49d1ec 100644 --- a/synapse/storage/data_stores/main/schema/delta/20/dummy.sql +++ b/synapse/storage/databases/main/schema/delta/20/dummy.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/20/pushers.py b/synapse/storage/databases/main/schema/delta/20/pushers.py
index 3edfcfd783..3edfcfd783 100644 --- a/synapse/storage/data_stores/main/schema/delta/20/pushers.py +++ b/synapse/storage/databases/main/schema/delta/20/pushers.py
diff --git a/synapse/storage/data_stores/main/schema/delta/21/end_to_end_keys.sql b/synapse/storage/databases/main/schema/delta/21/end_to_end_keys.sql
index 4c2fb20b77..4c2fb20b77 100644 --- a/synapse/storage/data_stores/main/schema/delta/21/end_to_end_keys.sql +++ b/synapse/storage/databases/main/schema/delta/21/end_to_end_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/21/receipts.sql b/synapse/storage/databases/main/schema/delta/21/receipts.sql
index d070845477..d070845477 100644 --- a/synapse/storage/data_stores/main/schema/delta/21/receipts.sql +++ b/synapse/storage/databases/main/schema/delta/21/receipts.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/22/receipts_index.sql b/synapse/storage/databases/main/schema/delta/22/receipts_index.sql
index bfc0b3bcaa..bfc0b3bcaa 100644 --- a/synapse/storage/data_stores/main/schema/delta/22/receipts_index.sql +++ b/synapse/storage/databases/main/schema/delta/22/receipts_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/22/user_threepids_unique.sql b/synapse/storage/databases/main/schema/delta/22/user_threepids_unique.sql
index 87edfa454c..87edfa454c 100644 --- a/synapse/storage/data_stores/main/schema/delta/22/user_threepids_unique.sql +++ b/synapse/storage/databases/main/schema/delta/22/user_threepids_unique.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/24/stats_reporting.sql b/synapse/storage/databases/main/schema/delta/24/stats_reporting.sql
index acea7483bd..acea7483bd 100644 --- a/synapse/storage/data_stores/main/schema/delta/24/stats_reporting.sql +++ b/synapse/storage/databases/main/schema/delta/24/stats_reporting.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/25/fts.py b/synapse/storage/databases/main/schema/delta/25/fts.py
index ee675e71ff..ee675e71ff 100644 --- a/synapse/storage/data_stores/main/schema/delta/25/fts.py +++ b/synapse/storage/databases/main/schema/delta/25/fts.py
diff --git a/synapse/storage/data_stores/main/schema/delta/25/guest_access.sql b/synapse/storage/databases/main/schema/delta/25/guest_access.sql
index 1ea389b471..1ea389b471 100644 --- a/synapse/storage/data_stores/main/schema/delta/25/guest_access.sql +++ b/synapse/storage/databases/main/schema/delta/25/guest_access.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/25/history_visibility.sql b/synapse/storage/databases/main/schema/delta/25/history_visibility.sql
index f468fc1897..f468fc1897 100644 --- a/synapse/storage/data_stores/main/schema/delta/25/history_visibility.sql +++ b/synapse/storage/databases/main/schema/delta/25/history_visibility.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/25/tags.sql b/synapse/storage/databases/main/schema/delta/25/tags.sql
index 7a32ce68e4..7a32ce68e4 100644 --- a/synapse/storage/data_stores/main/schema/delta/25/tags.sql +++ b/synapse/storage/databases/main/schema/delta/25/tags.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/26/account_data.sql b/synapse/storage/databases/main/schema/delta/26/account_data.sql
index e395de2b5e..e395de2b5e 100644 --- a/synapse/storage/data_stores/main/schema/delta/26/account_data.sql +++ b/synapse/storage/databases/main/schema/delta/26/account_data.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/27/account_data.sql b/synapse/storage/databases/main/schema/delta/27/account_data.sql
index bf0558b5b3..bf0558b5b3 100644 --- a/synapse/storage/data_stores/main/schema/delta/27/account_data.sql +++ b/synapse/storage/databases/main/schema/delta/27/account_data.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/27/forgotten_memberships.sql b/synapse/storage/databases/main/schema/delta/27/forgotten_memberships.sql
index e2094f37fe..e2094f37fe 100644 --- a/synapse/storage/data_stores/main/schema/delta/27/forgotten_memberships.sql +++ b/synapse/storage/databases/main/schema/delta/27/forgotten_memberships.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/27/ts.py b/synapse/storage/databases/main/schema/delta/27/ts.py
index b7972cfa8e..b7972cfa8e 100644 --- a/synapse/storage/data_stores/main/schema/delta/27/ts.py +++ b/synapse/storage/databases/main/schema/delta/27/ts.py
diff --git a/synapse/storage/data_stores/main/schema/delta/28/event_push_actions.sql b/synapse/storage/databases/main/schema/delta/28/event_push_actions.sql
index 4d519849df..4d519849df 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/event_push_actions.sql +++ b/synapse/storage/databases/main/schema/delta/28/event_push_actions.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/28/events_room_stream.sql b/synapse/storage/databases/main/schema/delta/28/events_room_stream.sql
index 36609475f1..36609475f1 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/events_room_stream.sql +++ b/synapse/storage/databases/main/schema/delta/28/events_room_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/28/public_roms_index.sql b/synapse/storage/databases/main/schema/delta/28/public_roms_index.sql
index 6c1fd68c5b..6c1fd68c5b 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/public_roms_index.sql +++ b/synapse/storage/databases/main/schema/delta/28/public_roms_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/28/receipts_user_id_index.sql b/synapse/storage/databases/main/schema/delta/28/receipts_user_id_index.sql
index cb84c69baa..cb84c69baa 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/receipts_user_id_index.sql +++ b/synapse/storage/databases/main/schema/delta/28/receipts_user_id_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/28/upgrade_times.sql b/synapse/storage/databases/main/schema/delta/28/upgrade_times.sql
index 3e4a9ab455..3e4a9ab455 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/upgrade_times.sql +++ b/synapse/storage/databases/main/schema/delta/28/upgrade_times.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/28/users_is_guest.sql b/synapse/storage/databases/main/schema/delta/28/users_is_guest.sql
index 21d2b420bf..21d2b420bf 100644 --- a/synapse/storage/data_stores/main/schema/delta/28/users_is_guest.sql +++ b/synapse/storage/databases/main/schema/delta/28/users_is_guest.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/29/push_actions.sql b/synapse/storage/databases/main/schema/delta/29/push_actions.sql
index 84b21cf813..84b21cf813 100644 --- a/synapse/storage/data_stores/main/schema/delta/29/push_actions.sql +++ b/synapse/storage/databases/main/schema/delta/29/push_actions.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/alias_creator.sql b/synapse/storage/databases/main/schema/delta/30/alias_creator.sql
index c9d0dde638..c9d0dde638 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/alias_creator.sql +++ b/synapse/storage/databases/main/schema/delta/30/alias_creator.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/as_users.py b/synapse/storage/databases/main/schema/delta/30/as_users.py
index b42c02710a..b42c02710a 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/as_users.py +++ b/synapse/storage/databases/main/schema/delta/30/as_users.py
diff --git a/synapse/storage/data_stores/main/schema/delta/30/deleted_pushers.sql b/synapse/storage/databases/main/schema/delta/30/deleted_pushers.sql
index 712c454aa1..712c454aa1 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/deleted_pushers.sql +++ b/synapse/storage/databases/main/schema/delta/30/deleted_pushers.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/presence_stream.sql b/synapse/storage/databases/main/schema/delta/30/presence_stream.sql
index 606bbb037d..606bbb037d 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/presence_stream.sql +++ b/synapse/storage/databases/main/schema/delta/30/presence_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/public_rooms.sql b/synapse/storage/databases/main/schema/delta/30/public_rooms.sql
index f09db4faa6..f09db4faa6 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/public_rooms.sql +++ b/synapse/storage/databases/main/schema/delta/30/public_rooms.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/push_rule_stream.sql b/synapse/storage/databases/main/schema/delta/30/push_rule_stream.sql
index 735aa8d5f6..735aa8d5f6 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/push_rule_stream.sql +++ b/synapse/storage/databases/main/schema/delta/30/push_rule_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/30/threepid_guest_access_tokens.sql b/synapse/storage/databases/main/schema/delta/30/threepid_guest_access_tokens.sql
index 0dd2f1360c..0dd2f1360c 100644 --- a/synapse/storage/data_stores/main/schema/delta/30/threepid_guest_access_tokens.sql +++ b/synapse/storage/databases/main/schema/delta/30/threepid_guest_access_tokens.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/31/invites.sql b/synapse/storage/databases/main/schema/delta/31/invites.sql
index 2c57846d5a..2c57846d5a 100644 --- a/synapse/storage/data_stores/main/schema/delta/31/invites.sql +++ b/synapse/storage/databases/main/schema/delta/31/invites.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/31/local_media_repository_url_cache.sql b/synapse/storage/databases/main/schema/delta/31/local_media_repository_url_cache.sql
index 9efb4280eb..9efb4280eb 100644 --- a/synapse/storage/data_stores/main/schema/delta/31/local_media_repository_url_cache.sql +++ b/synapse/storage/databases/main/schema/delta/31/local_media_repository_url_cache.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/31/pushers.py b/synapse/storage/databases/main/schema/delta/31/pushers.py
index 9bb504aad5..9bb504aad5 100644 --- a/synapse/storage/data_stores/main/schema/delta/31/pushers.py +++ b/synapse/storage/databases/main/schema/delta/31/pushers.py
diff --git a/synapse/storage/data_stores/main/schema/delta/31/pushers_index.sql b/synapse/storage/databases/main/schema/delta/31/pushers_index.sql
index a82add88fd..a82add88fd 100644 --- a/synapse/storage/data_stores/main/schema/delta/31/pushers_index.sql +++ b/synapse/storage/databases/main/schema/delta/31/pushers_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/31/search_update.py b/synapse/storage/databases/main/schema/delta/31/search_update.py
index 63b757ade6..63b757ade6 100644 --- a/synapse/storage/data_stores/main/schema/delta/31/search_update.py +++ b/synapse/storage/databases/main/schema/delta/31/search_update.py
diff --git a/synapse/storage/data_stores/main/schema/delta/32/events.sql b/synapse/storage/databases/main/schema/delta/32/events.sql
index 1dd0f9e170..1dd0f9e170 100644 --- a/synapse/storage/data_stores/main/schema/delta/32/events.sql +++ b/synapse/storage/databases/main/schema/delta/32/events.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/32/openid.sql b/synapse/storage/databases/main/schema/delta/32/openid.sql
index 36f37b11c8..36f37b11c8 100644 --- a/synapse/storage/data_stores/main/schema/delta/32/openid.sql +++ b/synapse/storage/databases/main/schema/delta/32/openid.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/32/pusher_throttle.sql b/synapse/storage/databases/main/schema/delta/32/pusher_throttle.sql
index d86d30c13c..d86d30c13c 100644 --- a/synapse/storage/data_stores/main/schema/delta/32/pusher_throttle.sql +++ b/synapse/storage/databases/main/schema/delta/32/pusher_throttle.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/32/remove_indices.sql b/synapse/storage/databases/main/schema/delta/32/remove_indices.sql
index 2de50d408c..2de50d408c 100644 --- a/synapse/storage/data_stores/main/schema/delta/32/remove_indices.sql +++ b/synapse/storage/databases/main/schema/delta/32/remove_indices.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/32/reports.sql b/synapse/storage/databases/main/schema/delta/32/reports.sql
index d13609776f..d13609776f 100644 --- a/synapse/storage/data_stores/main/schema/delta/32/reports.sql +++ b/synapse/storage/databases/main/schema/delta/32/reports.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/33/access_tokens_device_index.sql b/synapse/storage/databases/main/schema/delta/33/access_tokens_device_index.sql
index 61ad3fe3e8..61ad3fe3e8 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/access_tokens_device_index.sql +++ b/synapse/storage/databases/main/schema/delta/33/access_tokens_device_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/33/devices.sql b/synapse/storage/databases/main/schema/delta/33/devices.sql
index eca7268d82..eca7268d82 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/devices.sql +++ b/synapse/storage/databases/main/schema/delta/33/devices.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys.sql b/synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys.sql
index aa4a3b9f2f..aa4a3b9f2f 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys.sql +++ b/synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql b/synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql
index 6671573398..6671573398 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql +++ b/synapse/storage/databases/main/schema/delta/33/devices_for_e2e_keys_clear_unknown_device.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/33/event_fields.py b/synapse/storage/databases/main/schema/delta/33/event_fields.py
index a3e81eeac7..a3e81eeac7 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/event_fields.py +++ b/synapse/storage/databases/main/schema/delta/33/event_fields.py
diff --git a/synapse/storage/data_stores/main/schema/delta/33/remote_media_ts.py b/synapse/storage/databases/main/schema/delta/33/remote_media_ts.py
index a26057dfb6..a26057dfb6 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/remote_media_ts.py +++ b/synapse/storage/databases/main/schema/delta/33/remote_media_ts.py
diff --git a/synapse/storage/data_stores/main/schema/delta/33/user_ips_index.sql b/synapse/storage/databases/main/schema/delta/33/user_ips_index.sql
index 473f75a78e..473f75a78e 100644 --- a/synapse/storage/data_stores/main/schema/delta/33/user_ips_index.sql +++ b/synapse/storage/databases/main/schema/delta/33/user_ips_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/34/appservice_stream.sql b/synapse/storage/databases/main/schema/delta/34/appservice_stream.sql
index 69e16eda0f..69e16eda0f 100644 --- a/synapse/storage/data_stores/main/schema/delta/34/appservice_stream.sql +++ b/synapse/storage/databases/main/schema/delta/34/appservice_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/34/cache_stream.py b/synapse/storage/databases/main/schema/delta/34/cache_stream.py
index cf09e43e2b..cf09e43e2b 100644 --- a/synapse/storage/data_stores/main/schema/delta/34/cache_stream.py +++ b/synapse/storage/databases/main/schema/delta/34/cache_stream.py
diff --git a/synapse/storage/data_stores/main/schema/delta/34/device_inbox.sql b/synapse/storage/databases/main/schema/delta/34/device_inbox.sql
index e68844c74a..e68844c74a 100644 --- a/synapse/storage/data_stores/main/schema/delta/34/device_inbox.sql +++ b/synapse/storage/databases/main/schema/delta/34/device_inbox.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/34/push_display_name_rename.sql b/synapse/storage/databases/main/schema/delta/34/push_display_name_rename.sql
index 0d9fe1a99a..0d9fe1a99a 100644 --- a/synapse/storage/data_stores/main/schema/delta/34/push_display_name_rename.sql +++ b/synapse/storage/databases/main/schema/delta/34/push_display_name_rename.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/34/received_txn_purge.py b/synapse/storage/databases/main/schema/delta/34/received_txn_purge.py
index 67d505e68b..67d505e68b 100644 --- a/synapse/storage/data_stores/main/schema/delta/34/received_txn_purge.py +++ b/synapse/storage/databases/main/schema/delta/34/received_txn_purge.py
diff --git a/synapse/storage/data_stores/main/schema/delta/35/contains_url.sql b/synapse/storage/databases/main/schema/delta/35/contains_url.sql
index 6cd123027b..6cd123027b 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/contains_url.sql +++ b/synapse/storage/databases/main/schema/delta/35/contains_url.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/35/device_outbox.sql b/synapse/storage/databases/main/schema/delta/35/device_outbox.sql
index 17e6c43105..17e6c43105 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/device_outbox.sql +++ b/synapse/storage/databases/main/schema/delta/35/device_outbox.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/35/device_stream_id.sql b/synapse/storage/databases/main/schema/delta/35/device_stream_id.sql
index 7ab7d942e2..7ab7d942e2 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/device_stream_id.sql +++ b/synapse/storage/databases/main/schema/delta/35/device_stream_id.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/35/event_push_actions_index.sql b/synapse/storage/databases/main/schema/delta/35/event_push_actions_index.sql
index 2e836d8e9c..2e836d8e9c 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/event_push_actions_index.sql +++ b/synapse/storage/databases/main/schema/delta/35/event_push_actions_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/35/public_room_list_change_stream.sql b/synapse/storage/databases/main/schema/delta/35/public_room_list_change_stream.sql
index dd2bf2e28a..dd2bf2e28a 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/public_room_list_change_stream.sql +++ b/synapse/storage/databases/main/schema/delta/35/public_room_list_change_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/35/stream_order_to_extrem.sql b/synapse/storage/databases/main/schema/delta/35/stream_order_to_extrem.sql
index 2b945d8a57..2b945d8a57 100644 --- a/synapse/storage/data_stores/main/schema/delta/35/stream_order_to_extrem.sql +++ b/synapse/storage/databases/main/schema/delta/35/stream_order_to_extrem.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/36/readd_public_rooms.sql b/synapse/storage/databases/main/schema/delta/36/readd_public_rooms.sql
index 90d8fd18f9..90d8fd18f9 100644 --- a/synapse/storage/data_stores/main/schema/delta/36/readd_public_rooms.sql +++ b/synapse/storage/databases/main/schema/delta/36/readd_public_rooms.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/37/remove_auth_idx.py b/synapse/storage/databases/main/schema/delta/37/remove_auth_idx.py
index a377884169..a377884169 100644 --- a/synapse/storage/data_stores/main/schema/delta/37/remove_auth_idx.py +++ b/synapse/storage/databases/main/schema/delta/37/remove_auth_idx.py
diff --git a/synapse/storage/data_stores/main/schema/delta/37/user_threepids.sql b/synapse/storage/databases/main/schema/delta/37/user_threepids.sql
index cf7a90dd10..cf7a90dd10 100644 --- a/synapse/storage/data_stores/main/schema/delta/37/user_threepids.sql +++ b/synapse/storage/databases/main/schema/delta/37/user_threepids.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/38/postgres_fts_gist.sql b/synapse/storage/databases/main/schema/delta/38/postgres_fts_gist.sql
index 515e6b8e84..515e6b8e84 100644 --- a/synapse/storage/data_stores/main/schema/delta/38/postgres_fts_gist.sql +++ b/synapse/storage/databases/main/schema/delta/38/postgres_fts_gist.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/39/appservice_room_list.sql b/synapse/storage/databases/main/schema/delta/39/appservice_room_list.sql
index 74bdc49073..74bdc49073 100644 --- a/synapse/storage/data_stores/main/schema/delta/39/appservice_room_list.sql +++ b/synapse/storage/databases/main/schema/delta/39/appservice_room_list.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/39/device_federation_stream_idx.sql b/synapse/storage/databases/main/schema/delta/39/device_federation_stream_idx.sql
index 00be801e90..00be801e90 100644 --- a/synapse/storage/data_stores/main/schema/delta/39/device_federation_stream_idx.sql +++ b/synapse/storage/databases/main/schema/delta/39/device_federation_stream_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/39/event_push_index.sql b/synapse/storage/databases/main/schema/delta/39/event_push_index.sql
index de2ad93e5c..de2ad93e5c 100644 --- a/synapse/storage/data_stores/main/schema/delta/39/event_push_index.sql +++ b/synapse/storage/databases/main/schema/delta/39/event_push_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/39/federation_out_position.sql b/synapse/storage/databases/main/schema/delta/39/federation_out_position.sql
index 5af814290b..5af814290b 100644 --- a/synapse/storage/data_stores/main/schema/delta/39/federation_out_position.sql +++ b/synapse/storage/databases/main/schema/delta/39/federation_out_position.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/39/membership_profile.sql b/synapse/storage/databases/main/schema/delta/39/membership_profile.sql
index 1bf911c8ab..1bf911c8ab 100644 --- a/synapse/storage/data_stores/main/schema/delta/39/membership_profile.sql +++ b/synapse/storage/databases/main/schema/delta/39/membership_profile.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/40/current_state_idx.sql b/synapse/storage/databases/main/schema/delta/40/current_state_idx.sql
index 7ffa189f39..7ffa189f39 100644 --- a/synapse/storage/data_stores/main/schema/delta/40/current_state_idx.sql +++ b/synapse/storage/databases/main/schema/delta/40/current_state_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/40/device_inbox.sql b/synapse/storage/databases/main/schema/delta/40/device_inbox.sql
index b9fe1f0480..b9fe1f0480 100644 --- a/synapse/storage/data_stores/main/schema/delta/40/device_inbox.sql +++ b/synapse/storage/databases/main/schema/delta/40/device_inbox.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/40/device_list_streams.sql b/synapse/storage/databases/main/schema/delta/40/device_list_streams.sql
index dd6dcb65f1..dd6dcb65f1 100644 --- a/synapse/storage/data_stores/main/schema/delta/40/device_list_streams.sql +++ b/synapse/storage/databases/main/schema/delta/40/device_list_streams.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/40/event_push_summary.sql b/synapse/storage/databases/main/schema/delta/40/event_push_summary.sql
index 3918f0b794..3918f0b794 100644 --- a/synapse/storage/data_stores/main/schema/delta/40/event_push_summary.sql +++ b/synapse/storage/databases/main/schema/delta/40/event_push_summary.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/40/pushers.sql b/synapse/storage/databases/main/schema/delta/40/pushers.sql
index 054a223f14..054a223f14 100644 --- a/synapse/storage/data_stores/main/schema/delta/40/pushers.sql +++ b/synapse/storage/databases/main/schema/delta/40/pushers.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/41/device_list_stream_idx.sql b/synapse/storage/databases/main/schema/delta/41/device_list_stream_idx.sql
index b7bee8b692..b7bee8b692 100644 --- a/synapse/storage/data_stores/main/schema/delta/41/device_list_stream_idx.sql +++ b/synapse/storage/databases/main/schema/delta/41/device_list_stream_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/41/device_outbound_index.sql b/synapse/storage/databases/main/schema/delta/41/device_outbound_index.sql
index 62f0b9892b..62f0b9892b 100644 --- a/synapse/storage/data_stores/main/schema/delta/41/device_outbound_index.sql +++ b/synapse/storage/databases/main/schema/delta/41/device_outbound_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/41/event_search_event_id_idx.sql b/synapse/storage/databases/main/schema/delta/41/event_search_event_id_idx.sql
index 5d9cfecf36..5d9cfecf36 100644 --- a/synapse/storage/data_stores/main/schema/delta/41/event_search_event_id_idx.sql +++ b/synapse/storage/databases/main/schema/delta/41/event_search_event_id_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/41/ratelimit.sql b/synapse/storage/databases/main/schema/delta/41/ratelimit.sql
index a194bf0238..a194bf0238 100644 --- a/synapse/storage/data_stores/main/schema/delta/41/ratelimit.sql +++ b/synapse/storage/databases/main/schema/delta/41/ratelimit.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/42/current_state_delta.sql b/synapse/storage/databases/main/schema/delta/42/current_state_delta.sql
index d28851aff8..d28851aff8 100644 --- a/synapse/storage/data_stores/main/schema/delta/42/current_state_delta.sql +++ b/synapse/storage/databases/main/schema/delta/42/current_state_delta.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/42/device_list_last_id.sql b/synapse/storage/databases/main/schema/delta/42/device_list_last_id.sql
index 9ab8c14fa3..9ab8c14fa3 100644 --- a/synapse/storage/data_stores/main/schema/delta/42/device_list_last_id.sql +++ b/synapse/storage/databases/main/schema/delta/42/device_list_last_id.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/42/event_auth_state_only.sql b/synapse/storage/databases/main/schema/delta/42/event_auth_state_only.sql
index b8821ac759..b8821ac759 100644 --- a/synapse/storage/data_stores/main/schema/delta/42/event_auth_state_only.sql +++ b/synapse/storage/databases/main/schema/delta/42/event_auth_state_only.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/42/user_dir.py b/synapse/storage/databases/main/schema/delta/42/user_dir.py
index 506f326f4d..506f326f4d 100644 --- a/synapse/storage/data_stores/main/schema/delta/42/user_dir.py +++ b/synapse/storage/databases/main/schema/delta/42/user_dir.py
diff --git a/synapse/storage/data_stores/main/schema/delta/43/blocked_rooms.sql b/synapse/storage/databases/main/schema/delta/43/blocked_rooms.sql
index 0e3cd143ff..0e3cd143ff 100644 --- a/synapse/storage/data_stores/main/schema/delta/43/blocked_rooms.sql +++ b/synapse/storage/databases/main/schema/delta/43/blocked_rooms.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/43/quarantine_media.sql b/synapse/storage/databases/main/schema/delta/43/quarantine_media.sql
index 630907ec4f..630907ec4f 100644 --- a/synapse/storage/data_stores/main/schema/delta/43/quarantine_media.sql +++ b/synapse/storage/databases/main/schema/delta/43/quarantine_media.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/43/url_cache.sql b/synapse/storage/databases/main/schema/delta/43/url_cache.sql
index 45ebe020da..45ebe020da 100644 --- a/synapse/storage/data_stores/main/schema/delta/43/url_cache.sql +++ b/synapse/storage/databases/main/schema/delta/43/url_cache.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/43/user_share.sql b/synapse/storage/databases/main/schema/delta/43/user_share.sql
index ee7062abe4..ee7062abe4 100644 --- a/synapse/storage/data_stores/main/schema/delta/43/user_share.sql +++ b/synapse/storage/databases/main/schema/delta/43/user_share.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/44/expire_url_cache.sql b/synapse/storage/databases/main/schema/delta/44/expire_url_cache.sql
index b12f9b2ebf..b12f9b2ebf 100644 --- a/synapse/storage/data_stores/main/schema/delta/44/expire_url_cache.sql +++ b/synapse/storage/databases/main/schema/delta/44/expire_url_cache.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/45/group_server.sql b/synapse/storage/databases/main/schema/delta/45/group_server.sql
index b2333848a0..b2333848a0 100644 --- a/synapse/storage/data_stores/main/schema/delta/45/group_server.sql +++ b/synapse/storage/databases/main/schema/delta/45/group_server.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/45/profile_cache.sql b/synapse/storage/databases/main/schema/delta/45/profile_cache.sql
index e5ddc84df0..e5ddc84df0 100644 --- a/synapse/storage/data_stores/main/schema/delta/45/profile_cache.sql +++ b/synapse/storage/databases/main/schema/delta/45/profile_cache.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/drop_refresh_tokens.sql b/synapse/storage/databases/main/schema/delta/46/drop_refresh_tokens.sql
index 68c48a89a9..68c48a89a9 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/drop_refresh_tokens.sql +++ b/synapse/storage/databases/main/schema/delta/46/drop_refresh_tokens.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/drop_unique_deleted_pushers.sql b/synapse/storage/databases/main/schema/delta/46/drop_unique_deleted_pushers.sql
index bb307889c1..bb307889c1 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/drop_unique_deleted_pushers.sql +++ b/synapse/storage/databases/main/schema/delta/46/drop_unique_deleted_pushers.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/group_server.sql b/synapse/storage/databases/main/schema/delta/46/group_server.sql
index 097679bc9a..097679bc9a 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/group_server.sql +++ b/synapse/storage/databases/main/schema/delta/46/group_server.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/local_media_repository_url_idx.sql b/synapse/storage/databases/main/schema/delta/46/local_media_repository_url_idx.sql
index bbfc7f5d1a..bbfc7f5d1a 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/local_media_repository_url_idx.sql +++ b/synapse/storage/databases/main/schema/delta/46/local_media_repository_url_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/user_dir_null_room_ids.sql b/synapse/storage/databases/main/schema/delta/46/user_dir_null_room_ids.sql
index cb0d5a2576..cb0d5a2576 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/user_dir_null_room_ids.sql +++ b/synapse/storage/databases/main/schema/delta/46/user_dir_null_room_ids.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/46/user_dir_typos.sql b/synapse/storage/databases/main/schema/delta/46/user_dir_typos.sql
index d9505f8da1..d9505f8da1 100644 --- a/synapse/storage/data_stores/main/schema/delta/46/user_dir_typos.sql +++ b/synapse/storage/databases/main/schema/delta/46/user_dir_typos.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/47/last_access_media.sql b/synapse/storage/databases/main/schema/delta/47/last_access_media.sql
index f505fb22b5..f505fb22b5 100644 --- a/synapse/storage/data_stores/main/schema/delta/47/last_access_media.sql +++ b/synapse/storage/databases/main/schema/delta/47/last_access_media.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/47/postgres_fts_gin.sql b/synapse/storage/databases/main/schema/delta/47/postgres_fts_gin.sql
index 31d7a817eb..31d7a817eb 100644 --- a/synapse/storage/data_stores/main/schema/delta/47/postgres_fts_gin.sql +++ b/synapse/storage/databases/main/schema/delta/47/postgres_fts_gin.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/47/push_actions_staging.sql b/synapse/storage/databases/main/schema/delta/47/push_actions_staging.sql
index edccf4a96f..edccf4a96f 100644 --- a/synapse/storage/data_stores/main/schema/delta/47/push_actions_staging.sql +++ b/synapse/storage/databases/main/schema/delta/47/push_actions_staging.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/48/add_user_consent.sql b/synapse/storage/databases/main/schema/delta/48/add_user_consent.sql
index 5237491506..5237491506 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/add_user_consent.sql +++ b/synapse/storage/databases/main/schema/delta/48/add_user_consent.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/48/add_user_ips_last_seen_index.sql b/synapse/storage/databases/main/schema/delta/48/add_user_ips_last_seen_index.sql
index 9248b0b24a..9248b0b24a 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/add_user_ips_last_seen_index.sql +++ b/synapse/storage/databases/main/schema/delta/48/add_user_ips_last_seen_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/48/deactivated_users.sql b/synapse/storage/databases/main/schema/delta/48/deactivated_users.sql
index e9013a6969..e9013a6969 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/deactivated_users.sql +++ b/synapse/storage/databases/main/schema/delta/48/deactivated_users.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/48/group_unique_indexes.py b/synapse/storage/databases/main/schema/delta/48/group_unique_indexes.py
index 49f5f2c003..49f5f2c003 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/group_unique_indexes.py +++ b/synapse/storage/databases/main/schema/delta/48/group_unique_indexes.py
diff --git a/synapse/storage/data_stores/main/schema/delta/48/groups_joinable.sql b/synapse/storage/databases/main/schema/delta/48/groups_joinable.sql
index ce26eaf0c9..ce26eaf0c9 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/groups_joinable.sql +++ b/synapse/storage/databases/main/schema/delta/48/groups_joinable.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/48/profiles_batch.sql b/synapse/storage/databases/main/schema/delta/48/profiles_batch.sql
index e744c02fe8..e744c02fe8 100644 --- a/synapse/storage/data_stores/main/schema/delta/48/profiles_batch.sql +++ b/synapse/storage/databases/main/schema/delta/48/profiles_batch.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/49/add_user_consent_server_notice_sent.sql b/synapse/storage/databases/main/schema/delta/49/add_user_consent_server_notice_sent.sql
index 14dcf18d73..14dcf18d73 100644 --- a/synapse/storage/data_stores/main/schema/delta/49/add_user_consent_server_notice_sent.sql +++ b/synapse/storage/databases/main/schema/delta/49/add_user_consent_server_notice_sent.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/49/add_user_daily_visits.sql b/synapse/storage/databases/main/schema/delta/49/add_user_daily_visits.sql
index 3dd478196f..3dd478196f 100644 --- a/synapse/storage/data_stores/main/schema/delta/49/add_user_daily_visits.sql +++ b/synapse/storage/databases/main/schema/delta/49/add_user_daily_visits.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/49/add_user_ips_last_seen_only_index.sql b/synapse/storage/databases/main/schema/delta/49/add_user_ips_last_seen_only_index.sql
index 3a4ed59b5b..3a4ed59b5b 100644 --- a/synapse/storage/data_stores/main/schema/delta/49/add_user_ips_last_seen_only_index.sql +++ b/synapse/storage/databases/main/schema/delta/49/add_user_ips_last_seen_only_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/50/add_creation_ts_users_index.sql b/synapse/storage/databases/main/schema/delta/50/add_creation_ts_users_index.sql
index c93ae47532..c93ae47532 100644 --- a/synapse/storage/data_stores/main/schema/delta/50/add_creation_ts_users_index.sql +++ b/synapse/storage/databases/main/schema/delta/50/add_creation_ts_users_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/50/erasure_store.sql b/synapse/storage/databases/main/schema/delta/50/erasure_store.sql
index 5d8641a9ab..5d8641a9ab 100644 --- a/synapse/storage/data_stores/main/schema/delta/50/erasure_store.sql +++ b/synapse/storage/databases/main/schema/delta/50/erasure_store.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/50/make_event_content_nullable.py b/synapse/storage/databases/main/schema/delta/50/make_event_content_nullable.py
index b1684a8441..b1684a8441 100644 --- a/synapse/storage/data_stores/main/schema/delta/50/make_event_content_nullable.py +++ b/synapse/storage/databases/main/schema/delta/50/make_event_content_nullable.py
diff --git a/synapse/storage/data_stores/main/schema/delta/50/profiles_deactivated_users.sql b/synapse/storage/databases/main/schema/delta/50/profiles_deactivated_users.sql
index c8893ecbe8..96051ac179 100644 --- a/synapse/storage/data_stores/main/schema/delta/50/profiles_deactivated_users.sql +++ b/synapse/storage/databases/main/schema/delta/50/profiles_deactivated_users.sql
@@ -20,4 +20,4 @@ * so we can't easily join between the two tables. Plus, the batch number * realy ought to represent data in this table that has changed. */ -ALTER TABLE profiles ADD COLUMN active SMALLINT DEFAULT 1 NOT NULL; +ALTER TABLE profiles ADD COLUMN active SMALLINT DEFAULT 1 NOT NULL; \ No newline at end of file diff --git a/synapse/storage/data_stores/main/schema/delta/51/e2e_room_keys.sql b/synapse/storage/databases/main/schema/delta/51/e2e_room_keys.sql
index c0e66a697d..c0e66a697d 100644 --- a/synapse/storage/data_stores/main/schema/delta/51/e2e_room_keys.sql +++ b/synapse/storage/databases/main/schema/delta/51/e2e_room_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/51/monthly_active_users.sql b/synapse/storage/databases/main/schema/delta/51/monthly_active_users.sql
index c9d537d5a3..c9d537d5a3 100644 --- a/synapse/storage/data_stores/main/schema/delta/51/monthly_active_users.sql +++ b/synapse/storage/databases/main/schema/delta/51/monthly_active_users.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/52/add_event_to_state_group_index.sql b/synapse/storage/databases/main/schema/delta/52/add_event_to_state_group_index.sql
index 91e03d13e1..91e03d13e1 100644 --- a/synapse/storage/data_stores/main/schema/delta/52/add_event_to_state_group_index.sql +++ b/synapse/storage/databases/main/schema/delta/52/add_event_to_state_group_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/52/device_list_streams_unique_idx.sql b/synapse/storage/databases/main/schema/delta/52/device_list_streams_unique_idx.sql
index bfa49e6f92..bfa49e6f92 100644 --- a/synapse/storage/data_stores/main/schema/delta/52/device_list_streams_unique_idx.sql +++ b/synapse/storage/databases/main/schema/delta/52/device_list_streams_unique_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/52/e2e_room_keys.sql b/synapse/storage/databases/main/schema/delta/52/e2e_room_keys.sql
index db687cccae..db687cccae 100644 --- a/synapse/storage/data_stores/main/schema/delta/52/e2e_room_keys.sql +++ b/synapse/storage/databases/main/schema/delta/52/e2e_room_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/add_user_type_to_users.sql b/synapse/storage/databases/main/schema/delta/53/add_user_type_to_users.sql
index 88ec2f83e5..88ec2f83e5 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/add_user_type_to_users.sql +++ b/synapse/storage/databases/main/schema/delta/53/add_user_type_to_users.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/drop_sent_transactions.sql b/synapse/storage/databases/main/schema/delta/53/drop_sent_transactions.sql
index e372f5a44a..e372f5a44a 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/drop_sent_transactions.sql +++ b/synapse/storage/databases/main/schema/delta/53/drop_sent_transactions.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/event_format_version.sql b/synapse/storage/databases/main/schema/delta/53/event_format_version.sql
index 1d977c2834..1d977c2834 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/event_format_version.sql +++ b/synapse/storage/databases/main/schema/delta/53/event_format_version.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/user_dir_populate.sql b/synapse/storage/databases/main/schema/delta/53/user_dir_populate.sql
index ffcc896b58..ffcc896b58 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/user_dir_populate.sql +++ b/synapse/storage/databases/main/schema/delta/53/user_dir_populate.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/user_ips_index.sql b/synapse/storage/databases/main/schema/delta/53/user_ips_index.sql
index b812c5794f..b812c5794f 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/user_ips_index.sql +++ b/synapse/storage/databases/main/schema/delta/53/user_ips_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/user_share.sql b/synapse/storage/databases/main/schema/delta/53/user_share.sql
index 5831b1a6f8..5831b1a6f8 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/user_share.sql +++ b/synapse/storage/databases/main/schema/delta/53/user_share.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/user_threepid_id.sql b/synapse/storage/databases/main/schema/delta/53/user_threepid_id.sql
index 80c2c573b6..80c2c573b6 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/user_threepid_id.sql +++ b/synapse/storage/databases/main/schema/delta/53/user_threepid_id.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/53/users_in_public_rooms.sql b/synapse/storage/databases/main/schema/delta/53/users_in_public_rooms.sql
index f7827ca6d2..f7827ca6d2 100644 --- a/synapse/storage/data_stores/main/schema/delta/53/users_in_public_rooms.sql +++ b/synapse/storage/databases/main/schema/delta/53/users_in_public_rooms.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/account_validity_with_renewal.sql b/synapse/storage/databases/main/schema/delta/54/account_validity_with_renewal.sql
index 0adb2ad55e..0adb2ad55e 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/account_validity_with_renewal.sql +++ b/synapse/storage/databases/main/schema/delta/54/account_validity_with_renewal.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/add_validity_to_server_keys.sql b/synapse/storage/databases/main/schema/delta/54/add_validity_to_server_keys.sql
index c01aa9d2d9..c01aa9d2d9 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/add_validity_to_server_keys.sql +++ b/synapse/storage/databases/main/schema/delta/54/add_validity_to_server_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/delete_forward_extremities.sql b/synapse/storage/databases/main/schema/delta/54/delete_forward_extremities.sql
index b062ec840c..b062ec840c 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/delete_forward_extremities.sql +++ b/synapse/storage/databases/main/schema/delta/54/delete_forward_extremities.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/drop_legacy_tables.sql b/synapse/storage/databases/main/schema/delta/54/drop_legacy_tables.sql
index dbbe682697..dbbe682697 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/drop_legacy_tables.sql +++ b/synapse/storage/databases/main/schema/delta/54/drop_legacy_tables.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/drop_presence_list.sql b/synapse/storage/databases/main/schema/delta/54/drop_presence_list.sql
index e6ee70c623..e6ee70c623 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/drop_presence_list.sql +++ b/synapse/storage/databases/main/schema/delta/54/drop_presence_list.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/relations.sql b/synapse/storage/databases/main/schema/delta/54/relations.sql
index 134862b870..134862b870 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/relations.sql +++ b/synapse/storage/databases/main/schema/delta/54/relations.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/stats.sql b/synapse/storage/databases/main/schema/delta/54/stats.sql
index 652e58308e..652e58308e 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/stats.sql +++ b/synapse/storage/databases/main/schema/delta/54/stats.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/54/stats2.sql b/synapse/storage/databases/main/schema/delta/54/stats2.sql
index 3b2d48447f..3b2d48447f 100644 --- a/synapse/storage/data_stores/main/schema/delta/54/stats2.sql +++ b/synapse/storage/databases/main/schema/delta/54/stats2.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/55/access_token_expiry.sql b/synapse/storage/databases/main/schema/delta/55/access_token_expiry.sql
index 4590604bfd..4590604bfd 100644 --- a/synapse/storage/data_stores/main/schema/delta/55/access_token_expiry.sql +++ b/synapse/storage/databases/main/schema/delta/55/access_token_expiry.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/55/profile_replication_status_index.sql b/synapse/storage/databases/main/schema/delta/55/profile_replication_status_index.sql
index 18a0f7e10c..7542ab8cbd 100644 --- a/synapse/storage/data_stores/main/schema/delta/55/profile_replication_status_index.sql +++ b/synapse/storage/databases/main/schema/delta/55/profile_replication_status_index.sql
@@ -13,5 +13,4 @@ * limitations under the License. */ -INSERT INTO background_updates (update_name, progress_json) VALUES - ('profile_replication_status_host_index', '{}'); +CREATE UNIQUE INDEX profile_replication_status_idx ON profile_replication_status(host); \ No newline at end of file diff --git a/synapse/storage/data_stores/main/schema/delta/55/track_threepid_validations.sql b/synapse/storage/databases/main/schema/delta/55/track_threepid_validations.sql
index a8eced2e0a..a8eced2e0a 100644 --- a/synapse/storage/data_stores/main/schema/delta/55/track_threepid_validations.sql +++ b/synapse/storage/databases/main/schema/delta/55/track_threepid_validations.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/55/users_alter_deactivated.sql b/synapse/storage/databases/main/schema/delta/55/users_alter_deactivated.sql
index dabdde489b..dabdde489b 100644 --- a/synapse/storage/data_stores/main/schema/delta/55/users_alter_deactivated.sql +++ b/synapse/storage/databases/main/schema/delta/55/users_alter_deactivated.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/add_spans_to_device_lists.sql b/synapse/storage/databases/main/schema/delta/56/add_spans_to_device_lists.sql
index 41807eb1e7..41807eb1e7 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/add_spans_to_device_lists.sql +++ b/synapse/storage/databases/main/schema/delta/56/add_spans_to_device_lists.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership.sql b/synapse/storage/databases/main/schema/delta/56/current_state_events_membership.sql
index 473018676f..473018676f 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership.sql +++ b/synapse/storage/databases/main/schema/delta/56/current_state_events_membership.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership_mk2.sql b/synapse/storage/databases/main/schema/delta/56/current_state_events_membership_mk2.sql
index 3133d42d4a..3133d42d4a 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/current_state_events_membership_mk2.sql +++ b/synapse/storage/databases/main/schema/delta/56/current_state_events_membership_mk2.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/delete_keys_from_deleted_backups.sql b/synapse/storage/databases/main/schema/delta/56/delete_keys_from_deleted_backups.sql
index 1d2ddb1b1a..1d2ddb1b1a 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/delete_keys_from_deleted_backups.sql +++ b/synapse/storage/databases/main/schema/delta/56/delete_keys_from_deleted_backups.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/destinations_failure_ts.sql b/synapse/storage/databases/main/schema/delta/56/destinations_failure_ts.sql
index f00889290b..f00889290b 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/destinations_failure_ts.sql +++ b/synapse/storage/databases/main/schema/delta/56/destinations_failure_ts.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/destinations_retry_interval_type.sql.postgres b/synapse/storage/databases/main/schema/delta/56/destinations_retry_interval_type.sql.postgres
index b9bbb18a91..b9bbb18a91 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/destinations_retry_interval_type.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/56/destinations_retry_interval_type.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/56/device_stream_id_insert.sql b/synapse/storage/databases/main/schema/delta/56/device_stream_id_insert.sql
index c2f557fde9..c2f557fde9 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/device_stream_id_insert.sql +++ b/synapse/storage/databases/main/schema/delta/56/device_stream_id_insert.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/devices_last_seen.sql b/synapse/storage/databases/main/schema/delta/56/devices_last_seen.sql
index dfa902d0ba..dfa902d0ba 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/devices_last_seen.sql +++ b/synapse/storage/databases/main/schema/delta/56/devices_last_seen.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/drop_unused_event_tables.sql b/synapse/storage/databases/main/schema/delta/56/drop_unused_event_tables.sql
index 9f09922c67..9f09922c67 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/drop_unused_event_tables.sql +++ b/synapse/storage/databases/main/schema/delta/56/drop_unused_event_tables.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/event_expiry.sql b/synapse/storage/databases/main/schema/delta/56/event_expiry.sql
index 81a36a8b1d..81a36a8b1d 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/event_expiry.sql +++ b/synapse/storage/databases/main/schema/delta/56/event_expiry.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql b/synapse/storage/databases/main/schema/delta/56/event_labels.sql
index 5e29c1da19..ccf287971c 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql +++ b/synapse/storage/databases/main/schema/delta/56/event_labels.sql
@@ -13,7 +13,7 @@ * limitations under the License. */ --- room_id and topoligical_ordering are denormalised from the events table in order to +-- room_id and topological_ordering are denormalised from the events table in order to -- make the index work. CREATE TABLE IF NOT EXISTS event_labels ( event_id TEXT, diff --git a/synapse/storage/data_stores/main/schema/delta/56/event_labels_background_update.sql b/synapse/storage/databases/main/schema/delta/56/event_labels_background_update.sql
index 5f5e0499ae..5f5e0499ae 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/event_labels_background_update.sql +++ b/synapse/storage/databases/main/schema/delta/56/event_labels_background_update.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/fix_room_keys_index.sql b/synapse/storage/databases/main/schema/delta/56/fix_room_keys_index.sql
index 014cb3b538..014cb3b538 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/fix_room_keys_index.sql +++ b/synapse/storage/databases/main/schema/delta/56/fix_room_keys_index.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/hidden_devices.sql b/synapse/storage/databases/main/schema/delta/56/hidden_devices.sql
index 67f8b20297..67f8b20297 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/hidden_devices.sql +++ b/synapse/storage/databases/main/schema/delta/56/hidden_devices.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/hidden_devices_fix.sql.sqlite b/synapse/storage/databases/main/schema/delta/56/hidden_devices_fix.sql.sqlite
index e8b1fd35d8..e8b1fd35d8 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/hidden_devices_fix.sql.sqlite +++ b/synapse/storage/databases/main/schema/delta/56/hidden_devices_fix.sql.sqlite
diff --git a/synapse/storage/data_stores/main/schema/delta/56/nuke_empty_communities_from_db.sql b/synapse/storage/databases/main/schema/delta/56/nuke_empty_communities_from_db.sql
index 4f24c1405d..4f24c1405d 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/nuke_empty_communities_from_db.sql +++ b/synapse/storage/databases/main/schema/delta/56/nuke_empty_communities_from_db.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/public_room_list_idx.sql b/synapse/storage/databases/main/schema/delta/56/public_room_list_idx.sql
index 7be31ffebb..7be31ffebb 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/public_room_list_idx.sql +++ b/synapse/storage/databases/main/schema/delta/56/public_room_list_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor.sql b/synapse/storage/databases/main/schema/delta/56/redaction_censor.sql
index ea95db0ed7..ea95db0ed7 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor.sql +++ b/synapse/storage/databases/main/schema/delta/56/redaction_censor.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor2.sql b/synapse/storage/databases/main/schema/delta/56/redaction_censor2.sql
index 49ce35d794..49ce35d794 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor2.sql +++ b/synapse/storage/databases/main/schema/delta/56/redaction_censor2.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres b/synapse/storage/databases/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres
index 67471f3ef5..67471f3ef5 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/56/redaction_censor3_fix_update.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor4.sql b/synapse/storage/databases/main/schema/delta/56/redaction_censor4.sql
index b7550f6f4e..b7550f6f4e 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/redaction_censor4.sql +++ b/synapse/storage/databases/main/schema/delta/56/redaction_censor4.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql b/synapse/storage/databases/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql
index aeb17813d3..aeb17813d3 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql +++ b/synapse/storage/databases/main/schema/delta/56/remove_tombstoned_rooms_from_directory.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/room_key_etag.sql b/synapse/storage/databases/main/schema/delta/56/room_key_etag.sql
index 7d70dd071e..7d70dd071e 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/room_key_etag.sql +++ b/synapse/storage/databases/main/schema/delta/56/room_key_etag.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/room_membership_idx.sql b/synapse/storage/databases/main/schema/delta/56/room_membership_idx.sql
index 92ab1f5e65..92ab1f5e65 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/room_membership_idx.sql +++ b/synapse/storage/databases/main/schema/delta/56/room_membership_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/55/room_retention.sql b/synapse/storage/databases/main/schema/delta/56/room_retention.sql
index ee6cdf7a14..ee6cdf7a14 100644 --- a/synapse/storage/data_stores/main/schema/delta/55/room_retention.sql +++ b/synapse/storage/databases/main/schema/delta/56/room_retention.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/signing_keys.sql b/synapse/storage/databases/main/schema/delta/56/signing_keys.sql
index 5c5fffcafb..5c5fffcafb 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/signing_keys.sql +++ b/synapse/storage/databases/main/schema/delta/56/signing_keys.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/signing_keys_nonunique_signatures.sql b/synapse/storage/databases/main/schema/delta/56/signing_keys_nonunique_signatures.sql
index 0aa90ebf0c..0aa90ebf0c 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/signing_keys_nonunique_signatures.sql +++ b/synapse/storage/databases/main/schema/delta/56/signing_keys_nonunique_signatures.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/stats_separated.sql b/synapse/storage/databases/main/schema/delta/56/stats_separated.sql
index bbdde121e8..bbdde121e8 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/stats_separated.sql +++ b/synapse/storage/databases/main/schema/delta/56/stats_separated.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/unique_user_filter_index.py b/synapse/storage/databases/main/schema/delta/56/unique_user_filter_index.py
index 1de8b54961..1de8b54961 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/unique_user_filter_index.py +++ b/synapse/storage/databases/main/schema/delta/56/unique_user_filter_index.py
diff --git a/synapse/storage/data_stores/main/schema/delta/56/user_external_ids.sql b/synapse/storage/databases/main/schema/delta/56/user_external_ids.sql
index 91390c4527..91390c4527 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/user_external_ids.sql +++ b/synapse/storage/databases/main/schema/delta/56/user_external_ids.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/56/users_in_public_rooms_idx.sql b/synapse/storage/databases/main/schema/delta/56/users_in_public_rooms_idx.sql
index 149f8be8b6..149f8be8b6 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/users_in_public_rooms_idx.sql +++ b/synapse/storage/databases/main/schema/delta/56/users_in_public_rooms_idx.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/57/delete_old_current_state_events.sql b/synapse/storage/databases/main/schema/delta/57/delete_old_current_state_events.sql
index aec06c8261..aec06c8261 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/delete_old_current_state_events.sql +++ b/synapse/storage/databases/main/schema/delta/57/delete_old_current_state_events.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/57/device_list_remote_cache_stale.sql b/synapse/storage/databases/main/schema/delta/57/device_list_remote_cache_stale.sql
index c3b6de2099..c3b6de2099 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/device_list_remote_cache_stale.sql +++ b/synapse/storage/databases/main/schema/delta/57/device_list_remote_cache_stale.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/57/local_current_membership.py b/synapse/storage/databases/main/schema/delta/57/local_current_membership.py
index 63b5acdcf7..63b5acdcf7 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/local_current_membership.py +++ b/synapse/storage/databases/main/schema/delta/57/local_current_membership.py
diff --git a/synapse/storage/data_stores/main/schema/delta/57/remove_sent_outbound_pokes.sql b/synapse/storage/databases/main/schema/delta/57/remove_sent_outbound_pokes.sql
index 133d80af35..133d80af35 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/remove_sent_outbound_pokes.sql +++ b/synapse/storage/databases/main/schema/delta/57/remove_sent_outbound_pokes.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column.sql b/synapse/storage/databases/main/schema/delta/57/rooms_version_column.sql
index 352a66f5b0..352a66f5b0 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column.sql +++ b/synapse/storage/databases/main/schema/delta/57/rooms_version_column.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.postgres b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.postgres
index c601cff6de..c601cff6de 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.sqlite b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.sqlite
index 335c6f2074..335c6f2074 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_2.sql.sqlite +++ b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_2.sql.sqlite
diff --git a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.postgres b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.postgres
index 92aaadde0d..92aaadde0d 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.sqlite b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.sqlite
index e19dab97cb..e19dab97cb 100644 --- a/synapse/storage/data_stores/main/schema/delta/57/rooms_version_column_3.sql.sqlite +++ b/synapse/storage/databases/main/schema/delta/57/rooms_version_column_3.sql.sqlite
diff --git a/synapse/storage/data_stores/main/schema/delta/58/02remove_dup_outbound_pokes.sql b/synapse/storage/databases/main/schema/delta/58/02remove_dup_outbound_pokes.sql
index fdc39e9ba5..fdc39e9ba5 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/02remove_dup_outbound_pokes.sql +++ b/synapse/storage/databases/main/schema/delta/58/02remove_dup_outbound_pokes.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/58/03persist_ui_auth.sql b/synapse/storage/databases/main/schema/delta/58/03persist_ui_auth.sql
index dcb593fc2d..dcb593fc2d 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/03persist_ui_auth.sql +++ b/synapse/storage/databases/main/schema/delta/58/03persist_ui_auth.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/58/05cache_instance.sql.postgres b/synapse/storage/databases/main/schema/delta/58/05cache_instance.sql.postgres
index aa46eb0e10..aa46eb0e10 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/05cache_instance.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/58/05cache_instance.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/58/06dlols_unique_idx.py b/synapse/storage/databases/main/schema/delta/58/06dlols_unique_idx.py
index d353f2bcb3..d353f2bcb3 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/06dlols_unique_idx.py +++ b/synapse/storage/databases/main/schema/delta/58/06dlols_unique_idx.py
diff --git a/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.postgres b/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.postgres new file mode 100644
index 0000000000..b64926e9c9 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.postgres
@@ -0,0 +1,33 @@ +/* Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +/* + * This adds the method to the unique key constraint of the thumbnail databases. + * Otherwise you can't have a scaled and a cropped thumbnail with the same + * resolution, which happens quite often with dynamic thumbnailing. + * This is the postgres specific migration modifying the table with a background + * migration. + */ + +-- add new index that includes method to local media +INSERT INTO background_updates (update_name, progress_json) VALUES + ('local_media_repository_thumbnails_method_idx', '{}'); + +-- add new index that includes method to remote media +INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES + ('remote_media_repository_thumbnails_method_idx', '{}', 'local_media_repository_thumbnails_method_idx'); + +-- drop old index +INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES + ('media_repository_drop_index_wo_method', '{}', 'remote_media_repository_thumbnails_method_idx'); + diff --git a/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.sqlite b/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.sqlite new file mode 100644
index 0000000000..1d0c04b53a --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/07add_method_to_thumbnail_constraint.sql.sqlite
@@ -0,0 +1,44 @@ +/* Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +/* + * This adds the method to the unique key constraint of the thumbnail databases. + * Otherwise you can't have a scaled and a cropped thumbnail with the same + * resolution, which happens quite often with dynamic thumbnailing. + * This is a sqlite specific migration, since sqlite can't modify the unique + * constraint of a table without recreating it. + */ + +CREATE TABLE local_media_repository_thumbnails_new ( media_id TEXT, thumbnail_width INTEGER, thumbnail_height INTEGER, thumbnail_type TEXT, thumbnail_method TEXT, thumbnail_length INTEGER, UNIQUE ( media_id, thumbnail_width, thumbnail_height, thumbnail_type, thumbnail_method ) ); + +INSERT INTO local_media_repository_thumbnails_new + SELECT media_id, thumbnail_width, thumbnail_height, thumbnail_type, thumbnail_method, thumbnail_length + FROM local_media_repository_thumbnails; + +DROP TABLE local_media_repository_thumbnails; + +ALTER TABLE local_media_repository_thumbnails_new RENAME TO local_media_repository_thumbnails; + +CREATE INDEX local_media_repository_thumbnails_media_id ON local_media_repository_thumbnails (media_id); + + + +CREATE TABLE IF NOT EXISTS remote_media_cache_thumbnails_new ( media_origin TEXT, media_id TEXT, thumbnail_width INTEGER, thumbnail_height INTEGER, thumbnail_method TEXT, thumbnail_type TEXT, thumbnail_length INTEGER, filesystem_id TEXT, UNIQUE ( media_origin, media_id, thumbnail_width, thumbnail_height, thumbnail_type, thumbnail_method ) ); + +INSERT INTO remote_media_cache_thumbnails_new + SELECT media_origin, media_id, thumbnail_width, thumbnail_height, thumbnail_method, thumbnail_type, thumbnail_length, filesystem_id + FROM remote_media_cache_thumbnails; + +DROP TABLE remote_media_cache_thumbnails; + +ALTER TABLE remote_media_cache_thumbnails_new RENAME TO remote_media_cache_thumbnails; diff --git a/synapse/storage/databases/main/schema/delta/58/07persist_ui_auth_ips.sql b/synapse/storage/databases/main/schema/delta/58/07persist_ui_auth_ips.sql new file mode 100644
index 0000000000..4cc96a5341 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/07persist_ui_auth_ips.sql
@@ -0,0 +1,25 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- A table of the IP address and user-agent used to complete each step of a +-- user-interactive authentication session. +CREATE TABLE IF NOT EXISTS ui_auth_sessions_ips( + session_id TEXT NOT NULL, + ip TEXT NOT NULL, + user_agent TEXT NOT NULL, + UNIQUE (session_id, ip, user_agent), + FOREIGN KEY (session_id) + REFERENCES ui_auth_sessions (session_id) +); diff --git a/synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres b/synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres
index 597f2ffd3d..597f2ffd3d 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres +++ b/synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite b/synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite
index 69db89ac0e..69db89ac0e 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite +++ b/synapse/storage/databases/main/schema/delta/58/08_media_safe_from_quarantine.sql.sqlite
diff --git a/synapse/storage/databases/main/schema/delta/58/09shadow_ban.sql b/synapse/storage/databases/main/schema/delta/58/09shadow_ban.sql new file mode 100644
index 0000000000..260b009b48 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/09shadow_ban.sql
@@ -0,0 +1,18 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- A shadow-banned user may be told that their requests succeeded when they were +-- actually ignored. +ALTER TABLE users ADD COLUMN shadow_banned BOOLEAN; diff --git a/synapse/storage/databases/main/schema/delta/58/10_pushrules_enabled_delete_obsolete.sql b/synapse/storage/databases/main/schema/delta/58/10_pushrules_enabled_delete_obsolete.sql new file mode 100644
index 0000000000..847aebd85e --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/10_pushrules_enabled_delete_obsolete.sql
@@ -0,0 +1,28 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +/** + Delete stuck 'enabled' bits that correspond to deleted or non-existent push rules. + We ignore rules that are server-default rules because they are not defined + in the `push_rules` table. +**/ + +DELETE FROM push_rules_enable WHERE + rule_id NOT LIKE 'global/%/.m.rule.%' + AND NOT EXISTS ( + SELECT 1 FROM push_rules + WHERE push_rules.user_name = push_rules_enable.user_name + AND push_rules.rule_id = push_rules_enable.rule_id + ); diff --git a/synapse/storage/data_stores/main/schema/delta/58/10drop_local_rejections_stream.sql b/synapse/storage/databases/main/schema/delta/58/10drop_local_rejections_stream.sql
index eb57203e46..eb57203e46 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/10drop_local_rejections_stream.sql +++ b/synapse/storage/databases/main/schema/delta/58/10drop_local_rejections_stream.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/58/10federation_pos_instance_name.sql b/synapse/storage/databases/main/schema/delta/58/10federation_pos_instance_name.sql
index 1cc2633aad..1cc2633aad 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/10federation_pos_instance_name.sql +++ b/synapse/storage/databases/main/schema/delta/58/10federation_pos_instance_name.sql
diff --git a/synapse/storage/data_stores/main/schema/delta/58/11user_id_seq.py b/synapse/storage/databases/main/schema/delta/58/11user_id_seq.py
index 2011f6bceb..4310ec12ce 100644 --- a/synapse/storage/data_stores/main/schema/delta/58/11user_id_seq.py +++ b/synapse/storage/databases/main/schema/delta/58/11user_id_seq.py
@@ -16,7 +16,7 @@ Adds a postgres SEQUENCE for generating guest user IDs. """ -from synapse.storage.data_stores.main.registration import ( +from synapse.storage.databases.main.registration import ( find_max_generated_user_id_localpart, ) from synapse.storage.engines import PostgresEngine diff --git a/synapse/storage/databases/main/schema/delta/58/12room_stats.sql b/synapse/storage/databases/main/schema/delta/58/12room_stats.sql new file mode 100644
index 0000000000..cade5dcca8 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/12room_stats.sql
@@ -0,0 +1,32 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- Recalculate the stats for all rooms after the fix to joined_members erroneously +-- incrementing on per-room profile changes. + +-- Note that the populate_stats_process_rooms background update is already set to +-- run if you're upgrading from Synapse <1.0.0. + +-- Additionally, if you've upgraded to v1.18.0 (which doesn't include this fix), +-- this bg job runs, and then update to v1.19.0, you'd end up with only half of +-- your rooms having room stats recalculated after this fix was in place. + +-- So we've switched the old `populate_stats_process_rooms` background job to a +-- no-op, and then kick off a bg job with a new name, but with the same +-- functionality as the old one. This effectively restarts the background job +-- from the beginning, without running it twice in a row, supporting both +-- upgrade usecases. +INSERT INTO background_updates (update_name, progress_json) VALUES + ('populate_stats_process_rooms_2', '{}'); diff --git a/synapse/storage/databases/main/schema/delta/58/13remove_presence_allow_inbound.sql b/synapse/storage/databases/main/schema/delta/58/13remove_presence_allow_inbound.sql new file mode 100644
index 0000000000..15421b99ac --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/13remove_presence_allow_inbound.sql
@@ -0,0 +1,17 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- This table is no longer used. +DROP TABLE IF EXISTS presence_allow_inbound; diff --git a/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql b/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql new file mode 100644
index 0000000000..98ff76d709 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql
@@ -0,0 +1,16 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +ALTER TABLE events ADD COLUMN instance_name TEXT; diff --git a/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql.postgres b/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql.postgres new file mode 100644
index 0000000000..c31f9af82a --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/14events_instance_name.sql.postgres
@@ -0,0 +1,28 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +CREATE SEQUENCE IF NOT EXISTS events_stream_seq; + +SELECT setval('events_stream_seq', ( + SELECT COALESCE(MAX(stream_ordering), 1) FROM events +)); + +CREATE SEQUENCE IF NOT EXISTS events_backfill_stream_seq; + +-- If the server has never backfilled a room then doing `-MIN(...)` will give +-- a negative result, hence why we do `GREATEST(...)` +SELECT setval('events_backfill_stream_seq', ( + SELECT GREATEST(COALESCE(-MIN(stream_ordering), 1), 1) FROM events +)); diff --git a/synapse/storage/databases/main/schema/delta/58/15_catchup_destination_rooms.sql b/synapse/storage/databases/main/schema/delta/58/15_catchup_destination_rooms.sql new file mode 100644
index 0000000000..ebfbed7925 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/15_catchup_destination_rooms.sql
@@ -0,0 +1,42 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +-- This schema delta alters the schema to enable 'catching up' remote homeservers +-- after there has been a connectivity problem for any reason. + +-- This stores, for each (destination, room) pair, the stream_ordering of the +-- latest event for that destination. +CREATE TABLE IF NOT EXISTS destination_rooms ( + -- the destination in question. + destination TEXT NOT NULL REFERENCES destinations (destination), + -- the ID of the room in question + room_id TEXT NOT NULL REFERENCES rooms (room_id), + -- the stream_ordering of the event + stream_ordering BIGINT NOT NULL, + PRIMARY KEY (destination, room_id) + -- We don't declare a foreign key on stream_ordering here because that'd mean + -- we'd need to either maintain an index (expensive) or do a table scan of + -- destination_rooms whenever we delete an event (also potentially expensive). + -- In addition to that, a foreign key on stream_ordering would be redundant + -- as this row doesn't need to refer to a specific event; if the event gets + -- deleted then it doesn't affect the validity of the stream_ordering here. +); + +-- This index is needed to make it so that a deletion of a room (in the rooms +-- table) can be efficient, as otherwise a table scan would need to be performed +-- to check that no destination_rooms rows point to the room to be deleted. +-- Also: it makes it efficient to delete all the entries for a given room ID, +-- such as when purging a room. +CREATE INDEX IF NOT EXISTS destination_rooms_room_id + ON destination_rooms (room_id); diff --git a/synapse/storage/databases/main/schema/delta/58/15unread_count.sql b/synapse/storage/databases/main/schema/delta/58/15unread_count.sql new file mode 100644
index 0000000000..317fba8a5d --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/15unread_count.sql
@@ -0,0 +1,26 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- We're hijacking the push actions to store unread messages and unread counts (specified +-- in MSC2654) because doing otherwise would result in either performance issues or +-- reimplementing a consequent bit of the push actions. + +-- Add columns to event_push_actions and event_push_actions_staging to track unread +-- messages and calculate unread counts. +ALTER TABLE event_push_actions_staging ADD COLUMN unread SMALLINT; +ALTER TABLE event_push_actions ADD COLUMN unread SMALLINT; + +-- Add column to event_push_summary +ALTER TABLE event_push_summary ADD COLUMN unread_count BIGINT; \ No newline at end of file diff --git a/synapse/storage/databases/main/schema/delta/58/16populate_stats_process_rooms_fix.sql b/synapse/storage/databases/main/schema/delta/58/16populate_stats_process_rooms_fix.sql new file mode 100644
index 0000000000..55f5d0f732 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/16populate_stats_process_rooms_fix.sql
@@ -0,0 +1,22 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C. + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ +-- This delta file fixes a regression introduced by 58/12room_stats.sql, removing the hacky +-- populate_stats_process_rooms_2 background job and restores the functionality under the +-- original name. +-- See https://github.com/matrix-org/synapse/issues/8238 for details + +DELETE FROM background_updates WHERE update_name = 'populate_stats_process_rooms'; +UPDATE background_updates SET update_name = 'populate_stats_process_rooms' + WHERE update_name = 'populate_stats_process_rooms_2'; diff --git a/synapse/storage/databases/main/schema/delta/58/17_catchup_last_successful.sql b/synapse/storage/databases/main/schema/delta/58/17_catchup_last_successful.sql new file mode 100644
index 0000000000..a67aa5e500 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/17_catchup_last_successful.sql
@@ -0,0 +1,21 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +-- This column tracks the stream_ordering of the event that was most recently +-- successfully transmitted to the destination. +-- A value of NULL means that we have not sent an event successfully yet +-- (at least, not since the introduction of this column). +ALTER TABLE destinations + ADD COLUMN last_successful_stream_ordering BIGINT; diff --git a/synapse/storage/databases/main/schema/delta/58/18stream_positions.sql b/synapse/storage/databases/main/schema/delta/58/18stream_positions.sql new file mode 100644
index 0000000000..985fd949a2 --- /dev/null +++ b/synapse/storage/databases/main/schema/delta/58/18stream_positions.sql
@@ -0,0 +1,22 @@ +/* Copyright 2020 The Matrix.org Foundation C.I.C + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +CREATE TABLE stream_positions ( + stream_name TEXT NOT NULL, + instance_name TEXT NOT NULL, + stream_id BIGINT NOT NULL +); + +CREATE UNIQUE INDEX stream_positions_idx ON stream_positions(stream_name, instance_name); diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/application_services.sql b/synapse/storage/databases/main/schema/full_schemas/16/application_services.sql
index 883fcd10b2..883fcd10b2 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/application_services.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/application_services.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/event_edges.sql b/synapse/storage/databases/main/schema/full_schemas/16/event_edges.sql
index 10ce2aa7a0..10ce2aa7a0 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/event_edges.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/event_edges.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/event_signatures.sql b/synapse/storage/databases/main/schema/full_schemas/16/event_signatures.sql
index 95826da431..95826da431 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/event_signatures.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/event_signatures.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/im.sql b/synapse/storage/databases/main/schema/full_schemas/16/im.sql
index a1a2aa8e5b..a1a2aa8e5b 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/im.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/im.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/keys.sql b/synapse/storage/databases/main/schema/full_schemas/16/keys.sql
index 11cdffdbb3..11cdffdbb3 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/keys.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/keys.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/media_repository.sql b/synapse/storage/databases/main/schema/full_schemas/16/media_repository.sql
index 8f3759bb2a..8f3759bb2a 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/media_repository.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/media_repository.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/presence.sql b/synapse/storage/databases/main/schema/full_schemas/16/presence.sql
index 01d2d8f833..01d2d8f833 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/presence.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/presence.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/profiles.sql b/synapse/storage/databases/main/schema/full_schemas/16/profiles.sql
index c04f4747d9..c04f4747d9 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/profiles.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/profiles.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/push.sql b/synapse/storage/databases/main/schema/full_schemas/16/push.sql
index e44465cf45..e44465cf45 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/push.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/push.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/redactions.sql b/synapse/storage/databases/main/schema/full_schemas/16/redactions.sql
index 318f0d9aa5..318f0d9aa5 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/redactions.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/redactions.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/room_aliases.sql b/synapse/storage/databases/main/schema/full_schemas/16/room_aliases.sql
index d47da3b12f..d47da3b12f 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/room_aliases.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/room_aliases.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/state.sql b/synapse/storage/databases/main/schema/full_schemas/16/state.sql
index 96391a8f0e..96391a8f0e 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/state.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/state.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/transactions.sql b/synapse/storage/databases/main/schema/full_schemas/16/transactions.sql
index 17e67bedac..17e67bedac 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/transactions.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/transactions.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/16/users.sql b/synapse/storage/databases/main/schema/full_schemas/16/users.sql
index f013aa8b18..f013aa8b18 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/16/users.sql +++ b/synapse/storage/databases/main/schema/full_schemas/16/users.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.postgres b/synapse/storage/databases/main/schema/full_schemas/54/full.sql.postgres
index 20c5af2eb7..20c5af2eb7 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.postgres +++ b/synapse/storage/databases/main/schema/full_schemas/54/full.sql.postgres
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.sqlite b/synapse/storage/databases/main/schema/full_schemas/54/full.sql.sqlite
index e28ec3fa45..e28ec3fa45 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/54/full.sql.sqlite +++ b/synapse/storage/databases/main/schema/full_schemas/54/full.sql.sqlite
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/54/stream_positions.sql b/synapse/storage/databases/main/schema/full_schemas/54/stream_positions.sql
index 91d21b2921..91d21b2921 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/54/stream_positions.sql +++ b/synapse/storage/databases/main/schema/full_schemas/54/stream_positions.sql
diff --git a/synapse/storage/data_stores/main/schema/full_schemas/README.md b/synapse/storage/databases/main/schema/full_schemas/README.md
index c00f287190..c00f287190 100644 --- a/synapse/storage/data_stores/main/schema/full_schemas/README.md +++ b/synapse/storage/databases/main/schema/full_schemas/README.md
diff --git a/synapse/storage/data_stores/main/search.py b/synapse/storage/databases/main/search.py
index d52228297c..e34fce6281 100644 --- a/synapse/storage/data_stores/main/search.py +++ b/synapse/storage/databases/main/search.py
@@ -16,13 +16,13 @@ import logging import re from collections import namedtuple - -from twisted.internet import defer +from typing import List, Optional, Set from synapse.api.errors import SynapseError +from synapse.events import EventBase from synapse.storage._base import SQLBaseStore, db_to_json, make_in_list_sql_clause -from synapse.storage.data_stores.main.events_worker import EventRedactBehaviour -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.events_worker import EventRedactBehaviour from synapse.storage.engines import PostgresEngine, Sqlite3Engine logger = logging.getLogger(__name__) @@ -88,16 +88,16 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): EVENT_SEARCH_USE_GIST_POSTGRES_NAME = "event_search_postgres_gist" EVENT_SEARCH_USE_GIN_POSTGRES_NAME = "event_search_postgres_gin" - def __init__(self, database: Database, db_conn, hs): - super(SearchBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) if not hs.config.enable_search: return - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_SEARCH_UPDATE_NAME, self._background_reindex_search ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_SEARCH_ORDER_UPDATE_NAME, self._background_reindex_search_order ) @@ -106,16 +106,15 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): # a GIN index. However, it's possible that some people might still have # the background update queued, so we register a handler to clear the # background update. - self.db.updates.register_noop_background_update( + self.db_pool.updates.register_noop_background_update( self.EVENT_SEARCH_USE_GIST_POSTGRES_NAME ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.EVENT_SEARCH_USE_GIN_POSTGRES_NAME, self._background_reindex_gin_search ) - @defer.inlineCallbacks - def _background_reindex_search(self, progress, batch_size): + async def _background_reindex_search(self, progress, batch_size): # we work through the events table from highest stream id to lowest target_min_stream_id = progress["target_min_stream_id_inclusive"] max_stream_id = progress["max_stream_id_exclusive"] @@ -140,7 +139,7 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): # store_search_entries_txn with a generator function, but that # would mean having two cursors open on the database at once. # Instead we just build a list of results. - rows = self.db.cursor_to_dict(txn) + rows = self.db_pool.cursor_to_dict(txn) if not rows: return 0 @@ -200,23 +199,24 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): "rows_inserted": rows_inserted + len(event_search_rows), } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.EVENT_SEARCH_UPDATE_NAME, progress ) return len(event_search_rows) - result = yield self.db.runInteraction( + result = await self.db_pool.runInteraction( self.EVENT_SEARCH_UPDATE_NAME, reindex_search_txn ) if not result: - yield self.db.updates._end_background_update(self.EVENT_SEARCH_UPDATE_NAME) + await self.db_pool.updates._end_background_update( + self.EVENT_SEARCH_UPDATE_NAME + ) return result - @defer.inlineCallbacks - def _background_reindex_gin_search(self, progress, batch_size): + async def _background_reindex_gin_search(self, progress, batch_size): """This handles old synapses which used GIST indexes, if any; converting them back to be GIN as per the actual schema. """ @@ -253,15 +253,14 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): conn.set_session(autocommit=False) if isinstance(self.database_engine, PostgresEngine): - yield self.db.runWithConnection(create_index) + await self.db_pool.runWithConnection(create_index) - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.EVENT_SEARCH_USE_GIN_POSTGRES_NAME ) return 1 - @defer.inlineCallbacks - def _background_reindex_search_order(self, progress, batch_size): + async def _background_reindex_search_order(self, progress, batch_size): target_min_stream_id = progress["target_min_stream_id_inclusive"] max_stream_id = progress["max_stream_id_exclusive"] rows_inserted = progress.get("rows_inserted", 0) @@ -286,14 +285,14 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): ) conn.set_session(autocommit=False) - yield self.db.runWithConnection(create_index) + await self.db_pool.runWithConnection(create_index) pg = dict(progress) pg["have_added_indexes"] = True - yield self.db.runInteraction( + await self.db_pool.runInteraction( self.EVENT_SEARCH_ORDER_UPDATE_NAME, - self.db.updates._background_update_progress_txn, + self.db_pool.updates._background_update_progress_txn, self.EVENT_SEARCH_ORDER_UPDATE_NAME, pg, ) @@ -323,18 +322,18 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): "have_added_indexes": True, } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.EVENT_SEARCH_ORDER_UPDATE_NAME, progress ) return len(rows), True - num_rows, finished = yield self.db.runInteraction( + num_rows, finished = await self.db_pool.runInteraction( self.EVENT_SEARCH_ORDER_UPDATE_NAME, reindex_search_txn ) if not finished: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.EVENT_SEARCH_ORDER_UPDATE_NAME ) @@ -342,11 +341,10 @@ class SearchBackgroundUpdateStore(SearchWorkerStore): class SearchStore(SearchBackgroundUpdateStore): - def __init__(self, database: Database, db_conn, hs): - super(SearchStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - @defer.inlineCallbacks - def search_msgs(self, room_ids, search_term, keys): + async def search_msgs(self, room_ids, search_term, keys): """Performs a full text search over events with given keys. Args: @@ -423,15 +421,15 @@ class SearchStore(SearchBackgroundUpdateStore): # entire table from the database. sql += " ORDER BY rank DESC LIMIT 500" - results = yield self.db.execute( - "search_msgs", self.db.cursor_to_dict, sql, *args + results = await self.db_pool.execute( + "search_msgs", self.db_pool.cursor_to_dict, sql, *args ) results = list(filter(lambda row: row["room_id"] in room_ids, results)) # We set redact_behaviour to BLOCK here to prevent redacted events being returned in # search results (which is a data leak) - events = yield self.get_events_as_list( + events = await self.get_events_as_list( [r["event_id"] for r in results], redact_behaviour=EventRedactBehaviour.BLOCK, ) @@ -440,12 +438,12 @@ class SearchStore(SearchBackgroundUpdateStore): highlights = None if isinstance(self.database_engine, PostgresEngine): - highlights = yield self._find_highlights_in_postgres(search_query, events) + highlights = await self._find_highlights_in_postgres(search_query, events) count_sql += " GROUP BY room_id" - count_results = yield self.db.execute( - "search_rooms_count", self.db.cursor_to_dict, count_sql, *count_args + count_results = await self.db_pool.execute( + "search_rooms_count", self.db_pool.cursor_to_dict, count_sql, *count_args ) count = sum(row["count"] for row in count_results if row["room_id"] in room_ids) @@ -460,19 +458,25 @@ class SearchStore(SearchBackgroundUpdateStore): "count": count, } - @defer.inlineCallbacks - def search_rooms(self, room_ids, search_term, keys, limit, pagination_token=None): + async def search_rooms( + self, + room_ids: List[str], + search_term: str, + keys: List[str], + limit, + pagination_token: Optional[str] = None, + ) -> List[dict]: """Performs a full text search over events with given keys. Args: - room_id (list): The room_ids to search in - search_term (str): Search term to search for - keys (list): List of keys to search in, currently supports - "content.body", "content.name", "content.topic" - pagination_token (str): A pagination token previously returned + room_ids: The room_ids to search in + search_term: Search term to search for + keys: List of keys to search in, currently supports "content.body", + "content.name", "content.topic" + pagination_token: A pagination token previously returned Returns: - list of dicts + Each match as a dictionary. """ clauses = [] @@ -575,15 +579,15 @@ class SearchStore(SearchBackgroundUpdateStore): args.append(limit) - results = yield self.db.execute( - "search_rooms", self.db.cursor_to_dict, sql, *args + results = await self.db_pool.execute( + "search_rooms", self.db_pool.cursor_to_dict, sql, *args ) results = list(filter(lambda row: row["room_id"] in room_ids, results)) # We set redact_behaviour to BLOCK here to prevent redacted events being returned in # search results (which is a data leak) - events = yield self.get_events_as_list( + events = await self.get_events_as_list( [r["event_id"] for r in results], redact_behaviour=EventRedactBehaviour.BLOCK, ) @@ -592,12 +596,12 @@ class SearchStore(SearchBackgroundUpdateStore): highlights = None if isinstance(self.database_engine, PostgresEngine): - highlights = yield self._find_highlights_in_postgres(search_query, events) + highlights = await self._find_highlights_in_postgres(search_query, events) count_sql += " GROUP BY room_id" - count_results = yield self.db.execute( - "search_rooms_count", self.db.cursor_to_dict, count_sql, *count_args + count_results = await self.db_pool.execute( + "search_rooms_count", self.db_pool.cursor_to_dict, count_sql, *count_args ) count = sum(row["count"] for row in count_results if row["room_id"] in room_ids) @@ -617,7 +621,9 @@ class SearchStore(SearchBackgroundUpdateStore): "count": count, } - def _find_highlights_in_postgres(self, search_query, events): + async def _find_highlights_in_postgres( + self, search_query: str, events: List[EventBase] + ) -> Set[str]: """Given a list of events and a search term, return a list of words that match from the content of the event. @@ -625,11 +631,11 @@ class SearchStore(SearchBackgroundUpdateStore): highlight the matching parts. Args: - search_query (str) - events (list): A list of events + search_query + events: A list of events Returns: - deferred : A set of strings. + A set of strings. """ def f(txn): @@ -682,7 +688,7 @@ class SearchStore(SearchBackgroundUpdateStore): return highlight_words - return self.db.runInteraction("_find_highlights", f) + return await self.db_pool.runInteraction("_find_highlights", f) def _to_postgres_options(options_dict): diff --git a/synapse/storage/data_stores/main/signatures.py b/synapse/storage/databases/main/signatures.py
index 36244d9f5d..c8c67953e4 100644 --- a/synapse/storage/data_stores/main/signatures.py +++ b/synapse/storage/databases/main/signatures.py
@@ -13,11 +13,12 @@ # See the License for the specific language governing permissions and # limitations under the License. -from unpaddedbase64 import encode_base64 +from typing import Dict, Iterable, List, Tuple -from twisted.internet import defer +from unpaddedbase64 import encode_base64 from synapse.storage._base import SQLBaseStore +from synapse.storage.types import Cursor from synapse.util.caches.descriptors import cached, cachedList @@ -31,18 +32,38 @@ class SignatureWorkerStore(SQLBaseStore): @cachedList( cached_method_name="get_event_reference_hash", list_name="event_ids", num_args=1 ) - def get_event_reference_hashes(self, event_ids): + async def get_event_reference_hashes( + self, event_ids: Iterable[str] + ) -> Dict[str, Dict[str, bytes]]: + """Get all hashes for given events. + + Args: + event_ids: The event IDs to get hashes for. + + Returns: + A mapping of event ID to a mapping of algorithm to hash. + """ + def f(txn): return { event_id: self._get_event_reference_hashes_txn(txn, event_id) for event_id in event_ids } - return self.db.runInteraction("get_event_reference_hashes", f) + return await self.db_pool.runInteraction("get_event_reference_hashes", f) + + async def add_event_hashes( + self, event_ids: Iterable[str] + ) -> List[Tuple[str, Dict[str, str]]]: + """ - @defer.inlineCallbacks - def add_event_hashes(self, event_ids): - hashes = yield self.get_event_reference_hashes(event_ids) + Args: + event_ids: The event IDs + + Returns: + A list of tuples of event ID and a mapping of algorithm to base-64 encoded hash. + """ + hashes = await self.get_event_reference_hashes(event_ids) hashes = { e_id: {k: encode_base64(v) for k, v in h.items() if k == "sha256"} for e_id, h in hashes.items() @@ -50,13 +71,15 @@ class SignatureWorkerStore(SQLBaseStore): return list(hashes.items()) - def _get_event_reference_hashes_txn(self, txn, event_id): + def _get_event_reference_hashes_txn( + self, txn: Cursor, event_id: str + ) -> Dict[str, bytes]: """Get all the hashes for a given PDU. Args: - txn (cursor): - event_id (str): Id for the Event. + txn: + event_id: Id for the Event. Returns: - A dict[unicode, bytes] of algorithm -> hash. + A mapping of algorithm -> hash. """ query = ( "SELECT algorithm, hash" diff --git a/synapse/storage/data_stores/main/state.py b/synapse/storage/databases/main/state.py
index bb38a04ede..3c1e33819b 100644 --- a/synapse/storage/data_stores/main/state.py +++ b/synapse/storage/databases/main/state.py
@@ -16,17 +16,18 @@ import collections.abc import logging from collections import namedtuple - -from twisted.internet import defer +from typing import Iterable, Optional, Set from synapse.api.constants import EventTypes, Membership from synapse.api.errors import NotFoundError, UnsupportedRoomVersionError from synapse.api.room_versions import KNOWN_ROOM_VERSIONS, RoomVersion +from synapse.events import EventBase from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.data_stores.main.roommember import RoomMemberWorkerStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.storage.databases.main.roommember import RoomMemberWorkerStore from synapse.storage.state import StateFilter +from synapse.types import StateMap from synapse.util.caches import intern_string from synapse.util.caches.descriptors import cached, cachedList @@ -54,8 +55,8 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): """The parts of StateGroupStore that can be called from workers. """ - def __init__(self, database: Database, db_conn, hs): - super(StateGroupWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) async def get_room_version(self, room_id: str) -> RoomVersion: """Get the room_version of a given room @@ -93,7 +94,7 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): # We really should have an entry in the rooms table for every room we # care about, but let's be a bit paranoid (at least while the background # update is happening) to avoid breaking existing rooms. - version = await self.db.simple_select_one_onecol( + version = await self.db_pool.simple_select_one_onecol( table="rooms", keyvalues={"room_id": room_id}, retcol="room_version", @@ -108,28 +109,27 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): create_event = await self.get_create_event_for_room(room_id) return create_event.content.get("room_version", "1") - @defer.inlineCallbacks - def get_room_predecessor(self, room_id): + async def get_room_predecessor(self, room_id: str) -> Optional[dict]: """Get the predecessor of an upgraded room if it exists. Otherwise return None. Args: - room_id (str) + room_id: The room ID. Returns: - Deferred[dict|None]: A dictionary containing the structure of the predecessor - field from the room's create event. The structure is subject to other servers, - but it is expected to be: - * room_id (str): The room ID of the predecessor room - * event_id (str): The ID of the tombstone event in the predecessor room + A dictionary containing the structure of the predecessor + field from the room's create event. The structure is subject to other servers, + but it is expected to be: + * room_id (str): The room ID of the predecessor room + * event_id (str): The ID of the tombstone event in the predecessor room - None if a predecessor key is not found, or is not a dictionary. + None if a predecessor key is not found, or is not a dictionary. Raises: NotFoundError if the given room is unknown """ # Retrieve the room's create event - create_event = yield self.get_create_event_for_room(room_id) + create_event = await self.get_create_event_for_room(room_id) # Retrieve the predecessor key of the create event predecessor = create_event.content.get("predecessor", None) @@ -140,20 +140,19 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): return predecessor - @defer.inlineCallbacks - def get_create_event_for_room(self, room_id): + async def get_create_event_for_room(self, room_id: str) -> EventBase: """Get the create state event for a room. Args: - room_id (str) + room_id: The room ID. Returns: - Deferred[EventBase]: The room creation event. + The room creation event. Raises: NotFoundError if the room is unknown """ - state_ids = yield self.get_current_state_ids(room_id) + state_ids = await self.get_current_state_ids(room_id) create_id = state_ids.get((EventTypes.Create, "")) # If we can't find the create event, assume we've hit a dead end @@ -161,19 +160,19 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): raise NotFoundError("Unknown room %s" % (room_id,)) # Retrieve the room's create event and return - create_event = yield self.get_event(create_id) + create_event = await self.get_event(create_id) return create_event @cached(max_entries=100000, iterable=True) - def get_current_state_ids(self, room_id): + async def get_current_state_ids(self, room_id: str) -> StateMap[str]: """Get the current state event ids for a room based on the current_state_events table. Args: - room_id (str) + room_id: The room to get the state IDs of. Returns: - deferred: dict of (type, state_key) -> event_id + The current state of the room. """ def _get_current_state_ids_txn(txn): @@ -186,14 +185,14 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): return {(intern_string(r[0]), intern_string(r[1])): r[2] for r in txn} - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_current_state_ids", _get_current_state_ids_txn ) # FIXME: how should this be cached? - def get_filtered_current_state_ids( + async def get_filtered_current_state_ids( self, room_id: str, state_filter: StateFilter = StateFilter.all() - ): + ) -> StateMap[str]: """Get the current state event of a given type for a room based on the current_state_events table. This may not be as up-to-date as the result of doing a fresh state resolution as per state_handler.get_current_state @@ -204,14 +203,14 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): from the database. Returns: - defer.Deferred[StateMap[str]]: Map from type/state_key to event ID. + Map from type/state_key to event ID. """ where_clause, where_args = state_filter.make_sql_filter_clause() if not where_clause: # We delegate to the cached version - return self.get_current_state_ids(room_id) + return await self.get_current_state_ids(room_id) def _get_filtered_current_state_ids_txn(txn): results = {} @@ -233,22 +232,21 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): return results - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_filtered_current_state_ids", _get_filtered_current_state_ids_txn ) - @defer.inlineCallbacks - def get_canonical_alias_for_room(self, room_id): + async def get_canonical_alias_for_room(self, room_id: str) -> Optional[str]: """Get canonical alias for room, if any Args: - room_id (str) + room_id: The room ID Returns: - Deferred[str|None]: The canonical alias, if any + The canonical alias, if any """ - state = yield self.get_filtered_current_state_ids( + state = await self.get_filtered_current_state_ids( room_id, StateFilter.from_types([(EventTypes.CanonicalAlias, "")]) ) @@ -256,15 +254,15 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): if not event_id: return - event = yield self.get_event(event_id, allow_none=True) + event = await self.get_event(event_id, allow_none=True) if not event: return return event.content.get("canonical_alias") @cached(max_entries=50000) - def _get_state_group_for_event(self, event_id): - return self.db.simple_select_one_onecol( + async def _get_state_group_for_event(self, event_id: str) -> Optional[int]: + return await self.db_pool.simple_select_one_onecol( table="event_to_state_groups", keyvalues={"event_id": event_id}, retcol="state_group", @@ -276,12 +274,11 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): cached_method_name="_get_state_group_for_event", list_name="event_ids", num_args=1, - inlineCallbacks=True, ) - def _get_state_group_for_events(self, event_ids): + async def _get_state_group_for_events(self, event_ids): """Returns mapping event_id -> state_group """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="event_to_state_groups", column="event_id", iterable=event_ids, @@ -292,19 +289,19 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore): return {row["event_id"]: row["state_group"] for row in rows} - @defer.inlineCallbacks - def get_referenced_state_groups(self, state_groups): + async def get_referenced_state_groups( + self, state_groups: Iterable[int] + ) -> Set[int]: """Check if the state groups are referenced by events. Args: - state_groups (Iterable[int]) + state_groups Returns: - Deferred[set[int]]: The subset of state groups that are - referenced. + The subset of state groups that are referenced. """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="event_to_state_groups", column="state_group", iterable=state_groups, @@ -322,25 +319,25 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): EVENT_STATE_GROUP_INDEX_UPDATE_NAME = "event_to_state_groups_sg_index" DELETE_CURRENT_STATE_UPDATE_NAME = "delete_old_current_state_events" - def __init__(self, database: Database, db_conn, hs): - super(MainStateBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.server_name = hs.hostname - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( self.CURRENT_STATE_INDEX_UPDATE_NAME, index_name="current_state_events_member_index", table="current_state_events", columns=["state_key"], where_clause="type='m.room.member'", ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( self.EVENT_STATE_GROUP_INDEX_UPDATE_NAME, index_name="event_to_state_groups_sg_index", table="event_to_state_groups", columns=["state_group"], ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.DELETE_CURRENT_STATE_UPDATE_NAME, self._background_remove_left_rooms, ) @@ -432,7 +429,7 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): # potentially stale, since there may have been a period where the # server didn't share a room with the remote user and therefore may # have missed any device updates. - rows = self.db.simple_select_many_txn( + rows = self.db_pool.simple_select_many_txn( txn, table="current_state_events", column="room_id", @@ -444,7 +441,7 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): potentially_left_users = {row["state_key"] for row in rows} # Now lets actually delete the rooms from the DB. - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="current_state_events", column="room_id", @@ -452,7 +449,7 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): keyvalues={}, ) - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="event_forward_extremities", column="room_id", @@ -460,7 +457,7 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): keyvalues={}, ) - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.DELETE_CURRENT_STATE_UPDATE_NAME, {"last_room_id": room_ids[-1]}, @@ -468,12 +465,12 @@ class MainStateBackgroundUpdateStore(RoomMemberWorkerStore): return False, potentially_left_users - finished, potentially_left_users = await self.db.runInteraction( + finished, potentially_left_users = await self.db_pool.runInteraction( "_background_remove_left_rooms", _background_remove_left_rooms_txn ) if finished: - await self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.DELETE_CURRENT_STATE_UPDATE_NAME ) @@ -508,5 +505,5 @@ class StateStore(StateGroupWorkerStore, MainStateBackgroundUpdateStore): * `state_groups_state`: Maps state group to state events. """ - def __init__(self, database: Database, db_conn, hs): - super(StateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) diff --git a/synapse/storage/data_stores/main/state_deltas.py b/synapse/storage/databases/main/state_deltas.py
index 725e12507f..356623fc6e 100644 --- a/synapse/storage/data_stores/main/state_deltas.py +++ b/synapse/storage/databases/main/state_deltas.py
@@ -14,8 +14,7 @@ # limitations under the License. import logging - -from twisted.internet import defer +from typing import Any, Dict, List, Tuple from synapse.storage._base import SQLBaseStore @@ -23,7 +22,9 @@ logger = logging.getLogger(__name__) class StateDeltasStore(SQLBaseStore): - def get_current_state_deltas(self, prev_stream_id: int, max_stream_id: int): + async def get_current_state_deltas( + self, prev_stream_id: int, max_stream_id: int + ) -> Tuple[int, List[Dict[str, Any]]]: """Fetch a list of room state changes since the given stream id Each entry in the result contains the following fields: @@ -37,12 +38,12 @@ class StateDeltasStore(SQLBaseStore): if it's new state. Args: - prev_stream_id (int): point to get changes since (exclusive) - max_stream_id (int): the point that we know has been correctly persisted + prev_stream_id: point to get changes since (exclusive) + max_stream_id: the point that we know has been correctly persisted - ie, an upper limit to return changes from. Returns: - Deferred[tuple[int, list[dict]]: A tuple consisting of: + A tuple consisting of: - the stream id which these results go up to - list of current_state_delta_stream rows. If it is empty, we are up to date. @@ -58,7 +59,7 @@ class StateDeltasStore(SQLBaseStore): # if the CSDs haven't changed between prev_stream_id and now, we # know for certain that they haven't changed between prev_stream_id and # max_stream_id. - return defer.succeed((max_stream_id, [])) + return (max_stream_id, []) def get_current_state_deltas_txn(txn): # First we calculate the max stream id that will give us less than @@ -100,22 +101,22 @@ class StateDeltasStore(SQLBaseStore): ORDER BY stream_id ASC """ txn.execute(sql, (prev_stream_id, clipped_stream_id)) - return clipped_stream_id, self.db.cursor_to_dict(txn) + return clipped_stream_id, self.db_pool.cursor_to_dict(txn) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_current_state_deltas", get_current_state_deltas_txn ) def _get_max_stream_id_in_current_state_deltas_txn(self, txn): - return self.db.simple_select_one_onecol_txn( + return self.db_pool.simple_select_one_onecol_txn( txn, table="current_state_delta_stream", keyvalues={}, retcol="COALESCE(MAX(stream_id), -1)", ) - def get_max_stream_id_in_current_state_deltas(self): - return self.db.runInteraction( + async def get_max_stream_id_in_current_state_deltas(self): + return await self.db_pool.runInteraction( "get_max_stream_id_in_current_state_deltas", self._get_max_stream_id_in_current_state_deltas_txn, ) diff --git a/synapse/storage/data_stores/main/stats.py b/synapse/storage/databases/main/stats.py
index 380c1ec7da..5beb302be3 100644 --- a/synapse/storage/data_stores/main/stats.py +++ b/synapse/storage/databases/main/stats.py
@@ -15,14 +15,15 @@ # limitations under the License. import logging +from collections import Counter from itertools import chain +from typing import Any, Dict, List, Optional, Tuple -from twisted.internet import defer from twisted.internet.defer import DeferredLock from synapse.api.constants import EventTypes, Membership -from synapse.storage.data_stores.main.state_deltas import StateDeltasStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.state_deltas import StateDeltasStore from synapse.storage.engines import PostgresEngine from synapse.util.caches.descriptors import cached @@ -59,8 +60,8 @@ TYPE_TO_ORIGIN_TABLE = {"room": ("rooms", "room_id"), "user": ("users", "name")} class StatsStore(StateDeltasStore): - def __init__(self, database: Database, db_conn, hs): - super(StatsStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.server_name = hs.hostname self.clock = self.hs.get_clock() @@ -69,17 +70,17 @@ class StatsStore(StateDeltasStore): self.stats_delta_processing_lock = DeferredLock() - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_stats_process_rooms", self._populate_stats_process_rooms ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_stats_process_users", self._populate_stats_process_users ) # we no longer need to perform clean-up, but we will give ourselves # the potential to reintroduce it in the future – so documentation # will still encourage the use of this no-op handler. - self.db.updates.register_noop_background_update("populate_stats_cleanup") - self.db.updates.register_noop_background_update("populate_stats_prepare") + self.db_pool.updates.register_noop_background_update("populate_stats_cleanup") + self.db_pool.updates.register_noop_background_update("populate_stats_prepare") def quantise_stats_time(self, ts): """ @@ -97,13 +98,14 @@ class StatsStore(StateDeltasStore): """ return (ts // self.stats_bucket_size) * self.stats_bucket_size - @defer.inlineCallbacks - def _populate_stats_process_users(self, progress, batch_size): + async def _populate_stats_process_users(self, progress, batch_size): """ This is a background update which regenerates statistics for users. """ if not self.stats_enabled: - yield self.db.updates._end_background_update("populate_stats_process_users") + await self.db_pool.updates._end_background_update( + "populate_stats_process_users" + ) return 1 last_user_id = progress.get("last_user_id", "") @@ -118,35 +120,36 @@ class StatsStore(StateDeltasStore): txn.execute(sql, (last_user_id, batch_size)) return [r for r, in txn] - users_to_work_on = yield self.db.runInteraction( + users_to_work_on = await self.db_pool.runInteraction( "_populate_stats_process_users", _get_next_batch ) # No more rooms -- complete the transaction. if not users_to_work_on: - yield self.db.updates._end_background_update("populate_stats_process_users") + await self.db_pool.updates._end_background_update( + "populate_stats_process_users" + ) return 1 for user_id in users_to_work_on: - yield self._calculate_and_set_initial_state_for_user(user_id) + await self._calculate_and_set_initial_state_for_user(user_id) progress["last_user_id"] = user_id - yield self.db.runInteraction( + await self.db_pool.runInteraction( "populate_stats_process_users", - self.db.updates._background_update_progress_txn, + self.db_pool.updates._background_update_progress_txn, "populate_stats_process_users", progress, ) return len(users_to_work_on) - @defer.inlineCallbacks - def _populate_stats_process_rooms(self, progress, batch_size): - """ - This is a background update which regenerates statistics for rooms. - """ + async def _populate_stats_process_rooms(self, progress, batch_size): + """This is a background update which regenerates statistics for rooms.""" if not self.stats_enabled: - yield self.db.updates._end_background_update("populate_stats_process_rooms") + await self.db_pool.updates._end_background_update( + "populate_stats_process_rooms" + ) return 1 last_room_id = progress.get("last_room_id", "") @@ -161,48 +164,69 @@ class StatsStore(StateDeltasStore): txn.execute(sql, (last_room_id, batch_size)) return [r for r, in txn] - rooms_to_work_on = yield self.db.runInteraction( + rooms_to_work_on = await self.db_pool.runInteraction( "populate_stats_rooms_get_batch", _get_next_batch ) # No more rooms -- complete the transaction. if not rooms_to_work_on: - yield self.db.updates._end_background_update("populate_stats_process_rooms") + await self.db_pool.updates._end_background_update( + "populate_stats_process_rooms" + ) return 1 for room_id in rooms_to_work_on: - yield self._calculate_and_set_initial_state_for_room(room_id) + await self._calculate_and_set_initial_state_for_room(room_id) progress["last_room_id"] = room_id - yield self.db.runInteraction( + await self.db_pool.runInteraction( "_populate_stats_process_rooms", - self.db.updates._background_update_progress_txn, + self.db_pool.updates._background_update_progress_txn, "populate_stats_process_rooms", progress, ) return len(rooms_to_work_on) - def get_stats_positions(self): + async def get_stats_positions(self) -> int: """ Returns the stats processor positions. """ - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="stats_incremental_position", keyvalues={}, retcol="stream_id", desc="stats_incremental_position", ) - def update_room_state(self, room_id, fields): - """ + async def update_room_state(self, room_id: str, fields: Dict[str, Any]) -> None: + """Update the state of a room. + + fields can contain the following keys with string values: + * join_rules + * history_visibility + * encryption + * name + * topic + * avatar + * canonical_alias + * guest_access + + A is_federatable key can also be included with a boolean value. + Args: - room_id (str) - fields (dict[str:Any]) + room_id: The room ID to update the state of. + fields: The fields to update. This can include a partial list of the + above fields to only update some room information. """ - - # For whatever reason some of the fields may contain null bytes, which - # postgres isn't a fan of, so we replace those fields with null. + # Ensure that the values to update are valid, they should be strings and + # not contain any null bytes. + # + # Invalid data gets overwritten with null. + # + # Note that a missing value should not be overwritten (it keeps the + # previous value). + sentinel = object() for col in ( "join_rules", "history_visibility", @@ -211,33 +235,36 @@ class StatsStore(StateDeltasStore): "topic", "avatar", "canonical_alias", + "guest_access", ): - field = fields.get(col) - if field and "\0" in field: + field = fields.get(col, sentinel) + if field is not sentinel and (not isinstance(field, str) or "\0" in field): fields[col] = None - return self.db.simple_upsert( + await self.db_pool.simple_upsert( table="room_stats_state", keyvalues={"room_id": room_id}, values=fields, desc="update_room_state", ) - def get_statistics_for_subject(self, stats_type, stats_id, start, size=100): + async def get_statistics_for_subject( + self, stats_type: str, stats_id: str, start: str, size: int = 100 + ) -> List[dict]: """ Get statistics for a given subject. Args: - stats_type (str): The type of subject - stats_id (str): The ID of the subject (e.g. room_id or user_id) - start (int): Pagination start. Number of entries, not timestamp. - size (int): How many entries to return. + stats_type: The type of subject + stats_id: The ID of the subject (e.g. room_id or user_id) + start: Pagination start. Number of entries, not timestamp. + size: How many entries to return. Returns: - Deferred[list[dict]], where the dict has the keys of + A list of dicts, where the dict has the keys of ABSOLUTE_STATS_FIELDS[stats_type], and "bucket_size" and "end_ts". """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_statistics_for_subject", self._get_statistics_for_subject_txn, stats_type, @@ -258,7 +285,7 @@ class StatsStore(StateDeltasStore): ABSOLUTE_STATS_FIELDS[stats_type] + PER_SLICE_FIELDS[stats_type] ) - slice_list = self.db.simple_select_list_paginate_txn( + slice_list = self.db_pool.simple_select_list_paginate_txn( txn, table + "_historical", "end_ts", @@ -272,7 +299,7 @@ class StatsStore(StateDeltasStore): return slice_list @cached() - def get_earliest_token_for_stats(self, stats_type, id): + async def get_earliest_token_for_stats(self, stats_type: str, id: str) -> int: """ Fetch the "earliest token". This is used by the room stats delta processor to ignore deltas that have been processed between the @@ -280,29 +307,28 @@ class StatsStore(StateDeltasStore): being calculated. Returns: - Deferred[int] + The earliest token. """ table, id_col = TYPE_TO_TABLE[stats_type] - return self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( "%s_current" % (table,), keyvalues={id_col: id}, retcol="completed_delta_stream_id", allow_none=True, ) - def bulk_update_stats_delta(self, ts, updates, stream_id): + async def bulk_update_stats_delta( + self, ts: int, updates: Dict[str, Dict[str, Dict[str, Counter]]], stream_id: int + ) -> None: """Bulk update stats tables for a given stream_id and updates the stats incremental position. Args: - ts (int): Current timestamp in ms - updates(dict[str, dict[str, dict[str, Counter]]]): The updates to - commit as a mapping stats_type -> stats_id -> field -> delta. - stream_id (int): Current position. - - Returns: - Deferred + ts: Current timestamp in ms + updates: The updates to commit as a mapping of + stats_type -> stats_id -> field -> delta. + stream_id: Current position. """ def _bulk_update_stats_delta_txn(txn): @@ -320,45 +346,44 @@ class StatsStore(StateDeltasStore): complete_with_stream_id=stream_id, ) - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="stats_incremental_position", keyvalues={}, updatevalues={"stream_id": stream_id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "bulk_update_stats_delta", _bulk_update_stats_delta_txn ) - def update_stats_delta( + async def update_stats_delta( self, - ts, - stats_type, - stats_id, - fields, - complete_with_stream_id, - absolute_field_overrides=None, - ): + ts: int, + stats_type: str, + stats_id: str, + fields: Dict[str, int], + complete_with_stream_id: Optional[int], + absolute_field_overrides: Optional[Dict[str, int]] = None, + ) -> None: """ Updates the statistics for a subject, with a delta (difference/relative change). Args: - ts (int): timestamp of the change - stats_type (str): "room" or "user" – the kind of subject - stats_id (str): the subject's ID (room ID or user ID) - fields (dict[str, int]): Deltas of stats values. - complete_with_stream_id (int, optional): + ts: timestamp of the change + stats_type: "room" or "user" – the kind of subject + stats_id: the subject's ID (room ID or user ID) + fields: Deltas of stats values. + complete_with_stream_id: If supplied, converts an incomplete row into a complete row, with the supplied stream_id marked as the stream_id where the row was completed. - absolute_field_overrides (dict[str, int]): Current stats values - (i.e. not deltas) of absolute fields. - Does not work with per-slice fields. + absolute_field_overrides: Current stats values (i.e. not deltas) of + absolute fields. Does not work with per-slice fields. """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "update_stats_delta", self._update_stats_delta_txn, ts, @@ -493,17 +518,17 @@ class StatsStore(StateDeltasStore): else: self.database_engine.lock_table(txn, table) retcols = list(chain(absolutes.keys(), additive_relatives.keys())) - current_row = self.db.simple_select_one_txn( + current_row = self.db_pool.simple_select_one_txn( txn, table, keyvalues, retcols, allow_none=True ) if current_row is None: merged_dict = {**keyvalues, **absolutes, **additive_relatives} - self.db.simple_insert_txn(txn, table, merged_dict) + self.db_pool.simple_insert_txn(txn, table, merged_dict) else: for (key, val) in additive_relatives.items(): current_row[key] += val current_row.update(absolutes) - self.db.simple_update_one_txn(txn, table, keyvalues, current_row) + self.db_pool.simple_update_one_txn(txn, table, keyvalues, current_row) def _upsert_copy_from_table_with_additive_relatives_txn( self, @@ -590,11 +615,11 @@ class StatsStore(StateDeltasStore): txn.execute(sql, qargs) else: self.database_engine.lock_table(txn, into_table) - src_row = self.db.simple_select_one_txn( + src_row = self.db_pool.simple_select_one_txn( txn, src_table, keyvalues, copy_columns ) all_dest_keyvalues = {**keyvalues, **extra_dst_keyvalues} - dest_current_row = self.db.simple_select_one_txn( + dest_current_row = self.db_pool.simple_select_one_txn( txn, into_table, keyvalues=all_dest_keyvalues, @@ -610,25 +635,28 @@ class StatsStore(StateDeltasStore): **src_row, **additive_relatives, } - self.db.simple_insert_txn(txn, into_table, merged_dict) + self.db_pool.simple_insert_txn(txn, into_table, merged_dict) else: for (key, val) in additive_relatives.items(): src_row[key] = dest_current_row[key] + val - self.db.simple_update_txn(txn, into_table, all_dest_keyvalues, src_row) + self.db_pool.simple_update_txn( + txn, into_table, all_dest_keyvalues, src_row + ) - def get_changes_room_total_events_and_bytes(self, min_pos, max_pos): + async def get_changes_room_total_events_and_bytes( + self, min_pos: int, max_pos: int + ) -> Dict[str, Dict[str, int]]: """Fetches the counts of events in the given range of stream IDs. Args: - min_pos (int) - max_pos (int) + min_pos + max_pos Returns: - Deferred[dict[str, dict[str, int]]]: Mapping of room ID to field - changes. + Mapping of room ID to field changes. """ - return self.db.runInteraction( + return await self.db_pool.runInteraction( "stats_incremental_total_events_and_bytes", self.get_changes_room_total_events_and_bytes_txn, min_pos, @@ -696,22 +724,22 @@ class StatsStore(StateDeltasStore): return room_deltas, user_deltas - @defer.inlineCallbacks - def _calculate_and_set_initial_state_for_room(self, room_id): + async def _calculate_and_set_initial_state_for_room( + self, room_id: str + ) -> Tuple[dict, dict, int]: """Calculate and insert an entry into room_stats_current. Args: - room_id (str) + room_id: The room ID under calculation. Returns: - Deferred[tuple[dict, dict, int]]: A tuple of room state, membership - counts and stream position. + A tuple of room state, membership counts and stream position. """ def _fetch_current_state_stats(txn): pos = self.get_room_max_stream_ordering() - rows = self.db.simple_select_many_txn( + rows = self.db_pool.simple_select_many_txn( txn, table="current_state_events", column="type", @@ -767,11 +795,11 @@ class StatsStore(StateDeltasStore): current_state_events_count, users_in_room, pos, - ) = yield self.db.runInteraction( + ) = await self.db_pool.runInteraction( "get_initial_state_for_room", _fetch_current_state_stats ) - state_event_map = yield self.get_events(event_ids, get_prev_content=False) + state_event_map = await self.get_events(event_ids, get_prev_content=False) room_state = { "join_rules": None, @@ -806,11 +834,11 @@ class StatsStore(StateDeltasStore): event.content.get("m.federate", True) is True ) - yield self.update_room_state(room_id, room_state) + await self.update_room_state(room_id, room_state) local_users_in_room = [u for u in users_in_room if self.hs.is_mine_id(u)] - yield self.update_stats_delta( + await self.update_stats_delta( ts=self.clock.time_msec(), stats_type="room", stats_id=room_id, @@ -826,8 +854,7 @@ class StatsStore(StateDeltasStore): }, ) - @defer.inlineCallbacks - def _calculate_and_set_initial_state_for_user(self, user_id): + async def _calculate_and_set_initial_state_for_user(self, user_id): def _calculate_and_set_initial_state_for_user_txn(txn): pos = self._get_max_stream_id_in_current_state_deltas_txn(txn) @@ -842,12 +869,12 @@ class StatsStore(StateDeltasStore): (count,) = txn.fetchone() return count, pos - joined_rooms, pos = yield self.db.runInteraction( + joined_rooms, pos = await self.db_pool.runInteraction( "calculate_and_set_initial_state_for_user", _calculate_and_set_initial_state_for_user_txn, ) - yield self.update_stats_delta( + await self.update_stats_delta( ts=self.clock.time_msec(), stats_type="user", stats_id=user_id, diff --git a/synapse/storage/data_stores/main/stream.py b/synapse/storage/databases/main/stream.py
index 10d39b3699..37249f1e3f 100644 --- a/synapse/storage/data_stores/main/stream.py +++ b/synapse/storage/databases/main/stream.py
@@ -35,21 +35,30 @@ what sort order was used: - topological tokems: "t%d-%d", where the integers map to the topological and stream ordering columns respectively. """ - import abc import logging from collections import namedtuple +from typing import TYPE_CHECKING, Dict, List, Optional, Set, Tuple from twisted.internet import defer +from synapse.api.filtering import Filter +from synapse.events import EventBase from synapse.logging.context import make_deferred_yieldable, run_in_background from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.events_worker import EventsWorkerStore -from synapse.storage.database import Database, make_in_list_sql_clause -from synapse.storage.engines import PostgresEngine -from synapse.types import RoomStreamToken +from synapse.storage.database import ( + DatabasePool, + LoggingTransaction, + make_in_list_sql_clause, +) +from synapse.storage.databases.main.events_worker import EventsWorkerStore +from synapse.storage.engines import BaseDatabaseEngine, PostgresEngine +from synapse.types import Collection, PersistedEventPosition, RoomStreamToken from synapse.util.caches.stream_change_cache import StreamChangeCache +if TYPE_CHECKING: + from synapse.server import HomeServer + logger = logging.getLogger(__name__) @@ -67,8 +76,12 @@ _EventDictReturn = namedtuple( def generate_pagination_where_clause( - direction, column_names, from_token, to_token, engine -): + direction: str, + column_names: Tuple[str, str], + from_token: Optional[Tuple[Optional[int], int]], + to_token: Optional[Tuple[Optional[int], int]], + engine: BaseDatabaseEngine, +) -> str: """Creates an SQL expression to bound the columns by the pagination tokens. @@ -89,21 +102,19 @@ def generate_pagination_where_clause( token, but include those that match the to token. Args: - direction (str): Whether we're paginating backwards("b") or - forwards ("f"). - column_names (tuple[str, str]): The column names to bound. Must *not* - be user defined as these get inserted directly into the SQL - statement without escapes. - from_token (tuple[int, int]|None): The start point for the pagination. - This is an exclusive minimum bound if direction is "f", and an - inclusive maximum bound if direction is "b". - to_token (tuple[int, int]|None): The endpoint point for the pagination. - This is an inclusive maximum bound if direction is "f", and an - exclusive minimum bound if direction is "b". + direction: Whether we're paginating backwards("b") or forwards ("f"). + column_names: The column names to bound. Must *not* be user defined as + these get inserted directly into the SQL statement without escapes. + from_token: The start point for the pagination. This is an exclusive + minimum bound if direction is "f", and an inclusive maximum bound if + direction is "b". + to_token: The endpoint point for the pagination. This is an inclusive + maximum bound if direction is "f", and an exclusive minimum bound if + direction is "b". engine: The database engine to generate the clauses for Returns: - str: The sql expression + The sql expression """ assert direction in ("b", "f") @@ -131,7 +142,12 @@ def generate_pagination_where_clause( return " AND ".join(where_clause) -def _make_generic_sql_bound(bound, column_names, values, engine): +def _make_generic_sql_bound( + bound: str, + column_names: Tuple[str, str], + values: Tuple[Optional[int], int], + engine: BaseDatabaseEngine, +) -> str: """Create an SQL expression that bounds the given column names by the values, e.g. create the equivalent of `(1, 2) < (col1, col2)`. @@ -141,18 +157,18 @@ def _make_generic_sql_bound(bound, column_names, values, engine): out manually. Args: - bound (str): The comparison operator to use. One of ">", "<", ">=", + bound: The comparison operator to use. One of ">", "<", ">=", "<=", where the values are on the left and columns on the right. - names (tuple[str, str]): The column names. Must *not* be user defined + names: The column names. Must *not* be user defined as these get inserted directly into the SQL statement without escapes. - values (tuple[int|None, int]): The values to bound the columns by. If + values: The values to bound the columns by. If the first value is None then only creates a bound on the second column. engine: The database engine to generate the SQL for Returns: - str + The SQL statement """ assert bound in (">", "<", ">=", "<=") @@ -192,7 +208,7 @@ def _make_generic_sql_bound(bound, column_names, values, engine): ) -def filter_to_clause(event_filter): +def filter_to_clause(event_filter: Optional[Filter]) -> Tuple[str, List[str]]: # NB: This may create SQL clauses that don't optimise well (and we don't # have indices on all possible clauses). E.g. it may create # "room_id == X AND room_id != X", which postgres doesn't optimise. @@ -242,16 +258,14 @@ def filter_to_clause(event_filter): return " AND ".join(clauses), args -class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): +class StreamWorkerStore(EventsWorkerStore, SQLBaseStore, metaclass=abc.ABCMeta): """This is an abstract base class where subclasses must implement `get_room_max_stream_ordering` and `get_room_min_stream_ordering` which can be called in the initializer. """ - __metaclass__ = abc.ABCMeta - - def __init__(self, database: Database, db_conn, hs): - super(StreamWorkerStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs: "HomeServer"): + super().__init__(database, db_conn, hs) self._instance_name = hs.get_instance_name() self._send_federation = hs.should_send_federation() @@ -264,7 +278,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): self._need_to_reset_federation_stream_positions = self._send_federation events_max = self.get_room_max_stream_ordering() - event_cache_prefill, min_event_val = self.db.get_cache_dict( + event_cache_prefill, min_event_val = self.db_pool.get_cache_dict( db_conn, "events", entity_column="room_id", @@ -283,40 +297,44 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): self._stream_order_on_start = self.get_room_max_stream_ordering() @abc.abstractmethod - def get_room_max_stream_ordering(self): + def get_room_max_stream_ordering(self) -> int: raise NotImplementedError() @abc.abstractmethod - def get_room_min_stream_ordering(self): + def get_room_min_stream_ordering(self) -> int: raise NotImplementedError() - @defer.inlineCallbacks - def get_room_events_stream_for_rooms( - self, room_ids, from_key, to_key, limit=0, order="DESC" - ): + def get_room_max_token(self) -> RoomStreamToken: + return RoomStreamToken(None, self.get_room_max_stream_ordering()) + + async def get_room_events_stream_for_rooms( + self, + room_ids: Collection[str], + from_key: RoomStreamToken, + to_key: RoomStreamToken, + limit: int = 0, + order: str = "DESC", + ) -> Dict[str, Tuple[List[EventBase], RoomStreamToken]]: """Get new room events in stream ordering since `from_key`. Args: - room_id (str) - from_key (str): Token from which no events are returned before - to_key (str): Token from which no events are returned after. (This + room_ids + from_key: Token from which no events are returned before + to_key: Token from which no events are returned after. (This is typically the current stream token) - limit (int): Maximum number of events to return - order (str): Either "DESC" or "ASC". Determines which events are + limit: Maximum number of events to return + order: Either "DESC" or "ASC". Determines which events are returned when the result is limited. If "DESC" then the most recent `limit` events are returned, otherwise returns the oldest `limit` events. Returns: - Deferred[dict[str,tuple[list[FrozenEvent], str]]] - A map from room id to a tuple containing: - - list of recent events in the room - - stream ordering key for the start of the chunk of events returned. + A map from room id to a tuple containing: + - list of recent events in the room + - stream ordering key for the start of the chunk of events returned. """ - from_id = RoomStreamToken.parse_stream_token(from_key).stream - - room_ids = yield self._events_stream_cache.get_entities_changed( - room_ids, from_id + room_ids = self._events_stream_cache.get_entities_changed( + room_ids, from_key.stream ) if not room_ids: @@ -325,7 +343,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): results = {} room_ids = list(room_ids) for rm_ids in (room_ids[i : i + 20] for i in range(0, len(room_ids), 20)): - res = yield make_deferred_yieldable( + res = await make_deferred_yieldable( defer.gatherResults( [ run_in_background( @@ -345,53 +363,51 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return results - def get_rooms_that_changed(self, room_ids, from_key): + def get_rooms_that_changed( + self, room_ids: Collection[str], from_key: RoomStreamToken + ) -> Set[str]: """Given a list of rooms and a token, return rooms where there may have been changes. - - Args: - room_ids (list) - from_key (str): The room_key portion of a StreamToken """ - from_key = RoomStreamToken.parse_stream_token(from_key).stream + from_id = from_key.stream return { room_id for room_id in room_ids - if self._events_stream_cache.has_entity_changed(room_id, from_key) + if self._events_stream_cache.has_entity_changed(room_id, from_id) } - @defer.inlineCallbacks - def get_room_events_stream_for_room( - self, room_id, from_key, to_key, limit=0, order="DESC" - ): - + async def get_room_events_stream_for_room( + self, + room_id: str, + from_key: RoomStreamToken, + to_key: RoomStreamToken, + limit: int = 0, + order: str = "DESC", + ) -> Tuple[List[EventBase], RoomStreamToken]: """Get new room events in stream ordering since `from_key`. Args: - room_id (str) - from_key (str): Token from which no events are returned before - to_key (str): Token from which no events are returned after. (This + room_id + from_key: Token from which no events are returned before + to_key: Token from which no events are returned after. (This is typically the current stream token) - limit (int): Maximum number of events to return - order (str): Either "DESC" or "ASC". Determines which events are + limit: Maximum number of events to return + order: Either "DESC" or "ASC". Determines which events are returned when the result is limited. If "DESC" then the most recent `limit` events are returned, otherwise returns the oldest `limit` events. Returns: - Deferred[tuple[list[FrozenEvent], str]]: Returns the list of - events (in ascending order) and the token from the start of - the chunk of events returned. + The list of events (in ascending order) and the token from the start + of the chunk of events returned. """ if from_key == to_key: return [], from_key - from_id = RoomStreamToken.parse_stream_token(from_key).stream - to_id = RoomStreamToken.parse_stream_token(to_key).stream + from_id = from_key.stream + to_id = to_key.stream - has_changed = yield self._events_stream_cache.has_entity_changed( - room_id, from_id - ) + has_changed = self._events_stream_cache.has_entity_changed(room_id, from_id) if not has_changed: return [], from_key @@ -409,9 +425,9 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): rows = [_EventDictReturn(row[0], None, row[1]) for row in txn] return rows - rows = yield self.db.runInteraction("get_room_events_stream_for_room", f) + rows = await self.db_pool.runInteraction("get_room_events_stream_for_room", f) - ret = yield self.get_events_as_list( + ret = await self.get_events_as_list( [r.event_id for r in rows], get_prev_content=True ) @@ -421,7 +437,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): ret.reverse() if rows: - key = "s%d" % min(r.stream_ordering for r in rows) + key = RoomStreamToken(None, min(r.stream_ordering for r in rows)) else: # Assume we didn't get anything because there was nothing to # get. @@ -429,10 +445,11 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return ret, key - @defer.inlineCallbacks - def get_membership_changes_for_user(self, user_id, from_key, to_key): - from_id = RoomStreamToken.parse_stream_token(from_key).stream - to_id = RoomStreamToken.parse_stream_token(to_key).stream + async def get_membership_changes_for_user( + self, user_id: str, from_key: RoomStreamToken, to_key: RoomStreamToken + ) -> List[EventBase]: + from_id = from_key.stream + to_id = to_key.stream if from_key == to_key: return [] @@ -459,9 +476,9 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return rows - rows = yield self.db.runInteraction("get_membership_changes_for_user", f) + rows = await self.db_pool.runInteraction("get_membership_changes_for_user", f) - ret = yield self.get_events_as_list( + ret = await self.get_events_as_list( [r.event_id for r in rows], get_prev_content=True ) @@ -469,27 +486,26 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return ret - @defer.inlineCallbacks - def get_recent_events_for_room(self, room_id, limit, end_token): + async def get_recent_events_for_room( + self, room_id: str, limit: int, end_token: RoomStreamToken + ) -> Tuple[List[EventBase], RoomStreamToken]: """Get the most recent events in the room in topological ordering. Args: - room_id (str) - limit (int) - end_token (str): The stream token representing now. + room_id + limit + end_token: The stream token representing now. Returns: - Deferred[tuple[list[FrozenEvent], str]]: Returns a list of - events and a token pointing to the start of the returned - events. - The events returned are in ascending order. + A list of events and a token pointing to the start of the returned + events. The events returned are in ascending order. """ - rows, token = yield self.get_recent_event_ids_for_room( + rows, token = await self.get_recent_event_ids_for_room( room_id, limit, end_token ) - events = yield self.get_events_as_list( + events = await self.get_events_as_list( [r.event_id for r in rows], get_prev_content=True ) @@ -497,28 +513,25 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return (events, token) - @defer.inlineCallbacks - def get_recent_event_ids_for_room(self, room_id, limit, end_token): + async def get_recent_event_ids_for_room( + self, room_id: str, limit: int, end_token: RoomStreamToken + ) -> Tuple[List[_EventDictReturn], RoomStreamToken]: """Get the most recent events in the room in topological ordering. Args: - room_id (str) - limit (int) - end_token (str): The stream token representing now. + room_id + limit + end_token: The stream token representing now. Returns: - Deferred[tuple[list[_EventDictReturn], str]]: Returns a list of - _EventDictReturn and a token pointing to the start of the returned - events. - The events returned are in ascending order. + A list of _EventDictReturn and a token pointing to the start of the + returned events. The events returned are in ascending order. """ # Allow a zero limit here, and no-op. if limit == 0: return [], end_token - end_token = RoomStreamToken.parse(end_token) - - rows, token = yield self.db.runInteraction( + rows, token = await self.db_pool.runInteraction( "get_recent_event_ids_for_room", self._paginate_room_events_txn, room_id, @@ -531,16 +544,17 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return rows, token - def get_room_event_before_stream_ordering(self, room_id, stream_ordering): + async def get_room_event_before_stream_ordering( + self, room_id: str, stream_ordering: int + ) -> Tuple[int, int, str]: """Gets details of the first event in a room at or before a stream ordering Args: - room_id (str): - stream_ordering (int): + room_id: + stream_ordering: Returns: - Deferred[(int, int, str)]: - (stream ordering, topological ordering, event_id) + A tuple of (stream ordering, topological ordering, event_id) """ def _f(txn): @@ -555,76 +569,99 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): txn.execute(sql, (room_id, stream_ordering)) return txn.fetchone() - return self.db.runInteraction("get_room_event_before_stream_ordering", _f) + return await self.db_pool.runInteraction( + "get_room_event_before_stream_ordering", _f + ) - @defer.inlineCallbacks - def get_room_events_max_id(self, room_id=None): + async def get_room_events_max_id(self, room_id: Optional[str] = None) -> str: """Returns the current token for rooms stream. By default, it returns the current global stream token. Specifying a `room_id` causes it to return the current room specific topological token. """ - token = yield self.get_room_max_stream_ordering() + token = self.get_room_max_stream_ordering() if room_id is None: return "s%d" % (token,) else: - topo = yield self.db.runInteraction( + topo = await self.db_pool.runInteraction( "_get_max_topological_txn", self._get_max_topological_txn, room_id ) return "t%d-%d" % (topo, token) - def get_stream_token_for_event(self, event_id): - """The stream token for an event + async def get_stream_id_for_event(self, event_id: str) -> int: + """The stream ID for an event Args: - event_id(str): The id of the event to look up a stream token for. + event_id: The id of the event to look up a stream token for. Raises: StoreError if the event wasn't in the database. Returns: - A deferred "s%d" stream token. + A stream ID. """ - return self.db.simple_select_one_onecol( - table="events", keyvalues={"event_id": event_id}, retcol="stream_ordering" - ).addCallback(lambda row: "s%d" % (row,)) + return await self.db_pool.runInteraction( + "get_stream_id_for_event", self.get_stream_id_for_event_txn, event_id, + ) + + def get_stream_id_for_event_txn( + self, txn: LoggingTransaction, event_id: str, allow_none=False, + ) -> int: + return self.db_pool.simple_select_one_onecol_txn( + txn=txn, + table="events", + keyvalues={"event_id": event_id}, + retcol="stream_ordering", + allow_none=allow_none, + ) - def get_topological_token_for_event(self, event_id): + async def get_position_for_event(self, event_id: str) -> PersistedEventPosition: + """Get the persisted position for an event + """ + row = await self.db_pool.simple_select_one( + table="events", + keyvalues={"event_id": event_id}, + retcols=("stream_ordering", "instance_name"), + desc="get_position_for_event", + ) + + return PersistedEventPosition( + row["instance_name"] or "master", row["stream_ordering"] + ) + + async def get_topological_token_for_event(self, event_id: str) -> RoomStreamToken: """The stream token for an event Args: - event_id(str): The id of the event to look up a stream token for. + event_id: The id of the event to look up a stream token for. Raises: StoreError if the event wasn't in the database. Returns: - A deferred "t%d-%d" topological token. + A `RoomStreamToken` topological token. """ - return self.db.simple_select_one( + row = await self.db_pool.simple_select_one( table="events", keyvalues={"event_id": event_id}, retcols=("stream_ordering", "topological_ordering"), desc="get_topological_token_for_event", - ).addCallback( - lambda row: "t%d-%d" % (row["topological_ordering"], row["stream_ordering"]) ) + return RoomStreamToken(row["topological_ordering"], row["stream_ordering"]) - def get_max_topological_token(self, room_id, stream_key): - """Get the max topological token in a room before the given stream + async def get_current_topological_token(self, room_id: str, stream_key: int) -> int: + """Gets the topological token in a room after or at the given stream ordering. Args: - room_id (str) - stream_key (int) - - Returns: - Deferred[int] + room_id + stream_key """ sql = ( - "SELECT coalesce(max(topological_ordering), 0) FROM events" - " WHERE room_id = ? AND stream_ordering < ?" + "SELECT coalesce(MIN(topological_ordering), 0) FROM events" + " WHERE room_id = ? AND stream_ordering >= ?" + ) + row = await self.db_pool.execute( + "get_current_topological_token", None, sql, room_id, stream_key ) - return self.db.execute( - "get_max_topological_token", None, sql, room_id, stream_key - ).addCallback(lambda r: r[0][0] if r else 0) + return row[0][0] if row else 0 - def _get_max_topological_txn(self, txn, room_id): + def _get_max_topological_txn(self, txn: LoggingTransaction, room_id: str) -> int: txn.execute( "SELECT MAX(topological_ordering) FROM events WHERE room_id = ?", (room_id,), @@ -634,16 +671,18 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return rows[0][0] if rows else 0 @staticmethod - def _set_before_and_after(events, rows, topo_order=True): + def _set_before_and_after( + events: List[EventBase], rows: List[_EventDictReturn], topo_order: bool = True + ): """Inserts ordering information to events' internal metadata from the DB rows. Args: - events (list[FrozenEvent]) - rows (list[_EventDictReturn]) - topo_order (bool): Whether the events were ordered topologically - or by stream ordering. If true then all rows should have a non - null topological_ordering. + events + rows + topo_order: Whether the events were ordered topologically or by stream + ordering. If true then all rows should have a non null + topological_ordering. """ for event, row in zip(events, rows): stream = row.stream_ordering @@ -652,29 +691,23 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): else: topo = None internal = event.internal_metadata - internal.before = str(RoomStreamToken(topo, stream - 1)) - internal.after = str(RoomStreamToken(topo, stream)) + internal.before = RoomStreamToken(topo, stream - 1) + internal.after = RoomStreamToken(topo, stream) internal.order = (int(topo) if topo else 0, int(stream)) - @defer.inlineCallbacks - def get_events_around( - self, room_id, event_id, before_limit, after_limit, event_filter=None - ): + async def get_events_around( + self, + room_id: str, + event_id: str, + before_limit: int, + after_limit: int, + event_filter: Optional[Filter] = None, + ) -> dict: """Retrieve events and pagination tokens around a given event in a room. - - Args: - room_id (str) - event_id (str) - before_limit (int) - after_limit (int) - event_filter (Filter|None) - - Returns: - dict """ - results = yield self.db.runInteraction( + results = await self.db_pool.runInteraction( "get_events_around", self._get_events_around_txn, room_id, @@ -684,11 +717,11 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): event_filter, ) - events_before = yield self.get_events_as_list( + events_before = await self.get_events_as_list( list(results["before"]["event_ids"]), get_prev_content=True ) - events_after = yield self.get_events_as_list( + events_after = await self.get_events_as_list( list(results["after"]["event_ids"]), get_prev_content=True ) @@ -700,29 +733,38 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): } def _get_events_around_txn( - self, txn, room_id, event_id, before_limit, after_limit, event_filter - ): + self, + txn: LoggingTransaction, + room_id: str, + event_id: str, + before_limit: int, + after_limit: int, + event_filter: Optional[Filter], + ) -> dict: """Retrieves event_ids and pagination tokens around a given event in a room. Args: - room_id (str) - event_id (str) - before_limit (int) - after_limit (int) - event_filter (Filter|None) + room_id + event_id + before_limit + after_limit + event_filter Returns: dict """ - results = self.db.simple_select_one_txn( + results = self.db_pool.simple_select_one_txn( txn, "events", keyvalues={"event_id": event_id, "room_id": room_id}, retcols=["stream_ordering", "topological_ordering"], ) + # This cannot happen as `allow_none=False`. + assert results is not None + # Paginating backwards includes the event at the token, but paginating # forward doesn't. before_token = RoomStreamToken( @@ -758,22 +800,23 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): "after": {"event_ids": events_after, "token": end_token}, } - @defer.inlineCallbacks - def get_all_new_events_stream(self, from_id, current_id, limit): + async def get_all_new_events_stream( + self, from_id: int, current_id: int, limit: int + ) -> Tuple[int, List[EventBase]]: """Get all new events Returns all events with from_id < stream_ordering <= current_id. Args: - from_id (int): the stream_ordering of the last event we processed - current_id (int): the stream_ordering of the most recently processed event - limit (int): the maximum number of events to return + from_id: the stream_ordering of the last event we processed + current_id: the stream_ordering of the most recently processed event + limit: the maximum number of events to return Returns: - Deferred[Tuple[int, list[FrozenEvent]]]: A tuple of (next_id, events), where - `next_id` is the next value to pass as `from_id` (it will either be the - stream_ordering of the last returned event, or, if fewer than `limit` events - were found, `current_id`. + A tuple of (next_id, events), where `next_id` is the next value to + pass as `from_id` (it will either be the stream_ordering of the + last returned event, or, if fewer than `limit` events were found, + the `current_id`). """ def get_all_new_events_stream_txn(txn): @@ -795,43 +838,43 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): return upper_bound, [row[1] for row in rows] - upper_bound, event_ids = yield self.db.runInteraction( + upper_bound, event_ids = await self.db_pool.runInteraction( "get_all_new_events_stream", get_all_new_events_stream_txn ) - events = yield self.get_events_as_list(event_ids) + events = await self.get_events_as_list(event_ids) return upper_bound, events async def get_federation_out_pos(self, typ: str) -> int: if self._need_to_reset_federation_stream_positions: - await self.db.runInteraction( + await self.db_pool.runInteraction( "_reset_federation_positions_txn", self._reset_federation_positions_txn ) self._need_to_reset_federation_stream_positions = False - return await self.db.simple_select_one_onecol( + return await self.db_pool.simple_select_one_onecol( table="federation_stream_position", retcol="stream_id", keyvalues={"type": typ, "instance_name": self._instance_name}, desc="get_federation_out_pos", ) - async def update_federation_out_pos(self, typ, stream_id): + async def update_federation_out_pos(self, typ: str, stream_id: int) -> None: if self._need_to_reset_federation_stream_positions: - await self.db.runInteraction( + await self.db_pool.runInteraction( "_reset_federation_positions_txn", self._reset_federation_positions_txn ) self._need_to_reset_federation_stream_positions = False - return await self.db.simple_update_one( + await self.db_pool.simple_update_one( table="federation_stream_position", keyvalues={"type": typ, "instance_name": self._instance_name}, updatevalues={"stream_id": stream_id}, desc="update_federation_out_pos", ) - def _reset_federation_positions_txn(self, txn): + def _reset_federation_positions_txn(self, txn: LoggingTransaction) -> None: """Fiddles with the `federation_stream_position` table to make it match the configured federation sender instances during start up. """ @@ -854,7 +897,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): elif self._instance_name not in configured_instances: return - instances_in_table = self.db.simple_select_onecol_txn( + instances_in_table = self.db_pool.simple_select_onecol_txn( txn, table="federation_stream_position", keyvalues={}, @@ -870,7 +913,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): GROUP BY type """ txn.execute(sql) - min_positions = dict(txn) # Map from type -> min position + min_positions = {typ: pos for typ, pos in txn} # Map from type -> min position # Ensure we do actually have some values here assert set(min_positions) == {"federation", "events"} @@ -885,46 +928,44 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): txn.execute(sql % (clause,), args) for typ, stream_id in min_positions.items(): - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="federation_stream_position", keyvalues={"type": typ, "instance_name": self._instance_name}, values={"stream_id": stream_id}, ) - def has_room_changed_since(self, room_id, stream_id): + def has_room_changed_since(self, room_id: str, stream_id: int) -> bool: return self._events_stream_cache.has_entity_changed(room_id, stream_id) def _paginate_room_events_txn( self, - txn, - room_id, - from_token, - to_token=None, - direction="b", - limit=-1, - event_filter=None, - ): + txn: LoggingTransaction, + room_id: str, + from_token: RoomStreamToken, + to_token: Optional[RoomStreamToken] = None, + direction: str = "b", + limit: int = -1, + event_filter: Optional[Filter] = None, + ) -> Tuple[List[_EventDictReturn], RoomStreamToken]: """Returns list of events before or after a given token. Args: txn - room_id (str) - from_token (RoomStreamToken): The token used to stream from - to_token (RoomStreamToken|None): A token which if given limits the - results to only those before - direction(char): Either 'b' or 'f' to indicate whether we are - paginating forwards or backwards from `from_key`. - limit (int): The maximum number of events to return. - event_filter (Filter|None): If provided filters the events to + room_id + from_token: The token used to stream from + to_token: A token which if given limits the results to only those before + direction: Either 'b' or 'f' to indicate whether we are paginating + forwards or backwards from `from_key`. + limit: The maximum number of events to return. + event_filter: If provided filters the events to those that match the filter. Returns: - Deferred[tuple[list[_EventDictReturn], str]]: Returns the results - as a list of _EventDictReturn and a token that points to the end - of the result set. If no events are returned then the end of the - stream has been reached (i.e. there are no events between - `from_token` and `to_token`), or `limit` is zero. + A list of _EventDictReturn and a token that points to the end of the + result set. If no events are returned then the end of the stream has + been reached (i.e. there are no events between `from_token` and + `to_token`), or `limit` is zero. """ assert int(limit) >= 0 @@ -941,8 +982,8 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): bounds = generate_pagination_where_clause( direction=direction, column_names=("topological_ordering", "stream_ordering"), - from_token=from_token, - to_token=to_token, + from_token=from_token.as_tuple(), + to_token=to_token.as_tuple() if to_token else None, engine=self.database_engine, ) @@ -1006,37 +1047,36 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): # TODO (erikj): We should work out what to do here instead. next_token = to_token if to_token else from_token - return rows, str(next_token) + return rows, next_token - @defer.inlineCallbacks - def paginate_room_events( - self, room_id, from_key, to_key=None, direction="b", limit=-1, event_filter=None - ): + async def paginate_room_events( + self, + room_id: str, + from_key: RoomStreamToken, + to_key: Optional[RoomStreamToken] = None, + direction: str = "b", + limit: int = -1, + event_filter: Optional[Filter] = None, + ) -> Tuple[List[EventBase], RoomStreamToken]: """Returns list of events before or after a given token. Args: - room_id (str) - from_key (str): The token used to stream from - to_key (str|None): A token which if given limits the results to - only those before - direction(char): Either 'b' or 'f' to indicate whether we are - paginating forwards or backwards from `from_key`. - limit (int): The maximum number of events to return. - event_filter (Filter|None): If provided filters the events to - those that match the filter. + room_id + from_key: The token used to stream from + to_key: A token which if given limits the results to only those before + direction: Either 'b' or 'f' to indicate whether we are paginating + forwards or backwards from `from_key`. + limit: The maximum number of events to return. + event_filter: If provided filters the events to those that match the filter. Returns: - tuple[list[FrozenEvent], str]: Returns the results as a list of - events and a token that points to the end of the result set. If no - events are returned then the end of the stream has been reached - (i.e. there are no events between `from_key` and `to_key`). + The results as a list of events and a token that points to the end + of the result set. If no events are returned then the end of the + stream has been reached (i.e. there are no events between `from_key` + and `to_key`). """ - from_key = RoomStreamToken.parse(from_key) - if to_key: - to_key = RoomStreamToken.parse(to_key) - - rows, token = yield self.db.runInteraction( + rows, token = await self.db_pool.runInteraction( "paginate_room_events", self._paginate_room_events_txn, room_id, @@ -1047,7 +1087,7 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): event_filter, ) - events = yield self.get_events_as_list( + events = await self.get_events_as_list( [r.event_id for r in rows], get_prev_content=True ) @@ -1057,8 +1097,8 @@ class StreamWorkerStore(EventsWorkerStore, SQLBaseStore): class StreamStore(StreamWorkerStore): - def get_room_max_stream_ordering(self): + def get_room_max_stream_ordering(self) -> int: return self._stream_id_gen.get_current_token() - def get_room_min_stream_ordering(self): + def get_room_min_stream_ordering(self) -> int: return self._backfill_id_gen.get_current_token() diff --git a/synapse/storage/data_stores/main/tags.py b/synapse/storage/databases/main/tags.py
index bd7227773a..9f120d3cb6 100644 --- a/synapse/storage/data_stores/main/tags.py +++ b/synapse/storage/databases/main/tags.py
@@ -15,14 +15,12 @@ # limitations under the License. import logging -from typing import List, Tuple - -from canonicaljson import json - -from twisted.internet import defer +from typing import Dict, List, Tuple from synapse.storage._base import db_to_json -from synapse.storage.data_stores.main.account_data import AccountDataWorkerStore +from synapse.storage.databases.main.account_data import AccountDataWorkerStore +from synapse.types import JsonDict +from synapse.util import json_encoder from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) @@ -30,30 +28,26 @@ logger = logging.getLogger(__name__) class TagsWorkerStore(AccountDataWorkerStore): @cached() - def get_tags_for_user(self, user_id): + async def get_tags_for_user(self, user_id: str) -> Dict[str, Dict[str, JsonDict]]: """Get all the tags for a user. Args: - user_id(str): The user to get the tags for. + user_id: The user to get the tags for. Returns: - A deferred dict mapping from room_id strings to dicts mapping from - tag strings to tag content. + A mapping from room_id strings to dicts mapping from tag strings to + tag content. """ - deferred = self.db.simple_select_list( + rows = await self.db_pool.simple_select_list( "room_tags", {"user_id": user_id}, ["room_id", "tag", "content"] ) - @deferred.addCallback - def tags_by_room(rows): - tags_by_room = {} - for row in rows: - room_tags = tags_by_room.setdefault(row["room_id"], {}) - room_tags[row["tag"]] = db_to_json(row["content"]) - return tags_by_room - - return deferred + tags_by_room = {} # type: Dict[str, Dict[str, JsonDict]] + for row in rows: + room_tags = tags_by_room.setdefault(row["room_id"], {}) + room_tags[row["tag"]] = db_to_json(row["content"]) + return tags_by_room async def get_all_updated_tags( self, instance_name: str, last_id: int, current_id: int, limit: int @@ -92,7 +86,7 @@ class TagsWorkerStore(AccountDataWorkerStore): txn.execute(sql, (last_id, current_id, limit)) return txn.fetchall() - tag_ids = await self.db.runInteraction( + tag_ids = await self.db_pool.runInteraction( "get_all_updated_tags", get_all_updated_tags_txn ) @@ -103,7 +97,7 @@ class TagsWorkerStore(AccountDataWorkerStore): txn.execute(sql, (user_id, room_id)) tags = [] for tag, content in txn: - tags.append(json.dumps(tag) + ":" + content) + tags.append(json_encoder.encode(tag) + ":" + content) tag_json = "{" + ",".join(tags) + "}" results.append((stream_id, (user_id, room_id, tag_json))) @@ -112,7 +106,7 @@ class TagsWorkerStore(AccountDataWorkerStore): batch_size = 50 results = [] for i in range(0, len(tag_ids), batch_size): - tags = await self.db.runInteraction( + tags = await self.db_pool.runInteraction( "get_all_updated_tag_content", get_tag_content, tag_ids[i : i + batch_size], @@ -127,17 +121,19 @@ class TagsWorkerStore(AccountDataWorkerStore): return results, upto_token, limited - @defer.inlineCallbacks - def get_updated_tags(self, user_id, stream_id): + async def get_updated_tags( + self, user_id: str, stream_id: int + ) -> Dict[str, Dict[str, JsonDict]]: """Get all the tags for the rooms where the tags have changed since the given version Args: user_id(str): The user to get the tags for. stream_id(int): The earliest update to get for the user. + Returns: - A deferred dict mapping from room_id strings to lists of tag - strings for all the rooms that changed since the stream_id token. + A mapping from room_id strings to lists of tag strings for all the + rooms that changed since the stream_id token. """ def get_updated_tags_txn(txn): @@ -155,52 +151,58 @@ class TagsWorkerStore(AccountDataWorkerStore): if not changed: return {} - room_ids = yield self.db.runInteraction( + room_ids = await self.db_pool.runInteraction( "get_updated_tags", get_updated_tags_txn ) results = {} if room_ids: - tags_by_room = yield self.get_tags_for_user(user_id) + tags_by_room = await self.get_tags_for_user(user_id) for room_id in room_ids: results[room_id] = tags_by_room.get(room_id, {}) return results - def get_tags_for_room(self, user_id, room_id): + async def get_tags_for_room( + self, user_id: str, room_id: str + ) -> Dict[str, JsonDict]: """Get all the tags for the given room + Args: - user_id(str): The user to get tags for - room_id(str): The room to get tags for + user_id: The user to get tags for + room_id: The room to get tags for + Returns: - A deferred list of string tags. + A mapping of tags to tag content. """ - return self.db.simple_select_list( + rows = await self.db_pool.simple_select_list( table="room_tags", keyvalues={"user_id": user_id, "room_id": room_id}, retcols=("tag", "content"), desc="get_tags_for_room", - ).addCallback( - lambda rows: {row["tag"]: db_to_json(row["content"]) for row in rows} ) + return {row["tag"]: db_to_json(row["content"]) for row in rows} class TagsStore(TagsWorkerStore): - @defer.inlineCallbacks - def add_tag_to_room(self, user_id, room_id, tag, content): + async def add_tag_to_room( + self, user_id: str, room_id: str, tag: str, content: JsonDict + ) -> int: """Add a tag to a room for a user. + Args: - user_id(str): The user to add a tag for. - room_id(str): The room to add a tag for. - tag(str): The tag name to add. - content(dict): A json object to associate with the tag. + user_id: The user to add a tag for. + room_id: The room to add a tag for. + tag: The tag name to add. + content: A json object to associate with the tag. + Returns: - A deferred that completes once the tag has been added. + The next account data ID. """ - content_json = json.dumps(content) + content_json = json_encoder.encode(content) def add_tag_txn(txn, next_id): - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="room_tags", keyvalues={"user_id": user_id, "room_id": room_id, "tag": tag}, @@ -208,19 +210,18 @@ class TagsStore(TagsWorkerStore): ) self._update_revision_txn(txn, user_id, room_id, next_id) - with self._account_data_id_gen.get_next() as next_id: - yield self.db.runInteraction("add_tag", add_tag_txn, next_id) + async with self._account_data_id_gen.get_next() as next_id: + await self.db_pool.runInteraction("add_tag", add_tag_txn, next_id) self.get_tags_for_user.invalidate((user_id,)) - result = self._account_data_id_gen.get_current_token() - return result + return self._account_data_id_gen.get_current_token() - @defer.inlineCallbacks - def remove_tag_from_room(self, user_id, room_id, tag): + async def remove_tag_from_room(self, user_id: str, room_id: str, tag: str) -> int: """Remove a tag from a room for a user. + Returns: - A deferred that completes once the tag has been removed + The next account data ID. """ def remove_tag_txn(txn, next_id): @@ -231,22 +232,23 @@ class TagsStore(TagsWorkerStore): txn.execute(sql, (user_id, room_id, tag)) self._update_revision_txn(txn, user_id, room_id, next_id) - with self._account_data_id_gen.get_next() as next_id: - yield self.db.runInteraction("remove_tag", remove_tag_txn, next_id) + async with self._account_data_id_gen.get_next() as next_id: + await self.db_pool.runInteraction("remove_tag", remove_tag_txn, next_id) self.get_tags_for_user.invalidate((user_id,)) - result = self._account_data_id_gen.get_current_token() - return result + return self._account_data_id_gen.get_current_token() - def _update_revision_txn(self, txn, user_id, room_id, next_id): + def _update_revision_txn( + self, txn, user_id: str, room_id: str, next_id: int + ) -> None: """Update the latest revision of the tags for the given user and room. Args: txn: The database cursor - user_id(str): The ID of the user. - room_id(str): The ID of the room. - next_id(int): The the revision to advance to. + user_id: The ID of the user. + room_id: The ID of the room. + next_id: The the revision to advance to. """ txn.call_after( diff --git a/synapse/storage/databases/main/transactions.py b/synapse/storage/databases/main/transactions.py new file mode 100644
index 0000000000..97aed1500e --- /dev/null +++ b/synapse/storage/databases/main/transactions.py
@@ -0,0 +1,476 @@ +# -*- coding: utf-8 -*- +# Copyright 2014-2016 OpenMarket Ltd +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import logging +from collections import namedtuple +from typing import Iterable, List, Optional, Tuple + +from canonicaljson import encode_canonical_json + +from synapse.metrics.background_process_metrics import run_as_background_process +from synapse.storage._base import SQLBaseStore, db_to_json +from synapse.storage.database import DatabasePool, LoggingTransaction +from synapse.storage.engines import PostgresEngine, Sqlite3Engine +from synapse.types import JsonDict +from synapse.util.caches.expiringcache import ExpiringCache + +db_binary_type = memoryview + +logger = logging.getLogger(__name__) + + +_TransactionRow = namedtuple( + "_TransactionRow", + ("id", "transaction_id", "destination", "ts", "response_code", "response_json"), +) + +_UpdateTransactionRow = namedtuple( + "_TransactionRow", ("response_code", "response_json") +) + +SENTINEL = object() + + +class TransactionStore(SQLBaseStore): + """A collection of queries for handling PDUs. + """ + + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) + + self._clock.looping_call(self._start_cleanup_transactions, 30 * 60 * 1000) + + self._destination_retry_cache = ExpiringCache( + cache_name="get_destination_retry_timings", + clock=self._clock, + expiry_ms=5 * 60 * 1000, + ) + + async def get_received_txn_response( + self, transaction_id: str, origin: str + ) -> Optional[Tuple[int, JsonDict]]: + """For an incoming transaction from a given origin, check if we have + already responded to it. If so, return the response code and response + body (as a dict). + + Args: + transaction_id + origin + + Returns: + None if we have not previously responded to this transaction or a + 2-tuple of (int, dict) + """ + + return await self.db_pool.runInteraction( + "get_received_txn_response", + self._get_received_txn_response, + transaction_id, + origin, + ) + + def _get_received_txn_response(self, txn, transaction_id, origin): + result = self.db_pool.simple_select_one_txn( + txn, + table="received_transactions", + keyvalues={"transaction_id": transaction_id, "origin": origin}, + retcols=( + "transaction_id", + "origin", + "ts", + "response_code", + "response_json", + "has_been_referenced", + ), + allow_none=True, + ) + + if result and result["response_code"]: + return result["response_code"], db_to_json(result["response_json"]) + + else: + return None + + async def set_received_txn_response( + self, transaction_id: str, origin: str, code: int, response_dict: JsonDict + ) -> None: + """Persist the response we returned for an incoming transaction, and + should return for subsequent transactions with the same transaction_id + and origin. + + Args: + transaction_id: The incoming transaction ID. + origin: The origin server. + code: The response code. + response_dict: The response, to be encoded into JSON. + """ + + await self.db_pool.simple_insert( + table="received_transactions", + values={ + "transaction_id": transaction_id, + "origin": origin, + "response_code": code, + "response_json": db_binary_type(encode_canonical_json(response_dict)), + "ts": self._clock.time_msec(), + }, + or_ignore=True, + desc="set_received_txn_response", + ) + + async def get_destination_retry_timings(self, destination): + """Gets the current retry timings (if any) for a given destination. + + Args: + destination (str) + + Returns: + None if not retrying + Otherwise a dict for the retry scheme + """ + + result = self._destination_retry_cache.get(destination, SENTINEL) + if result is not SENTINEL: + return result + + result = await self.db_pool.runInteraction( + "get_destination_retry_timings", + self._get_destination_retry_timings, + destination, + ) + + # We don't hugely care about race conditions between getting and + # invalidating the cache, since we time out fairly quickly anyway. + self._destination_retry_cache[destination] = result + return result + + def _get_destination_retry_timings(self, txn, destination): + result = self.db_pool.simple_select_one_txn( + txn, + table="destinations", + keyvalues={"destination": destination}, + retcols=("destination", "failure_ts", "retry_last_ts", "retry_interval"), + allow_none=True, + ) + + # check we have a row and retry_last_ts is not null or zero + # (retry_last_ts can't be negative) + if result and result["retry_last_ts"]: + return result + else: + return None + + async def set_destination_retry_timings( + self, + destination: str, + failure_ts: Optional[int], + retry_last_ts: int, + retry_interval: int, + ) -> None: + """Sets the current retry timings for a given destination. + Both timings should be zero if retrying is no longer occuring. + + Args: + destination + failure_ts: when the server started failing (ms since epoch) + retry_last_ts: time of last retry attempt in unix epoch ms + retry_interval: how long until next retry in ms + """ + + self._destination_retry_cache.pop(destination, None) + return await self.db_pool.runInteraction( + "set_destination_retry_timings", + self._set_destination_retry_timings, + destination, + failure_ts, + retry_last_ts, + retry_interval, + ) + + def _set_destination_retry_timings( + self, txn, destination, failure_ts, retry_last_ts, retry_interval + ): + + if self.database_engine.can_native_upsert: + # Upsert retry time interval if retry_interval is zero (i.e. we're + # resetting it) or greater than the existing retry interval. + + sql = """ + INSERT INTO destinations ( + destination, failure_ts, retry_last_ts, retry_interval + ) + VALUES (?, ?, ?, ?) + ON CONFLICT (destination) DO UPDATE SET + failure_ts = EXCLUDED.failure_ts, + retry_last_ts = EXCLUDED.retry_last_ts, + retry_interval = EXCLUDED.retry_interval + WHERE + EXCLUDED.retry_interval = 0 + OR destinations.retry_interval IS NULL + OR destinations.retry_interval < EXCLUDED.retry_interval + """ + + txn.execute(sql, (destination, failure_ts, retry_last_ts, retry_interval)) + + return + + self.database_engine.lock_table(txn, "destinations") + + # We need to be careful here as the data may have changed from under us + # due to a worker setting the timings. + + prev_row = self.db_pool.simple_select_one_txn( + txn, + table="destinations", + keyvalues={"destination": destination}, + retcols=("failure_ts", "retry_last_ts", "retry_interval"), + allow_none=True, + ) + + if not prev_row: + self.db_pool.simple_insert_txn( + txn, + table="destinations", + values={ + "destination": destination, + "failure_ts": failure_ts, + "retry_last_ts": retry_last_ts, + "retry_interval": retry_interval, + }, + ) + elif ( + retry_interval == 0 + or prev_row["retry_interval"] is None + or prev_row["retry_interval"] < retry_interval + ): + self.db_pool.simple_update_one_txn( + txn, + "destinations", + keyvalues={"destination": destination}, + updatevalues={ + "failure_ts": failure_ts, + "retry_last_ts": retry_last_ts, + "retry_interval": retry_interval, + }, + ) + + def _start_cleanup_transactions(self): + return run_as_background_process( + "cleanup_transactions", self._cleanup_transactions + ) + + async def _cleanup_transactions(self) -> None: + now = self._clock.time_msec() + month_ago = now - 30 * 24 * 60 * 60 * 1000 + + def _cleanup_transactions_txn(txn): + txn.execute("DELETE FROM received_transactions WHERE ts < ?", (month_ago,)) + + await self.db_pool.runInteraction( + "_cleanup_transactions", _cleanup_transactions_txn + ) + + async def store_destination_rooms_entries( + self, destinations: Iterable[str], room_id: str, stream_ordering: int, + ) -> None: + """ + Updates or creates `destination_rooms` entries in batch for a single event. + + Args: + destinations: list of destinations + room_id: the room_id of the event + stream_ordering: the stream_ordering of the event + """ + + return await self.db_pool.runInteraction( + "store_destination_rooms_entries", + self._store_destination_rooms_entries_txn, + destinations, + room_id, + stream_ordering, + ) + + def _store_destination_rooms_entries_txn( + self, + txn: LoggingTransaction, + destinations: Iterable[str], + room_id: str, + stream_ordering: int, + ) -> None: + + # ensure we have a `destinations` row for this destination, as there is + # a foreign key constraint. + if isinstance(self.database_engine, PostgresEngine): + q = """ + INSERT INTO destinations (destination) + VALUES (?) + ON CONFLICT DO NOTHING; + """ + elif isinstance(self.database_engine, Sqlite3Engine): + q = """ + INSERT OR IGNORE INTO destinations (destination) + VALUES (?); + """ + else: + raise RuntimeError("Unknown database engine") + + txn.execute_batch(q, ((destination,) for destination in destinations)) + + rows = [(destination, room_id) for destination in destinations] + + self.db_pool.simple_upsert_many_txn( + txn, + "destination_rooms", + ["destination", "room_id"], + rows, + ["stream_ordering"], + [(stream_ordering,)] * len(rows), + ) + + async def get_destination_last_successful_stream_ordering( + self, destination: str + ) -> Optional[int]: + """ + Gets the stream ordering of the PDU most-recently successfully sent + to the specified destination, or None if this information has not been + tracked yet. + + Args: + destination: the destination to query + """ + return await self.db_pool.simple_select_one_onecol( + "destinations", + {"destination": destination}, + "last_successful_stream_ordering", + allow_none=True, + desc="get_last_successful_stream_ordering", + ) + + async def set_destination_last_successful_stream_ordering( + self, destination: str, last_successful_stream_ordering: int + ) -> None: + """ + Marks that we have successfully sent the PDUs up to and including the + one specified. + + Args: + destination: the destination we have successfully sent to + last_successful_stream_ordering: the stream_ordering of the most + recent successfully-sent PDU + """ + return await self.db_pool.simple_upsert( + "destinations", + keyvalues={"destination": destination}, + values={"last_successful_stream_ordering": last_successful_stream_ordering}, + desc="set_last_successful_stream_ordering", + ) + + async def get_catch_up_room_event_ids( + self, destination: str, last_successful_stream_ordering: int, + ) -> List[str]: + """ + Returns at most 50 event IDs and their corresponding stream_orderings + that correspond to the oldest events that have not yet been sent to + the destination. + + Args: + destination: the destination in question + last_successful_stream_ordering: the stream_ordering of the + most-recently successfully-transmitted event to the destination + + Returns: + list of event_ids + """ + return await self.db_pool.runInteraction( + "get_catch_up_room_event_ids", + self._get_catch_up_room_event_ids_txn, + destination, + last_successful_stream_ordering, + ) + + @staticmethod + def _get_catch_up_room_event_ids_txn( + txn: LoggingTransaction, destination: str, last_successful_stream_ordering: int, + ) -> List[str]: + q = """ + SELECT event_id FROM destination_rooms + JOIN events USING (stream_ordering) + WHERE destination = ? + AND stream_ordering > ? + ORDER BY stream_ordering + LIMIT 50 + """ + txn.execute( + q, (destination, last_successful_stream_ordering), + ) + event_ids = [row[0] for row in txn] + return event_ids + + async def get_catch_up_outstanding_destinations( + self, after_destination: Optional[str] + ) -> List[str]: + """ + Gets at most 25 destinations which have outstanding PDUs to be caught up, + and are not being backed off from + Args: + after_destination: + If provided, all destinations must be lexicographically greater + than this one. + + Returns: + list of up to 25 destinations with outstanding catch-up. + These are the lexicographically first destinations which are + lexicographically greater than after_destination (if provided). + """ + time = self.hs.get_clock().time_msec() + + return await self.db_pool.runInteraction( + "get_catch_up_outstanding_destinations", + self._get_catch_up_outstanding_destinations_txn, + time, + after_destination, + ) + + @staticmethod + def _get_catch_up_outstanding_destinations_txn( + txn: LoggingTransaction, now_time_ms: int, after_destination: Optional[str] + ) -> List[str]: + q = """ + SELECT destination FROM destinations + WHERE destination IN ( + SELECT destination FROM destination_rooms + WHERE destination_rooms.stream_ordering > + destinations.last_successful_stream_ordering + ) + AND destination > ? + AND ( + retry_last_ts IS NULL OR + retry_last_ts + retry_interval < ? + ) + ORDER BY destination + LIMIT 25 + """ + txn.execute( + q, + ( + # everything is lexicographically greater than "" so this gives + # us the first batch of up to 25. + after_destination or "", + now_time_ms, + ), + ) + + destinations = [row[0] for row in txn] + return destinations diff --git a/synapse/storage/data_stores/main/ui_auth.py b/synapse/storage/databases/main/ui_auth.py
index 5f1b919748..3b9211a6d2 100644 --- a/synapse/storage/data_stores/main/ui_auth.py +++ b/synapse/storage/databases/main/ui_auth.py
@@ -12,18 +12,18 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from typing import Any, Dict, Optional, Union +from typing import Any, Dict, List, Optional, Tuple, Union import attr -from canonicaljson import json from synapse.api.errors import StoreError from synapse.storage._base import SQLBaseStore, db_to_json +from synapse.storage.database import LoggingTransaction from synapse.types import JsonDict -from synapse.util import stringutils as stringutils +from synapse.util import json_encoder, stringutils -@attr.s +@attr.s(slots=True) class UIAuthSessionData: session_id = attr.ib(type=str) # The dictionary from the client root level, not the 'auth' key. @@ -72,7 +72,7 @@ class UIAuthWorkerStore(SQLBaseStore): StoreError if a unique session ID cannot be generated. """ # The clientdict gets stored as JSON. - clientdict_json = json.dumps(clientdict) + clientdict_json = json_encoder.encode(clientdict) # autogen a session ID and try to create it. We may clash, so just # try a few times till one goes through, giving up eventually. @@ -81,7 +81,7 @@ class UIAuthWorkerStore(SQLBaseStore): session_id = stringutils.random_string(24) try: - await self.db.simple_insert( + await self.db_pool.simple_insert( table="ui_auth_sessions", values={ "session_id": session_id, @@ -97,7 +97,7 @@ class UIAuthWorkerStore(SQLBaseStore): return UIAuthSessionData( session_id, clientdict, uri, method, description ) - except self.db.engine.module.IntegrityError: + except self.db_pool.engine.module.IntegrityError: attempts += 1 raise StoreError(500, "Couldn't generate a session ID.") @@ -111,7 +111,7 @@ class UIAuthWorkerStore(SQLBaseStore): Raises: StoreError if the session is not found. """ - result = await self.db.simple_select_one( + result = await self.db_pool.simple_select_one( table="ui_auth_sessions", keyvalues={"session_id": session_id}, retcols=("clientdict", "uri", "method", "description"), @@ -140,13 +140,13 @@ class UIAuthWorkerStore(SQLBaseStore): # Note that we need to allow for the same stage to complete multiple # times here so that registration is idempotent. try: - await self.db.simple_upsert( + await self.db_pool.simple_upsert( table="ui_auth_sessions_credentials", keyvalues={"session_id": session_id, "stage_type": stage_type}, - values={"result": json.dumps(result)}, + values={"result": json_encoder.encode(result)}, desc="mark_ui_auth_stage_complete", ) - except self.db.engine.module.IntegrityError: + except self.db_pool.engine.module.IntegrityError: raise StoreError(400, "Unknown session ID: %s" % (session_id,)) async def get_completed_ui_auth_stages( @@ -162,7 +162,7 @@ class UIAuthWorkerStore(SQLBaseStore): that auth-type. """ results = {} - for row in await self.db.simple_select_list( + for row in await self.db_pool.simple_select_list( table="ui_auth_sessions_credentials", keyvalues={"session_id": session_id}, retcols=("stage_type", "result"), @@ -184,9 +184,9 @@ class UIAuthWorkerStore(SQLBaseStore): The dictionary from the client root level, not the 'auth' key. """ # The clientdict gets stored as JSON. - clientdict_json = json.dumps(clientdict) + clientdict_json = json_encoder.encode(clientdict) - await self.db.simple_update_one( + await self.db_pool.simple_update_one( table="ui_auth_sessions", keyvalues={"session_id": session_id}, updatevalues={"clientdict": clientdict_json}, @@ -206,7 +206,7 @@ class UIAuthWorkerStore(SQLBaseStore): Raises: StoreError if the session cannot be found. """ - await self.db.runInteraction( + await self.db_pool.runInteraction( "set_ui_auth_session_data", self._set_ui_auth_session_data_txn, session_id, @@ -214,24 +214,26 @@ class UIAuthWorkerStore(SQLBaseStore): value, ) - def _set_ui_auth_session_data_txn(self, txn, session_id: str, key: str, value: Any): + def _set_ui_auth_session_data_txn( + self, txn: LoggingTransaction, session_id: str, key: str, value: Any + ): # Get the current value. - result = self.db.simple_select_one_txn( + result = self.db_pool.simple_select_one_txn( txn, table="ui_auth_sessions", keyvalues={"session_id": session_id}, retcols=("serverdict",), - ) + ) # type: Dict[str, Any] # type: ignore # Update it and add it back to the database. serverdict = db_to_json(result["serverdict"]) serverdict[key] = value - self.db.simple_update_one_txn( + self.db_pool.simple_update_one_txn( txn, table="ui_auth_sessions", keyvalues={"session_id": session_id}, - updatevalues={"serverdict": json.dumps(serverdict)}, + updatevalues={"serverdict": json_encoder.encode(serverdict)}, ) async def get_ui_auth_session_data( @@ -247,7 +249,7 @@ class UIAuthWorkerStore(SQLBaseStore): Raises: StoreError if the session cannot be found. """ - result = await self.db.simple_select_one( + result = await self.db_pool.simple_select_one( table="ui_auth_sessions", keyvalues={"session_id": session_id}, retcols=("serverdict",), @@ -258,9 +260,37 @@ class UIAuthWorkerStore(SQLBaseStore): return serverdict.get(key, default) + async def add_user_agent_ip_to_ui_auth_session( + self, session_id: str, user_agent: str, ip: str, + ): + """Add the given user agent / IP to the tracking table + """ + await self.db_pool.simple_upsert( + table="ui_auth_sessions_ips", + keyvalues={"session_id": session_id, "user_agent": user_agent, "ip": ip}, + values={}, + desc="add_user_agent_ip_to_ui_auth_session", + ) + + async def get_user_agents_ips_to_ui_auth_session( + self, session_id: str, + ) -> List[Tuple[str, str]]: + """Get the given user agents / IPs used during the ui auth process + + Returns: + List of user_agent/ip pairs + """ + rows = await self.db_pool.simple_select_list( + table="ui_auth_sessions_ips", + keyvalues={"session_id": session_id}, + retcols=("user_agent", "ip"), + desc="get_user_agents_ips_to_ui_auth_session", + ) + return [(row["user_agent"], row["ip"]) for row in rows] + class UIAuthStore(UIAuthWorkerStore): - def delete_old_ui_auth_sessions(self, expiration_time: int): + async def delete_old_ui_auth_sessions(self, expiration_time: int) -> None: """ Remove sessions which were last used earlier than the expiration time. @@ -269,20 +299,31 @@ class UIAuthStore(UIAuthWorkerStore): This is an epoch time in milliseconds. """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_old_ui_auth_sessions", self._delete_old_ui_auth_sessions_txn, expiration_time, ) - def _delete_old_ui_auth_sessions_txn(self, txn, expiration_time: int): + def _delete_old_ui_auth_sessions_txn( + self, txn: LoggingTransaction, expiration_time: int + ): # Get the expired sessions. sql = "SELECT session_id FROM ui_auth_sessions WHERE creation_time <= ?" txn.execute(sql, [expiration_time]) session_ids = [r[0] for r in txn.fetchall()] + # Delete the corresponding IP/user agents. + self.db_pool.simple_delete_many_txn( + txn, + table="ui_auth_sessions_ips", + column="session_id", + iterable=session_ids, + keyvalues={}, + ) + # Delete the corresponding completed credentials. - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="ui_auth_sessions_credentials", column="session_id", @@ -291,7 +332,7 @@ class UIAuthStore(UIAuthWorkerStore): ) # Finally, delete the sessions. - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="ui_auth_sessions", column="session_id", diff --git a/synapse/storage/data_stores/main/user_directory.py b/synapse/storage/databases/main/user_directory.py
index 942e51fd3a..5a390ff2f6 100644 --- a/synapse/storage/data_stores/main/user_directory.py +++ b/synapse/storage/databases/main/user_directory.py
@@ -15,13 +15,12 @@ import logging import re - -from twisted.internet import defer +from typing import Any, Dict, Iterable, Optional, Set, Tuple from synapse.api.constants import EventTypes, JoinRules -from synapse.storage.data_stores.main.state import StateFilter -from synapse.storage.data_stores.main.state_deltas import StateDeltasStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.main.state import StateFilter +from synapse.storage.databases.main.state_deltas import StateDeltasStore from synapse.storage.engines import PostgresEngine, Sqlite3Engine from synapse.types import get_domain_from_id, get_localpart_from_id from synapse.util.caches.descriptors import cached @@ -38,29 +37,28 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): # add_users_who_share_private_rooms? SHARE_PRIVATE_WORKING_SET = 500 - def __init__(self, database: Database, db_conn, hs): - super(UserDirectoryBackgroundUpdateStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) self.server_name = hs.hostname - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_user_directory_createtables", self._populate_user_directory_createtables, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_user_directory_process_rooms", self._populate_user_directory_process_rooms, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_user_directory_process_users", self._populate_user_directory_process_users, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( "populate_user_directory_cleanup", self._populate_user_directory_cleanup ) - @defer.inlineCallbacks - def _populate_user_directory_createtables(self, progress, batch_size): + async def _populate_user_directory_createtables(self, progress, batch_size): # Get all the rooms that we want to process. def _make_staging_area(txn): @@ -85,7 +83,7 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): """ txn.execute(sql) rooms = [{"room_id": x[0], "events": x[1]} for x in txn.fetchall()] - self.db.simple_insert_many_txn(txn, TEMP_TABLE + "_rooms", rooms) + self.db_pool.simple_insert_many_txn(txn, TEMP_TABLE + "_rooms", rooms) del rooms # If search all users is on, get all the users we want to add. @@ -100,43 +98,45 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): txn.execute("SELECT name FROM users") users = [{"user_id": x[0]} for x in txn.fetchall()] - self.db.simple_insert_many_txn(txn, TEMP_TABLE + "_users", users) + self.db_pool.simple_insert_many_txn(txn, TEMP_TABLE + "_users", users) - new_pos = yield self.get_max_stream_id_in_current_state_deltas() - yield self.db.runInteraction( + new_pos = await self.get_max_stream_id_in_current_state_deltas() + await self.db_pool.runInteraction( "populate_user_directory_temp_build", _make_staging_area ) - yield self.db.simple_insert(TEMP_TABLE + "_position", {"position": new_pos}) + await self.db_pool.simple_insert( + TEMP_TABLE + "_position", {"position": new_pos} + ) - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( "populate_user_directory_createtables" ) return 1 - @defer.inlineCallbacks - def _populate_user_directory_cleanup(self, progress, batch_size): + async def _populate_user_directory_cleanup(self, progress, batch_size): """ Update the user directory stream position, then clean up the old tables. """ - position = yield self.db.simple_select_one_onecol( + position = await self.db_pool.simple_select_one_onecol( TEMP_TABLE + "_position", None, "position" ) - yield self.update_user_directory_stream_pos(position) + await self.update_user_directory_stream_pos(position) def _delete_staging_area(txn): txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_rooms") txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_users") txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_position") - yield self.db.runInteraction( + await self.db_pool.runInteraction( "populate_user_directory_cleanup", _delete_staging_area ) - yield self.db.updates._end_background_update("populate_user_directory_cleanup") + await self.db_pool.updates._end_background_update( + "populate_user_directory_cleanup" + ) return 1 - @defer.inlineCallbacks - def _populate_user_directory_process_rooms(self, progress, batch_size): + async def _populate_user_directory_process_rooms(self, progress, batch_size): """ Args: progress (dict) @@ -147,7 +147,7 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): # If we don't have progress filed, delete everything. if not progress: - yield self.delete_all_from_user_dir() + await self.delete_all_from_user_dir() def _get_next_batch(txn): # Only fetch 250 rooms, so we don't fetch too many at once, even @@ -172,13 +172,13 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): return rooms_to_work_on - rooms_to_work_on = yield self.db.runInteraction( + rooms_to_work_on = await self.db_pool.runInteraction( "populate_user_directory_temp_read", _get_next_batch ) # No more rooms -- complete the transaction. if not rooms_to_work_on: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( "populate_user_directory_process_rooms" ) return 1 @@ -191,21 +191,19 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): processed_event_count = 0 for room_id, event_count in rooms_to_work_on: - is_in_room = yield self.is_host_joined(room_id, self.server_name) + is_in_room = await self.is_host_joined(room_id, self.server_name) if is_in_room: - is_public = yield self.is_room_world_readable_or_publicly_joinable( + is_public = await self.is_room_world_readable_or_publicly_joinable( room_id ) - users_with_profile = yield defer.ensureDeferred( - state.get_current_users_in_room(room_id) - ) + users_with_profile = await state.get_current_users_in_room(room_id) user_ids = set(users_with_profile) # Update each user in the user directory. for user_id, profile in users_with_profile.items(): - yield self.update_profile_in_user_dir( + await self.update_profile_in_user_dir( user_id, profile.display_name, profile.avatar_url ) @@ -219,7 +217,7 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): to_insert.add(user_id) if to_insert: - yield self.add_users_in_public_rooms(room_id, to_insert) + await self.add_users_in_public_rooms(room_id, to_insert) to_insert.clear() else: for user_id in user_ids: @@ -239,22 +237,24 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): # If it gets too big, stop and write to the database # to prevent storing too much in RAM. if len(to_insert) >= self.SHARE_PRIVATE_WORKING_SET: - yield self.add_users_who_share_private_room( + await self.add_users_who_share_private_room( room_id, to_insert ) to_insert.clear() if to_insert: - yield self.add_users_who_share_private_room(room_id, to_insert) + await self.add_users_who_share_private_room(room_id, to_insert) to_insert.clear() # We've finished a room. Delete it from the table. - yield self.db.simple_delete_one(TEMP_TABLE + "_rooms", {"room_id": room_id}) + await self.db_pool.simple_delete_one( + TEMP_TABLE + "_rooms", {"room_id": room_id} + ) # Update the remaining counter. progress["remaining"] -= 1 - yield self.db.runInteraction( + await self.db_pool.runInteraction( "populate_user_directory", - self.db.updates._background_update_progress_txn, + self.db_pool.updates._background_update_progress_txn, "populate_user_directory_process_rooms", progress, ) @@ -267,13 +267,12 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): return processed_event_count - @defer.inlineCallbacks - def _populate_user_directory_process_users(self, progress, batch_size): + async def _populate_user_directory_process_users(self, progress, batch_size): """ If search_all_users is enabled, add all of the users to the user directory. """ if not self.hs.config.user_directory_search_all_users: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( "populate_user_directory_process_users" ) return 1 @@ -299,13 +298,13 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): return users_to_work_on - users_to_work_on = yield self.db.runInteraction( + users_to_work_on = await self.db_pool.runInteraction( "populate_user_directory_temp_read", _get_next_batch ) # No more users -- complete the transaction. if not users_to_work_on: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( "populate_user_directory_process_users" ) return 1 @@ -316,26 +315,27 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): ) for user_id in users_to_work_on: - profile = yield self.get_profileinfo(get_localpart_from_id(user_id)) - yield self.update_profile_in_user_dir( + profile = await self.get_profileinfo(get_localpart_from_id(user_id)) + await self.update_profile_in_user_dir( user_id, profile.display_name, profile.avatar_url ) # We've finished processing a user. Delete it from the table. - yield self.db.simple_delete_one(TEMP_TABLE + "_users", {"user_id": user_id}) + await self.db_pool.simple_delete_one( + TEMP_TABLE + "_users", {"user_id": user_id} + ) # Update the remaining counter. progress["remaining"] -= 1 - yield self.db.runInteraction( + await self.db_pool.runInteraction( "populate_user_directory", - self.db.updates._background_update_progress_txn, + self.db_pool.updates._background_update_progress_txn, "populate_user_directory_process_users", progress, ) return len(users_to_work_on) - @defer.inlineCallbacks - def is_room_world_readable_or_publicly_joinable(self, room_id): + async def is_room_world_readable_or_publicly_joinable(self, room_id): """Check if the room is either world_readable or publically joinable """ @@ -345,33 +345,40 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): (EventTypes.RoomHistoryVisibility, ""), ) - current_state_ids = yield self.get_filtered_current_state_ids( + current_state_ids = await self.get_filtered_current_state_ids( room_id, StateFilter.from_types(types_to_filter) ) join_rules_id = current_state_ids.get((EventTypes.JoinRules, "")) if join_rules_id: - join_rule_ev = yield self.get_event(join_rules_id, allow_none=True) + join_rule_ev = await self.get_event(join_rules_id, allow_none=True) if join_rule_ev: if join_rule_ev.content.get("join_rule") == JoinRules.PUBLIC: return True hist_vis_id = current_state_ids.get((EventTypes.RoomHistoryVisibility, "")) if hist_vis_id: - hist_vis_ev = yield self.get_event(hist_vis_id, allow_none=True) + hist_vis_ev = await self.get_event(hist_vis_id, allow_none=True) if hist_vis_ev: if hist_vis_ev.content.get("history_visibility") == "world_readable": return True return False - def update_profile_in_user_dir(self, user_id, display_name, avatar_url): + async def update_profile_in_user_dir( + self, user_id: str, display_name: str, avatar_url: str + ) -> None: """ Update or add a user's profile in the user directory. """ + # If the display name or avatar URL are unexpected types, overwrite them. + if not isinstance(display_name, str): + display_name = None + if not isinstance(avatar_url, str): + avatar_url = None def _update_profile_in_user_dir_txn(txn): - new_entry = self.db.simple_upsert_txn( + new_entry = self.db_pool.simple_upsert_txn( txn, table="user_directory", keyvalues={"user_id": user_id}, @@ -445,7 +452,7 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): ) elif isinstance(self.database_engine, Sqlite3Engine): value = "%s %s" % (user_id, display_name) if display_name else user_id - self.db.simple_upsert_txn( + self.db_pool.simple_upsert_txn( txn, table="user_directory_search", keyvalues={"user_id": user_id}, @@ -458,21 +465,23 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): txn.call_after(self.get_user_in_directory.invalidate, (user_id,)) - return self.db.runInteraction( + await self.db_pool.runInteraction( "update_profile_in_user_dir", _update_profile_in_user_dir_txn ) - def add_users_who_share_private_room(self, room_id, user_id_tuples): + async def add_users_who_share_private_room( + self, room_id: str, user_id_tuples: Iterable[Tuple[str, str]] + ) -> None: """Insert entries into the users_who_share_private_rooms table. The first user should be a local user. Args: - room_id (str) - user_id_tuples([(str, str)]): iterable of 2-tuple of user IDs. + room_id + user_id_tuples: iterable of 2-tuple of user IDs. """ def _add_users_who_share_room_txn(txn): - self.db.simple_upsert_many_txn( + self.db_pool.simple_upsert_many_txn( txn, table="users_who_share_private_rooms", key_names=["user_id", "other_user_id", "room_id"], @@ -484,22 +493,24 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): value_values=None, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "add_users_who_share_room", _add_users_who_share_room_txn ) - def add_users_in_public_rooms(self, room_id, user_ids): + async def add_users_in_public_rooms( + self, room_id: str, user_ids: Iterable[str] + ) -> None: """Insert entries into the users_who_share_private_rooms table. The first user should be a local user. Args: - room_id (str) - user_ids (list[str]) + room_id + user_ids """ def _add_users_in_public_rooms_txn(txn): - self.db.simple_upsert_many_txn( + self.db_pool.simple_upsert_many_txn( txn, table="users_in_public_rooms", key_names=["user_id", "room_id"], @@ -508,11 +519,11 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): value_values=None, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "add_users_in_public_rooms", _add_users_in_public_rooms_txn ) - def delete_all_from_user_dir(self): + async def delete_all_from_user_dir(self) -> None: """Delete the entire user directory """ @@ -523,13 +534,13 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): txn.execute("DELETE FROM users_who_share_private_rooms") txn.call_after(self.get_user_in_directory.invalidate_all) - return self.db.runInteraction( + await self.db_pool.runInteraction( "delete_all_from_user_dir", _delete_all_from_user_dir_txn ) @cached() - def get_user_in_directory(self, user_id): - return self.db.simple_select_one( + async def get_user_in_directory(self, user_id: str) -> Optional[Dict[str, Any]]: + return await self.db_pool.simple_select_one( table="user_directory", keyvalues={"user_id": user_id}, retcols=("display_name", "avatar_url"), @@ -537,8 +548,8 @@ class UserDirectoryBackgroundUpdateStore(StateDeltasStore): desc="get_user_in_directory", ) - def update_user_directory_stream_pos(self, stream_id): - return self.db.simple_update_one( + async def update_user_directory_stream_pos(self, stream_id: str) -> None: + await self.db_pool.simple_update_one( table="user_directory_stream_pos", keyvalues={}, updatevalues={"stream_id": stream_id}, @@ -552,47 +563,48 @@ class UserDirectoryStore(UserDirectoryBackgroundUpdateStore): # add_users_who_share_private_rooms? SHARE_PRIVATE_WORKING_SET = 500 - def __init__(self, database: Database, db_conn, hs): - super(UserDirectoryStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) - def remove_from_user_dir(self, user_id): + async def remove_from_user_dir(self, user_id: str) -> None: def _remove_from_user_dir_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="user_directory", keyvalues={"user_id": user_id} ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="user_directory_search", keyvalues={"user_id": user_id} ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_in_public_rooms", keyvalues={"user_id": user_id} ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_who_share_private_rooms", keyvalues={"user_id": user_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_who_share_private_rooms", keyvalues={"other_user_id": user_id}, ) txn.call_after(self.get_user_in_directory.invalidate, (user_id,)) - return self.db.runInteraction("remove_from_user_dir", _remove_from_user_dir_txn) + await self.db_pool.runInteraction( + "remove_from_user_dir", _remove_from_user_dir_txn + ) - @defer.inlineCallbacks - def get_users_in_dir_due_to_room(self, room_id): + async def get_users_in_dir_due_to_room(self, room_id): """Get all user_ids that are in the room directory because they're in the given room_id """ - user_ids_share_pub = yield self.db.simple_select_onecol( + user_ids_share_pub = await self.db_pool.simple_select_onecol( table="users_in_public_rooms", keyvalues={"room_id": room_id}, retcol="user_id", desc="get_users_in_dir_due_to_room", ) - user_ids_share_priv = yield self.db.simple_select_onecol( + user_ids_share_priv = await self.db_pool.simple_select_onecol( table="users_who_share_private_rooms", keyvalues={"room_id": room_id}, retcol="other_user_id", @@ -604,39 +616,38 @@ class UserDirectoryStore(UserDirectoryBackgroundUpdateStore): return user_ids - def remove_user_who_share_room(self, user_id, room_id): + async def remove_user_who_share_room(self, user_id: str, room_id: str) -> None: """ Deletes entries in the users_who_share_*_rooms table. The first user should be a local user. Args: - user_id (str) - room_id (str) + user_id + room_id """ def _remove_user_who_share_room_txn(txn): - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_who_share_private_rooms", keyvalues={"user_id": user_id, "room_id": room_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_who_share_private_rooms", keyvalues={"other_user_id": user_id, "room_id": room_id}, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="users_in_public_rooms", keyvalues={"user_id": user_id, "room_id": room_id}, ) - return self.db.runInteraction( + await self.db_pool.runInteraction( "remove_user_who_share_room", _remove_user_who_share_room_txn ) - @defer.inlineCallbacks - def get_user_dir_rooms_user_is_in(self, user_id): + async def get_user_dir_rooms_user_is_in(self, user_id): """ Returns the rooms that a user is in. @@ -646,14 +657,14 @@ class UserDirectoryStore(UserDirectoryBackgroundUpdateStore): Returns: list: user_id """ - rows = yield self.db.simple_select_onecol( + rows = await self.db_pool.simple_select_onecol( table="users_who_share_private_rooms", keyvalues={"user_id": user_id}, retcol="room_id", desc="get_rooms_user_is_in", ) - pub_rows = yield self.db.simple_select_onecol( + pub_rows = await self.db_pool.simple_select_onecol( table="users_in_public_rooms", keyvalues={"user_id": user_id}, retcol="room_id", @@ -664,42 +675,57 @@ class UserDirectoryStore(UserDirectoryBackgroundUpdateStore): users.update(rows) return list(users) - @defer.inlineCallbacks - def get_rooms_in_common_for_users(self, user_id, other_user_id): - """Given two user_ids find out the list of rooms they share. + @cached() + async def get_shared_rooms_for_users( + self, user_id: str, other_user_id: str + ) -> Set[str]: """ - sql = """ - SELECT room_id FROM ( - SELECT c.room_id FROM current_state_events AS c - INNER JOIN room_memberships AS m USING (event_id) - WHERE type = 'm.room.member' - AND m.membership = 'join' - AND state_key = ? - ) AS f1 INNER JOIN ( - SELECT c.room_id FROM current_state_events AS c - INNER JOIN room_memberships AS m USING (event_id) - WHERE type = 'm.room.member' - AND m.membership = 'join' - AND state_key = ? - ) f2 USING (room_id) + Returns the rooms that a local user shares with another local or remote user. + + Args: + user_id: The MXID of a local user + other_user_id: The MXID of the other user + + Returns: + A set of room ID's that the users share. """ - rows = yield self.db.execute( - "get_rooms_in_common_for_users", None, sql, user_id, other_user_id + def _get_shared_rooms_for_users_txn(txn): + txn.execute( + """ + SELECT p1.room_id + FROM users_in_public_rooms as p1 + INNER JOIN users_in_public_rooms as p2 + ON p1.room_id = p2.room_id + AND p1.user_id = ? + AND p2.user_id = ? + UNION + SELECT room_id + FROM users_who_share_private_rooms + WHERE + user_id = ? + AND other_user_id = ? + """, + (user_id, other_user_id, user_id, other_user_id), + ) + rows = self.db_pool.cursor_to_dict(txn) + return rows + + rows = await self.db_pool.runInteraction( + "get_shared_rooms_for_users", _get_shared_rooms_for_users_txn ) - return [room_id for room_id, in rows] + return {row["room_id"] for row in rows} - def get_user_directory_stream_pos(self): - return self.db.simple_select_one_onecol( + async def get_user_directory_stream_pos(self) -> int: + return await self.db_pool.simple_select_one_onecol( table="user_directory_stream_pos", keyvalues={}, retcol="stream_id", desc="get_user_directory_stream_pos", ) - @defer.inlineCallbacks - def search_user_dir(self, user_id, search_term, limit): + async def search_user_dir(self, user_id, search_term, limit): """Searches for users in directory Returns: @@ -796,8 +822,8 @@ class UserDirectoryStore(UserDirectoryBackgroundUpdateStore): # This should be unreachable. raise Exception("Unrecognized database engine") - results = yield self.db.execute( - "search_user_dir", self.db.cursor_to_dict, sql, *args + results = await self.db_pool.execute( + "search_user_dir", self.db_pool.cursor_to_dict, sql, *args ) limited = len(results) > limit diff --git a/synapse/storage/data_stores/main/user_erasure_store.py b/synapse/storage/databases/main/user_erasure_store.py
index d3038ff06d..f9575b1f1f 100644 --- a/synapse/storage/data_stores/main/user_erasure_store.py +++ b/synapse/storage/databases/main/user_erasure_store.py
@@ -13,35 +13,32 @@ # See the License for the specific language governing permissions and # limitations under the License. -import operator - from synapse.storage._base import SQLBaseStore from synapse.util.caches.descriptors import cached, cachedList class UserErasureWorkerStore(SQLBaseStore): @cached() - def is_user_erased(self, user_id): + async def is_user_erased(self, user_id: str) -> bool: """ Check if the given user id has requested erasure Args: - user_id (str): full user id to check + user_id: full user id to check Returns: - Deferred[bool]: True if the user has requested erasure + True if the user has requested erasure """ - return self.db.simple_select_onecol( + result = await self.db_pool.simple_select_onecol( table="erased_users", keyvalues={"user_id": user_id}, retcol="1", desc="is_user_erased", - ).addCallback(operator.truth) + ) + return bool(result) - @cachedList( - cached_method_name="is_user_erased", list_name="user_ids", inlineCallbacks=True - ) - def are_users_erased(self, user_ids): + @cachedList(cached_method_name="is_user_erased", list_name="user_ids") + async def are_users_erased(self, user_ids): """ Checks which users in a list have requested erasure @@ -49,14 +46,14 @@ class UserErasureWorkerStore(SQLBaseStore): user_ids (iterable[str]): full user id to check Returns: - Deferred[dict[str, bool]]: + dict[str, bool]: for each user, whether the user has requested erasure. """ # this serves the dual purpose of (a) making sure we can do len and # iterate it multiple times, and (b) avoiding duplicates. user_ids = tuple(set(user_ids)) - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="erased_users", column="user_id", iterable=user_ids, @@ -65,12 +62,11 @@ class UserErasureWorkerStore(SQLBaseStore): ) erased_users = {row["user_id"] for row in rows} - res = {u: u in erased_users for u in user_ids} - return res + return {u: u in erased_users for u in user_ids} class UserErasureStore(UserErasureWorkerStore): - def mark_user_erased(self, user_id: str) -> None: + async def mark_user_erased(self, user_id: str) -> None: """Indicate that user_id wishes their message history to be erased. Args: @@ -88,9 +84,9 @@ class UserErasureStore(UserErasureWorkerStore): self._invalidate_cache_and_stream(txn, self.is_user_erased, (user_id,)) - return self.db.runInteraction("mark_user_erased", f) + await self.db_pool.runInteraction("mark_user_erased", f) - def mark_user_not_erased(self, user_id: str) -> None: + async def mark_user_not_erased(self, user_id: str) -> None: """Indicate that user_id is no longer erased. Args: @@ -104,10 +100,10 @@ class UserErasureStore(UserErasureWorkerStore): return # They are there, delete them. - self.simple_delete_one_txn( + self.db_pool.simple_delete_one_txn( txn, "erased_users", keyvalues={"user_id": user_id} ) self._invalidate_cache_and_stream(txn, self.is_user_erased, (user_id,)) - return self.db.runInteraction("mark_user_not_erased", f) + await self.db_pool.runInteraction("mark_user_not_erased", f) diff --git a/synapse/storage/data_stores/state/__init__.py b/synapse/storage/databases/state/__init__.py
index 86e09f6229..c90d022899 100644 --- a/synapse/storage/data_stores/state/__init__.py +++ b/synapse/storage/databases/state/__init__.py
@@ -13,4 +13,4 @@ # See the License for the specific language governing permissions and # limitations under the License. -from synapse.storage.data_stores.state.store import StateGroupDataStore # noqa: F401 +from synapse.storage.databases.state.store import StateGroupDataStore # noqa: F401 diff --git a/synapse/storage/data_stores/state/bg_updates.py b/synapse/storage/databases/state/bg_updates.py
index be1fe97d79..acb24e33af 100644 --- a/synapse/storage/data_stores/state/bg_updates.py +++ b/synapse/storage/databases/state/bg_updates.py
@@ -15,10 +15,8 @@ import logging -from twisted.internet import defer - from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.engines import PostgresEngine from synapse.storage.state import StateFilter @@ -62,7 +60,7 @@ class StateGroupBackgroundUpdateStore(SQLBaseStore): count = 0 while next_group: - next_group = self.db.simple_select_one_onecol_txn( + next_group = self.db_pool.simple_select_one_onecol_txn( txn, table="state_group_edges", keyvalues={"state_group": next_group}, @@ -165,7 +163,7 @@ class StateGroupBackgroundUpdateStore(SQLBaseStore): ): break - next_group = self.db.simple_select_one_onecol_txn( + next_group = self.db_pool.simple_select_one_onecol_txn( txn, table="state_group_edges", keyvalues={"state_group": next_group}, @@ -182,24 +180,23 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): STATE_GROUP_INDEX_UPDATE_NAME = "state_group_state_type_index" STATE_GROUPS_ROOM_INDEX_UPDATE_NAME = "state_groups_room_id_idx" - def __init__(self, database: Database, db_conn, hs): - super(StateBackgroundUpdateStore, self).__init__(database, db_conn, hs) - self.db.updates.register_background_update_handler( + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) + self.db_pool.updates.register_background_update_handler( self.STATE_GROUP_DEDUPLICATION_UPDATE_NAME, self._background_deduplicate_state, ) - self.db.updates.register_background_update_handler( + self.db_pool.updates.register_background_update_handler( self.STATE_GROUP_INDEX_UPDATE_NAME, self._background_index_state ) - self.db.updates.register_background_index_update( + self.db_pool.updates.register_background_index_update( self.STATE_GROUPS_ROOM_INDEX_UPDATE_NAME, index_name="state_groups_room_id_idx", table="state_groups", columns=["room_id"], ) - @defer.inlineCallbacks - def _background_deduplicate_state(self, progress, batch_size): + async def _background_deduplicate_state(self, progress, batch_size): """This background update will slowly deduplicate state by reencoding them as deltas. """ @@ -212,7 +209,7 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): batch_size = max(1, int(batch_size / BATCH_SIZE_SCALE_FACTOR)) if max_group is None: - rows = yield self.db.execute( + rows = await self.db_pool.execute( "_background_deduplicate_state", None, "SELECT coalesce(max(id), 0) FROM state_groups", @@ -282,13 +279,13 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): if prev_state.get(key, None) != value } - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="state_group_edges", keyvalues={"state_group": state_group}, ) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="state_group_edges", values={ @@ -297,13 +294,13 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): }, ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="state_groups_state", keyvalues={"state_group": state_group}, ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="state_groups_state", values=[ @@ -324,25 +321,24 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): "max_group": max_group, } - self.db.updates._background_update_progress_txn( + self.db_pool.updates._background_update_progress_txn( txn, self.STATE_GROUP_DEDUPLICATION_UPDATE_NAME, progress ) return False, batch_size - finished, result = yield self.db.runInteraction( + finished, result = await self.db_pool.runInteraction( self.STATE_GROUP_DEDUPLICATION_UPDATE_NAME, reindex_txn ) if finished: - yield self.db.updates._end_background_update( + await self.db_pool.updates._end_background_update( self.STATE_GROUP_DEDUPLICATION_UPDATE_NAME ) return result * BATCH_SIZE_SCALE_FACTOR - @defer.inlineCallbacks - def _background_index_state(self, progress, batch_size): + async def _background_index_state(self, progress, batch_size): def reindex_txn(conn): conn.rollback() if isinstance(self.database_engine, PostgresEngine): @@ -365,8 +361,10 @@ class StateBackgroundUpdateStore(StateGroupBackgroundUpdateStore): ) txn.execute("DROP INDEX IF EXISTS state_groups_state_id") - yield self.db.runWithConnection(reindex_txn) + await self.db_pool.runWithConnection(reindex_txn) - yield self.db.updates._end_background_update(self.STATE_GROUP_INDEX_UPDATE_NAME) + await self.db_pool.updates._end_background_update( + self.STATE_GROUP_INDEX_UPDATE_NAME + ) return 1 diff --git a/synapse/storage/data_stores/state/schema/delta/23/drop_state_index.sql b/synapse/storage/databases/state/schema/delta/23/drop_state_index.sql
index ae09fa0065..ae09fa0065 100644 --- a/synapse/storage/data_stores/state/schema/delta/23/drop_state_index.sql +++ b/synapse/storage/databases/state/schema/delta/23/drop_state_index.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/30/state_stream.sql b/synapse/storage/databases/state/schema/delta/30/state_stream.sql
index e85699e82e..e85699e82e 100644 --- a/synapse/storage/data_stores/state/schema/delta/30/state_stream.sql +++ b/synapse/storage/databases/state/schema/delta/30/state_stream.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/32/remove_state_indices.sql b/synapse/storage/databases/state/schema/delta/32/remove_state_indices.sql
index 1450313bfa..1450313bfa 100644 --- a/synapse/storage/data_stores/state/schema/delta/32/remove_state_indices.sql +++ b/synapse/storage/databases/state/schema/delta/32/remove_state_indices.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/35/add_state_index.sql b/synapse/storage/databases/state/schema/delta/35/add_state_index.sql
index 33980d02f0..33980d02f0 100644 --- a/synapse/storage/data_stores/state/schema/delta/35/add_state_index.sql +++ b/synapse/storage/databases/state/schema/delta/35/add_state_index.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/35/state.sql b/synapse/storage/databases/state/schema/delta/35/state.sql
index 0f1fa68a89..0f1fa68a89 100644 --- a/synapse/storage/data_stores/state/schema/delta/35/state.sql +++ b/synapse/storage/databases/state/schema/delta/35/state.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/35/state_dedupe.sql b/synapse/storage/databases/state/schema/delta/35/state_dedupe.sql
index 97e5067ef4..97e5067ef4 100644 --- a/synapse/storage/data_stores/state/schema/delta/35/state_dedupe.sql +++ b/synapse/storage/databases/state/schema/delta/35/state_dedupe.sql
diff --git a/synapse/storage/data_stores/state/schema/delta/47/state_group_seq.py b/synapse/storage/databases/state/schema/delta/47/state_group_seq.py
index 9fd1ccf6f7..9fd1ccf6f7 100644 --- a/synapse/storage/data_stores/state/schema/delta/47/state_group_seq.py +++ b/synapse/storage/databases/state/schema/delta/47/state_group_seq.py
diff --git a/synapse/storage/data_stores/state/schema/delta/56/state_group_room_idx.sql b/synapse/storage/databases/state/schema/delta/56/state_group_room_idx.sql
index 7916ef18b2..7916ef18b2 100644 --- a/synapse/storage/data_stores/state/schema/delta/56/state_group_room_idx.sql +++ b/synapse/storage/databases/state/schema/delta/56/state_group_room_idx.sql
diff --git a/synapse/storage/data_stores/state/schema/full_schemas/54/full.sql b/synapse/storage/databases/state/schema/full_schemas/54/full.sql
index 35f97d6b3d..35f97d6b3d 100644 --- a/synapse/storage/data_stores/state/schema/full_schemas/54/full.sql +++ b/synapse/storage/databases/state/schema/full_schemas/54/full.sql
diff --git a/synapse/storage/data_stores/state/schema/full_schemas/54/sequence.sql.postgres b/synapse/storage/databases/state/schema/full_schemas/54/sequence.sql.postgres
index fcd926c9fb..fcd926c9fb 100644 --- a/synapse/storage/data_stores/state/schema/full_schemas/54/sequence.sql.postgres +++ b/synapse/storage/databases/state/schema/full_schemas/54/sequence.sql.postgres
diff --git a/synapse/storage/data_stores/state/store.py b/synapse/storage/databases/state/store.py
index 128c09a2cf..0e31cc811a 100644 --- a/synapse/storage/data_stores/state/store.py +++ b/synapse/storage/databases/state/store.py
@@ -17,16 +17,14 @@ import logging from collections import namedtuple from typing import Dict, Iterable, List, Set, Tuple -from twisted.internet import defer - from synapse.api.constants import EventTypes from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.state.bg_updates import StateBackgroundUpdateStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.databases.state.bg_updates import StateBackgroundUpdateStore from synapse.storage.state import StateFilter from synapse.storage.types import Cursor from synapse.storage.util.sequence import build_sequence_generator -from synapse.types import StateMap +from synapse.types import MutableStateMap, StateMap from synapse.util.caches.descriptors import cached from synapse.util.caches.dictionary_cache import DictionaryCache @@ -53,8 +51,8 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): """A data store for fetching/storing state groups. """ - def __init__(self, database: Database, db_conn, hs): - super(StateGroupDataStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) # Originally the state store used a single DictionaryCache to cache the # event IDs for the state types in a given state group to avoid hammering @@ -101,9 +99,12 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): self._state_group_seq_gen = build_sequence_generator( self.database_engine, get_max_state_group_txn, "state_group_id_seq" ) + self._state_group_seq_gen.check_consistency( + db_conn, table="state_groups", id_column="id" + ) @cached(max_entries=10000, iterable=True) - def get_state_group_delta(self, state_group): + async def get_state_group_delta(self, state_group): """Given a state group try to return a previous group and a delta between the old and the new. @@ -112,7 +113,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): """ def _get_state_group_delta_txn(txn): - prev_group = self.db.simple_select_one_onecol_txn( + prev_group = self.db_pool.simple_select_one_onecol_txn( txn, table="state_group_edges", keyvalues={"state_group": state_group}, @@ -123,7 +124,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): if not prev_group: return _GetStateGroupDelta(None, None) - delta_ids = self.db.simple_select_list_txn( + delta_ids = self.db_pool.simple_select_list_txn( txn, table="state_groups_state", keyvalues={"state_group": state_group}, @@ -135,14 +136,13 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): {(row["type"], row["state_key"]): row["event_id"] for row in delta_ids}, ) - return self.db.runInteraction( + return await self.db_pool.runInteraction( "get_state_group_delta", _get_state_group_delta_txn ) - @defer.inlineCallbacks - def _get_state_groups_from_groups( + async def _get_state_groups_from_groups( self, groups: List[int], state_filter: StateFilter - ): + ) -> Dict[int, StateMap[str]]: """Returns the state groups for a given set of groups from the database, filtering on types of state events. @@ -151,13 +151,13 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): state_filter: The state filter used to fetch state from the database. Returns: - Deferred[Dict[int, StateMap[str]]]: Dict of state group to state map. + Dict of state group to state map. """ results = {} chunks = [groups[i : i + 100] for i in range(0, len(groups), 100)] for chunk in chunks: - res = yield self.db.runInteraction( + res = await self.db_pool.runInteraction( "_get_state_groups_from_groups", self._get_state_groups_from_groups_txn, chunk, @@ -206,10 +206,9 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): return state_filter.filter_state(state_dict_ids), not missing_types - @defer.inlineCallbacks - def _get_state_for_groups( + async def _get_state_for_groups( self, groups: Iterable[int], state_filter: StateFilter = StateFilter.all() - ): + ) -> Dict[int, MutableStateMap[str]]: """Gets the state at each of a list of state groups, optionally filtering by type/state_key @@ -219,7 +218,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): state_filter: The state filter used to fetch state from the database. Returns: - Deferred[Dict[int, StateMap[str]]]: Dict of state group to state map. + Dict of state group to state map. """ member_filter, non_member_filter = state_filter.get_member_split() @@ -228,14 +227,11 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): ( non_member_state, incomplete_groups_nm, - ) = yield self._get_state_for_groups_using_cache( + ) = self._get_state_for_groups_using_cache( groups, self._state_group_cache, state_filter=non_member_filter ) - ( - member_state, - incomplete_groups_m, - ) = yield self._get_state_for_groups_using_cache( + (member_state, incomplete_groups_m,) = self._get_state_for_groups_using_cache( groups, self._state_group_members_cache, state_filter=member_filter ) @@ -256,7 +252,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): # Help the cache hit ratio by expanding the filter a bit db_state_filter = state_filter.return_expanded() - group_to_state_dict = yield self._get_state_groups_from_groups( + group_to_state_dict = await self._get_state_groups_from_groups( list(incomplete_groups), state_filter=db_state_filter ) @@ -372,9 +368,9 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): fetched_keys=non_member_types, ) - def store_state_group( + async def store_state_group( self, event_id, room_id, prev_group, delta_ids, current_state_ids - ): + ) -> int: """Store a new set of state, returning a newly assigned state group. Args: @@ -388,7 +384,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): to event_id. Returns: - Deferred[int]: The state group ID + The state group ID """ def _store_state_group_txn(txn): @@ -398,7 +394,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): state_group = self._state_group_seq_gen.get_next_id_txn(txn) - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="state_groups", values={"id": state_group, "room_id": room_id, "event_id": event_id}, @@ -407,7 +403,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): # We persist as a delta if we can, while also ensuring the chain # of deltas isn't tooo long, as otherwise read performance degrades. if prev_group: - is_in_db = self.db.simple_select_one_onecol_txn( + is_in_db = self.db_pool.simple_select_one_onecol_txn( txn, table="state_groups", keyvalues={"id": prev_group}, @@ -422,13 +418,13 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): potential_hops = self._count_state_group_hops_txn(txn, prev_group) if prev_group and potential_hops < MAX_STATE_DELTA_HOPS: - self.db.simple_insert_txn( + self.db_pool.simple_insert_txn( txn, table="state_group_edges", values={"state_group": state_group, "prev_state_group": prev_group}, ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="state_groups_state", values=[ @@ -443,7 +439,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): ], ) else: - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="state_groups_state", values=[ @@ -489,11 +485,13 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): return state_group - return self.db.runInteraction("store_state_group", _store_state_group_txn) + return await self.db_pool.runInteraction( + "store_state_group", _store_state_group_txn + ) - def purge_unreferenced_state_groups( + async def purge_unreferenced_state_groups( self, room_id: str, state_groups_to_delete - ) -> defer.Deferred: + ) -> None: """Deletes no longer referenced state groups and de-deltas any state groups that reference them. @@ -504,7 +502,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): to delete. """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "purge_unreferenced_state_groups", self._purge_unreferenced_state_groups, room_id, @@ -516,7 +514,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): "[purge] found %i state groups to delete", len(state_groups_to_delete) ) - rows = self.db.simple_select_many_txn( + rows = self.db_pool.simple_select_many_txn( txn, table="state_group_edges", column="prev_state_group", @@ -543,15 +541,15 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): curr_state = self._get_state_groups_from_groups_txn(txn, [sg]) curr_state = curr_state[sg] - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="state_groups_state", keyvalues={"state_group": sg} ) - self.db.simple_delete_txn( + self.db_pool.simple_delete_txn( txn, table="state_group_edges", keyvalues={"state_group": sg} ) - self.db.simple_insert_many_txn( + self.db_pool.simple_insert_many_txn( txn, table="state_groups_state", values=[ @@ -576,19 +574,19 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): ((sg,) for sg in state_groups_to_delete), ) - @defer.inlineCallbacks - def get_previous_state_groups(self, state_groups): + async def get_previous_state_groups( + self, state_groups: Iterable[int] + ) -> Dict[int, int]: """Fetch the previous groups of the given state groups. Args: - state_groups (Iterable[int]) + state_groups Returns: - Deferred[dict[int, int]]: mapping from state group to previous - state group. + A mapping from state group to previous state group. """ - rows = yield self.db.simple_select_many_batch( + rows = await self.db_pool.simple_select_many_batch( table="state_group_edges", column="prev_state_group", iterable=state_groups, @@ -599,7 +597,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): return {row["state_group"]: row["prev_state_group"] for row in rows} - def purge_room_state(self, room_id, state_groups_to_delete): + async def purge_room_state(self, room_id, state_groups_to_delete): """Deletes all record of a room from state tables Args: @@ -607,7 +605,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): state_groups_to_delete (list[int]): State groups to delete """ - return self.db.runInteraction( + await self.db_pool.runInteraction( "purge_room_state", self._purge_room_state_txn, room_id, @@ -618,7 +616,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): # first we have to delete the state groups states logger.info("[purge] removing %s from state_groups_state", room_id) - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="state_groups_state", column="state_group", @@ -629,7 +627,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): # ... and the state group edges logger.info("[purge] removing %s from state_group_edges", room_id) - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="state_group_edges", column="state_group", @@ -640,7 +638,7 @@ class StateGroupDataStore(StateBackgroundUpdateStore, SQLBaseStore): # ... and the state groups logger.info("[purge] removing %s from state_groups", room_id) - self.db.simple_delete_many_txn( + self.db_pool.simple_delete_many_txn( txn, table="state_groups", column="id", diff --git a/synapse/storage/engines/_base.py b/synapse/storage/engines/_base.py
index 908cbc79e3..d6d632dc10 100644 --- a/synapse/storage/engines/_base.py +++ b/synapse/storage/engines/_base.py
@@ -97,3 +97,20 @@ class BaseDatabaseEngine(Generic[ConnectionType], metaclass=abc.ABCMeta): """Gets a string giving the server version. For example: '3.22.0' """ ... + + @abc.abstractmethod + def in_transaction(self, conn: Connection) -> bool: + """Whether the connection is currently in a transaction. + """ + ... + + @abc.abstractmethod + def attempt_to_set_autocommit(self, conn: Connection, autocommit: bool): + """Attempt to set the connections autocommit mode. + + When True queries are run outside of transactions. + + Note: This has no effect on SQLite3, so callers still need to + commit/rollback the connections. + """ + ... diff --git a/synapse/storage/engines/postgres.py b/synapse/storage/engines/postgres.py
index ff39281f85..7719ac32f7 100644 --- a/synapse/storage/engines/postgres.py +++ b/synapse/storage/engines/postgres.py
@@ -15,7 +15,8 @@ import logging -from ._base import BaseDatabaseEngine, IncorrectDatabaseSetup +from synapse.storage.engines._base import BaseDatabaseEngine, IncorrectDatabaseSetup +from synapse.storage.types import Connection logger = logging.getLogger(__name__) @@ -119,6 +120,7 @@ class PostgresEngine(BaseDatabaseEngine): cursor.execute("SET synchronous_commit TO OFF") cursor.close() + db_conn.commit() @property def can_native_upsert(self): @@ -171,3 +173,9 @@ class PostgresEngine(BaseDatabaseEngine): return "%i.%i" % (numver / 10000, numver % 10000) else: return "%i.%i.%i" % (numver / 10000, (numver % 10000) / 100, numver % 100) + + def in_transaction(self, conn: Connection) -> bool: + return conn.status != self.module.extensions.STATUS_READY # type: ignore + + def attempt_to_set_autocommit(self, conn: Connection, autocommit: bool): + return conn.set_session(autocommit=autocommit) # type: ignore diff --git a/synapse/storage/engines/sqlite.py b/synapse/storage/engines/sqlite.py
index 8a0f8c89d1..5db0f0b520 100644 --- a/synapse/storage/engines/sqlite.py +++ b/synapse/storage/engines/sqlite.py
@@ -17,6 +17,7 @@ import threading import typing from synapse.storage.engines import BaseDatabaseEngine +from synapse.storage.types import Connection if typing.TYPE_CHECKING: import sqlite3 # noqa: F401 @@ -86,6 +87,7 @@ class Sqlite3Engine(BaseDatabaseEngine["sqlite3.Connection"]): db_conn.create_function("rank", 1, _rank) db_conn.execute("PRAGMA foreign_keys = ON;") + db_conn.commit() def is_deadlock(self, error): return False @@ -105,6 +107,14 @@ class Sqlite3Engine(BaseDatabaseEngine["sqlite3.Connection"]): """ return "%i.%i.%i" % self.module.sqlite_version_info + def in_transaction(self, conn: Connection) -> bool: + return conn.in_transaction # type: ignore + + def attempt_to_set_autocommit(self, conn: Connection, autocommit: bool): + # Twisted doesn't let us set attributes on the connections, so we can't + # set the connection to autocommit mode. + pass + # Following functions taken from: https://github.com/coleifer/peewee diff --git a/synapse/storage/keys.py b/synapse/storage/keys.py
index 4769b21529..afd10f7bae 100644 --- a/synapse/storage/keys.py +++ b/synapse/storage/keys.py
@@ -22,6 +22,6 @@ logger = logging.getLogger(__name__) @attr.s(slots=True, frozen=True) -class FetchKeyResult(object): +class FetchKeyResult: verify_key = attr.ib() # VerifyKey: the key itself valid_until_ts = attr.ib() # int: how long we can use this key for diff --git a/synapse/storage/persist_events.py b/synapse/storage/persist_events.py
index 78fbdcdee8..72939f3984 100644 --- a/synapse/storage/persist_events.py +++ b/synapse/storage/persist_events.py
@@ -18,20 +18,20 @@ import itertools import logging from collections import deque, namedtuple -from typing import Iterable, List, Optional, Set, Tuple +from typing import Dict, Iterable, List, Optional, Set, Tuple from prometheus_client import Counter, Histogram from twisted.internet import defer from synapse.api.constants import EventTypes, Membership -from synapse.events import FrozenEvent +from synapse.events import EventBase from synapse.events.snapshot import EventContext from synapse.logging.context import PreserveLoggingContext, make_deferred_yieldable from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.storage.data_stores import DataStores -from synapse.storage.data_stores.main.events import DeltaState -from synapse.types import StateMap +from synapse.storage.databases import Databases +from synapse.storage.databases.main.events import DeltaState +from synapse.types import Collection, PersistedEventPosition, RoomStreamToken, StateMap from synapse.util.async_helpers import ObservableDeferred from synapse.util.metrics import Measure @@ -69,7 +69,7 @@ stale_forward_extremities_counter = Histogram( ) -class _EventPeristenceQueue(object): +class _EventPeristenceQueue: """Queues up events so that they can be persisted in bulk with only one concurrent transaction per room. """ @@ -172,32 +172,34 @@ class _EventPeristenceQueue(object): pass -class EventsPersistenceStorage(object): +class EventsPersistenceStorage: """High level interface for handling persisting newly received events. Takes care of batching up events by room, and calculating the necessary current state and forward extremity changes. """ - def __init__(self, hs, stores: DataStores): + def __init__(self, hs, stores: Databases): # We ultimately want to split out the state store from the main store, # so we use separate variables here even though they point to the same # store for now. self.main_store = stores.main self.state_store = stores.state + + assert stores.persist_events self.persist_events_store = stores.persist_events self._clock = hs.get_clock() + self._instance_name = hs.get_instance_name() self.is_mine_id = hs.is_mine_id self._event_persist_queue = _EventPeristenceQueue() self._state_resolution_handler = hs.get_state_resolution_handler() - @defer.inlineCallbacks - def persist_events( + async def persist_events( self, - events_and_contexts: List[Tuple[FrozenEvent, EventContext]], + events_and_contexts: Iterable[Tuple[EventBase, EventContext]], backfilled: bool = False, - ): + ) -> RoomStreamToken: """ Write events to the database Args: @@ -207,9 +209,9 @@ class EventsPersistenceStorage(object): which might update the current state etc. Returns: - Deferred[int]: the stream ordering of the latest persisted event + the stream ordering of the latest persisted event """ - partitioned = {} + partitioned = {} # type: Dict[str, List[Tuple[EventBase, EventContext]]] for event, ctx in events_and_contexts: partitioned.setdefault(event.room_id, []).append((event, ctx)) @@ -223,22 +225,19 @@ class EventsPersistenceStorage(object): for room_id in partitioned: self._maybe_start_persisting(room_id) - yield make_deferred_yieldable( + await make_deferred_yieldable( defer.gatherResults(deferreds, consumeErrors=True) ) - max_persisted_id = yield self.main_store.get_current_events_token() - - return max_persisted_id + return self.main_store.get_room_max_token() - @defer.inlineCallbacks - def persist_event( - self, event: FrozenEvent, context: EventContext, backfilled: bool = False - ): + async def persist_event( + self, event: EventBase, context: EventContext, backfilled: bool = False + ) -> Tuple[PersistedEventPosition, RoomStreamToken]: """ Returns: - Deferred[Tuple[int, int]]: the stream ordering of ``event``, - and the stream ordering of the latest persisted event + The stream ordering of `event`, and the stream ordering of the + latest persisted event """ deferred = self._event_persist_queue.add_to_queue( event.room_id, [(event, context)], backfilled=backfilled @@ -246,10 +245,12 @@ class EventsPersistenceStorage(object): self._maybe_start_persisting(event.room_id) - yield make_deferred_yieldable(deferred) + await make_deferred_yieldable(deferred) + + event_stream_id = event.internal_metadata.stream_ordering - max_persisted_id = yield self.main_store.get_current_events_token() - return (event.internal_metadata.stream_ordering, max_persisted_id) + pos = PersistedEventPosition(self._instance_name, event_stream_id) + return pos, self.main_store.get_room_max_token() def _maybe_start_persisting(self, room_id: str): async def persisting_queue(item): @@ -262,7 +263,7 @@ class EventsPersistenceStorage(object): async def _persist_events( self, - events_and_contexts: List[Tuple[FrozenEvent, EventContext]], + events_and_contexts: List[Tuple[EventBase, EventContext]], backfilled: bool = False, ): """Calculates the change to current state and forward extremities, and @@ -309,7 +310,9 @@ class EventsPersistenceStorage(object): # Work out the new "current state" for each room. # We do this by working out what the new extremities are and then # calculating the state from that. - events_by_room = {} + events_by_room = ( + {} + ) # type: Dict[str, List[Tuple[EventBase, EventContext]]] for event, context in chunk: events_by_room.setdefault(event.room_id, []).append( (event, context) @@ -439,8 +442,8 @@ class EventsPersistenceStorage(object): async def _calculate_new_extremities( self, room_id: str, - event_contexts: List[Tuple[FrozenEvent, EventContext]], - latest_event_ids: List[str], + event_contexts: List[Tuple[EventBase, EventContext]], + latest_event_ids: Collection[str], ): """Calculates the new forward extremities for a room given events to persist. @@ -474,7 +477,7 @@ class EventsPersistenceStorage(object): # Remove any events which are prev_events of any existing events. existing_prevs = await self.persist_events_store._get_events_which_are_prevs( result - ) + ) # type: Collection[str] result.difference_update(existing_prevs) # Finally handle the case where the new events have soft-failed prev @@ -497,7 +500,7 @@ class EventsPersistenceStorage(object): async def _get_new_state_after_events( self, room_id: str, - events_context: List[Tuple[FrozenEvent, EventContext]], + events_context: List[Tuple[EventBase, EventContext]], old_latest_event_ids: Iterable[str], new_latest_event_ids: Iterable[str], ) -> Tuple[Optional[StateMap[str]], Optional[StateMap[str]]]: @@ -683,7 +686,7 @@ class EventsPersistenceStorage(object): async def _is_server_still_joined( self, room_id: str, - ev_ctx_rm: List[Tuple[FrozenEvent, EventContext]], + ev_ctx_rm: List[Tuple[EventBase, EventContext]], delta: DeltaState, current_state: Optional[StateMap[str]], potentially_left_users: Set[str], diff --git a/synapse/storage/prepare_database.py b/synapse/storage/prepare_database.py
index 9cc3b51fe6..4957e77f4c 100644 --- a/synapse/storage/prepare_database.py +++ b/synapse/storage/prepare_database.py
@@ -19,12 +19,15 @@ import logging import os import re from collections import Counter -from typing import TextIO +from typing import Optional, TextIO import attr +from synapse.config.homeserver import HomeServerConfig +from synapse.storage.engines import BaseDatabaseEngine from synapse.storage.engines.postgres import PostgresEngine -from synapse.storage.types import Cursor +from synapse.storage.types import Connection, Cursor +from synapse.types import Collection logger = logging.getLogger(__name__) @@ -47,8 +50,29 @@ class UpgradeDatabaseException(PrepareDatabaseException): pass -def prepare_database(db_conn, database_engine, config, data_stores=["main", "state"]): - """Prepares a database for usage. Will either create all necessary tables +OUTDATED_SCHEMA_ON_WORKER_ERROR = ( + "Expected database schema version %i but got %i: run the main synapse process to " + "upgrade the database schema before starting worker processes." +) + +EMPTY_DATABASE_ON_WORKER_ERROR = ( + "Uninitialised database: run the main synapse process to prepare the database " + "schema before starting worker processes." +) + +UNAPPLIED_DELTA_ON_WORKER_ERROR = ( + "Database schema delta %s has not been applied: run the main synapse process to " + "upgrade the database schema before starting worker processes." +) + + +def prepare_database( + db_conn: Connection, + database_engine: BaseDatabaseEngine, + config: Optional[HomeServerConfig], + databases: Collection[str] = ["main", "state"], +): + """Prepares a physical database for usage. Will either create all necessary tables or upgrade from an older schema version. If `config` is None then prepare_database will assert that no upgrade is @@ -57,40 +81,68 @@ def prepare_database(db_conn, database_engine, config, data_stores=["main", "sta Args: db_conn: database_engine: - config (synapse.config.homeserver.HomeServerConfig|None): + config : application config, or None if we are connecting to an existing database which we expect to be configured already - data_stores (list[str]): The name of the data stores that will be used - with this database. Defaults to all data stores. + databases: The name of the databases that will be used + with this physical database. Defaults to all databases. """ try: cur = db_conn.cursor() + + # sqlite does not automatically start transactions for DDL / SELECT statements, + # so we start one before running anything. This ensures that any upgrades + # are either applied completely, or not at all. + # + # (psycopg2 automatically starts a transaction as soon as we run any statements + # at all, so this is redundant but harmless there.) + cur.execute("BEGIN TRANSACTION") + + logger.info("%r: Checking existing schema version", databases) version_info = _get_or_create_schema_state(cur, database_engine) if version_info: user_version, delta_files, upgraded = version_info + logger.info( + "%r: Existing schema is %i (+%i deltas)", + databases, + user_version, + len(delta_files), + ) + # config should only be None when we are preparing an in-memory SQLite db, + # which should be empty. if config is None: - if user_version != SCHEMA_VERSION: - # If we don't pass in a config file then we are expecting to - # have already upgraded the DB. - raise UpgradeDatabaseException( - "Expected database schema version %i but got %i" - % (SCHEMA_VERSION, user_version) - ) - else: - _upgrade_existing_database( - cur, - user_version, - delta_files, - upgraded, - database_engine, - config, - data_stores=data_stores, + raise ValueError( + "config==None in prepare_database, but databse is not empty" + ) + + # if it's a worker app, refuse to upgrade the database, to avoid multiple + # workers doing it at once. + if config.worker_app is not None and user_version != SCHEMA_VERSION: + raise UpgradeDatabaseException( + OUTDATED_SCHEMA_ON_WORKER_ERROR % (SCHEMA_VERSION, user_version) ) + + _upgrade_existing_database( + cur, + user_version, + delta_files, + upgraded, + database_engine, + config, + databases=databases, + ) else: - _setup_new_database(cur, database_engine, data_stores=data_stores) + logger.info("%r: Initialising new database", databases) + + # if it's a worker app, refuse to upgrade the database, to avoid multiple + # workers doing it at once. + if config and config.worker_app is not None: + raise UpgradeDatabaseException(EMPTY_DATABASE_ON_WORKER_ERROR) + + _setup_new_database(cur, database_engine, databases=databases) # check if any of our configured dynamic modules want a database if config is not None: @@ -103,9 +155,9 @@ def prepare_database(db_conn, database_engine, config, data_stores=["main", "sta raise -def _setup_new_database(cur, database_engine, data_stores): - """Sets up the database by finding a base set of "full schemas" and then - applying any necessary deltas, including schemas from the given data +def _setup_new_database(cur, database_engine, databases): + """Sets up the physical database by finding a base set of "full schemas" and + then applying any necessary deltas, including schemas from the given data stores. The "full_schemas" directory has subdirectories named after versions. This @@ -138,8 +190,8 @@ def _setup_new_database(cur, database_engine, data_stores): Args: cur (Cursor): a database cursor database_engine (DatabaseEngine) - data_stores (list[str]): The names of the data stores to instantiate - on the given database. + databases (list[str]): The names of the databases to instantiate + on the given physical database. """ # We're about to set up a brand new database so we check that its @@ -176,13 +228,13 @@ def _setup_new_database(cur, database_engine, data_stores): directories.extend( os.path.join( dir_path, - "data_stores", - data_store, + "databases", + database, "schema", "full_schemas", str(max_current_ver), ) - for data_store in data_stores + for database in databases ) directory_entries = [] @@ -219,7 +271,7 @@ def _setup_new_database(cur, database_engine, data_stores): upgraded=False, database_engine=database_engine, config=None, - data_stores=data_stores, + databases=databases, is_empty=True, ) @@ -231,10 +283,10 @@ def _upgrade_existing_database( upgraded, database_engine, config, - data_stores, + databases, is_empty=False, ): - """Upgrades an existing database. + """Upgrades an existing physical database. Delta files can either be SQL stored in *.sql files, or python modules in *.py. @@ -285,8 +337,8 @@ def _upgrade_existing_database( config (synapse.config.homeserver.HomeServerConfig|None): None if we are initialising a blank database, otherwise the application config - data_stores (list[str]): The names of the data stores to instantiate - on the given database. + databases (list[str]): The names of the databases to instantiate + on the given physical database. is_empty (bool): Is this a blank database? I.e. do we need to run the upgrade portions of the delta scripts. """ @@ -295,6 +347,8 @@ def _upgrade_existing_database( else: assert config + is_worker = config and config.worker_app is not None + if current_version > SCHEMA_VERSION: raise ValueError( "Cannot use this database as it is too " @@ -303,8 +357,8 @@ def _upgrade_existing_database( # some of the deltas assume that config.server_name is set correctly, so now # is a good time to run the sanity check. - if not is_empty and "main" in data_stores: - from synapse.storage.data_stores.main import check_database_before_upgrade + if not is_empty and "main" in databases: + from synapse.storage.databases.main import check_database_before_upgrade check_database_before_upgrade(cur, database_engine, config) @@ -322,7 +376,7 @@ def _upgrade_existing_database( specific_engine_extensions = (".sqlite", ".postgres") for v in range(start_ver, SCHEMA_VERSION + 1): - logger.info("Upgrading schema to v%d", v) + logger.info("Applying schema deltas for v%d", v) # We need to search both the global and per data store schema # directories for schema updates. @@ -330,11 +384,9 @@ def _upgrade_existing_database( # First we find the directories to search in delta_dir = os.path.join(dir_path, "schema", "delta", str(v)) directories = [delta_dir] - for data_store in data_stores: + for database in databases: directories.append( - os.path.join( - dir_path, "data_stores", data_store, "schema", "delta", str(v) - ) + os.path.join(dir_path, "databases", database, "schema", "delta", str(v)) ) # Used to check if we have any duplicate file names @@ -384,9 +436,15 @@ def _upgrade_existing_database( continue root_name, ext = os.path.splitext(file_name) + if ext == ".py": # This is a python upgrade module. We need to import into some # package and then execute its `run_upgrade` function. + if is_worker: + raise PrepareDatabaseException( + UNAPPLIED_DELTA_ON_WORKER_ERROR % relative_path + ) + module_name = "synapse.storage.v%d_%s" % (v, root_name) with open(absolute_path) as python_file: module = imp.load_source(module_name, absolute_path, python_file) @@ -401,10 +459,18 @@ def _upgrade_existing_database( continue elif ext == ".sql": # A plain old .sql file, just read and execute it + if is_worker: + raise PrepareDatabaseException( + UNAPPLIED_DELTA_ON_WORKER_ERROR % relative_path + ) logger.info("Applying schema %s", relative_path) executescript(cur, absolute_path) elif ext == specific_engine_extension and root_name.endswith(".sql"): # A .sql file specific to our engine; just read and execute it + if is_worker: + raise PrepareDatabaseException( + UNAPPLIED_DELTA_ON_WORKER_ERROR % relative_path + ) logger.info("Applying engine-specific schema %s", relative_path) executescript(cur, absolute_path) elif ext in specific_engine_extensions and root_name.endswith(".sql"): @@ -434,6 +500,8 @@ def _upgrade_existing_database( (v, True), ) + logger.info("Schema now up to date") + def _apply_module_schemas(txn, database_engine, config): """Apply the module schemas for the dynamic modules, if any @@ -570,8 +638,8 @@ def _get_or_create_schema_state(txn, database_engine): return None -@attr.s() -class _DirectoryListing(object): +@attr.s(slots=True) +class _DirectoryListing: """Helper class to store schema file name and the absolute path to it. diff --git a/synapse/storage/purge_events.py b/synapse/storage/purge_events.py
index fdc0abf5cf..bfa0a9fd06 100644 --- a/synapse/storage/purge_events.py +++ b/synapse/storage/purge_events.py
@@ -15,62 +15,60 @@ import itertools import logging - -from twisted.internet import defer +from typing import Set logger = logging.getLogger(__name__) -class PurgeEventsStorage(object): +class PurgeEventsStorage: """High level interface for purging rooms and event history. """ def __init__(self, hs, stores): self.stores = stores - @defer.inlineCallbacks - def purge_room(self, room_id: str): + async def purge_room(self, room_id: str): """Deletes all record of a room """ - state_groups_to_delete = yield self.stores.main.purge_room(room_id) - yield self.stores.state.purge_room_state(room_id, state_groups_to_delete) + state_groups_to_delete = await self.stores.main.purge_room(room_id) + await self.stores.state.purge_room_state(room_id, state_groups_to_delete) - @defer.inlineCallbacks - def purge_history(self, room_id, token, delete_local_events): + async def purge_history( + self, room_id: str, token: str, delete_local_events: bool + ) -> None: """Deletes room history before a certain point Args: - room_id (str): + room_id: The room ID - token (str): A topological token to delete events before + token: A topological token to delete events before - delete_local_events (bool): + delete_local_events: if True, we will delete local events as well as remote ones (instead of just marking them as outliers and deleting their state groups). """ - state_groups = yield self.stores.main.purge_history( + state_groups = await self.stores.main.purge_history( room_id, token, delete_local_events ) logger.info("[purge] finding state groups that can be deleted") - sg_to_delete = yield self._find_unreferenced_groups(state_groups) + sg_to_delete = await self._find_unreferenced_groups(state_groups) - yield self.stores.state.purge_unreferenced_state_groups(room_id, sg_to_delete) + await self.stores.state.purge_unreferenced_state_groups(room_id, sg_to_delete) - @defer.inlineCallbacks - def _find_unreferenced_groups(self, state_groups): + async def _find_unreferenced_groups(self, state_groups: Set[int]) -> Set[int]: """Used when purging history to figure out which state groups can be deleted. Args: - state_groups (set[int]): Set of state groups referenced by events + state_groups: Set of state groups referenced by events that are going to be deleted. Returns: - Deferred[set[int]] The set of state groups that can be deleted. + The set of state groups that can be deleted. """ # Graph of state group -> previous group graph = {} @@ -93,7 +91,7 @@ class PurgeEventsStorage(object): current_search = set(itertools.islice(next_to_search, 100)) next_to_search -= current_search - referenced = yield self.stores.main.get_referenced_state_groups( + referenced = await self.stores.main.get_referenced_state_groups( current_search ) referenced_groups |= referenced @@ -102,7 +100,7 @@ class PurgeEventsStorage(object): # groups that are referenced. current_search -= referenced - edges = yield self.stores.state.get_previous_state_groups(current_search) + edges = await self.stores.state.get_previous_state_groups(current_search) prevs = set(edges.values()) # We don't bother re-handling groups we've already seen diff --git a/synapse/storage/relations.py b/synapse/storage/relations.py
index d471ec9860..cec96ad6a7 100644 --- a/synapse/storage/relations.py +++ b/synapse/storage/relations.py
@@ -22,8 +22,8 @@ from synapse.api.errors import SynapseError logger = logging.getLogger(__name__) -@attr.s -class PaginationChunk(object): +@attr.s(slots=True) +class PaginationChunk: """Returned by relation pagination APIs. Attributes: @@ -51,7 +51,7 @@ class PaginationChunk(object): @attr.s(frozen=True, slots=True) -class RelationPaginationToken(object): +class RelationPaginationToken: """Pagination token for relation pagination API. As the results are in topological order, we can use the @@ -82,7 +82,7 @@ class RelationPaginationToken(object): @attr.s(frozen=True, slots=True) -class AggregationPaginationToken(object): +class AggregationPaginationToken: """Pagination token for relation aggregation pagination API. As the results are order by count and then MAX(stream_ordering) of the diff --git a/synapse/storage/roommember.py b/synapse/storage/roommember.py
index 8c4a83a840..f152f63321 100644 --- a/synapse/storage/roommember.py +++ b/synapse/storage/roommember.py
@@ -25,7 +25,7 @@ RoomsForUser = namedtuple( ) GetRoomsForUserWithStreamOrdering = namedtuple( - "_GetRoomsForUserWithStreamOrdering", ("room_id", "stream_ordering") + "_GetRoomsForUserWithStreamOrdering", ("room_id", "event_pos") ) diff --git a/synapse/storage/state.py b/synapse/storage/state.py
index dc568476f4..08a69f2f96 100644 --- a/synapse/storage/state.py +++ b/synapse/storage/state.py
@@ -14,14 +14,13 @@ # limitations under the License. import logging -from typing import Iterable, List, TypeVar +from typing import Awaitable, Dict, Iterable, List, Optional, Set, Tuple, TypeVar import attr -from twisted.internet import defer - from synapse.api.constants import EventTypes -from synapse.types import StateMap +from synapse.events import EventBase +from synapse.types import MutableStateMap, StateMap logger = logging.getLogger(__name__) @@ -30,20 +29,20 @@ T = TypeVar("T") @attr.s(slots=True) -class StateFilter(object): +class StateFilter: """A filter used when querying for state. Attributes: - types (dict[str, set[str]|None]): Map from type to set of state keys (or - None). This specifies which state_keys for the given type to fetch - from the DB. If None then all events with that type are fetched. If - the set is empty then no events with that type are fetched. - include_others (bool): Whether to fetch events with types that do not + types: Map from type to set of state keys (or None). This specifies + which state_keys for the given type to fetch from the DB. If None + then all events with that type are fetched. If the set is empty + then no events with that type are fetched. + include_others: Whether to fetch events with types that do not appear in `types`. """ - types = attr.ib() - include_others = attr.ib(default=False) + types = attr.ib(type=Dict[str, Optional[Set[str]]]) + include_others = attr.ib(default=False, type=bool) def __attrs_post_init__(self): # If `include_others` is set we canonicalise the filter by removing @@ -52,36 +51,35 @@ class StateFilter(object): self.types = {k: v for k, v in self.types.items() if v is not None} @staticmethod - def all(): + def all() -> "StateFilter": """Creates a filter that fetches everything. Returns: - StateFilter + The new state filter. """ return StateFilter(types={}, include_others=True) @staticmethod - def none(): + def none() -> "StateFilter": """Creates a filter that fetches nothing. Returns: - StateFilter + The new state filter. """ return StateFilter(types={}, include_others=False) @staticmethod - def from_types(types): + def from_types(types: Iterable[Tuple[str, Optional[str]]]) -> "StateFilter": """Creates a filter that only fetches the given types Args: - types (Iterable[tuple[str, str|None]]): A list of type and state - keys to fetch. A state_key of None fetches everything for - that type + types: A list of type and state keys to fetch. A state_key of None + fetches everything for that type Returns: - StateFilter + The new state filter. """ - type_dict = {} + type_dict = {} # type: Dict[str, Optional[Set[str]]] for typ, s in types: if typ in type_dict: if type_dict[typ] is None: @@ -91,24 +89,24 @@ class StateFilter(object): type_dict[typ] = None continue - type_dict.setdefault(typ, set()).add(s) + type_dict.setdefault(typ, set()).add(s) # type: ignore return StateFilter(types=type_dict) @staticmethod - def from_lazy_load_member_list(members): + def from_lazy_load_member_list(members: Iterable[str]) -> "StateFilter": """Creates a filter that returns all non-member events, plus the member events for the given users Args: - members (iterable[str]): Set of user IDs + members: Set of user IDs Returns: - StateFilter + The new state filter """ return StateFilter(types={EventTypes.Member: set(members)}, include_others=True) - def return_expanded(self): + def return_expanded(self) -> "StateFilter": """Creates a new StateFilter where type wild cards have been removed (except for memberships). The returned filter is a superset of the current one, i.e. anything that passes the current filter will pass @@ -130,7 +128,7 @@ class StateFilter(object): return all non-member events Returns: - StateFilter + The new state filter. """ if self.is_full(): @@ -167,7 +165,7 @@ class StateFilter(object): include_others=True, ) - def make_sql_filter_clause(self): + def make_sql_filter_clause(self) -> Tuple[str, List[str]]: """Converts the filter to an SQL clause. For example: @@ -179,13 +177,12 @@ class StateFilter(object): Returns: - tuple[str, list]: The SQL string (may be empty) and arguments. An - empty SQL string is returned when the filter matches everything - (i.e. is "full"). + The SQL string (may be empty) and arguments. An empty SQL string is + returned when the filter matches everything (i.e. is "full"). """ where_clause = "" - where_args = [] + where_args = [] # type: List[str] if self.is_full(): return where_clause, where_args @@ -221,7 +218,7 @@ class StateFilter(object): return where_clause, where_args - def max_entries_returned(self): + def max_entries_returned(self) -> Optional[int]: """Returns the maximum number of entries this filter will return if known, otherwise returns None. @@ -260,33 +257,33 @@ class StateFilter(object): return filtered_state - def is_full(self): + def is_full(self) -> bool: """Whether this filter fetches everything or not Returns: - bool + True if the filter fetches everything. """ return self.include_others and not self.types - def has_wildcards(self): + def has_wildcards(self) -> bool: """Whether the filter includes wildcards or is attempting to fetch specific state. Returns: - bool + True if the filter includes wildcards. """ return self.include_others or any( state_keys is None for state_keys in self.types.values() ) - def concrete_types(self): + def concrete_types(self) -> List[Tuple[str, str]]: """Returns a list of concrete type/state_keys (i.e. not None) that will be fetched. This will be a complete list if `has_wildcards` returns False, but otherwise will be a subset (or even empty). Returns: - list[tuple[str,str]] + A list of type/state_keys tuples. """ return [ (t, s) @@ -295,7 +292,7 @@ class StateFilter(object): for s in state_keys ] - def get_member_split(self): + def get_member_split(self) -> Tuple["StateFilter", "StateFilter"]: """Return the filter split into two: one which assumes it's exclusively matching against member state, and one which assumes it's matching against non member state. @@ -307,7 +304,7 @@ class StateFilter(object): state caches). Returns: - tuple[StateFilter, StateFilter]: The member and non member filters + The member and non member filters """ if EventTypes.Member in self.types: @@ -329,73 +326,80 @@ class StateFilter(object): return member_filter, non_member_filter -class StateGroupStorage(object): +class StateGroupStorage: """High level interface to fetching state for event. """ def __init__(self, hs, stores): self.stores = stores - def get_state_group_delta(self, state_group: int): + async def get_state_group_delta(self, state_group: int): """Given a state group try to return a previous group and a delta between the old and the new. + Args: + state_group: The state group used to retrieve state deltas. + Returns: - Deferred[Tuple[Optional[int], Optional[StateMap[str]]]]: + Tuple[Optional[int], Optional[StateMap[str]]]: (prev_group, delta_ids) """ - return self.stores.state.get_state_group_delta(state_group) + return await self.stores.state.get_state_group_delta(state_group) - @defer.inlineCallbacks - def get_state_groups_ids(self, _room_id, event_ids): + async def get_state_groups_ids( + self, _room_id: str, event_ids: Iterable[str] + ) -> Dict[int, MutableStateMap[str]]: """Get the event IDs of all the state for the state groups for the given events Args: - _room_id (str): id of the room for these events - event_ids (iterable[str]): ids of the events + _room_id: id of the room for these events + event_ids: ids of the events Returns: - Deferred[dict[int, StateMap[str]]]: - dict of state_group_id -> (dict of (type, state_key) -> event id) + dict of state_group_id -> (dict of (type, state_key) -> event id) """ if not event_ids: return {} - event_to_groups = yield self.stores.main._get_state_group_for_events(event_ids) + event_to_groups = await self.stores.main._get_state_group_for_events(event_ids) groups = set(event_to_groups.values()) - group_to_state = yield self.stores.state._get_state_for_groups(groups) + group_to_state = await self.stores.state._get_state_for_groups(groups) return group_to_state - @defer.inlineCallbacks - def get_state_ids_for_group(self, state_group): + async def get_state_ids_for_group(self, state_group: int) -> StateMap[str]: """Get the event IDs of all the state in the given state group Args: - state_group (int) + state_group: A state group for which we want to get the state IDs. Returns: - Deferred[dict]: Resolves to a map of (type, state_key) -> event_id + Resolves to a map of (type, state_key) -> event_id """ - group_to_state = yield self._get_state_for_groups((state_group,)) + group_to_state = await self._get_state_for_groups((state_group,)) return group_to_state[state_group] - @defer.inlineCallbacks - def get_state_groups(self, room_id, event_ids): + async def get_state_groups( + self, room_id: str, event_ids: Iterable[str] + ) -> Dict[int, List[EventBase]]: """ Get the state groups for the given list of event_ids + + Args: + room_id: ID of the room for these events. + event_ids: The event IDs to retrieve state for. + Returns: - Deferred[dict[int, list[EventBase]]]: - dict of state_group_id -> list of state events. + dict of state_group_id -> list of state events. """ if not event_ids: return {} - group_to_ids = yield self.get_state_groups_ids(room_id, event_ids) + group_to_ids = await self.get_state_groups_ids(room_id, event_ids) - state_event_map = yield self.stores.main.get_events( + state_event_map = await self.stores.main.get_events( [ ev_id for group_ids in group_to_ids.values() @@ -415,7 +419,7 @@ class StateGroupStorage(object): def _get_state_groups_from_groups( self, groups: List[int], state_filter: StateFilter - ): + ) -> Awaitable[Dict[int, StateMap[str]]]: """Returns the state groups for a given set of groups, filtering on types of state events. @@ -423,31 +427,34 @@ class StateGroupStorage(object): groups: list of state group IDs to query state_filter: The state filter used to fetch state from the database. + Returns: - Deferred[Dict[int, StateMap[str]]]: Dict of state group to state map. + Dict of state group to state map. """ return self.stores.state._get_state_groups_from_groups(groups, state_filter) - @defer.inlineCallbacks - def get_state_for_events(self, event_ids, state_filter=StateFilter.all()): + async def get_state_for_events( + self, event_ids: List[str], state_filter: StateFilter = StateFilter.all() + ): """Given a list of event_ids and type tuples, return a list of state dicts for each event. + Args: - event_ids (list[string]) - state_filter (StateFilter): The state filter used to fetch state - from the database. + event_ids: The events to fetch the state of. + state_filter: The state filter used to fetch state. + Returns: - deferred: A dict of (event_id) -> (type, state_key) -> [state_events] + A dict of (event_id) -> (type, state_key) -> [state_events] """ - event_to_groups = yield self.stores.main._get_state_group_for_events(event_ids) + event_to_groups = await self.stores.main._get_state_group_for_events(event_ids) groups = set(event_to_groups.values()) - group_to_state = yield self.stores.state._get_state_for_groups( + group_to_state = await self.stores.state._get_state_for_groups( groups, state_filter ) - state_event_map = yield self.stores.main.get_events( + state_event_map = await self.stores.main.get_events( [ev_id for sd in group_to_state.values() for ev_id in sd.values()], get_prev_content=False, ) @@ -463,24 +470,24 @@ class StateGroupStorage(object): return {event: event_to_state[event] for event in event_ids} - @defer.inlineCallbacks - def get_state_ids_for_events(self, event_ids, state_filter=StateFilter.all()): + async def get_state_ids_for_events( + self, event_ids: List[str], state_filter: StateFilter = StateFilter.all() + ): """ Get the state dicts corresponding to a list of events, containing the event_ids of the state events (as opposed to the events themselves) Args: - event_ids(list(str)): events whose state should be returned - state_filter (StateFilter): The state filter used to fetch state - from the database. + event_ids: events whose state should be returned + state_filter: The state filter used to fetch state from the database. Returns: - A deferred dict from event_id -> (type, state_key) -> event_id + A dict from event_id -> (type, state_key) -> event_id """ - event_to_groups = yield self.stores.main._get_state_group_for_events(event_ids) + event_to_groups = await self.stores.main._get_state_group_for_events(event_ids) groups = set(event_to_groups.values()) - group_to_state = yield self.stores.state._get_state_for_groups( + group_to_state = await self.stores.state._get_state_for_groups( groups, state_filter ) @@ -491,72 +498,77 @@ class StateGroupStorage(object): return {event: event_to_state[event] for event in event_ids} - @defer.inlineCallbacks - def get_state_for_event(self, event_id, state_filter=StateFilter.all()): + async def get_state_for_event( + self, event_id: str, state_filter: StateFilter = StateFilter.all() + ): """ Get the state dict corresponding to a particular event Args: - event_id(str): event whose state should be returned - state_filter (StateFilter): The state filter used to fetch state - from the database. + event_id: event whose state should be returned + state_filter: The state filter used to fetch state from the database. Returns: - A deferred dict from (type, state_key) -> state_event + A dict from (type, state_key) -> state_event """ - state_map = yield self.get_state_for_events([event_id], state_filter) + state_map = await self.get_state_for_events([event_id], state_filter) return state_map[event_id] - @defer.inlineCallbacks - def get_state_ids_for_event(self, event_id, state_filter=StateFilter.all()): + async def get_state_ids_for_event( + self, event_id: str, state_filter: StateFilter = StateFilter.all() + ): """ Get the state dict corresponding to a particular event Args: - event_id(str): event whose state should be returned - state_filter (StateFilter): The state filter used to fetch state - from the database. + event_id: event whose state should be returned + state_filter: The state filter used to fetch state from the database. Returns: - A deferred dict from (type, state_key) -> state_event + A dict from (type, state_key) -> state_event """ - state_map = yield self.get_state_ids_for_events([event_id], state_filter) + state_map = await self.get_state_ids_for_events([event_id], state_filter) return state_map[event_id] def _get_state_for_groups( self, groups: Iterable[int], state_filter: StateFilter = StateFilter.all() - ): + ) -> Awaitable[Dict[int, MutableStateMap[str]]]: """Gets the state at each of a list of state groups, optionally filtering by type/state_key Args: - groups (iterable[int]): list of state groups for which we want - to get the state. - state_filter (StateFilter): The state filter used to fetch state + groups: list of state groups for which we want to get the state. + state_filter: The state filter used to fetch state. from the database. + Returns: - Deferred[dict[int, StateMap[str]]]: Dict of state group to state map. + Dict of state group to state map. """ return self.stores.state._get_state_for_groups(groups, state_filter) - def store_state_group( - self, event_id, room_id, prev_group, delta_ids, current_state_ids - ): + async def store_state_group( + self, + event_id: str, + room_id: str, + prev_group: Optional[int], + delta_ids: Optional[dict], + current_state_ids: dict, + ) -> int: """Store a new set of state, returning a newly assigned state group. Args: - event_id (str): The event ID for which the state was calculated - room_id (str) - prev_group (int|None): A previous state group for the room, optional. - delta_ids (dict|None): The delta between state at `prev_group` and + event_id: The event ID for which the state was calculated. + room_id: ID of the room for which the state was calculated. + prev_group: A previous state group for the room, optional. + delta_ids: The delta between state at `prev_group` and `current_state_ids`, if `prev_group` was given. Same format as `current_state_ids`. - current_state_ids (dict): The state to store. Map of (type, state_key) + current_state_ids: The state to store. Map of (type, state_key) to event_id. Returns: - Deferred[int]: The state group ID + The state group ID """ - return self.stores.state.store_state_group( + return await self.stores.state.store_state_group( event_id, room_id, prev_group, delta_ids, current_state_ids ) diff --git a/synapse/storage/util/id_generators.py b/synapse/storage/util/id_generators.py
index 787cebfbec..eccd2d5b7b 100644 --- a/synapse/storage/util/id_generators.py +++ b/synapse/storage/util/id_generators.py
@@ -12,19 +12,25 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - -import contextlib +import heapq +import logging import threading from collections import deque -from typing import Dict, Set, Tuple +from contextlib import contextmanager +from typing import Dict, List, Optional, Set, Union +import attr from typing_extensions import Deque -from synapse.storage.database import Database, LoggingTransaction +from synapse.metrics.background_process_metrics import run_as_background_process +from synapse.storage.database import DatabasePool, LoggingTransaction +from synapse.storage.types import Cursor from synapse.storage.util.sequence import PostgresSequenceGenerator +logger = logging.getLogger(__name__) + -class IdGenerator(object): +class IdGenerator: def __init__(self, db_conn, table, column): self._lock = threading.Lock() self._next_id = _load_current_id(db_conn, table, column) @@ -47,6 +53,8 @@ def _load_current_id(db_conn, table, column, step=1): Returns: int """ + # debug logging for https://github.com/matrix-org/synapse/issues/7968 + logger.info("initialising stream generator for %s(%s)", table, column) cur = db_conn.cursor() if step == 1: cur.execute("SELECT MAX(%s) FROM %s" % (column, table)) @@ -58,7 +66,7 @@ def _load_current_id(db_conn, table, column, step=1): return (max if step > 0 else min)(current_id, step) -class StreamIdGenerator(object): +class StreamIdGenerator: """Used to generate new stream ids when persisting events while keeping track of which transactions have been completed. @@ -80,7 +88,7 @@ class StreamIdGenerator(object): upwards, -1 to grow downwards. Usage: - with stream_id_gen.get_next() as stream_id: + async with stream_id_gen.get_next() as stream_id: # ... persist event ... """ @@ -98,7 +106,7 @@ class StreamIdGenerator(object): def get_next(self): """ Usage: - with stream_id_gen.get_next() as stream_id: + async with stream_id_gen.get_next() as stream_id: # ... persist event ... """ with self._lock: @@ -107,7 +115,7 @@ class StreamIdGenerator(object): self._unfinished_ids.append(next_id) - @contextlib.contextmanager + @contextmanager def manager(): try: yield next_id @@ -115,12 +123,12 @@ class StreamIdGenerator(object): with self._lock: self._unfinished_ids.remove(next_id) - return manager() + return _AsyncCtxManagerWrapper(manager()) def get_next_mult(self, n): """ Usage: - with stream_id_gen.get_next(n) as stream_ids: + async with stream_id_gen.get_next(n) as stream_ids: # ... persist events ... """ with self._lock: @@ -134,7 +142,7 @@ class StreamIdGenerator(object): for next_id in next_ids: self._unfinished_ids.append(next_id) - @contextlib.contextmanager + @contextmanager def manager(): try: yield next_ids @@ -143,7 +151,7 @@ class StreamIdGenerator(object): for next_id in next_ids: self._unfinished_ids.remove(next_id) - return manager() + return _AsyncCtxManagerWrapper(manager()) def get_current_token(self): """Returns the maximum stream id such that all stream ids less than or @@ -158,63 +166,13 @@ class StreamIdGenerator(object): return self._current + def get_current_token_for_writer(self, instance_name: str) -> int: + """Returns the position of the given writer. -class ChainedIdGenerator(object): - """Used to generate new stream ids where the stream must be kept in sync - with another stream. It generates pairs of IDs, the first element is an - integer ID for this stream, the second element is the ID for the stream - that this stream needs to be kept in sync with.""" - - def __init__(self, chained_generator, db_conn, table, column): - self.chained_generator = chained_generator - self._table = table - self._lock = threading.Lock() - self._current_max = _load_current_id(db_conn, table, column) - self._unfinished_ids = deque() # type: Deque[Tuple[int, int]] - - def get_next(self): + For streams with single writers this is equivalent to + `get_current_token`. """ - Usage: - with stream_id_gen.get_next() as (stream_id, chained_id): - # ... persist event ... - """ - with self._lock: - self._current_max += 1 - next_id = self._current_max - chained_id = self.chained_generator.get_current_token() - - self._unfinished_ids.append((next_id, chained_id)) - - @contextlib.contextmanager - def manager(): - try: - yield (next_id, chained_id) - finally: - with self._lock: - self._unfinished_ids.remove((next_id, chained_id)) - - return manager() - - def get_current_token(self): - """Returns the maximum stream id such that all stream ids less than or - equal to it have been successfully persisted. - """ - with self._lock: - if self._unfinished_ids: - stream_id, chained_id = self._unfinished_ids[0] - return stream_id - 1, chained_id - - return self._current_max, self.chained_generator.get_current_token() - - def advance(self, token: int): - """Stub implementation for advancing the token when receiving updates - over replication; raises an exception as this instance should be the - only source of updates. - """ - - raise Exception( - "Attempted to advance token on source for table %r", self._table - ) + return self.get_current_token() class MultiWriterIdGenerator: @@ -228,89 +186,199 @@ class MultiWriterIdGenerator: Args: db_conn db + stream_name: A name for the stream. instance_name: The name of this instance. table: Database table associated with stream. instance_column: Column that stores the row's writer's instance name id_column: Column that stores the stream ID. sequence_name: The name of the postgres sequence used to generate new IDs. + writers: A list of known writers to use to populate current positions + on startup. Can be empty if nothing uses `get_current_token` or + `get_positions` (e.g. caches stream). + positive: Whether the IDs are positive (true) or negative (false). + When using negative IDs we go backwards from -1 to -2, -3, etc. """ def __init__( self, db_conn, - db: Database, + db: DatabasePool, + stream_name: str, instance_name: str, table: str, instance_column: str, id_column: str, sequence_name: str, + writers: List[str], + positive: bool = True, ): self._db = db + self._stream_name = stream_name self._instance_name = instance_name + self._positive = positive + self._writers = writers + self._return_factor = 1 if positive else -1 # We lock as some functions may be called from DB threads. self._lock = threading.Lock() - self._current_positions = self._load_current_ids( - db_conn, table, instance_column, id_column - ) + # Note: If we are a negative stream then we still store all the IDs as + # positive to make life easier for us, and simply negate the IDs when we + # return them. + self._current_positions = {} # type: Dict[str, int] # Set of local IDs that we're still processing. The current position # should be less than the minimum of this set (if not empty). self._unfinished_ids = set() # type: Set[int] + # Set of local IDs that we've processed that are larger than the current + # position, due to there being smaller unpersisted IDs. + self._finished_ids = set() # type: Set[int] + + # We track the max position where we know everything before has been + # persisted. This is done by a) looking at the min across all instances + # and b) noting that if we have seen a run of persisted positions + # without gaps (e.g. 5, 6, 7) then we can skip forward (e.g. to 7). + # + # Note: There is no guarentee that the IDs generated by the sequence + # will be gapless; gaps can form when e.g. a transaction was rolled + # back. This means that sometimes we won't be able to skip forward the + # position even though everything has been persisted. However, since + # gaps should be relatively rare it's still worth doing the book keeping + # that allows us to skip forwards when there are gapless runs of + # positions. + # + # We start at 1 here as a) the first generated stream ID will be 2, and + # b) other parts of the code assume that stream IDs are strictly greater + # than 0. + self._persisted_upto_position = ( + min(self._current_positions.values()) if self._current_positions else 1 + ) + self._known_persisted_positions = [] # type: List[int] + self._sequence_gen = PostgresSequenceGenerator(sequence_name) + # We check that the table and sequence haven't diverged. + self._sequence_gen.check_consistency( + db_conn, table=table, id_column=id_column, positive=positive + ) + + # This goes and fills out the above state from the database. + self._load_current_ids(db_conn, table, instance_column, id_column) + def _load_current_ids( self, db_conn, table: str, instance_column: str, id_column: str - ) -> Dict[str, int]: - sql = """ - SELECT %(instance)s, MAX(%(id)s) FROM %(table)s - GROUP BY %(instance)s - """ % { - "instance": instance_column, - "id": id_column, - "table": table, - } - + ): cur = db_conn.cursor() - cur.execute(sql) - # `cur` is an iterable over returned rows, which are 2-tuples. - current_positions = dict(cur) + # Load the current positions of all writers for the stream. + if self._writers: + # We delete any stale entries in the positions table. This is + # important if we add back a writer after a long time; we want to + # consider that a "new" writer, rather than using the old stale + # entry here. + sql = """ + DELETE FROM stream_positions + WHERE + stream_name = ? + AND instance_name != ALL(?) + """ + sql = self._db.engine.convert_param_style(sql) + cur.execute(sql, (self._stream_name, self._writers)) + + sql = """ + SELECT instance_name, stream_id FROM stream_positions + WHERE stream_name = ? + """ + sql = self._db.engine.convert_param_style(sql) + + cur.execute(sql, (self._stream_name,)) + + self._current_positions = { + instance: stream_id * self._return_factor + for instance, stream_id in cur + if instance in self._writers + } + + # We set the `_persisted_upto_position` to be the minimum of all current + # positions. If empty we use the max stream ID from the DB table. + min_stream_id = min(self._current_positions.values(), default=None) + + if min_stream_id is None: + # We add a GREATEST here to ensure that the result is always + # positive. (This can be a problem for e.g. backfill streams where + # the server has never backfilled). + sql = """ + SELECT GREATEST(COALESCE(%(agg)s(%(id)s), 1), 1) + FROM %(table)s + """ % { + "id": id_column, + "table": table, + "agg": "MAX" if self._positive else "-MIN", + } + cur.execute(sql) + (stream_id,) = cur.fetchone() + self._persisted_upto_position = stream_id + else: + # If we have a min_stream_id then we pull out everything greater + # than it from the DB so that we can prefill + # `_known_persisted_positions` and get a more accurate + # `_persisted_upto_position`. + # + # We also check if any of the later rows are from this instance, in + # which case we use that for this instance's current position. This + # is to handle the case where we didn't finish persisting to the + # stream positions table before restart (or the stream position + # table otherwise got out of date). + + sql = """ + SELECT %(instance)s, %(id)s FROM %(table)s + WHERE ? %(cmp)s %(id)s + """ % { + "id": id_column, + "table": table, + "instance": instance_column, + "cmp": "<=" if self._positive else ">=", + } + sql = self._db.engine.convert_param_style(sql) + cur.execute(sql, (min_stream_id * self._return_factor,)) + + self._persisted_upto_position = min_stream_id + + with self._lock: + for (instance, stream_id,) in cur: + stream_id = self._return_factor * stream_id + self._add_persisted_position(stream_id) + + if instance == self._instance_name: + self._current_positions[instance] = stream_id cur.close() - return current_positions - - def _load_next_id_txn(self, txn): + def _load_next_id_txn(self, txn) -> int: return self._sequence_gen.get_next_id_txn(txn) - async def get_next(self): + def _load_next_mult_id_txn(self, txn, n: int) -> List[int]: + return self._sequence_gen.get_next_mult_txn(txn, n) + + def get_next(self): """ Usage: - with await stream_id_gen.get_next() as stream_id: + async with stream_id_gen.get_next() as stream_id: # ... persist event ... """ - next_id = await self._db.runInteraction("_load_next_id", self._load_next_id_txn) - # Assert the fetched ID is actually greater than what we currently - # believe the ID to be. If not, then the sequence and table have got - # out of sync somehow. - assert self.get_current_token() < next_id + return _MultiWriterCtxManager(self) - with self._lock: - self._unfinished_ids.add(next_id) - - @contextlib.contextmanager - def manager(): - try: - yield next_id - finally: - self._mark_id_as_finished(next_id) + def get_next_mult(self, n: int): + """ + Usage: + async with stream_id_gen.get_next_mult(5) as stream_ids: + # ... persist events ... + """ - return manager() + return _MultiWriterCtxManager(self, n) def get_next_txn(self, txn: LoggingTransaction): """ @@ -328,49 +396,253 @@ class MultiWriterIdGenerator: txn.call_after(self._mark_id_as_finished, next_id) txn.call_on_exception(self._mark_id_as_finished, next_id) - return next_id + # Update the `stream_positions` table with newly updated stream + # ID (unless self._writers is not set in which case we don't + # bother, as nothing will read it). + # + # We only do this on the success path so that the persisted current + # position points to a persited row with the correct instance name. + if self._writers: + txn.call_after( + run_as_background_process, + "MultiWriterIdGenerator._update_table", + self._db.runInteraction, + "MultiWriterIdGenerator._update_table", + self._update_stream_positions_table_txn, + ) + + return self._return_factor * next_id def _mark_id_as_finished(self, next_id: int): """The ID has finished being processed so we should advance the - current poistion if possible. + current position if possible. """ with self._lock: self._unfinished_ids.discard(next_id) + self._finished_ids.add(next_id) + + new_cur = None - # Figure out if its safe to advance the position by checking there - # aren't any lower allocated IDs that are yet to finish. - if all(c > next_id for c in self._unfinished_ids): + if self._unfinished_ids: + # If there are unfinished IDs then the new position will be the + # largest finished ID less than the minimum unfinished ID. + + finished = set() + + min_unfinshed = min(self._unfinished_ids) + for s in self._finished_ids: + if s < min_unfinshed: + if new_cur is None or new_cur < s: + new_cur = s + else: + finished.add(s) + + # We clear these out since they're now all less than the new + # position. + self._finished_ids = finished + else: + # There are no unfinished IDs so the new position is simply the + # largest finished one. + new_cur = max(self._finished_ids) + + # We clear these out since they're now all less than the new + # position. + self._finished_ids.clear() + + if new_cur: curr = self._current_positions.get(self._instance_name, 0) - self._current_positions[self._instance_name] = max(curr, next_id) + self._current_positions[self._instance_name] = max(curr, new_cur) - def get_current_token(self, instance_name: str = None) -> int: - """Gets the current position of a named writer (defaults to current - instance). + self._add_persisted_position(next_id) - Returns 0 if we don't have a position for the named writer (likely due - to it being a new writer). + def get_current_token(self) -> int: + """Returns the maximum stream id such that all stream ids less than or + equal to it have been successfully persisted. """ - if instance_name is None: - instance_name = self._instance_name + return self.get_persisted_upto_position() + + def get_current_token_for_writer(self, instance_name: str) -> int: + """Returns the position of the given writer. + """ + # If we don't have an entry for the given instance name, we assume it's a + # new writer. + # + # For new writers we assume their initial position to be the current + # persisted up to position. This stops Synapse from doing a full table + # scan when a new writer announces itself over replication. with self._lock: - return self._current_positions.get(instance_name, 0) + return self._return_factor * self._current_positions.get( + instance_name, self._persisted_upto_position + ) def get_positions(self) -> Dict[str, int]: """Get a copy of the current positon map. + + Note that this won't necessarily include all configured writers if some + writers haven't written anything yet. """ with self._lock: - return dict(self._current_positions) + return { + name: self._return_factor * i + for name, i in self._current_positions.items() + } def advance(self, instance_name: str, new_id: int): """Advance the postion of the named writer to the given ID, if greater than existing entry. """ + new_id *= self._return_factor + with self._lock: self._current_positions[instance_name] = max( new_id, self._current_positions.get(instance_name, 0) ) + + self._add_persisted_position(new_id) + + def get_persisted_upto_position(self) -> int: + """Get the max position where all previous positions have been + persisted. + + Note: In the worst case scenario this will be equal to the minimum + position across writers. This means that the returned position here can + lag if one writer doesn't write very often. + """ + + with self._lock: + return self._return_factor * self._persisted_upto_position + + def _add_persisted_position(self, new_id: int): + """Record that we have persisted a position. + + This is used to keep the `_current_positions` up to date. + """ + + # We require that the lock is locked by caller + assert self._lock.locked() + + heapq.heappush(self._known_persisted_positions, new_id) + + # We move the current min position up if the minimum current positions + # of all instances is higher (since by definition all positions less + # that that have been persisted). + min_curr = min(self._current_positions.values(), default=0) + self._persisted_upto_position = max(min_curr, self._persisted_upto_position) + + # We now iterate through the seen positions, discarding those that are + # less than the current min positions, and incrementing the min position + # if its exactly one greater. + # + # This is also where we discard items from `_known_persisted_positions` + # (to ensure the list doesn't infinitely grow). + while self._known_persisted_positions: + if self._known_persisted_positions[0] <= self._persisted_upto_position: + heapq.heappop(self._known_persisted_positions) + elif ( + self._known_persisted_positions[0] == self._persisted_upto_position + 1 + ): + heapq.heappop(self._known_persisted_positions) + self._persisted_upto_position += 1 + else: + # There was a gap in seen positions, so there is nothing more to + # do. + break + + def _update_stream_positions_table_txn(self, txn: Cursor): + """Update the `stream_positions` table with newly persisted position. + """ + + if not self._writers: + return + + # We upsert the value, ensuring on conflict that we always increase the + # value (or decrease if stream goes backwards). + sql = """ + INSERT INTO stream_positions (stream_name, instance_name, stream_id) + VALUES (?, ?, ?) + ON CONFLICT (stream_name, instance_name) + DO UPDATE SET + stream_id = %(agg)s(stream_positions.stream_id, EXCLUDED.stream_id) + """ % { + "agg": "GREATEST" if self._positive else "LEAST", + } + + pos = (self.get_current_token_for_writer(self._instance_name),) + txn.execute(sql, (self._stream_name, self._instance_name, pos)) + + +@attr.s(slots=True) +class _AsyncCtxManagerWrapper: + """Helper class to convert a plain context manager to an async one. + + This is mainly useful if you have a plain context manager but the interface + requires an async one. + """ + + inner = attr.ib() + + async def __aenter__(self): + return self.inner.__enter__() + + async def __aexit__(self, exc_type, exc, tb): + return self.inner.__exit__(exc_type, exc, tb) + + +@attr.s(slots=True) +class _MultiWriterCtxManager: + """Async context manager returned by MultiWriterIdGenerator + """ + + id_gen = attr.ib(type=MultiWriterIdGenerator) + multiple_ids = attr.ib(type=Optional[int], default=None) + stream_ids = attr.ib(type=List[int], factory=list) + + async def __aenter__(self) -> Union[int, List[int]]: + # It's safe to run this in autocommit mode as fetching values from a + # sequence ignores transaction semantics anyway. + self.stream_ids = await self.id_gen._db.runInteraction( + "_load_next_mult_id", + self.id_gen._load_next_mult_id_txn, + self.multiple_ids or 1, + db_autocommit=True, + ) + + with self.id_gen._lock: + self.id_gen._unfinished_ids.update(self.stream_ids) + + if self.multiple_ids is None: + return self.stream_ids[0] * self.id_gen._return_factor + else: + return [i * self.id_gen._return_factor for i in self.stream_ids] + + async def __aexit__(self, exc_type, exc, tb): + for i in self.stream_ids: + self.id_gen._mark_id_as_finished(i) + + if exc_type is not None: + return False + + # Update the `stream_positions` table with newly updated stream + # ID (unless self._writers is not set in which case we don't + # bother, as nothing will read it). + # + # We only do this on the success path so that the persisted current + # position points to a persisted row with the correct instance name. + # + # We do this in autocommit mode as a) the upsert works correctly outside + # transactions and b) reduces the amount of time the rows are locked + # for. If we don't do this then we'll often hit serialization errors due + # to the fact we default to REPEATABLE READ isolation levels. + if self.id_gen._writers: + await self.id_gen._db.runInteraction( + "MultiWriterIdGenerator._update_table", + self.id_gen._update_stream_positions_table_txn, + db_autocommit=True, + ) + + return False diff --git a/synapse/storage/util/sequence.py b/synapse/storage/util/sequence.py
index 63dfea4220..2dd95e2709 100644 --- a/synapse/storage/util/sequence.py +++ b/synapse/storage/util/sequence.py
@@ -13,11 +13,34 @@ # See the License for the specific language governing permissions and # limitations under the License. import abc +import logging import threading -from typing import Callable, Optional +from typing import Callable, List, Optional -from synapse.storage.engines import BaseDatabaseEngine, PostgresEngine -from synapse.storage.types import Cursor +from synapse.storage.engines import ( + BaseDatabaseEngine, + IncorrectDatabaseSetup, + PostgresEngine, +) +from synapse.storage.types import Connection, Cursor + +logger = logging.getLogger(__name__) + + +_INCONSISTENT_SEQUENCE_ERROR = """ +Postgres sequence '%(seq)s' is inconsistent with associated +table '%(table)s'. This can happen if Synapse has been downgraded and +then upgraded again, or due to a bad migration. + +To fix this error, shut down Synapse (including any and all workers) +and run the following SQL: + + SELECT setval('%(seq)s', ( + %(max_id_sql)s + )); + +See docs/postgres.md for more information. +""" class SequenceGenerator(metaclass=abc.ABCMeta): @@ -28,6 +51,19 @@ class SequenceGenerator(metaclass=abc.ABCMeta): """Gets the next ID in the sequence""" ... + @abc.abstractmethod + def check_consistency( + self, db_conn: Connection, table: str, id_column: str, positive: bool = True + ): + """Should be called during start up to test that the current value of + the sequence is greater than or equal to the maximum ID in the table. + + This is to handle various cases where the sequence value can get out + of sync with the table, e.g. if Synapse gets rolled back to a previous + version and the rolled forwards again. + """ + ... + class PostgresSequenceGenerator(SequenceGenerator): """An implementation of SequenceGenerator which uses a postgres sequence""" @@ -39,6 +75,56 @@ class PostgresSequenceGenerator(SequenceGenerator): txn.execute("SELECT nextval(?)", (self._sequence_name,)) return txn.fetchone()[0] + def get_next_mult_txn(self, txn: Cursor, n: int) -> List[int]: + txn.execute( + "SELECT nextval(?) FROM generate_series(1, ?)", (self._sequence_name, n) + ) + return [i for (i,) in txn] + + def check_consistency( + self, db_conn: Connection, table: str, id_column: str, positive: bool = True + ): + txn = db_conn.cursor() + + # First we get the current max ID from the table. + table_sql = "SELECT GREATEST(%(agg)s(%(id)s), 0) FROM %(table)s" % { + "id": id_column, + "table": table, + "agg": "MAX" if positive else "-MIN", + } + + txn.execute(table_sql) + row = txn.fetchone() + if not row: + # Table is empty, so nothing to do. + txn.close() + return + + # Now we fetch the current value from the sequence and compare with the + # above. + max_stream_id = row[0] + txn.execute( + "SELECT last_value, is_called FROM %(seq)s" % {"seq": self._sequence_name} + ) + last_value, is_called = txn.fetchone() + txn.close() + + # If `is_called` is False then `last_value` is actually the value that + # will be generated next, so we decrement to get the true "last value". + if not is_called: + last_value -= 1 + + if max_stream_id > last_value: + logger.warning( + "Postgres sequence %s is behind table %s: %d < %d", + last_value, + max_stream_id, + ) + raise IncorrectDatabaseSetup( + _INCONSISTENT_SEQUENCE_ERROR + % {"seq": self._sequence_name, "table": table, "max_id_sql": table_sql} + ) + GetFirstCallbackType = Callable[[Cursor], int] @@ -75,6 +161,12 @@ class LocalSequenceGenerator(SequenceGenerator): self._current_max_id += 1 return self._current_max_id + def check_consistency( + self, db_conn: Connection, table: str, id_column: str, positive: bool = True + ): + # There is nothing to do for in memory sequences + pass + def build_sequence_generator( database_engine: BaseDatabaseEngine, diff --git a/synapse/streams/config.py b/synapse/streams/config.py
index ca7c16ff65..fdda21d165 100644 --- a/synapse/streams/config.py +++ b/synapse/streams/config.py
@@ -12,11 +12,15 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import logging +from typing import Optional + +import attr from synapse.api.errors import SynapseError from synapse.http.servlet import parse_integer, parse_string +from synapse.http.site import SynapseRequest +from synapse.storage.databases.main import DataStore from synapse.types import StreamToken logger = logging.getLogger(__name__) @@ -25,38 +29,23 @@ logger = logging.getLogger(__name__) MAX_LIMIT = 1000 -class SourcePaginationConfig(object): - - """A configuration object which stores pagination parameters for a - specific event source.""" - - def __init__(self, from_key=None, to_key=None, direction="f", limit=None): - self.from_key = from_key - self.to_key = to_key - self.direction = "f" if direction == "f" else "b" - self.limit = min(int(limit), MAX_LIMIT) if limit is not None else None - - def __repr__(self): - return "StreamConfig(from_key=%r, to_key=%r, direction=%r, limit=%r)" % ( - self.from_key, - self.to_key, - self.direction, - self.limit, - ) - - -class PaginationConfig(object): - +@attr.s(slots=True) +class PaginationConfig: """A configuration object which stores pagination parameters.""" - def __init__(self, from_token=None, to_token=None, direction="f", limit=None): - self.from_token = from_token - self.to_token = to_token - self.direction = "f" if direction == "f" else "b" - self.limit = min(int(limit), MAX_LIMIT) if limit is not None else None + from_token = attr.ib(type=Optional[StreamToken]) + to_token = attr.ib(type=Optional[StreamToken]) + direction = attr.ib(type=str) + limit = attr.ib(type=Optional[int]) @classmethod - def from_request(cls, request, raise_invalid_params=True, default_limit=None): + async def from_request( + cls, + store: "DataStore", + request: SynapseRequest, + raise_invalid_params: bool = True, + default_limit: Optional[int] = None, + ) -> "PaginationConfig": direction = parse_string(request, "dir", default="f", allowed_values=["f", "b"]) from_tok = parse_string(request, "from") @@ -66,20 +55,23 @@ class PaginationConfig(object): if from_tok == "END": from_tok = None # For backwards compat. elif from_tok: - from_tok = StreamToken.from_string(from_tok) + from_tok = await StreamToken.from_string(store, from_tok) except Exception: raise SynapseError(400, "'from' parameter is invalid") try: if to_tok: - to_tok = StreamToken.from_string(to_tok) + to_tok = await StreamToken.from_string(store, to_tok) except Exception: raise SynapseError(400, "'to' parameter is invalid") limit = parse_integer(request, "limit", default=default_limit) - if limit and limit < 0: - raise SynapseError(400, "Limit must be 0 or above") + if limit: + if limit < 0: + raise SynapseError(400, "Limit must be 0 or above") + + limit = min(int(limit), MAX_LIMIT) try: return PaginationConfig(from_tok, to_tok, direction, limit) @@ -87,20 +79,10 @@ class PaginationConfig(object): logger.exception("Failed to create pagination config") raise SynapseError(400, "Invalid request.") - def __repr__(self): + def __repr__(self) -> str: return ("PaginationConfig(from_tok=%r, to_tok=%r, direction=%r, limit=%r)") % ( self.from_token, self.to_token, self.direction, self.limit, ) - - def get_source_config(self, source_name): - keyname = "%s_key" % source_name - - return SourcePaginationConfig( - from_key=getattr(self.from_token, keyname), - to_key=getattr(self.to_token, keyname) if self.to_token else None, - direction=self.direction, - limit=self.limit, - ) diff --git a/synapse/streams/events.py b/synapse/streams/events.py
index 5d3eddcfdc..92fd5d489f 100644 --- a/synapse/streams/events.py +++ b/synapse/streams/events.py
@@ -15,8 +15,6 @@ from typing import Any, Dict -from twisted.internet import defer - from synapse.handlers.account_data import AccountDataEventSource from synapse.handlers.presence import PresenceEventSource from synapse.handlers.receipts import ReceiptEventSource @@ -25,7 +23,7 @@ from synapse.handlers.typing import TypingNotificationEventSource from synapse.types import StreamToken -class EventSources(object): +class EventSources: SOURCE_TYPES = { "room": RoomEventSource, "presence": PresenceEventSource, @@ -40,19 +38,18 @@ class EventSources(object): } # type: Dict[str, Any] self.store = hs.get_datastore() - @defer.inlineCallbacks - def get_current_token(self): - push_rules_key, _ = self.store.get_push_rules_stream_token() + def get_current_token(self) -> StreamToken: + push_rules_key = self.store.get_max_push_rules_stream_id() to_device_key = self.store.get_to_device_stream_token() device_list_key = self.store.get_device_stream_token() groups_key = self.store.get_group_stream_token() token = StreamToken( - room_key=(yield self.sources["room"].get_current_key()), - presence_key=(yield self.sources["presence"].get_current_key()), - typing_key=(yield self.sources["typing"].get_current_key()), - receipt_key=(yield self.sources["receipt"].get_current_key()), - account_data_key=(yield self.sources["account_data"].get_current_key()), + room_key=self.sources["room"].get_current_key(), + presence_key=self.sources["presence"].get_current_key(), + typing_key=self.sources["typing"].get_current_key(), + receipt_key=self.sources["receipt"].get_current_key(), + account_data_key=self.sources["account_data"].get_current_key(), push_rules_key=push_rules_key, to_device_key=to_device_key, device_list_key=device_list_key, @@ -60,8 +57,7 @@ class EventSources(object): ) return token - @defer.inlineCallbacks - def get_current_token_for_pagination(self): + def get_current_token_for_pagination(self) -> StreamToken: """Get the current token for a given room to be used to paginate events. @@ -69,10 +65,10 @@ class EventSources(object): than `room`, since they are not used during pagination. Returns: - Deferred[StreamToken] + The current token for pagination. """ token = StreamToken( - room_key=(yield self.sources["room"].get_current_key()), + room_key=self.sources["room"].get_current_key(), presence_key=0, typing_key=0, receipt_key=0, diff --git a/synapse/types.py b/synapse/types.py
index 7cc523e4f8..bc73e3775d 100644 --- a/synapse/types.py +++ b/synapse/types.py
@@ -13,11 +13,22 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. +import abc import re import string import sys from collections import namedtuple -from typing import Any, Dict, Tuple, TypeVar +from typing import ( + TYPE_CHECKING, + Any, + Dict, + Mapping, + MutableMapping, + Optional, + Tuple, + Type, + TypeVar, +) import attr from signedjson.key import decode_verify_key_bytes @@ -26,6 +37,9 @@ from unpaddedbase64 import decode_base64 from synapse.api.errors import Codes, SynapseError +if TYPE_CHECKING: + from synapse.storage.databases.main import DataStore + # define a version of typing.Collection that works on python 3.5 if sys.version_info[:3] >= (3, 6, 0): from typing import Collection @@ -34,15 +48,16 @@ else: T_co = TypeVar("T_co", covariant=True) - class Collection(Iterable[T_co], Container[T_co], Sized): + class Collection(Iterable[T_co], Container[T_co], Sized): # type: ignore __slots__ = () # Define a state map type from type/state_key to T (usually an event ID or # event) T = TypeVar("T") -StateMap = Dict[Tuple[str, str], T] - +StateKey = Tuple[str, str] +StateMap = Mapping[StateKey, T] +MutableStateMap = MutableMapping[StateKey, T] # the type of a JSON-serialisable dict. This could be made stronger, but it will # do for now. @@ -51,7 +66,15 @@ JsonDict = Dict[str, Any] class Requester( namedtuple( - "Requester", ["user", "access_token_id", "is_guest", "device_id", "app_service"] + "Requester", + [ + "user", + "access_token_id", + "is_guest", + "shadow_banned", + "device_id", + "app_service", + ], ) ): """ @@ -62,6 +85,7 @@ class Requester( access_token_id (int|None): *ID* of the access token used for this request, or None if it came via the appservice API or similar is_guest (bool): True if the user making this request is a guest user + shadow_banned (bool): True if the user making this request has been shadow-banned. device_id (str|None): device_id which was set at authentication time app_service (ApplicationService|None): the AS requesting on behalf of the user """ @@ -77,6 +101,7 @@ class Requester( "user_id": self.user.to_string(), "access_token_id": self.access_token_id, "is_guest": self.is_guest, + "shadow_banned": self.shadow_banned, "device_id": self.device_id, "app_server_id": self.app_service.id if self.app_service else None, } @@ -101,13 +126,19 @@ class Requester( user=UserID.from_string(input["user_id"]), access_token_id=input["access_token_id"], is_guest=input["is_guest"], + shadow_banned=input["shadow_banned"], device_id=input["device_id"], app_service=appservice, ) def create_requester( - user_id, access_token_id=None, is_guest=False, device_id=None, app_service=None + user_id, + access_token_id=None, + is_guest=False, + shadow_banned=False, + device_id=None, + app_service=None, ): """ Create a new ``Requester`` object @@ -117,6 +148,7 @@ def create_requester( access_token_id (int|None): *ID* of the access token used for this request, or None if it came via the appservice API or similar is_guest (bool): True if the user making this request is a guest user + shadow_banned (bool): True if the user making this request is shadow-banned. device_id (str|None): device_id which was set at authentication time app_service (ApplicationService|None): the AS requesting on behalf of the user @@ -125,7 +157,9 @@ def create_requester( """ if not isinstance(user_id, UserID): user_id = UserID.from_string(user_id) - return Requester(user_id, access_token_id, is_guest, device_id, app_service) + return Requester( + user_id, access_token_id, is_guest, shadow_banned, device_id, app_service + ) def get_domain_from_id(string): @@ -142,7 +176,12 @@ def get_localpart_from_id(string): return string[1:idx] -class DomainSpecificString(namedtuple("DomainSpecificString", ("localpart", "domain"))): +DS = TypeVar("DS", bound="DomainSpecificString") + + +class DomainSpecificString( + namedtuple("DomainSpecificString", ("localpart", "domain")), metaclass=abc.ABCMeta +): """Common base class among ID/name strings that have a local part and a domain name, prefixed with a sigil. @@ -152,6 +191,8 @@ class DomainSpecificString(namedtuple("DomainSpecificString", ("localpart", "dom 'domain' : The domain part of the name """ + SIGIL = abc.abstractproperty() # type: str # type: ignore + # Deny iteration because it will bite you if you try to create a singleton # set by: # users = set(user) @@ -167,7 +208,7 @@ class DomainSpecificString(namedtuple("DomainSpecificString", ("localpart", "dom return self @classmethod - def from_string(cls, s: str): + def from_string(cls: Type[DS], s: str) -> DS: """Parse the string given by 's' into a structure object.""" if len(s) < 1 or s[0:1] != cls.SIGIL: raise SynapseError( @@ -191,12 +232,12 @@ class DomainSpecificString(namedtuple("DomainSpecificString", ("localpart", "dom # names on one HS return cls(localpart=parts[0], domain=domain) - def to_string(self): + def to_string(self) -> str: """Return a string encoding the fields of the structure object.""" return "%s%s:%s" % (self.SIGIL, self.localpart, self.domain) @classmethod - def is_valid(cls, s): + def is_valid(cls: Type[DS], s: str) -> bool: try: cls.from_string(s) return True @@ -236,8 +277,9 @@ class GroupID(DomainSpecificString): SIGIL = "+" @classmethod - def from_string(cls, s): - group_id = super(GroupID, cls).from_string(s) + def from_string(cls: Type[DS], s: str) -> DS: + group_id = super().from_string(s) # type: DS # type: ignore + if not group_id.localpart: raise SynapseError(400, "Group ID cannot be empty", Codes.INVALID_PARAM) @@ -347,86 +389,8 @@ def map_username_to_mxid_localpart(username, case_sensitive=False): return username.decode("ascii") -class StreamToken( - namedtuple( - "Token", - ( - "room_key", - "presence_key", - "typing_key", - "receipt_key", - "account_data_key", - "push_rules_key", - "to_device_key", - "device_list_key", - "groups_key", - ), - ) -): - _SEPARATOR = "_" - START = None # type: StreamToken - - @classmethod - def from_string(cls, string): - try: - keys = string.split(cls._SEPARATOR) - while len(keys) < len(cls._fields): - # i.e. old token from before receipt_key - keys.append("0") - return cls(*keys) - except Exception: - raise SynapseError(400, "Invalid Token") - - def to_string(self): - return self._SEPARATOR.join([str(k) for k in self]) - - @property - def room_stream_id(self): - # TODO(markjh): Awful hack to work around hacks in the presence tests - # which assume that the keys are integers. - if type(self.room_key) is int: - return self.room_key - else: - return int(self.room_key[1:].split("-")[-1]) - - def is_after(self, other): - """Does this token contain events that the other doesn't?""" - return ( - (other.room_stream_id < self.room_stream_id) - or (int(other.presence_key) < int(self.presence_key)) - or (int(other.typing_key) < int(self.typing_key)) - or (int(other.receipt_key) < int(self.receipt_key)) - or (int(other.account_data_key) < int(self.account_data_key)) - or (int(other.push_rules_key) < int(self.push_rules_key)) - or (int(other.to_device_key) < int(self.to_device_key)) - or (int(other.device_list_key) < int(self.device_list_key)) - or (int(other.groups_key) < int(self.groups_key)) - ) - - def copy_and_advance(self, key, new_value): - """Advance the given key in the token to a new value if and only if the - new value is after the old value. - """ - new_token = self.copy_and_replace(key, new_value) - if key == "room_key": - new_id = new_token.room_stream_id - old_id = self.room_stream_id - else: - new_id = int(getattr(new_token, key)) - old_id = int(getattr(self, key)) - if old_id < new_id: - return new_token - else: - return self - - def copy_and_replace(self, key, new_value): - return self._replace(**{key: new_value}) - - -StreamToken.START = StreamToken(*(["s0"] + ["0"] * (len(StreamToken._fields) - 1))) - - -class RoomStreamToken(namedtuple("_StreamToken", "topological stream")): +@attr.s(frozen=True, slots=True) +class RoomStreamToken: """Tokens are positions between events. The token "s1" comes after event 1. s0 s1 @@ -449,10 +413,14 @@ class RoomStreamToken(namedtuple("_StreamToken", "topological stream")): followed by the "stream_ordering" id of the event it comes after. """ - __slots__ = [] # type: list + topological = attr.ib( + type=Optional[int], + validator=attr.validators.optional(attr.validators.instance_of(int)), + ) + stream = attr.ib(type=int, validator=attr.validators.instance_of(int)) @classmethod - def parse(cls, string): + async def parse(cls, store: "DataStore", string: str) -> "RoomStreamToken": try: if string[0] == "s": return cls(topological=None, stream=int(string[1:])) @@ -464,7 +432,7 @@ class RoomStreamToken(namedtuple("_StreamToken", "topological stream")): raise SynapseError(400, "Invalid token %r" % (string,)) @classmethod - def parse_stream_token(cls, string): + def parse_stream_token(cls, string: str) -> "RoomStreamToken": try: if string[0] == "s": return cls(topological=None, stream=int(string[1:])) @@ -472,13 +440,130 @@ class RoomStreamToken(namedtuple("_StreamToken", "topological stream")): pass raise SynapseError(400, "Invalid token %r" % (string,)) - def __str__(self): + def copy_and_advance(self, other: "RoomStreamToken") -> "RoomStreamToken": + """Return a new token such that if an event is after both this token and + the other token, then its after the returned token too. + """ + + if self.topological or other.topological: + raise Exception("Can't advance topological tokens") + + max_stream = max(self.stream, other.stream) + + return RoomStreamToken(None, max_stream) + + def as_tuple(self) -> Tuple[Optional[int], int]: + return (self.topological, self.stream) + + async def to_string(self, store: "DataStore") -> str: if self.topological is not None: return "t%d-%d" % (self.topological, self.stream) else: return "s%d" % (self.stream,) +@attr.s(slots=True, frozen=True) +class StreamToken: + room_key = attr.ib( + type=RoomStreamToken, validator=attr.validators.instance_of(RoomStreamToken) + ) + presence_key = attr.ib(type=int) + typing_key = attr.ib(type=int) + receipt_key = attr.ib(type=int) + account_data_key = attr.ib(type=int) + push_rules_key = attr.ib(type=int) + to_device_key = attr.ib(type=int) + device_list_key = attr.ib(type=int) + groups_key = attr.ib(type=int) + + _SEPARATOR = "_" + START = None # type: StreamToken + + @classmethod + async def from_string(cls, store: "DataStore", string: str) -> "StreamToken": + try: + keys = string.split(cls._SEPARATOR) + while len(keys) < len(attr.fields(cls)): + # i.e. old token from before receipt_key + keys.append("0") + return cls( + await RoomStreamToken.parse(store, keys[0]), *(int(k) for k in keys[1:]) + ) + except Exception: + raise SynapseError(400, "Invalid Token") + + async def to_string(self, store: "DataStore") -> str: + return self._SEPARATOR.join( + [ + await self.room_key.to_string(store), + str(self.presence_key), + str(self.typing_key), + str(self.receipt_key), + str(self.account_data_key), + str(self.push_rules_key), + str(self.to_device_key), + str(self.device_list_key), + str(self.groups_key), + ] + ) + + @property + def room_stream_id(self): + return self.room_key.stream + + def copy_and_advance(self, key, new_value) -> "StreamToken": + """Advance the given key in the token to a new value if and only if the + new value is after the old value. + """ + if key == "room_key": + new_token = self.copy_and_replace( + "room_key", self.room_key.copy_and_advance(new_value) + ) + return new_token + + new_token = self.copy_and_replace(key, new_value) + new_id = int(getattr(new_token, key)) + old_id = int(getattr(self, key)) + + if old_id < new_id: + return new_token + else: + return self + + def copy_and_replace(self, key, new_value) -> "StreamToken": + return attr.evolve(self, **{key: new_value}) + + +StreamToken.START = StreamToken(RoomStreamToken(None, 0), 0, 0, 0, 0, 0, 0, 0, 0) + + +@attr.s(slots=True, frozen=True) +class PersistedEventPosition: + """Position of a newly persisted event with instance that persisted it. + + This can be used to test whether the event is persisted before or after a + RoomStreamToken. + """ + + instance_name = attr.ib(type=str) + stream = attr.ib(type=int) + + def persisted_after(self, token: RoomStreamToken) -> bool: + return token.stream < self.stream + + def to_room_stream_token(self) -> RoomStreamToken: + """Converts the position to a room stream token such that events + persisted in the same room after this position will be after the + returned `RoomStreamToken`. + + Note: no guarentees are made about ordering w.r.t. events in other + rooms. + """ + # Doing the naive thing satisfies the desired properties described in + # the docstring. + return RoomStreamToken(None, self.stream) + + class ThirdPartyInstanceID( namedtuple("ThirdPartyInstanceID", ("appservice_id", "network_id")) ): @@ -514,7 +599,7 @@ class ThirdPartyInstanceID( @attr.s(slots=True) -class ReadReceipt(object): +class ReadReceipt: """Information about a read-receipt""" room_id = attr.ib() diff --git a/synapse/util/__init__.py b/synapse/util/__init__.py
index c63256d3bd..d55b93d763 100644 --- a/synapse/util/__init__.py +++ b/synapse/util/__init__.py
@@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. +import json import logging import re @@ -25,14 +26,27 @@ from synapse.logging import context logger = logging.getLogger(__name__) +def _reject_invalid_json(val): + """Do not allow Infinity, -Infinity, or NaN values in JSON.""" + raise ValueError("Invalid JSON value: '%s'" % val) + + +# Create a custom encoder to reduce the whitespace produced by JSON encoding and +# ensure that valid JSON is produced. +json_encoder = json.JSONEncoder(allow_nan=False, separators=(",", ":")) + +# Create a custom decoder to reject Python extensions to JSON. +json_decoder = json.JSONDecoder(parse_constant=_reject_invalid_json) + + def unwrapFirstError(failure): # defer.gatherResults and DeferredLists wrap failures. failure.trap(defer.FirstError) return failure.value.subFailure -@attr.s -class Clock(object): +@attr.s(slots=True) +class Clock: """ A Clock wraps a Twisted reactor and provides utilities on top of it. diff --git a/synapse/util/async_helpers.py b/synapse/util/async_helpers.py
index f562770922..382f0cf3f0 100644 --- a/synapse/util/async_helpers.py +++ b/synapse/util/async_helpers.py
@@ -17,12 +17,25 @@ import collections import logging from contextlib import contextmanager -from typing import Dict, Sequence, Set, Union +from typing import ( + Any, + Callable, + Dict, + Hashable, + Iterable, + List, + Optional, + Set, + TypeVar, + Union, +) import attr +from typing_extensions import ContextManager from twisted.internet import defer from twisted.internet.defer import CancelledError +from twisted.internet.interfaces import IReactorTime from twisted.python import failure from synapse.logging.context import ( @@ -35,7 +48,7 @@ from synapse.util import Clock, unwrapFirstError logger = logging.getLogger(__name__) -class ObservableDeferred(object): +class ObservableDeferred: """Wraps a deferred object so that we can add observer deferreds. These observer deferreds do not affect the callback chain of the original deferred. @@ -53,7 +66,7 @@ class ObservableDeferred(object): __slots__ = ["_deferred", "_observers", "_result"] - def __init__(self, deferred, consumeErrors=False): + def __init__(self, deferred: defer.Deferred, consumeErrors: bool = False): object.__setattr__(self, "_deferred", deferred) object.__setattr__(self, "_result", None) object.__setattr__(self, "_observers", set()) @@ -110,25 +123,25 @@ class ObservableDeferred(object): success, res = self._result return defer.succeed(res) if success else defer.fail(res) - def observers(self): + def observers(self) -> List[defer.Deferred]: return self._observers - def has_called(self): + def has_called(self) -> bool: return self._result is not None - def has_succeeded(self): + def has_succeeded(self) -> bool: return self._result is not None and self._result[0] is True - def get_result(self): + def get_result(self) -> Any: return self._result[1] - def __getattr__(self, name): + def __getattr__(self, name: str) -> Any: return getattr(self._deferred, name) - def __setattr__(self, name, value): + def __setattr__(self, name: str, value: Any) -> None: setattr(self._deferred, name, value) - def __repr__(self): + def __repr__(self) -> str: return "<ObservableDeferred object at %s, result=%r, _deferred=%r>" % ( id(self), self._result, @@ -136,18 +149,20 @@ class ObservableDeferred(object): ) -def concurrently_execute(func, args, limit): - """Executes the function with each argument conncurrently while limiting +def concurrently_execute( + func: Callable, args: Iterable[Any], limit: int +) -> defer.Deferred: + """Executes the function with each argument concurrently while limiting the number of concurrent executions. Args: - func (func): Function to execute, should return a deferred or coroutine. - args (Iterable): List of arguments to pass to func, each invocation of func + func: Function to execute, should return a deferred or coroutine. + args: List of arguments to pass to func, each invocation of func gets a single argument. - limit (int): Maximum number of conccurent executions. + limit: Maximum number of conccurent executions. Returns: - deferred: Resolved when all function invocations have finished. + Deferred[list]: Resolved when all function invocations have finished. """ it = iter(args) @@ -166,14 +181,17 @@ def concurrently_execute(func, args, limit): ).addErrback(unwrapFirstError) -def yieldable_gather_results(func, iter, *args, **kwargs): +def yieldable_gather_results( + func: Callable, iter: Iterable, *args: Any, **kwargs: Any +) -> defer.Deferred: """Executes the function with each argument concurrently. Args: - func (func): Function to execute that returns a Deferred - iter (iter): An iterable that yields items that get passed as the first + func: Function to execute that returns a Deferred + iter: An iterable that yields items that get passed as the first argument to the function *args: Arguments to be passed to each call to func + **kwargs: Keyword arguments to be passed to each call to func Returns Deferred[list]: Resolved when all functions have been invoked, or errors if @@ -187,24 +205,37 @@ def yieldable_gather_results(func, iter, *args, **kwargs): ).addErrback(unwrapFirstError) -class Linearizer(object): +@attr.s(slots=True) +class _LinearizerEntry: + # The number of things executing. + count = attr.ib(type=int) + # Deferreds for the things blocked from executing. + deferreds = attr.ib(type=collections.OrderedDict) + + +class Linearizer: """Limits concurrent access to resources based on a key. Useful to ensure only a few things happen at a time on a given resource. Example: - with (yield limiter.queue("test_key")): + with await limiter.queue("test_key"): # do some work. """ - def __init__(self, name=None, max_count=1, clock=None): + def __init__( + self, + name: Optional[str] = None, + max_count: int = 1, + clock: Optional[Clock] = None, + ): """ Args: - max_count(int): The maximum number of concurrent accesses + max_count: The maximum number of concurrent accesses """ if name is None: - self.name = id(self) + self.name = id(self) # type: Union[str, int] else: self.name = name @@ -215,15 +246,10 @@ class Linearizer(object): self._clock = clock self.max_count = max_count - # key_to_defer is a map from the key to a 2 element list where - # the first element is the number of things executing, and - # the second element is an OrderedDict, where the keys are deferreds for the - # things blocked from executing. - self.key_to_defer = ( - {} - ) # type: Dict[str, Sequence[Union[int, Dict[defer.Deferred, int]]]] + # key_to_defer is a map from the key to a _LinearizerEntry. + self.key_to_defer = {} # type: Dict[Hashable, _LinearizerEntry] - def is_queued(self, key) -> bool: + def is_queued(self, key: Hashable) -> bool: """Checks whether there is a process queued up waiting """ entry = self.key_to_defer.get(key) @@ -233,25 +259,27 @@ class Linearizer(object): # There are waiting deferreds only in the OrderedDict of deferreds is # non-empty. - return bool(entry[1]) + return bool(entry.deferreds) - def queue(self, key): + def queue(self, key: Hashable) -> defer.Deferred: # we avoid doing defer.inlineCallbacks here, so that cancellation works correctly. # (https://twistedmatrix.com/trac/ticket/4632 meant that cancellations were not # propagated inside inlineCallbacks until Twisted 18.7) - entry = self.key_to_defer.setdefault(key, [0, collections.OrderedDict()]) + entry = self.key_to_defer.setdefault( + key, _LinearizerEntry(0, collections.OrderedDict()) + ) # If the number of things executing is greater than the maximum # then add a deferred to the list of blocked items # When one of the things currently executing finishes it will callback # this item so that it can continue executing. - if entry[0] >= self.max_count: + if entry.count >= self.max_count: res = self._await_lock(key) else: logger.debug( "Acquired uncontended linearizer lock %r for key %r", self.name, key ) - entry[0] += 1 + entry.count += 1 res = defer.succeed(None) # once we successfully get the lock, we need to return a context manager which @@ -266,15 +294,15 @@ class Linearizer(object): # We've finished executing so check if there are any things # blocked waiting to execute and start one of them - entry[0] -= 1 + entry.count -= 1 - if entry[1]: - (next_def, _) = entry[1].popitem(last=False) + if entry.deferreds: + (next_def, _) = entry.deferreds.popitem(last=False) # we need to run the next thing in the sentinel context. with PreserveLoggingContext(): next_def.callback(None) - elif entry[0] == 0: + elif entry.count == 0: # We were the last thing for this key: remove it from the # map. del self.key_to_defer[key] @@ -282,7 +310,7 @@ class Linearizer(object): res.addCallback(_ctx_manager) return res - def _await_lock(self, key): + def _await_lock(self, key: Hashable) -> defer.Deferred: """Helper for queue: adds a deferred to the queue Assumes that we've already checked that we've reached the limit of the number @@ -297,11 +325,11 @@ class Linearizer(object): logger.debug("Waiting to acquire linearizer lock %r for key %r", self.name, key) new_defer = make_deferred_yieldable(defer.Deferred()) - entry[1][new_defer] = 1 + entry.deferreds[new_defer] = 1 def cb(_r): logger.debug("Acquired linearizer lock %r for key %r", self.name, key) - entry[0] += 1 + entry.count += 1 # if the code holding the lock completes synchronously, then it # will recursively run the next claimant on the list. That can @@ -330,19 +358,19 @@ class Linearizer(object): ) # we just have to take ourselves back out of the queue. - del entry[1][new_defer] + del entry.deferreds[new_defer] return e new_defer.addCallbacks(cb, eb) return new_defer -class ReadWriteLock(object): - """A deferred style read write lock. +class ReadWriteLock: + """An async read write lock. Example: - with (yield read_write_lock.read("test_key")): + with await read_write_lock.read("test_key"): # do some work """ @@ -365,8 +393,7 @@ class ReadWriteLock(object): # Latest writer queued self.key_to_current_writer = {} # type: Dict[str, defer.Deferred] - @defer.inlineCallbacks - def read(self, key): + async def read(self, key: str) -> ContextManager: new_defer = defer.Deferred() curr_readers = self.key_to_current_readers.setdefault(key, set()) @@ -376,7 +403,8 @@ class ReadWriteLock(object): # We wait for the latest writer to finish writing. We can safely ignore # any existing readers... as they're readers. - yield make_deferred_yieldable(curr_writer) + if curr_writer: + await make_deferred_yieldable(curr_writer) @contextmanager def _ctx_manager(): @@ -388,8 +416,7 @@ class ReadWriteLock(object): return _ctx_manager() - @defer.inlineCallbacks - def write(self, key): + async def write(self, key: str) -> ContextManager: new_defer = defer.Deferred() curr_readers = self.key_to_current_readers.get(key, set()) @@ -405,7 +432,7 @@ class ReadWriteLock(object): curr_readers.clear() self.key_to_current_writer[key] = new_defer - yield make_deferred_yieldable(defer.gatherResults(to_wait_on)) + await make_deferred_yieldable(defer.gatherResults(to_wait_on)) @contextmanager def _ctx_manager(): @@ -419,14 +446,12 @@ class ReadWriteLock(object): return _ctx_manager() -def _cancelled_to_timed_out_error(value, timeout): - if isinstance(value, failure.Failure): - value.trap(CancelledError) - raise defer.TimeoutError(timeout, "Deferred") - return value +R = TypeVar("R") -def timeout_deferred(deferred, timeout, reactor, on_timeout_cancel=None): +def timeout_deferred( + deferred: defer.Deferred, timeout: float, reactor: IReactorTime, +) -> defer.Deferred: """The in built twisted `Deferred.addTimeout` fails to time out deferreds that have a canceller that throws exceptions. This method creates a new deferred that wraps and times out the given deferred, correctly handling @@ -434,27 +459,21 @@ def timeout_deferred(deferred, timeout, reactor, on_timeout_cancel=None): (See https://twistedmatrix.com/trac/ticket/9534) - NOTE: Unlike `Deferred.addTimeout`, this function returns a new deferred + NOTE: Unlike `Deferred.addTimeout`, this function returns a new deferred. - Args: - deferred (Deferred) - timeout (float): Timeout in seconds - reactor (twisted.interfaces.IReactorTime): The twisted reactor to use - on_timeout_cancel (callable): A callable which is called immediately - after the deferred times out, and not if this deferred is - otherwise cancelled before the timeout. + NOTE: the TimeoutError raised by the resultant deferred is + twisted.internet.defer.TimeoutError, which is *different* to the built-in + TimeoutError, as well as various other TimeoutErrors you might have imported. - It takes an arbitrary value, which is the value of the deferred at - that exact point in time (probably a CancelledError Failure), and - the timeout. + Args: + deferred: The Deferred to potentially timeout. + timeout: Timeout in seconds + reactor: The twisted reactor to use - The default callable (if none is provided) will translate a - CancelledError Failure into a defer.TimeoutError. Returns: - Deferred + A new Deferred, which will errback with defer.TimeoutError on timeout. """ - new_d = defer.Deferred() timed_out = [False] @@ -467,18 +486,23 @@ def timeout_deferred(deferred, timeout, reactor, on_timeout_cancel=None): except: # noqa: E722, if we throw any exception it'll break time outs logger.exception("Canceller failed during timeout") + # the cancel() call should have set off a chain of errbacks which + # will have errbacked new_d, but in case it hasn't, errback it now. + if not new_d.called: - new_d.errback(defer.TimeoutError(timeout, "Deferred")) + new_d.errback(defer.TimeoutError("Timed out after %gs" % (timeout,))) delayed_call = reactor.callLater(timeout, time_it_out) - def convert_cancelled(value): - if timed_out[0]: - to_call = on_timeout_cancel or _cancelled_to_timed_out_error - return to_call(value, timeout) + def convert_cancelled(value: failure.Failure): + # if the orgininal deferred was cancelled, and our timeout has fired, then + # the reason it was cancelled was due to our timeout. Turn the CancelledError + # into a TimeoutError. + if timed_out[0] and value.check(CancelledError): + raise defer.TimeoutError("Timed out after %gs" % (timeout,)) return value - deferred.addBoth(convert_cancelled) + deferred.addErrback(convert_cancelled) def cancel_timeout(result): # stop the pending call to cancel the deferred if it's been fired @@ -502,7 +526,7 @@ def timeout_deferred(deferred, timeout, reactor, on_timeout_cancel=None): @attr.s(slots=True, frozen=True) -class DoneAwaitable(object): +class DoneAwaitable: """Simple awaitable that returns the provided value. """ diff --git a/synapse/util/caches/__init__.py b/synapse/util/caches/__init__.py
index dd356bf156..8fc05be278 100644 --- a/synapse/util/caches/__init__.py +++ b/synapse/util/caches/__init__.py
@@ -42,8 +42,8 @@ response_cache_evicted = Gauge( response_cache_total = Gauge("synapse_util_caches_response_cache:total", "", ["name"]) -@attr.s -class CacheMetric(object): +@attr.s(slots=True) +class CacheMetric: _cache = attr.ib() _cache_type = attr.ib(type=str) diff --git a/synapse/util/caches/descriptors.py b/synapse/util/caches/descriptors.py
index 9b09c08b89..98b34f2223 100644 --- a/synapse/util/caches/descriptors.py +++ b/synapse/util/caches/descriptors.py
@@ -18,11 +18,10 @@ import functools import inspect import logging import threading -from typing import Any, Tuple, Union, cast +from typing import Any, Callable, Generic, Optional, Tuple, TypeVar, Union, cast from weakref import WeakValueDictionary from prometheus_client import Gauge -from typing_extensions import Protocol from twisted.internet import defer @@ -38,8 +37,10 @@ logger = logging.getLogger(__name__) CacheKey = Union[Tuple, Any] +F = TypeVar("F", bound=Callable[..., Any]) -class _CachedFunction(Protocol): + +class _CachedFunction(Generic[F]): invalidate = None # type: Any invalidate_all = None # type: Any invalidate_many = None # type: Any @@ -47,8 +48,11 @@ class _CachedFunction(Protocol): cache = None # type: Any num_args = None # type: Any - def __name__(self): - ... + __name__ = None # type: str + + # Note: This function signature is actually fiddled with by the synapse mypy + # plugin to a) make it a bound method, and b) remove any `cache_context` arg. + __call__ = None # type: F cache_pending_metric = Gauge( @@ -60,7 +64,7 @@ cache_pending_metric = Gauge( _CacheSentinel = object() -class CacheEntry(object): +class CacheEntry: __slots__ = ["deferred", "callbacks", "invalidated"] def __init__(self, deferred, callbacks): @@ -76,7 +80,7 @@ class CacheEntry(object): self.callbacks.clear() -class Cache(object): +class Cache: __slots__ = ( "cache", "name", @@ -123,7 +127,7 @@ class Cache(object): self.name = name self.keylen = keylen - self.thread = None + self.thread = None # type: Optional[threading.Thread] self.metrics = register_cache( "cache", name, @@ -192,7 +196,7 @@ class Cache(object): callbacks = [callback] if callback else [] self.check_thread() observable = ObservableDeferred(value, consumeErrors=True) - observer = defer.maybeDeferred(observable.observe) + observer = observable.observe() entry = CacheEntry(deferred=observable, callbacks=callbacks) existing_entry = self._pending_deferred_cache.pop(key, None) @@ -284,17 +288,10 @@ class Cache(object): self._pending_deferred_cache.clear() -class _CacheDescriptorBase(object): - def __init__( - self, orig: _CachedFunction, num_args, inlineCallbacks, cache_context=False - ): +class _CacheDescriptorBase: + def __init__(self, orig: _CachedFunction, num_args, cache_context=False): self.orig = orig - if inlineCallbacks: - self.function_to_call = defer.inlineCallbacks(orig) - else: - self.function_to_call = orig - arg_spec = inspect.getfullargspec(orig) all_args = arg_spec.args @@ -364,7 +361,7 @@ class CacheDescriptor(_CacheDescriptorBase): invalidated) by adding a special "cache_context" argument to the function and passing that as a kwarg to all caches called. For example:: - @cachedInlineCallbacks(cache_context=True) + @cached(cache_context=True) def foo(self, key, cache_context): r1 = yield self.bar1(key, on_invalidate=cache_context.invalidate) r2 = yield self.bar2(key, on_invalidate=cache_context.invalidate) @@ -382,17 +379,11 @@ class CacheDescriptor(_CacheDescriptorBase): max_entries=1000, num_args=None, tree=False, - inlineCallbacks=False, cache_context=False, iterable=False, ): - super(CacheDescriptor, self).__init__( - orig, - num_args=num_args, - inlineCallbacks=inlineCallbacks, - cache_context=cache_context, - ) + super().__init__(orig, num_args=num_args, cache_context=cache_context) self.max_entries = max_entries self.tree = tree @@ -465,9 +456,7 @@ class CacheDescriptor(_CacheDescriptorBase): observer = defer.succeed(cached_result_d) except KeyError: - ret = defer.maybeDeferred( - preserve_fn(self.function_to_call), obj, *args, **kwargs - ) + ret = defer.maybeDeferred(preserve_fn(self.orig), obj, *args, **kwargs) def onErr(f): cache.invalidate(cache_key) @@ -510,9 +499,7 @@ class CacheListDescriptor(_CacheDescriptorBase): of results. """ - def __init__( - self, orig, cached_method_name, list_name, num_args=None, inlineCallbacks=False - ): + def __init__(self, orig, cached_method_name, list_name, num_args=None): """ Args: orig (function) @@ -521,12 +508,8 @@ class CacheListDescriptor(_CacheDescriptorBase): num_args (int): number of positional arguments (excluding ``self``, but including list_name) to use as cache keys. Defaults to all named args of the function. - inlineCallbacks (bool): Whether orig is a generator that should - be wrapped by defer.inlineCallbacks """ - super(CacheListDescriptor, self).__init__( - orig, num_args=num_args, inlineCallbacks=inlineCallbacks - ) + super().__init__(orig, num_args=num_args) self.list_name = list_name @@ -631,7 +614,7 @@ class CacheListDescriptor(_CacheDescriptorBase): cached_defers.append( defer.maybeDeferred( - preserve_fn(self.function_to_call), **args_to_call + preserve_fn(self.orig), **args_to_call ).addCallbacks(complete_all, errback) ) @@ -683,9 +666,13 @@ class _CacheContext: def cached( - max_entries=1000, num_args=None, tree=False, cache_context=False, iterable=False -): - return lambda orig: CacheDescriptor( + max_entries: int = 1000, + num_args: Optional[int] = None, + tree: bool = False, + cache_context: bool = False, + iterable: bool = False, +) -> Callable[[F], _CachedFunction[F]]: + func = lambda orig: CacheDescriptor( orig, max_entries=max_entries, num_args=num_args, @@ -694,22 +681,12 @@ def cached( iterable=iterable, ) - -def cachedInlineCallbacks( - max_entries=1000, num_args=None, tree=False, cache_context=False, iterable=False -): - return lambda orig: CacheDescriptor( - orig, - max_entries=max_entries, - num_args=num_args, - tree=tree, - inlineCallbacks=True, - cache_context=cache_context, - iterable=iterable, - ) + return cast(Callable[[F], _CachedFunction[F]], func) -def cachedList(cached_method_name, list_name, num_args=None, inlineCallbacks=False): +def cachedList( + cached_method_name: str, list_name: str, num_args: Optional[int] = None +) -> Callable[[F], _CachedFunction[F]]: """Creates a descriptor that wraps a function in a `CacheListDescriptor`. Used to do batch lookups for an already created cache. A single argument @@ -719,18 +696,16 @@ def cachedList(cached_method_name, list_name, num_args=None, inlineCallbacks=Fal cache. Args: - cached_method_name (str): The name of the single-item lookup method. + cached_method_name: The name of the single-item lookup method. This is only used to find the cache to use. - list_name (str): The name of the argument that is the list to use to + list_name: The name of the argument that is the list to use to do batch lookups in the cache. - num_args (int): Number of arguments to use as the key in the cache + num_args: Number of arguments to use as the key in the cache (including list_name). Defaults to all named parameters. - inlineCallbacks (bool): Should the function be wrapped in an - `defer.inlineCallbacks`? Example: - class Example(object): + class Example: @cached(num_args=2) def do_something(self, first_arg): ... @@ -739,10 +714,11 @@ def cachedList(cached_method_name, list_name, num_args=None, inlineCallbacks=Fal def batch_do_something(self, first_arg, second_args): ... """ - return lambda orig: CacheListDescriptor( + func = lambda orig: CacheListDescriptor( orig, cached_method_name=cached_method_name, list_name=list_name, num_args=num_args, - inlineCallbacks=inlineCallbacks, ) + + return cast(Callable[[F], _CachedFunction[F]], func) diff --git a/synapse/util/caches/dictionary_cache.py b/synapse/util/caches/dictionary_cache.py
index 6834e6f3ae..8592b93689 100644 --- a/synapse/util/caches/dictionary_cache.py +++ b/synapse/util/caches/dictionary_cache.py
@@ -40,7 +40,7 @@ class DictionaryEntry(namedtuple("DictionaryEntry", ("full", "known_absent", "va return len(self.value) -class DictionaryCache(object): +class DictionaryCache: """Caches key -> dictionary lookups, supporting caching partial dicts, i.e. fetching a subset of dictionary keys for a particular key. """ @@ -53,7 +53,7 @@ class DictionaryCache(object): self.thread = None # caches_by_name[name] = self.cache - class Sentinel(object): + class Sentinel: __slots__ = [] self.sentinel = Sentinel() diff --git a/synapse/util/caches/expiringcache.py b/synapse/util/caches/expiringcache.py
index 89a3420f92..e15f7ee698 100644 --- a/synapse/util/caches/expiringcache.py +++ b/synapse/util/caches/expiringcache.py
@@ -26,7 +26,7 @@ logger = logging.getLogger(__name__) SENTINEL = object() -class ExpiringCache(object): +class ExpiringCache: def __init__( self, cache_name, @@ -190,7 +190,7 @@ class ExpiringCache(object): return False -class _CacheEntry(object): +class _CacheEntry: __slots__ = ["time", "value"] def __init__(self, time, value): diff --git a/synapse/util/caches/lrucache.py b/synapse/util/caches/lrucache.py
index df4ea5901d..4bc1a67b58 100644 --- a/synapse/util/caches/lrucache.py +++ b/synapse/util/caches/lrucache.py
@@ -30,7 +30,7 @@ def enumerate_leaves(node, depth): yield m -class _Node(object): +class _Node: __slots__ = ["prev_node", "next_node", "key", "value", "callbacks"] def __init__(self, prev_node, next_node, key, value, callbacks=set()): @@ -41,7 +41,7 @@ class _Node(object): self.callbacks = callbacks -class LruCache(object): +class LruCache: """ Least-recently-used cache. Supports del_multi only if cache_type=TreeCache diff --git a/synapse/util/caches/response_cache.py b/synapse/util/caches/response_cache.py
index a6c60888e5..df1a721add 100644 --- a/synapse/util/caches/response_cache.py +++ b/synapse/util/caches/response_cache.py
@@ -23,7 +23,7 @@ from synapse.util.caches import register_cache logger = logging.getLogger(__name__) -class ResponseCache(object): +class ResponseCache: """ This caches a deferred response. Until the deferred completes it will be returned from the cache. This means that if the client retries the request diff --git a/synapse/util/caches/treecache.py b/synapse/util/caches/treecache.py
index ecd9948e79..eb4d98f683 100644 --- a/synapse/util/caches/treecache.py +++ b/synapse/util/caches/treecache.py
@@ -3,7 +3,7 @@ from typing import Dict SENTINEL = object() -class TreeCache(object): +class TreeCache: """ Tree-based backing store for LruCache. Allows subtrees of data to be deleted efficiently. @@ -89,7 +89,7 @@ def iterate_tree_cache_entry(d): yield d -class _Entry(object): +class _Entry: __slots__ = ["value"] def __init__(self, value): diff --git a/synapse/util/caches/ttlcache.py b/synapse/util/caches/ttlcache.py
index 6437aa907e..3e180cafd3 100644 --- a/synapse/util/caches/ttlcache.py +++ b/synapse/util/caches/ttlcache.py
@@ -26,7 +26,7 @@ logger = logging.getLogger(__name__) SENTINEL = object() -class TTLCache(object): +class TTLCache: """A key/value cache implementation where each entry has its own TTL""" def __init__(self, cache_name, timer=time.time): @@ -154,7 +154,7 @@ class TTLCache(object): @attr.s(frozen=True, slots=True) -class _CacheEntry(object): +class _CacheEntry: """TTLCache entry""" # expiry_time is the first attribute, so that entries are sorted by expiry. diff --git a/synapse/util/daemonize.py b/synapse/util/daemonize.py new file mode 100644
index 0000000000..23393cf49b --- /dev/null +++ b/synapse/util/daemonize.py
@@ -0,0 +1,137 @@ +# -*- coding: utf-8 -*- +# Copyright (c) 2012, 2013, 2014 Ilya Otyutskiy <ilya.otyutskiy@icloud.com> +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import atexit +import fcntl +import logging +import os +import signal +import sys + + +def daemonize_process(pid_file: str, logger: logging.Logger, chdir: str = "/") -> None: + """daemonize the current process + + This calls fork(), and has the main process exit. When it returns we will be + running in the child process. + """ + + # If pidfile already exists, we should read pid from there; to overwrite it, if + # locking will fail, because locking attempt somehow purges the file contents. + if os.path.isfile(pid_file): + with open(pid_file, "r") as pid_fh: + old_pid = pid_fh.read() + + # Create a lockfile so that only one instance of this daemon is running at any time. + try: + lock_fh = open(pid_file, "w") + except IOError: + print("Unable to create the pidfile.") + sys.exit(1) + + try: + # Try to get an exclusive lock on the file. This will fail if another process + # has the file locked. + fcntl.flock(lock_fh, fcntl.LOCK_EX | fcntl.LOCK_NB) + except IOError: + print("Unable to lock on the pidfile.") + # We need to overwrite the pidfile if we got here. + # + # XXX better to avoid overwriting it, surely. this looks racey as the pid file + # could be created between us trying to read it and us trying to lock it. + with open(pid_file, "w") as pid_fh: + pid_fh.write(old_pid) + sys.exit(1) + + # Fork, creating a new process for the child. + process_id = os.fork() + + if process_id != 0: + # parent process: exit. + + # we use os._exit to avoid running the atexit handlers. In particular, that + # means we don't flush the logs. This is important because if we are using + # a MemoryHandler, we could have logs buffered which are now buffered in both + # the main and the child process, so if we let the main process flush the logs, + # we'll get two copies. + os._exit(0) + + # This is the child process. Continue. + + # Stop listening for signals that the parent process receives. + # This is done by getting a new process id. + # setpgrp() is an alternative to setsid(). + # setsid puts the process in a new parent group and detaches its controlling + # terminal. + + os.setsid() + + # point stdin, stdout, stderr at /dev/null + devnull = "/dev/null" + if hasattr(os, "devnull"): + # Python has set os.devnull on this system, use it instead as it might be + # different than /dev/null. + devnull = os.devnull + + devnull_fd = os.open(devnull, os.O_RDWR) + os.dup2(devnull_fd, 0) + os.dup2(devnull_fd, 1) + os.dup2(devnull_fd, 2) + os.close(devnull_fd) + + # now that we have redirected stderr to /dev/null, any uncaught exceptions will + # get sent to /dev/null, so make sure we log them. + # + # (we don't normally expect reactor.run to raise any exceptions, but this will + # also catch any other uncaught exceptions before we get that far.) + + def excepthook(type_, value, traceback): + logger.critical("Unhanded exception", exc_info=(type_, value, traceback)) + + sys.excepthook = excepthook + + # Set umask to default to safe file permissions when running as a root daemon. 027 + # is an octal number which we are typing as 0o27 for Python3 compatibility. + os.umask(0o27) + + # Change to a known directory. If this isn't done, starting a daemon in a + # subdirectory that needs to be deleted results in "directory busy" errors. + os.chdir(chdir) + + try: + lock_fh.write("%s" % (os.getpid())) + lock_fh.flush() + except IOError: + logger.error("Unable to write pid to the pidfile.") + print("Unable to write pid to the pidfile.") + sys.exit(1) + + # write a log line on SIGTERM. + def sigterm(signum, frame): + logger.warning("Caught signal %s. Stopping daemon." % signum) + sys.exit(0) + + signal.signal(signal.SIGTERM, sigterm) + + # Cleanup pid file at exit. + def exit(): + logger.warning("Stopping daemon.") + os.remove(pid_file) + sys.exit(0) + + atexit.register(exit) + + logger.warning("Starting daemon.") diff --git a/synapse/util/distributor.py b/synapse/util/distributor.py
index 22a857a306..f73e95393c 100644 --- a/synapse/util/distributor.py +++ b/synapse/util/distributor.py
@@ -16,8 +16,6 @@ import inspect import logging from twisted.internet import defer -from twisted.internet.defer import Deferred, fail, succeed -from twisted.python import failure from synapse.logging.context import make_deferred_yieldable, run_in_background from synapse.metrics.background_process_metrics import run_as_background_process @@ -29,12 +27,7 @@ def user_left_room(distributor, user, room_id): distributor.fire("user_left_room", user=user, room_id=room_id) -# XXX: this is no longer used. We should probably kill it. -def user_joined_room(distributor, user, room_id): - distributor.fire("user_joined_room", user=user, room_id=room_id) - - -class Distributor(object): +class Distributor: """A central dispatch point for loosely-connected pieces of code to register, observe, and fire signals. @@ -81,29 +74,7 @@ class Distributor(object): run_as_background_process(name, self.signals[name].fire, *args, **kwargs) -def maybeAwaitableDeferred(f, *args, **kw): - """ - Invoke a function that may or may not return a Deferred or an Awaitable. - - This is a modified version of twisted.internet.defer.maybeDeferred. - """ - try: - result = f(*args, **kw) - except Exception: - return fail(failure.Failure(captureVars=Deferred.debug)) - - if isinstance(result, Deferred): - return result - # Handle the additional case of an awaitable being returned. - elif inspect.isawaitable(result): - return defer.ensureDeferred(result) - elif isinstance(result, failure.Failure): - return fail(result) - else: - return succeed(result) - - -class Signal(object): +class Signal: """A Signal is a dispatch point that stores a list of callables as observers of it. @@ -132,22 +103,17 @@ class Signal(object): Returns a Deferred that will complete when all the observers have completed.""" - def do(observer): - def eb(failure): + async def do(observer): + try: + result = observer(*args, **kwargs) + if inspect.isawaitable(result): + result = await result + return result + except Exception as e: logger.warning( - "%s signal observer %s failed: %r", - self.name, - observer, - failure, - exc_info=( - failure.type, - failure.value, - failure.getTracebackObject(), - ), + "%s signal observer %s failed: %r", self.name, observer, e, ) - return maybeAwaitableDeferred(observer, *args, **kwargs).addErrback(eb) - deferreds = [run_in_background(do, o) for o in self.observers] return make_deferred_yieldable( diff --git a/synapse/util/file_consumer.py b/synapse/util/file_consumer.py
index 6a3f6177b1..733f5e26e6 100644 --- a/synapse/util/file_consumer.py +++ b/synapse/util/file_consumer.py
@@ -20,7 +20,7 @@ from twisted.internet import threads from synapse.logging.context import make_deferred_yieldable, run_in_background -class BackgroundFileConsumer(object): +class BackgroundFileConsumer: """A consumer that writes to a file like object. Supports both push and pull producers diff --git a/synapse/util/frozenutils.py b/synapse/util/frozenutils.py
index eab78dd256..bf094c9386 100644 --- a/synapse/util/frozenutils.py +++ b/synapse/util/frozenutils.py
@@ -13,7 +13,8 @@ # See the License for the specific language governing permissions and # limitations under the License. -from canonicaljson import json +import json + from frozendict import frozendict @@ -63,5 +64,8 @@ def _handle_frozendict(obj): ) -# A JSONEncoder which is capable of encoding frozendicts without barfing -frozendict_json_encoder = json.JSONEncoder(default=_handle_frozendict) +# A JSONEncoder which is capable of encoding frozendicts without barfing. +# Additionally reduce the whitespace produced by JSON encoding. +frozendict_json_encoder = json.JSONEncoder( + allow_nan=False, separators=(",", ":"), default=_handle_frozendict, +) diff --git a/synapse/util/jsonobject.py b/synapse/util/jsonobject.py
index 6dce03dd3a..50516926f3 100644 --- a/synapse/util/jsonobject.py +++ b/synapse/util/jsonobject.py
@@ -14,7 +14,7 @@ # limitations under the License. -class JsonEncodedObject(object): +class JsonEncodedObject: """ A common base class for defining protocol units that are represented as JSON. diff --git a/synapse/util/manhole.py b/synapse/util/manhole.py
index 631654f297..da24ba0470 100644 --- a/synapse/util/manhole.py +++ b/synapse/util/manhole.py
@@ -94,7 +94,7 @@ class SynapseManhole(ColoredManhole): """Overrides connectionMade to create our own ManholeInterpreter""" def connectionMade(self): - super(SynapseManhole, self).connectionMade() + super().connectionMade() # replace the manhole interpreter with our own impl self.interpreter = SynapseManholeInterpreter(self, self.namespace) diff --git a/synapse/util/metrics.py b/synapse/util/metrics.py
index ec61e14423..ffdea0de8d 100644 --- a/synapse/util/metrics.py +++ b/synapse/util/metrics.py
@@ -13,15 +13,17 @@ # See the License for the specific language governing permissions and # limitations under the License. -import inspect import logging from functools import wraps +from typing import Any, Callable, Optional, TypeVar, cast from prometheus_client import Counter -from twisted.internet import defer - -from synapse.logging.context import LoggingContext, current_context +from synapse.logging.context import ( + ContextResourceUsage, + LoggingContext, + current_context, +) from synapse.metrics import InFlightGauge logger = logging.getLogger(__name__) @@ -60,34 +62,42 @@ in_flight = InFlightGauge( sub_metrics=["real_time_max", "real_time_sum"], ) +T = TypeVar("T", bound=Callable[..., Any]) -def measure_func(name=None): - def wrapper(func): - block_name = func.__name__ if name is None else name - if inspect.iscoroutinefunction(func): +def measure_func(name: Optional[str] = None) -> Callable[[T], T]: + """ + Used to decorate an async function with a `Measure` context manager. + + Usage: + + @measure_func() + async def foo(...): + ... - @wraps(func) - async def measured_func(self, *args, **kwargs): - with Measure(self.clock, block_name): - r = await func(self, *args, **kwargs) - return r + Which is analogous to: - else: + async def foo(...): + with Measure(...): + ... - @wraps(func) - @defer.inlineCallbacks - def measured_func(self, *args, **kwargs): - with Measure(self.clock, block_name): - r = yield func(self, *args, **kwargs) - return r + """ - return measured_func + def wrapper(func: T) -> T: + block_name = func.__name__ if name is None else name + + @wraps(func) + async def measured_func(self, *args, **kwargs): + with Measure(self.clock, block_name): + r = await func(self, *args, **kwargs) + return r + + return cast(T, measured_func) return wrapper -class Measure(object): +class Measure: __slots__ = [ "clock", "name", @@ -98,27 +108,27 @@ class Measure(object): def __init__(self, clock, name): self.clock = clock self.name = name - self._logging_context = None + parent_context = current_context() + self._logging_context = LoggingContext( + "Measure[%s]" % (self.name,), parent_context + ) self.start = None - def __enter__(self): - if self._logging_context: + def __enter__(self) -> "Measure": + if self.start is not None: raise RuntimeError("Measure() objects cannot be re-used") self.start = self.clock.time() - parent_context = current_context() - self._logging_context = LoggingContext( - "Measure[%s]" % (self.name,), parent_context - ) self._logging_context.__enter__() in_flight.register((self.name,), self._update_in_flight) + return self def __exit__(self, exc_type, exc_val, exc_tb): - if not self._logging_context: + if self.start is None: raise RuntimeError("Measure() block exited without being entered") duration = self.clock.time() - self.start - usage = self._logging_context.get_resource_usage() + usage = self.get_resource_usage() in_flight.unregister((self.name,), self._update_in_flight) self._logging_context.__exit__(exc_type, exc_val, exc_tb) @@ -134,6 +144,13 @@ class Measure(object): except ValueError: logger.warning("Failed to save metrics! Usage: %s", usage) + def get_resource_usage(self) -> ContextResourceUsage: + """Get the resources used within this Measure block + + If the Measure block is still active, returns the resource usage so far. + """ + return self._logging_context.get_resource_usage() + def _update_in_flight(self, metrics): """Gets called when processing in flight metrics """ diff --git a/synapse/util/patch_inline_callbacks.py b/synapse/util/patch_inline_callbacks.py
index 54c046b6e1..72574d3af2 100644 --- a/synapse/util/patch_inline_callbacks.py +++ b/synapse/util/patch_inline_callbacks.py
@@ -13,8 +13,6 @@ # See the License for the specific language governing permissions and # limitations under the License. -from __future__ import print_function - import functools import sys from typing import Any, Callable, List diff --git a/synapse/util/ratelimitutils.py b/synapse/util/ratelimitutils.py
index e5efdfcd02..70d11e1ec3 100644 --- a/synapse/util/ratelimitutils.py +++ b/synapse/util/ratelimitutils.py
@@ -29,7 +29,7 @@ from synapse.logging.context import ( logger = logging.getLogger(__name__) -class FederationRateLimiter(object): +class FederationRateLimiter: def __init__(self, clock, config): """ Args: @@ -60,7 +60,7 @@ class FederationRateLimiter(object): return self.ratelimiters[host].ratelimit() -class _PerHostRatelimiter(object): +class _PerHostRatelimiter: def __init__(self, clock, config): """ Args: diff --git a/synapse/util/retryutils.py b/synapse/util/retryutils.py
index 8794317caa..a5cc9d0551 100644 --- a/synapse/util/retryutils.py +++ b/synapse/util/retryutils.py
@@ -15,8 +15,6 @@ import logging import random -from twisted.internet import defer - import synapse.logging.context from synapse.api.errors import CodeMessageException @@ -47,15 +45,14 @@ class NotRetryingDestination(Exception): """ msg = "Not retrying server %s." % (destination,) - super(NotRetryingDestination, self).__init__(msg) + super().__init__(msg) self.retry_last_ts = retry_last_ts self.retry_interval = retry_interval self.destination = destination -@defer.inlineCallbacks -def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs): +async def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs): """For a given destination check if we have previously failed to send a request there and are waiting before retrying the destination. If we are not ready to retry the destination, this will raise a @@ -73,9 +70,9 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs) Example usage: try: - limiter = yield get_retry_limiter(destination, clock, store) + limiter = await get_retry_limiter(destination, clock, store) with limiter: - response = yield do_request() + response = await do_request() except NotRetryingDestination: # We aren't ready to retry that destination. raise @@ -83,7 +80,7 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs) failure_ts = None retry_last_ts, retry_interval = (0, 0) - retry_timings = yield store.get_destination_retry_timings(destination) + retry_timings = await store.get_destination_retry_timings(destination) if retry_timings: failure_ts = retry_timings["failure_ts"] @@ -117,7 +114,7 @@ def get_retry_limiter(destination, clock, store, ignore_backoff=False, **kwargs) ) -class RetryDestinationLimiter(object): +class RetryDestinationLimiter: def __init__( self, destination, @@ -222,10 +219,9 @@ class RetryDestinationLimiter(object): if self.failure_ts is None: self.failure_ts = retry_last_ts - @defer.inlineCallbacks - def store_retry_timings(): + async def store_retry_timings(): try: - yield self.store.set_destination_retry_timings( + await self.store.set_destination_retry_timings( self.destination, self.failure_ts, retry_last_ts, diff --git a/synapse/util/stringutils.py b/synapse/util/stringutils.py
index 2e2b40a426..61d96a6c28 100644 --- a/synapse/util/stringutils.py +++ b/synapse/util/stringutils.py
@@ -24,9 +24,7 @@ from synapse.api.errors import Codes, SynapseError _string_with_symbols = string.digits + string.ascii_letters + ".,;:^&*-_+=#~@" # https://matrix.org/docs/spec/client_server/r0.6.0#post-matrix-client-r0-register-email-requesttoken -# Note: The : character is allowed here for older clients, but will be removed in a -# future release. Context: https://github.com/matrix-org/synapse/issues/6766 -client_secret_regex = re.compile(r"^[0-9a-zA-Z\.\=\_\-\:]+$") +client_secret_regex = re.compile(r"^[0-9a-zA-Z\.\=\_\-]+$") # random_string and random_string_with_symbols are used for a range of things, # some cryptographically important, some less so. We use SystemRandom to make sure diff --git a/synapse/util/wheel_timer.py b/synapse/util/wheel_timer.py
index 023beb5ede..be3b22469d 100644 --- a/synapse/util/wheel_timer.py +++ b/synapse/util/wheel_timer.py
@@ -14,7 +14,7 @@ # limitations under the License. -class _Entry(object): +class _Entry: __slots__ = ["end_key", "queue"] def __init__(self, end_key): @@ -22,7 +22,7 @@ class _Entry(object): self.queue = [] -class WheelTimer(object): +class WheelTimer: """Stores arbitrary objects that will be returned after their timers have expired. """ diff --git a/synapse/visibility.py b/synapse/visibility.py
index 0f042c5696..e3da7744d2 100644 --- a/synapse/visibility.py +++ b/synapse/visibility.py
@@ -16,8 +16,6 @@ import logging import operator -from twisted.internet import defer - from synapse.api.constants import EventTypes, Membership from synapse.events.utils import prune_event from synapse.storage import Storage @@ -39,8 +37,7 @@ MEMBERSHIP_PRIORITY = ( ) -@defer.inlineCallbacks -def filter_events_for_client( +async def filter_events_for_client( storage: Storage, user_id, events, @@ -67,19 +64,19 @@ def filter_events_for_client( also be called to check whether a user can see the state at a given point. Returns: - Deferred[list[synapse.events.EventBase]] + list[synapse.events.EventBase] """ # Filter out events that have been soft failed so that we don't relay them # to clients. events = [e for e in events if not e.internal_metadata.is_soft_failed()] types = ((EventTypes.RoomHistoryVisibility, ""), (EventTypes.Member, user_id)) - event_id_to_state = yield storage.state.get_state_for_events( + event_id_to_state = await storage.state.get_state_for_events( frozenset(e.event_id for e in events), state_filter=StateFilter.from_types(types), ) - ignore_dict_content = yield storage.main.get_global_account_data_by_type_for_user( + ignore_dict_content = await storage.main.get_global_account_data_by_type_for_user( "m.ignored_user_list", user_id ) @@ -90,7 +87,7 @@ def filter_events_for_client( else [] ) - erased_senders = yield storage.main.are_users_erased((e.sender for e in events)) + erased_senders = await storage.main.are_users_erased((e.sender for e in events)) if filter_send_to_client: room_ids = {e.room_id for e in events} @@ -99,7 +96,7 @@ def filter_events_for_client( for room_id in room_ids: retention_policies[ room_id - ] = yield storage.main.get_retention_policy_for_room(room_id) + ] = await storage.main.get_retention_policy_for_room(room_id) def allowed(event): """ @@ -254,8 +251,7 @@ def filter_events_for_client( return list(filtered_events) -@defer.inlineCallbacks -def filter_events_for_server( +async def filter_events_for_server( storage: Storage, server_name, events, @@ -277,7 +273,7 @@ def filter_events_for_server( backfill or not. Returns - Deferred[list[FrozenEvent]] + list[FrozenEvent] """ def is_sender_erased(event, erased_senders): @@ -321,7 +317,7 @@ def filter_events_for_server( # Lets check to see if all the events have a history visibility # of "shared" or "world_readable". If that's the case then we don't # need to check membership (as we know the server is in the room). - event_to_state_ids = yield storage.state.get_state_ids_for_events( + event_to_state_ids = await storage.state.get_state_ids_for_events( frozenset(e.event_id for e in events), state_filter=StateFilter.from_types( types=((EventTypes.RoomHistoryVisibility, ""),) @@ -339,14 +335,14 @@ def filter_events_for_server( if not visibility_ids: all_open = True else: - event_map = yield storage.main.get_events(visibility_ids) + event_map = await storage.main.get_events(visibility_ids) all_open = all( e.content.get("history_visibility") in (None, "shared", "world_readable") for e in event_map.values() ) if not check_history_visibility_only: - erased_senders = yield storage.main.are_users_erased((e.sender for e in events)) + erased_senders = await storage.main.are_users_erased((e.sender for e in events)) else: # We don't want to check whether users are erased, which is equivalent # to no users having been erased. @@ -375,7 +371,7 @@ def filter_events_for_server( # first, for each event we're wanting to return, get the event_ids # of the history vis and membership state at those events. - event_to_state_ids = yield storage.state.get_state_ids_for_events( + event_to_state_ids = await storage.state.get_state_ids_for_events( frozenset(e.event_id for e in events), state_filter=StateFilter.from_types( types=((EventTypes.RoomHistoryVisibility, ""), (EventTypes.Member, None)) @@ -405,7 +401,7 @@ def filter_events_for_server( return False return state_key[idx + 1 :] == server_name - event_map = yield storage.main.get_events( + event_map = await storage.main.get_events( [e_id for e_id, key in event_id_to_state_key.items() if include(key[0], key[1])] ) diff --git a/synmark/__init__.py b/synmark/__init__.py
index afe4fad8cb..53698bd5ab 100644 --- a/synmark/__init__.py +++ b/synmark/__init__.py
@@ -47,9 +47,9 @@ async def make_homeserver(reactor, config=None): stor = hs.get_datastore() # Run the database background updates. - if hasattr(stor.db.updates, "do_next_background_update"): - while not await stor.db.updates.has_completed_background_updates(): - await stor.db.updates.do_next_background_update(1) + if hasattr(stor.db_pool.updates, "do_next_background_update"): + while not await stor.db_pool.updates.has_completed_background_updates(): + await stor.db_pool.updates.do_next_background_update(1) def cleanup(): for i in cleanup_tasks: diff --git a/sytest-blacklist b/sytest-blacklist
index 7ceeaca8d6..a0e72eb155 100644 --- a/sytest-blacklist +++ b/sytest-blacklist
@@ -37,6 +37,14 @@ Inbound federation of state requires event_id as a mandatory paramater # Blacklisted until https://github.com/matrix-org/synapse/pull/6486 lands Can upload self-signing keys +# Blacklisted until MSC2753 is implemented +Local users can peek into world_readable rooms by room ID +We can't peek into rooms with shared history_visibility +We can't peek into rooms with invited history_visibility +We can't peek into rooms with joined history_visibility +Local users can peek by room alias +Peeked rooms only turn up in the sync for the device who peeked them + # Blacklisted due to https://github.com/matrix-org/synapse-dinsic/issues/43 Inviting an AS-hosted user asks the AS server Accesing an AS-hosted room alias asks the AS server diff --git a/tests/api/test_auth.py b/tests/api/test_auth.py
index 0bfb86bf1f..8ab56ec94c 100644 --- a/tests/api/test_auth.py +++ b/tests/api/test_auth.py
@@ -36,7 +36,7 @@ from tests import unittest from tests.utils import mock_getRawHeaders, setup_test_homeserver -class TestHandlers(object): +class TestHandlers: def __init__(self, hs): self.auth_handler = synapse.handlers.auth.AuthHandler(hs) @@ -62,12 +62,15 @@ class AuthTestCase(unittest.TestCase): # this is overridden for the appservice tests self.store.get_app_service_by_token = Mock(return_value=None) + self.store.insert_client_ip = Mock(return_value=defer.succeed(None)) self.store.is_support_user = Mock(return_value=defer.succeed(False)) @defer.inlineCallbacks def test_get_user_by_req_user_valid_token(self): user_info = {"name": self.test_user, "token_id": "ditto", "device_id": "device"} - self.store.get_user_by_access_token = Mock(return_value=user_info) + self.store.get_user_by_access_token = Mock( + return_value=defer.succeed(user_info) + ) request = Mock(args={}) request.args[b"access_token"] = [self.test_token] @@ -76,23 +79,25 @@ class AuthTestCase(unittest.TestCase): self.assertEquals(requester.user.to_string(), self.test_user) def test_get_user_by_req_user_bad_token(self): - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.args[b"access_token"] = [self.test_token] request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) f = self.failureResultOf(d, InvalidClientTokenError).value self.assertEqual(f.code, 401) self.assertEqual(f.errcode, "M_UNKNOWN_TOKEN") def test_get_user_by_req_user_missing_token(self): user_info = {"name": self.test_user, "token_id": "ditto"} - self.store.get_user_by_access_token = Mock(return_value=user_info) + self.store.get_user_by_access_token = Mock( + return_value=defer.succeed(user_info) + ) request = Mock(args={}) request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) f = self.failureResultOf(d, MissingClientTokenError).value self.assertEqual(f.code, 401) self.assertEqual(f.errcode, "M_MISSING_TOKEN") @@ -103,7 +108,7 @@ class AuthTestCase(unittest.TestCase): token="foobar", url="a_url", sender=self.test_user, ip_range_whitelist=None ) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.getClientIP.return_value = "127.0.0.1" @@ -123,7 +128,7 @@ class AuthTestCase(unittest.TestCase): ip_range_whitelist=IPSet(["192.168/16"]), ) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.getClientIP.return_value = "192.168.10.10" @@ -142,25 +147,25 @@ class AuthTestCase(unittest.TestCase): ip_range_whitelist=IPSet(["192.168/16"]), ) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.getClientIP.return_value = "131.111.8.42" request.args[b"access_token"] = [self.test_token] request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) f = self.failureResultOf(d, InvalidClientTokenError).value self.assertEqual(f.code, 401) self.assertEqual(f.errcode, "M_UNKNOWN_TOKEN") def test_get_user_by_req_appservice_bad_token(self): self.store.get_app_service_by_token = Mock(return_value=None) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.args[b"access_token"] = [self.test_token] request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) f = self.failureResultOf(d, InvalidClientTokenError).value self.assertEqual(f.code, 401) self.assertEqual(f.errcode, "M_UNKNOWN_TOKEN") @@ -168,11 +173,11 @@ class AuthTestCase(unittest.TestCase): def test_get_user_by_req_appservice_missing_token(self): app_service = Mock(token="foobar", url="a_url", sender=self.test_user) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) f = self.failureResultOf(d, MissingClientTokenError).value self.assertEqual(f.code, 401) self.assertEqual(f.errcode, "M_MISSING_TOKEN") @@ -185,7 +190,11 @@ class AuthTestCase(unittest.TestCase): ) app_service.is_interested_in_user = Mock(return_value=True) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + # This just needs to return a truth-y value. + self.store.get_user_by_id = Mock( + return_value=defer.succeed({"is_guest": False}) + ) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.getClientIP.return_value = "127.0.0.1" @@ -204,20 +213,22 @@ class AuthTestCase(unittest.TestCase): ) app_service.is_interested_in_user = Mock(return_value=False) self.store.get_app_service_by_token = Mock(return_value=app_service) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) request = Mock(args={}) request.getClientIP.return_value = "127.0.0.1" request.args[b"access_token"] = [self.test_token] request.args[b"user_id"] = [masquerading_user_id] request.requestHeaders.getRawHeaders = mock_getRawHeaders() - d = self.auth.get_user_by_req(request) + d = defer.ensureDeferred(self.auth.get_user_by_req(request)) self.failureResultOf(d, AuthError) @defer.inlineCallbacks def test_get_user_from_macaroon(self): self.store.get_user_by_access_token = Mock( - return_value={"name": "@baldrick:matrix.org", "device_id": "device"} + return_value=defer.succeed( + {"name": "@baldrick:matrix.org", "device_id": "device"} + ) ) user_id = "@baldrick:matrix.org" @@ -241,8 +252,8 @@ class AuthTestCase(unittest.TestCase): @defer.inlineCallbacks def test_get_guest_user_from_macaroon(self): - self.store.get_user_by_id = Mock(return_value={"is_guest": True}) - self.store.get_user_by_access_token = Mock(return_value=None) + self.store.get_user_by_id = Mock(return_value=defer.succeed({"is_guest": True})) + self.store.get_user_by_access_token = Mock(return_value=defer.succeed(None)) user_id = "@baldrick:matrix.org" macaroon = pymacaroons.Macaroon( @@ -282,16 +293,20 @@ class AuthTestCase(unittest.TestCase): def get_user(tok): if token != tok: - return None - return { - "name": USER_ID, - "is_guest": False, - "token_id": 1234, - "device_id": "DEVICE", - } + return defer.succeed(None) + return defer.succeed( + { + "name": USER_ID, + "is_guest": False, + "token_id": 1234, + "device_id": "DEVICE", + } + ) self.store.get_user_by_access_token = get_user - self.store.get_user_by_id = Mock(return_value={"is_guest": False}) + self.store.get_user_by_id = Mock( + return_value=defer.succeed({"is_guest": False}) + ) # check the token works request = Mock(args={}) diff --git a/tests/api/test_filtering.py b/tests/api/test_filtering.py
index 4e67503cf0..d2d535d23c 100644 --- a/tests/api/test_filtering.py +++ b/tests/api/test_filtering.py
@@ -369,14 +369,18 @@ class FilteringTestCase(unittest.TestCase): @defer.inlineCallbacks def test_filter_presence_match(self): user_filter_json = {"presence": {"types": ["m.*"]}} - filter_id = yield self.datastore.add_user_filter( - user_localpart=user_localpart, user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.datastore.add_user_filter( + user_localpart=user_localpart, user_filter=user_filter_json + ) ) event = MockEvent(sender="@foo:bar", type="m.profile") events = [event] - user_filter = yield self.filtering.get_user_filter( - user_localpart=user_localpart, filter_id=filter_id + user_filter = yield defer.ensureDeferred( + self.filtering.get_user_filter( + user_localpart=user_localpart, filter_id=filter_id + ) ) results = user_filter.filter_presence(events=events) @@ -386,8 +390,10 @@ class FilteringTestCase(unittest.TestCase): def test_filter_presence_no_match(self): user_filter_json = {"presence": {"types": ["m.*"]}} - filter_id = yield self.datastore.add_user_filter( - user_localpart=user_localpart + "2", user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.datastore.add_user_filter( + user_localpart=user_localpart + "2", user_filter=user_filter_json + ) ) event = MockEvent( event_id="$asdasd:localhost", @@ -396,8 +402,10 @@ class FilteringTestCase(unittest.TestCase): ) events = [event] - user_filter = yield self.filtering.get_user_filter( - user_localpart=user_localpart + "2", filter_id=filter_id + user_filter = yield defer.ensureDeferred( + self.filtering.get_user_filter( + user_localpart=user_localpart + "2", filter_id=filter_id + ) ) results = user_filter.filter_presence(events=events) @@ -406,14 +414,18 @@ class FilteringTestCase(unittest.TestCase): @defer.inlineCallbacks def test_filter_room_state_match(self): user_filter_json = {"room": {"state": {"types": ["m.*"]}}} - filter_id = yield self.datastore.add_user_filter( - user_localpart=user_localpart, user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.datastore.add_user_filter( + user_localpart=user_localpart, user_filter=user_filter_json + ) ) event = MockEvent(sender="@foo:bar", type="m.room.topic", room_id="!foo:bar") events = [event] - user_filter = yield self.filtering.get_user_filter( - user_localpart=user_localpart, filter_id=filter_id + user_filter = yield defer.ensureDeferred( + self.filtering.get_user_filter( + user_localpart=user_localpart, filter_id=filter_id + ) ) results = user_filter.filter_room_state(events=events) @@ -422,16 +434,20 @@ class FilteringTestCase(unittest.TestCase): @defer.inlineCallbacks def test_filter_room_state_no_match(self): user_filter_json = {"room": {"state": {"types": ["m.*"]}}} - filter_id = yield self.datastore.add_user_filter( - user_localpart=user_localpart, user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.datastore.add_user_filter( + user_localpart=user_localpart, user_filter=user_filter_json + ) ) event = MockEvent( sender="@foo:bar", type="org.matrix.custom.event", room_id="!foo:bar" ) events = [event] - user_filter = yield self.filtering.get_user_filter( - user_localpart=user_localpart, filter_id=filter_id + user_filter = yield defer.ensureDeferred( + self.filtering.get_user_filter( + user_localpart=user_localpart, filter_id=filter_id + ) ) results = user_filter.filter_room_state(events) @@ -457,16 +473,20 @@ class FilteringTestCase(unittest.TestCase): def test_add_filter(self): user_filter_json = {"room": {"state": {"types": ["m.*"]}}} - filter_id = yield self.filtering.add_user_filter( - user_localpart=user_localpart, user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.filtering.add_user_filter( + user_localpart=user_localpart, user_filter=user_filter_json + ) ) self.assertEquals(filter_id, 0) self.assertEquals( user_filter_json, ( - yield self.datastore.get_user_filter( - user_localpart=user_localpart, filter_id=0 + yield defer.ensureDeferred( + self.datastore.get_user_filter( + user_localpart=user_localpart, filter_id=0 + ) ) ), ) @@ -475,12 +495,16 @@ class FilteringTestCase(unittest.TestCase): def test_get_filter(self): user_filter_json = {"room": {"state": {"types": ["m.*"]}}} - filter_id = yield self.datastore.add_user_filter( - user_localpart=user_localpart, user_filter=user_filter_json + filter_id = yield defer.ensureDeferred( + self.datastore.add_user_filter( + user_localpart=user_localpart, user_filter=user_filter_json + ) ) - filter = yield self.filtering.get_user_filter( - user_localpart=user_localpart, filter_id=filter_id + filter = yield defer.ensureDeferred( + self.filtering.get_user_filter( + user_localpart=user_localpart, filter_id=filter_id + ) ) self.assertEquals(filter.get_filter_json(), user_filter_json) diff --git a/tests/api/test_ratelimiting.py b/tests/api/test_ratelimiting.py
index d580e729c5..1e1f30d790 100644 --- a/tests/api/test_ratelimiting.py +++ b/tests/api/test_ratelimiting.py
@@ -1,4 +1,6 @@ from synapse.api.ratelimiting import LimitExceededError, Ratelimiter +from synapse.appservice import ApplicationService +from synapse.types import create_requester from tests import unittest @@ -18,6 +20,77 @@ class TestRatelimiter(unittest.TestCase): self.assertTrue(allowed) self.assertEquals(20.0, time_allowed) + def test_allowed_user_via_can_requester_do_action(self): + user_requester = create_requester("@user:example.com") + limiter = Ratelimiter(clock=None, rate_hz=0.1, burst_count=1) + allowed, time_allowed = limiter.can_requester_do_action( + user_requester, _time_now_s=0 + ) + self.assertTrue(allowed) + self.assertEquals(10.0, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + user_requester, _time_now_s=5 + ) + self.assertFalse(allowed) + self.assertEquals(10.0, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + user_requester, _time_now_s=10 + ) + self.assertTrue(allowed) + self.assertEquals(20.0, time_allowed) + + def test_allowed_appservice_ratelimited_via_can_requester_do_action(self): + appservice = ApplicationService( + None, "example.com", id="foo", rate_limited=True, + ) + as_requester = create_requester("@user:example.com", app_service=appservice) + + limiter = Ratelimiter(clock=None, rate_hz=0.1, burst_count=1) + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=0 + ) + self.assertTrue(allowed) + self.assertEquals(10.0, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=5 + ) + self.assertFalse(allowed) + self.assertEquals(10.0, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=10 + ) + self.assertTrue(allowed) + self.assertEquals(20.0, time_allowed) + + def test_allowed_appservice_via_can_requester_do_action(self): + appservice = ApplicationService( + None, "example.com", id="foo", rate_limited=False, + ) + as_requester = create_requester("@user:example.com", app_service=appservice) + + limiter = Ratelimiter(clock=None, rate_hz=0.1, burst_count=1) + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=0 + ) + self.assertTrue(allowed) + self.assertEquals(-1, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=5 + ) + self.assertTrue(allowed) + self.assertEquals(-1, time_allowed) + + allowed, time_allowed = limiter.can_requester_do_action( + as_requester, _time_now_s=10 + ) + self.assertTrue(allowed) + self.assertEquals(-1, time_allowed) + def test_allowed_via_ratelimit(self): limiter = Ratelimiter(clock=None, rate_hz=0.1, burst_count=1) diff --git a/tests/appservice/test_appservice.py b/tests/appservice/test_appservice.py
index 4003869ed6..236b608d58 100644 --- a/tests/appservice/test_appservice.py +++ b/tests/appservice/test_appservice.py
@@ -50,13 +50,17 @@ class ApplicationServiceTestCase(unittest.TestCase): def test_regex_user_id_prefix_match(self): self.service.namespaces[ApplicationService.NS_USERS].append(_regex("@irc_.*")) self.event.sender = "@irc_foobar:matrix.org" - self.assertTrue((yield self.service.is_interested(self.event))) + self.assertTrue( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_regex_user_id_prefix_no_match(self): self.service.namespaces[ApplicationService.NS_USERS].append(_regex("@irc_.*")) self.event.sender = "@someone_else:matrix.org" - self.assertFalse((yield self.service.is_interested(self.event))) + self.assertFalse( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_regex_room_member_is_checked(self): @@ -64,7 +68,9 @@ class ApplicationServiceTestCase(unittest.TestCase): self.event.sender = "@someone_else:matrix.org" self.event.type = "m.room.member" self.event.state_key = "@irc_foobar:matrix.org" - self.assertTrue((yield self.service.is_interested(self.event))) + self.assertTrue( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_regex_room_id_match(self): @@ -72,7 +78,9 @@ class ApplicationServiceTestCase(unittest.TestCase): _regex("!some_prefix.*some_suffix:matrix.org") ) self.event.room_id = "!some_prefixs0m3th1nGsome_suffix:matrix.org" - self.assertTrue((yield self.service.is_interested(self.event))) + self.assertTrue( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_regex_room_id_no_match(self): @@ -80,19 +88,26 @@ class ApplicationServiceTestCase(unittest.TestCase): _regex("!some_prefix.*some_suffix:matrix.org") ) self.event.room_id = "!XqBunHwQIXUiqCaoxq:matrix.org" - self.assertFalse((yield self.service.is_interested(self.event))) + self.assertFalse( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_regex_alias_match(self): self.service.namespaces[ApplicationService.NS_ALIASES].append( _regex("#irc_.*:matrix.org") ) - self.store.get_aliases_for_room.return_value = [ - "#irc_foobar:matrix.org", - "#athing:matrix.org", - ] - self.store.get_users_in_room.return_value = [] - self.assertTrue((yield self.service.is_interested(self.event, self.store))) + self.store.get_aliases_for_room.return_value = defer.succeed( + ["#irc_foobar:matrix.org", "#athing:matrix.org"] + ) + self.store.get_users_in_room.return_value = defer.succeed([]) + self.assertTrue( + ( + yield defer.ensureDeferred( + self.service.is_interested(self.event, self.store) + ) + ) + ) def test_non_exclusive_alias(self): self.service.namespaces[ApplicationService.NS_ALIASES].append( @@ -135,12 +150,17 @@ class ApplicationServiceTestCase(unittest.TestCase): self.service.namespaces[ApplicationService.NS_ALIASES].append( _regex("#irc_.*:matrix.org") ) - self.store.get_aliases_for_room.return_value = [ - "#xmpp_foobar:matrix.org", - "#athing:matrix.org", - ] - self.store.get_users_in_room.return_value = [] - self.assertFalse((yield self.service.is_interested(self.event, self.store))) + self.store.get_aliases_for_room.return_value = defer.succeed( + ["#xmpp_foobar:matrix.org", "#athing:matrix.org"] + ) + self.store.get_users_in_room.return_value = defer.succeed([]) + self.assertFalse( + ( + yield defer.ensureDeferred( + self.service.is_interested(self.event, self.store) + ) + ) + ) @defer.inlineCallbacks def test_regex_multiple_matches(self): @@ -149,9 +169,17 @@ class ApplicationServiceTestCase(unittest.TestCase): ) self.service.namespaces[ApplicationService.NS_USERS].append(_regex("@irc_.*")) self.event.sender = "@irc_foobar:matrix.org" - self.store.get_aliases_for_room.return_value = ["#irc_barfoo:matrix.org"] - self.store.get_users_in_room.return_value = [] - self.assertTrue((yield self.service.is_interested(self.event, self.store))) + self.store.get_aliases_for_room.return_value = defer.succeed( + ["#irc_barfoo:matrix.org"] + ) + self.store.get_users_in_room.return_value = defer.succeed([]) + self.assertTrue( + ( + yield defer.ensureDeferred( + self.service.is_interested(self.event, self.store) + ) + ) + ) @defer.inlineCallbacks def test_interested_in_self(self): @@ -161,19 +189,24 @@ class ApplicationServiceTestCase(unittest.TestCase): self.event.type = "m.room.member" self.event.content = {"membership": "invite"} self.event.state_key = self.service.sender - self.assertTrue((yield self.service.is_interested(self.event))) + self.assertTrue( + (yield defer.ensureDeferred(self.service.is_interested(self.event))) + ) @defer.inlineCallbacks def test_member_list_match(self): self.service.namespaces[ApplicationService.NS_USERS].append(_regex("@irc_.*")) - self.store.get_users_in_room.return_value = [ - "@alice:here", - "@irc_fo:here", # AS user - "@bob:here", - ] - self.store.get_aliases_for_room.return_value = [] + # Note that @irc_fo:here is the AS user. + self.store.get_users_in_room.return_value = defer.succeed( + ["@alice:here", "@irc_fo:here", "@bob:here"] + ) + self.store.get_aliases_for_room.return_value = defer.succeed([]) self.event.sender = "@xmpp_foobar:matrix.org" self.assertTrue( - (yield self.service.is_interested(event=self.event, store=self.store)) + ( + yield defer.ensureDeferred( + self.service.is_interested(event=self.event, store=self.store) + ) + ) ) diff --git a/tests/appservice/test_scheduler.py b/tests/appservice/test_scheduler.py
index 52f89d3f83..68a4caabbf 100644 --- a/tests/appservice/test_scheduler.py +++ b/tests/appservice/test_scheduler.py
@@ -25,6 +25,7 @@ from synapse.appservice.scheduler import ( from synapse.logging.context import make_deferred_yieldable from tests import unittest +from tests.test_utils import make_awaitable from ..utils import MockClock @@ -52,11 +53,11 @@ class ApplicationServiceSchedulerTransactionCtrlTestCase(unittest.TestCase): self.store.get_appservice_state = Mock( return_value=defer.succeed(ApplicationServiceState.UP) ) - txn.send = Mock(return_value=defer.succeed(True)) + txn.send = Mock(return_value=make_awaitable(True)) self.store.create_appservice_txn = Mock(return_value=defer.succeed(txn)) # actual call - self.txnctrl.send(service, events) + self.successResultOf(defer.ensureDeferred(self.txnctrl.send(service, events))) self.store.create_appservice_txn.assert_called_once_with( service=service, events=events # txn made and saved @@ -77,7 +78,7 @@ class ApplicationServiceSchedulerTransactionCtrlTestCase(unittest.TestCase): self.store.create_appservice_txn = Mock(return_value=defer.succeed(txn)) # actual call - self.txnctrl.send(service, events) + self.successResultOf(defer.ensureDeferred(self.txnctrl.send(service, events))) self.store.create_appservice_txn.assert_called_once_with( service=service, events=events # txn made and saved @@ -98,11 +99,11 @@ class ApplicationServiceSchedulerTransactionCtrlTestCase(unittest.TestCase): return_value=defer.succeed(ApplicationServiceState.UP) ) self.store.set_appservice_state = Mock(return_value=defer.succeed(True)) - txn.send = Mock(return_value=defer.succeed(False)) # fails to send + txn.send = Mock(return_value=make_awaitable(False)) # fails to send self.store.create_appservice_txn = Mock(return_value=defer.succeed(txn)) # actual call - self.txnctrl.send(service, events) + self.successResultOf(defer.ensureDeferred(self.txnctrl.send(service, events))) self.store.create_appservice_txn.assert_called_once_with( service=service, events=events @@ -144,7 +145,8 @@ class ApplicationServiceSchedulerRecovererTestCase(unittest.TestCase): self.recoverer.recover() # shouldn't have called anything prior to waiting for exp backoff self.assertEquals(0, self.store.get_oldest_unsent_txn.call_count) - txn.send = Mock(return_value=True) + txn.send = Mock(return_value=make_awaitable(True)) + txn.complete.return_value = make_awaitable(None) # wait for exp backoff self.clock.advance_time(2) self.assertEquals(1, txn.send.call_count) @@ -169,7 +171,8 @@ class ApplicationServiceSchedulerRecovererTestCase(unittest.TestCase): self.recoverer.recover() self.assertEquals(0, self.store.get_oldest_unsent_txn.call_count) - txn.send = Mock(return_value=False) + txn.send = Mock(return_value=make_awaitable(False)) + txn.complete.return_value = make_awaitable(None) self.clock.advance_time(2) self.assertEquals(1, txn.send.call_count) self.assertEquals(0, txn.complete.call_count) @@ -182,7 +185,7 @@ class ApplicationServiceSchedulerRecovererTestCase(unittest.TestCase): self.assertEquals(3, txn.send.call_count) self.assertEquals(0, txn.complete.call_count) self.assertEquals(0, self.callback.call_count) - txn.send = Mock(return_value=True) # successfully send the txn + txn.send = Mock(return_value=make_awaitable(True)) # successfully send the txn pop_txn = True # returns the txn the first time, then no more. self.clock.advance_time(16) self.assertEquals(1, txn.send.call_count) # new mock reset call count diff --git a/tests/config/test_base.py b/tests/config/test_base.py new file mode 100644
index 0000000000..42ee5f56d9 --- /dev/null +++ b/tests/config/test_base.py
@@ -0,0 +1,82 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import os.path +import tempfile + +from synapse.config import ConfigError +from synapse.util.stringutils import random_string + +from tests import unittest + + +class BaseConfigTestCase(unittest.HomeserverTestCase): + def prepare(self, reactor, clock, hs): + self.hs = hs + + def test_loading_missing_templates(self): + # Use a temporary directory that exists on the system, but that isn't likely to + # contain template files + with tempfile.TemporaryDirectory() as tmp_dir: + # Attempt to load an HTML template from our custom template directory + template = self.hs.config.read_templates(["sso_error.html"], tmp_dir)[0] + + # If no errors, we should've gotten the default template instead + + # Render the template + a_random_string = random_string(5) + html_content = template.render({"error_description": a_random_string}) + + # Check that our string exists in the template + self.assertIn( + a_random_string, + html_content, + "Template file did not contain our test string", + ) + + def test_loading_custom_templates(self): + # Use a temporary directory that exists on the system + with tempfile.TemporaryDirectory() as tmp_dir: + # Create a temporary bogus template file + with tempfile.NamedTemporaryFile(dir=tmp_dir) as tmp_template: + # Get temporary file's filename + template_filename = os.path.basename(tmp_template.name) + + # Write a custom HTML template + contents = b"{{ test_variable }}" + tmp_template.write(contents) + tmp_template.flush() + + # Attempt to load the template from our custom template directory + template = ( + self.hs.config.read_templates([template_filename], tmp_dir) + )[0] + + # Render the template + a_random_string = random_string(5) + html_content = template.render({"test_variable": a_random_string}) + + # Check that our string exists in the template + self.assertIn( + a_random_string, + html_content, + "Template file did not contain our test string", + ) + + def test_loading_template_from_nonexistent_custom_directory(self): + with self.assertRaises(ConfigError): + self.hs.config.read_templates( + ["some_filename.html"], "a_nonexistent_directory" + ) diff --git a/tests/crypto/test_keyring.py b/tests/crypto/test_keyring.py
index f9ce609923..8ff1460c0d 100644 --- a/tests/crypto/test_keyring.py +++ b/tests/crypto/test_keyring.py
@@ -23,6 +23,7 @@ from nacl.signing import SigningKey from signedjson.key import encode_verify_key_base64, get_verify_key from twisted.internet import defer +from twisted.internet.defer import Deferred, ensureDeferred from synapse.api.errors import SynapseError from synapse.crypto import keyring @@ -33,16 +34,17 @@ from synapse.crypto.keyring import ( ) from synapse.logging.context import ( LoggingContext, - PreserveLoggingContext, current_context, make_deferred_yieldable, ) from synapse.storage.keys import FetchKeyResult from tests import unittest +from tests.test_utils import make_awaitable +from tests.unittest import logcontext_clean -class MockPerspectiveServer(object): +class MockPerspectiveServer: def __init__(self): self.server_name = "mock_server" self.key = signedjson.key.generate_signing_key(0) @@ -66,56 +68,42 @@ class MockPerspectiveServer(object): signedjson.sign.sign_json(res, self.server_name, self.key) +@logcontext_clean class KeyringTestCase(unittest.HomeserverTestCase): - def make_homeserver(self, reactor, clock): - self.mock_perspective_server = MockPerspectiveServer() - self.http_client = Mock() - - config = self.default_config() - config["trusted_key_servers"] = [ - { - "server_name": self.mock_perspective_server.server_name, - "verify_keys": self.mock_perspective_server.get_verify_keys(), - } - ] - - return self.setup_test_homeserver( - handlers=None, http_client=self.http_client, config=config - ) - - def check_context(self, _, expected): + def check_context(self, val, expected): self.assertEquals(getattr(current_context(), "request", None), expected) + return val def test_verify_json_objects_for_server_awaits_previous_requests(self): - key1 = signedjson.key.generate_signing_key(1) + mock_fetcher = keyring.KeyFetcher() + mock_fetcher.get_keys = Mock() + kr = keyring.Keyring(self.hs, key_fetchers=(mock_fetcher,)) - kr = keyring.Keyring(self.hs) + # a signed object that we are going to try to validate + key1 = signedjson.key.generate_signing_key(1) json1 = {} signedjson.sign.sign_json(json1, "server10", key1) - persp_resp = { - "server_keys": [ - self.mock_perspective_server.get_signed_key( - "server10", signedjson.key.get_verify_key(key1) - ) - ] - } - persp_deferred = defer.Deferred() + # start off a first set of lookups. We make the mock fetcher block until this + # deferred completes. + first_lookup_deferred = Deferred() - @defer.inlineCallbacks - def get_perspectives(**kwargs): - self.assertEquals(current_context().request, "11") - with PreserveLoggingContext(): - yield persp_deferred - return persp_resp + async def first_lookup_fetch(keys_to_fetch): + self.assertEquals(current_context().request, "context_11") + self.assertEqual(keys_to_fetch, {"server10": {get_key_id(key1): 0}}) - self.http_client.post_json.side_effect = get_perspectives + await make_deferred_yieldable(first_lookup_deferred) + return { + "server10": { + get_key_id(key1): FetchKeyResult(get_verify_key(key1), 100) + } + } + + mock_fetcher.get_keys.side_effect = first_lookup_fetch - # start off a first set of lookups - @defer.inlineCallbacks - def first_lookup(): - with LoggingContext("11") as context_11: - context_11.request = "11" + async def first_lookup(): + with LoggingContext("context_11") as context_11: + context_11.request = "context_11" res_deferreds = kr.verify_json_objects_for_server( [("server10", json1, 0, "test10"), ("server11", {}, 0, "test11")] @@ -124,7 +112,7 @@ class KeyringTestCase(unittest.HomeserverTestCase): # the unsigned json should be rejected pretty quickly self.assertTrue(res_deferreds[1].called) try: - yield res_deferreds[1] + await res_deferreds[1] self.assertFalse("unsigned json didn't cause a failure") except SynapseError: pass @@ -132,45 +120,51 @@ class KeyringTestCase(unittest.HomeserverTestCase): self.assertFalse(res_deferreds[0].called) res_deferreds[0].addBoth(self.check_context, None) - yield make_deferred_yieldable(res_deferreds[0]) + await make_deferred_yieldable(res_deferreds[0]) - # let verify_json_objects_for_server finish its work before we kill the - # logcontext - yield self.clock.sleep(0) + d0 = ensureDeferred(first_lookup()) - d0 = first_lookup() - - # wait a tick for it to send the request to the perspectives server - # (it first tries the datastore) - self.pump() - self.http_client.post_json.assert_called_once() + mock_fetcher.get_keys.assert_called_once() # a second request for a server with outstanding requests # should block rather than start a second call - @defer.inlineCallbacks - def second_lookup(): - with LoggingContext("12") as context_12: - context_12.request = "12" - self.http_client.post_json.reset_mock() - self.http_client.post_json.return_value = defer.Deferred() + + async def second_lookup_fetch(keys_to_fetch): + self.assertEquals(current_context().request, "context_12") + return { + "server10": { + get_key_id(key1): FetchKeyResult(get_verify_key(key1), 100) + } + } + + mock_fetcher.get_keys.reset_mock() + mock_fetcher.get_keys.side_effect = second_lookup_fetch + second_lookup_state = [0] + + async def second_lookup(): + with LoggingContext("context_12") as context_12: + context_12.request = "context_12" res_deferreds_2 = kr.verify_json_objects_for_server( [("server10", json1, 0, "test")] ) res_deferreds_2[0].addBoth(self.check_context, None) - yield make_deferred_yieldable(res_deferreds_2[0]) + second_lookup_state[0] = 1 + await make_deferred_yieldable(res_deferreds_2[0]) + second_lookup_state[0] = 2 - # let verify_json_objects_for_server finish its work before we kill the - # logcontext - yield self.clock.sleep(0) - - d2 = second_lookup() + d2 = ensureDeferred(second_lookup()) self.pump() - self.http_client.post_json.assert_not_called() + # the second request should be pending, but the fetcher should not yet have been + # called + self.assertEqual(second_lookup_state[0], 1) + mock_fetcher.get_keys.assert_not_called() # complete the first request - persp_deferred.callback(persp_resp) + first_lookup_deferred.callback(None) + + # and now both verifications should succeed. self.get_success(d0) self.get_success(d2) @@ -190,7 +184,7 @@ class KeyringTestCase(unittest.HomeserverTestCase): # should fail immediately on an unsigned object d = _verify_json_for_server(kr, "server9", {}, 0, "test unsigned") - self.failureResultOf(d, SynapseError) + self.get_failure(d, SynapseError) # should succeed on a signed object d = _verify_json_for_server(kr, "server9", json1, 500, "test signed") @@ -202,7 +196,7 @@ class KeyringTestCase(unittest.HomeserverTestCase): with a null `ts_valid_until_ms` """ mock_fetcher = keyring.KeyFetcher() - mock_fetcher.get_keys = Mock(return_value=defer.succeed({})) + mock_fetcher.get_keys = Mock(return_value=make_awaitable({})) kr = keyring.Keyring( self.hs, key_fetchers=(StoreKeyFetcher(self.hs), mock_fetcher) @@ -221,7 +215,7 @@ class KeyringTestCase(unittest.HomeserverTestCase): # should fail immediately on an unsigned object d = _verify_json_for_server(kr, "server9", {}, 0, "test unsigned") - self.failureResultOf(d, SynapseError) + self.get_failure(d, SynapseError) # should fail on a signed object with a non-zero minimum_valid_until_ms, # as it tries to refetch the keys and fails. @@ -245,17 +239,15 @@ class KeyringTestCase(unittest.HomeserverTestCase): """Two requests for the same key should be deduped.""" key1 = signedjson.key.generate_signing_key(1) - def get_keys(keys_to_fetch): + async def get_keys(keys_to_fetch): # there should only be one request object (with the max validity) self.assertEqual(keys_to_fetch, {"server1": {get_key_id(key1): 1500}}) - return defer.succeed( - { - "server1": { - get_key_id(key1): FetchKeyResult(get_verify_key(key1), 1200) - } + return { + "server1": { + get_key_id(key1): FetchKeyResult(get_verify_key(key1), 1200) } - ) + } mock_fetcher = keyring.KeyFetcher() mock_fetcher.get_keys = Mock(side_effect=get_keys) @@ -282,25 +274,19 @@ class KeyringTestCase(unittest.HomeserverTestCase): """If the first fetcher cannot provide a recent enough key, we fall back""" key1 = signedjson.key.generate_signing_key(1) - def get_keys1(keys_to_fetch): + async def get_keys1(keys_to_fetch): self.assertEqual(keys_to_fetch, {"server1": {get_key_id(key1): 1500}}) - return defer.succeed( - { - "server1": { - get_key_id(key1): FetchKeyResult(get_verify_key(key1), 800) - } - } - ) + return { + "server1": {get_key_id(key1): FetchKeyResult(get_verify_key(key1), 800)} + } - def get_keys2(keys_to_fetch): + async def get_keys2(keys_to_fetch): self.assertEqual(keys_to_fetch, {"server1": {get_key_id(key1): 1500}}) - return defer.succeed( - { - "server1": { - get_key_id(key1): FetchKeyResult(get_verify_key(key1), 1200) - } + return { + "server1": { + get_key_id(key1): FetchKeyResult(get_verify_key(key1), 1200) } - ) + } mock_fetcher1 = keyring.KeyFetcher() mock_fetcher1.get_keys = Mock(side_effect=get_keys1) @@ -325,6 +311,7 @@ class KeyringTestCase(unittest.HomeserverTestCase): mock_fetcher2.get_keys.assert_called_once() +@logcontext_clean class ServerKeyFetcherTestCase(unittest.HomeserverTestCase): def make_homeserver(self, reactor, clock): self.http_client = Mock() @@ -355,7 +342,7 @@ class ServerKeyFetcherTestCase(unittest.HomeserverTestCase): } signedjson.sign.sign_json(response, SERVER_NAME, testkey) - def get_json(destination, path, **kwargs): + async def get_json(destination, path, **kwargs): self.assertEqual(destination, SERVER_NAME) self.assertEqual(path, "/_matrix/key/v2/server/key1") return response @@ -444,7 +431,7 @@ class PerspectivesKeyFetcherTestCase(unittest.HomeserverTestCase): Tell the mock http client to expect a perspectives-server key query """ - def post_json(destination, path, data, **kwargs): + async def post_json(destination, path, data, **kwargs): self.assertEqual(destination, self.mock_perspective_server.server_name) self.assertEqual(path, "/_matrix/key/v2/query") @@ -580,14 +567,12 @@ class PerspectivesKeyFetcherTestCase(unittest.HomeserverTestCase): # remove the perspectives server's signature response = build_response() del response["signatures"][self.mock_perspective_server.server_name] - self.http_client.post_json.return_value = {"server_keys": [response]} keys = get_key_from_perspectives(response) self.assertEqual(keys, {}, "Expected empty dict with missing persp server sig") # remove the origin server's signature response = build_response() del response["signatures"][SERVER_NAME] - self.http_client.post_json.return_value = {"server_keys": [response]} keys = get_key_from_perspectives(response) self.assertEqual(keys, {}, "Expected empty dict with missing origin server sig") diff --git a/tests/federation/test_complexity.py b/tests/federation/test_complexity.py
index 0c9987be54..1471cc1a28 100644 --- a/tests/federation/test_complexity.py +++ b/tests/federation/test_complexity.py
@@ -15,14 +15,13 @@ from mock import Mock -from twisted.internet import defer - from synapse.api.errors import Codes, SynapseError from synapse.rest import admin from synapse.rest.client.v1 import login, room from synapse.types import UserID from tests import unittest +from tests.test_utils import make_awaitable class RoomComplexityTests(unittest.FederatingHomeserverTestCase): @@ -59,7 +58,7 @@ class RoomComplexityTests(unittest.FederatingHomeserverTestCase): # Artificially raise the complexity store = self.hs.get_datastore() - store.get_current_state_event_counts = lambda x: defer.succeed(500 * 1.23) + store.get_current_state_event_counts = lambda x: make_awaitable(500 * 1.23) # Get the room complexity again -- make sure it's our artificial value request, channel = self.make_request( @@ -78,9 +77,40 @@ class RoomComplexityTests(unittest.FederatingHomeserverTestCase): fed_transport = self.hs.get_federation_transport_client() # Mock out some things, because we don't want to test the whole join - fed_transport.client.get_json = Mock(return_value=defer.succeed({"v1": 9999})) + fed_transport.client.get_json = Mock(return_value=make_awaitable({"v1": 9999})) + handler.federation_handler.do_invite_join = Mock( + return_value=make_awaitable(("", 1)) + ) + + d = handler._remote_join( + None, + ["other.example.com"], + "roomid", + UserID.from_string(u1), + {"membership": "join"}, + ) + + self.pump() + + # The request failed with a SynapseError saying the resource limit was + # exceeded. + f = self.get_failure(d, SynapseError) + self.assertEqual(f.value.code, 400, f.value) + self.assertEqual(f.value.errcode, Codes.RESOURCE_LIMIT_EXCEEDED) + + def test_join_too_large_admin(self): + # Check whether an admin can join if option "admins_can_join" is undefined, + # this option defaults to false, so the join should fail. + + u1 = self.register_user("u1", "pass", admin=True) + + handler = self.hs.get_room_member_handler() + fed_transport = self.hs.get_federation_transport_client() + + # Mock out some things, because we don't want to test the whole join + fed_transport.client.get_json = Mock(return_value=make_awaitable({"v1": 9999})) handler.federation_handler.do_invite_join = Mock( - return_value=defer.succeed(("", 1)) + return_value=make_awaitable(("", 1)) ) d = handler._remote_join( @@ -116,13 +146,13 @@ class RoomComplexityTests(unittest.FederatingHomeserverTestCase): fed_transport = self.hs.get_federation_transport_client() # Mock out some things, because we don't want to test the whole join - fed_transport.client.get_json = Mock(return_value=defer.succeed(None)) + fed_transport.client.get_json = Mock(return_value=make_awaitable(None)) handler.federation_handler.do_invite_join = Mock( - return_value=defer.succeed(("", 1)) + return_value=make_awaitable(("", 1)) ) # Artificially raise the complexity - self.hs.get_datastore().get_current_state_event_counts = lambda x: defer.succeed( + self.hs.get_datastore().get_current_state_event_counts = lambda x: make_awaitable( 600 ) @@ -141,3 +171,81 @@ class RoomComplexityTests(unittest.FederatingHomeserverTestCase): f = self.get_failure(d, SynapseError) self.assertEqual(f.value.code, 400) self.assertEqual(f.value.errcode, Codes.RESOURCE_LIMIT_EXCEEDED) + + +class RoomComplexityAdminTests(unittest.FederatingHomeserverTestCase): + # Test the behavior of joining rooms which exceed the complexity if option + # limit_remote_rooms.admins_can_join is True. + + servlets = [ + admin.register_servlets, + room.register_servlets, + login.register_servlets, + ] + + def default_config(self): + config = super().default_config() + config["limit_remote_rooms"] = { + "enabled": True, + "complexity": 0.05, + "admins_can_join": True, + } + return config + + def test_join_too_large_no_admin(self): + # A user which is not an admin should not be able to join a remote room + # which is too complex. + + u1 = self.register_user("u1", "pass") + + handler = self.hs.get_room_member_handler() + fed_transport = self.hs.get_federation_transport_client() + + # Mock out some things, because we don't want to test the whole join + fed_transport.client.get_json = Mock(return_value=make_awaitable({"v1": 9999})) + handler.federation_handler.do_invite_join = Mock( + return_value=make_awaitable(("", 1)) + ) + + d = handler._remote_join( + None, + ["other.example.com"], + "roomid", + UserID.from_string(u1), + {"membership": "join"}, + ) + + self.pump() + + # The request failed with a SynapseError saying the resource limit was + # exceeded. + f = self.get_failure(d, SynapseError) + self.assertEqual(f.value.code, 400, f.value) + self.assertEqual(f.value.errcode, Codes.RESOURCE_LIMIT_EXCEEDED) + + def test_join_too_large_admin(self): + # An admin should be able to join rooms where a complexity check fails. + + u1 = self.register_user("u1", "pass", admin=True) + + handler = self.hs.get_room_member_handler() + fed_transport = self.hs.get_federation_transport_client() + + # Mock out some things, because we don't want to test the whole join + fed_transport.client.get_json = Mock(return_value=make_awaitable({"v1": 9999})) + handler.federation_handler.do_invite_join = Mock( + return_value=make_awaitable(("", 1)) + ) + + d = handler._remote_join( + None, + ["other.example.com"], + "roomid", + UserID.from_string(u1), + {"membership": "join"}, + ) + + self.pump() + + # The request success since the user is an admin + self.get_success(d) diff --git a/tests/federation/test_federation_catch_up.py b/tests/federation/test_federation_catch_up.py new file mode 100644
index 0000000000..1a3ccb263d --- /dev/null +++ b/tests/federation/test_federation_catch_up.py
@@ -0,0 +1,422 @@ +from typing import List, Tuple + +from mock import Mock + +from synapse.events import EventBase +from synapse.federation.sender import PerDestinationQueue, TransactionManager +from synapse.federation.units import Edu +from synapse.rest import admin +from synapse.rest.client.v1 import login, room + +from tests.test_utils import event_injection, make_awaitable +from tests.unittest import FederatingHomeserverTestCase, override_config + + +class FederationCatchUpTestCases(FederatingHomeserverTestCase): + servlets = [ + admin.register_servlets, + room.register_servlets, + login.register_servlets, + ] + + def make_homeserver(self, reactor, clock): + return self.setup_test_homeserver( + federation_transport_client=Mock(spec=["send_transaction"]), + ) + + def prepare(self, reactor, clock, hs): + # stub out get_current_hosts_in_room + state_handler = hs.get_state_handler() + + # This mock is crucial for destination_rooms to be populated. + state_handler.get_current_hosts_in_room = Mock( + return_value=make_awaitable(["test", "host2"]) + ) + + # whenever send_transaction is called, record the pdu data + self.pdus = [] + self.failed_pdus = [] + self.is_online = True + self.hs.get_federation_transport_client().send_transaction.side_effect = ( + self.record_transaction + ) + + async def record_transaction(self, txn, json_cb): + if self.is_online: + data = json_cb() + self.pdus.extend(data["pdus"]) + return {} + else: + data = json_cb() + self.failed_pdus.extend(data["pdus"]) + raise IOError("Failed to connect because this is a test!") + + def get_destination_room(self, room: str, destination: str = "host2") -> dict: + """ + Gets the destination_rooms entry for a (destination, room_id) pair. + + Args: + room: room ID + destination: what destination, default is "host2" + + Returns: + Dictionary of { event_id: str, stream_ordering: int } + """ + event_id, stream_ordering = self.get_success( + self.hs.get_datastore().db_pool.execute( + "test:get_destination_rooms", + None, + """ + SELECT event_id, stream_ordering + FROM destination_rooms dr + JOIN events USING (stream_ordering) + WHERE dr.destination = ? AND dr.room_id = ? + """, + destination, + room, + ) + )[0] + return {"event_id": event_id, "stream_ordering": stream_ordering} + + @override_config({"send_federation": True}) + def test_catch_up_destination_rooms_tracking(self): + """ + Tests that we populate the `destination_rooms` table as needed. + """ + self.register_user("u1", "you the one") + u1_token = self.login("u1", "you the one") + room = self.helper.create_room_as("u1", tok=u1_token) + + self.get_success( + event_injection.inject_member_event(self.hs, room, "@user:host2", "join") + ) + + event_id_1 = self.helper.send(room, "wombats!", tok=u1_token)["event_id"] + + row_1 = self.get_destination_room(room) + + event_id_2 = self.helper.send(room, "rabbits!", tok=u1_token)["event_id"] + + row_2 = self.get_destination_room(room) + + # check: events correctly registered in order + self.assertEqual(row_1["event_id"], event_id_1) + self.assertEqual(row_2["event_id"], event_id_2) + self.assertEqual(row_1["stream_ordering"], row_2["stream_ordering"] - 1) + + @override_config({"send_federation": True}) + def test_catch_up_last_successful_stream_ordering_tracking(self): + """ + Tests that we populate the `destination_rooms` table as needed. + """ + self.register_user("u1", "you the one") + u1_token = self.login("u1", "you the one") + room = self.helper.create_room_as("u1", tok=u1_token) + + # take the remote offline + self.is_online = False + + self.get_success( + event_injection.inject_member_event(self.hs, room, "@user:host2", "join") + ) + + self.helper.send(room, "wombats!", tok=u1_token) + self.pump() + + lsso_1 = self.get_success( + self.hs.get_datastore().get_destination_last_successful_stream_ordering( + "host2" + ) + ) + + self.assertIsNone( + lsso_1, + "There should be no last successful stream ordering for an always-offline destination", + ) + + # bring the remote online + self.is_online = True + + event_id_2 = self.helper.send(room, "rabbits!", tok=u1_token)["event_id"] + + lsso_2 = self.get_success( + self.hs.get_datastore().get_destination_last_successful_stream_ordering( + "host2" + ) + ) + row_2 = self.get_destination_room(room) + + self.assertEqual( + self.pdus[0]["content"]["body"], + "rabbits!", + "Test fault: didn't receive the right PDU", + ) + self.assertEqual( + row_2["event_id"], + event_id_2, + "Test fault: destination_rooms not updated correctly", + ) + self.assertEqual( + lsso_2, + row_2["stream_ordering"], + "Send succeeded but not marked as last_successful_stream_ordering", + ) + + @override_config({"send_federation": True}) # critical to federate + def test_catch_up_from_blank_state(self): + """ + Runs an overall test of federation catch-up from scratch. + Further tests will focus on more narrow aspects and edge-cases, but I + hope to provide an overall view with this test. + """ + # bring the other server online + self.is_online = True + + # let's make some events for the other server to receive + self.register_user("u1", "you the one") + u1_token = self.login("u1", "you the one") + room_1 = self.helper.create_room_as("u1", tok=u1_token) + room_2 = self.helper.create_room_as("u1", tok=u1_token) + + # also critical to federate + self.get_success( + event_injection.inject_member_event(self.hs, room_1, "@user:host2", "join") + ) + self.get_success( + event_injection.inject_member_event(self.hs, room_2, "@user:host2", "join") + ) + + self.helper.send_state( + room_1, event_type="m.room.topic", body={"topic": "wombat"}, tok=u1_token + ) + + # check: PDU received for topic event + self.assertEqual(len(self.pdus), 1) + self.assertEqual(self.pdus[0]["type"], "m.room.topic") + + # take the remote offline + self.is_online = False + + # send another event + self.helper.send(room_1, "hi user!", tok=u1_token) + + # check: things didn't go well since the remote is down + self.assertEqual(len(self.failed_pdus), 1) + self.assertEqual(self.failed_pdus[0]["content"]["body"], "hi user!") + + # let's delete the federation transmission queue + # (this pretends we are starting up fresh.) + self.assertFalse( + self.hs.get_federation_sender() + ._per_destination_queues["host2"] + .transmission_loop_running + ) + del self.hs.get_federation_sender()._per_destination_queues["host2"] + + # let's also clear any backoffs + self.get_success( + self.hs.get_datastore().set_destination_retry_timings("host2", None, 0, 0) + ) + + # bring the remote online and clear the received pdu list + self.is_online = True + self.pdus = [] + + # now we need to initiate a federation transaction somehow… + # to do that, let's send another event (because it's simple to do) + # (do it to another room otherwise the catch-up logic decides it doesn't + # need to catch up room_1 — something I overlooked when first writing + # this test) + self.helper.send(room_2, "wombats!", tok=u1_token) + + # we should now have received both PDUs + self.assertEqual(len(self.pdus), 2) + self.assertEqual(self.pdus[0]["content"]["body"], "hi user!") + self.assertEqual(self.pdus[1]["content"]["body"], "wombats!") + + def make_fake_destination_queue( + self, destination: str = "host2" + ) -> Tuple[PerDestinationQueue, List[EventBase]]: + """ + Makes a fake per-destination queue. + """ + transaction_manager = TransactionManager(self.hs) + per_dest_queue = PerDestinationQueue(self.hs, transaction_manager, destination) + results_list = [] + + async def fake_send( + destination_tm: str, + pending_pdus: List[EventBase], + _pending_edus: List[Edu], + ) -> bool: + assert destination == destination_tm + results_list.extend(pending_pdus) + return True # success! + + transaction_manager.send_new_transaction = fake_send + + return per_dest_queue, results_list + + @override_config({"send_federation": True}) + def test_catch_up_loop(self): + """ + Tests the behaviour of _catch_up_transmission_loop. + """ + + # ARRANGE: + # - a local user (u1) + # - 3 rooms which u1 is joined to (and remote user @user:host2 is + # joined to) + # - some events (1 to 5) in those rooms + # we have 'already sent' events 1 and 2 to host2 + per_dest_queue, sent_pdus = self.make_fake_destination_queue() + + self.register_user("u1", "you the one") + u1_token = self.login("u1", "you the one") + room_1 = self.helper.create_room_as("u1", tok=u1_token) + room_2 = self.helper.create_room_as("u1", tok=u1_token) + room_3 = self.helper.create_room_as("u1", tok=u1_token) + self.get_success( + event_injection.inject_member_event(self.hs, room_1, "@user:host2", "join") + ) + self.get_success( + event_injection.inject_member_event(self.hs, room_2, "@user:host2", "join") + ) + self.get_success( + event_injection.inject_member_event(self.hs, room_3, "@user:host2", "join") + ) + + # create some events + self.helper.send(room_1, "you hear me!!", tok=u1_token) + event_id_2 = self.helper.send(room_2, "wombats!", tok=u1_token)["event_id"] + self.helper.send(room_3, "Matrix!", tok=u1_token) + event_id_4 = self.helper.send(room_2, "rabbits!", tok=u1_token)["event_id"] + event_id_5 = self.helper.send(room_3, "Synapse!", tok=u1_token)["event_id"] + + # destination_rooms should already be populated, but let us pretend that we already + # sent (successfully) up to and including event id 2 + event_2 = self.get_success(self.hs.get_datastore().get_event(event_id_2)) + + # also fetch event 5 so we know its last_successful_stream_ordering later + event_5 = self.get_success(self.hs.get_datastore().get_event(event_id_5)) + + self.get_success( + self.hs.get_datastore().set_destination_last_successful_stream_ordering( + "host2", event_2.internal_metadata.stream_ordering + ) + ) + + # ACT + self.get_success(per_dest_queue._catch_up_transmission_loop()) + + # ASSERT, noticing in particular: + # - event 3 not sent out, because event 5 replaces it + # - order is least recent first, so event 5 comes after event 4 + # - catch-up is completed + self.assertEqual(len(sent_pdus), 2) + self.assertEqual(sent_pdus[0].event_id, event_id_4) + self.assertEqual(sent_pdus[1].event_id, event_id_5) + self.assertFalse(per_dest_queue._catching_up) + self.assertEqual( + per_dest_queue._last_successful_stream_ordering, + event_5.internal_metadata.stream_ordering, + ) + + @override_config({"send_federation": True}) + def test_catch_up_on_synapse_startup(self): + """ + Tests the behaviour of get_catch_up_outstanding_destinations and + _wake_destinations_needing_catchup. + """ + + # list of sorted server names (note that there are more servers than the batch + # size used in get_catch_up_outstanding_destinations). + server_names = ["server%02d" % number for number in range(42)] + ["zzzerver"] + + # ARRANGE: + # - a local user (u1) + # - a room which u1 is joined to (and remote users @user:serverXX are + # joined to) + + # mark the remotes as online + self.is_online = True + + self.register_user("u1", "you the one") + u1_token = self.login("u1", "you the one") + room_id = self.helper.create_room_as("u1", tok=u1_token) + + for server_name in server_names: + self.get_success( + event_injection.inject_member_event( + self.hs, room_id, "@user:%s" % server_name, "join" + ) + ) + + # create an event + self.helper.send(room_id, "deary me!", tok=u1_token) + + # ASSERT: + # - All servers are up to date so none should have outstanding catch-up + outstanding_when_successful = self.get_success( + self.hs.get_datastore().get_catch_up_outstanding_destinations(None) + ) + self.assertEqual(outstanding_when_successful, []) + + # ACT: + # - Make the remote servers unreachable + self.is_online = False + + # - Mark zzzerver as being backed-off from + now = self.clock.time_msec() + self.get_success( + self.hs.get_datastore().set_destination_retry_timings( + "zzzerver", now, now, 24 * 60 * 60 * 1000 # retry in 1 day + ) + ) + + # - Send an event + self.helper.send(room_id, "can anyone hear me?", tok=u1_token) + + # ASSERT (get_catch_up_outstanding_destinations): + # - all remotes are outstanding + # - they are returned in batches of 25, in order + outstanding_1 = self.get_success( + self.hs.get_datastore().get_catch_up_outstanding_destinations(None) + ) + + self.assertEqual(len(outstanding_1), 25) + self.assertEqual(outstanding_1, server_names[0:25]) + + outstanding_2 = self.get_success( + self.hs.get_datastore().get_catch_up_outstanding_destinations( + outstanding_1[-1] + ) + ) + self.assertNotIn("zzzerver", outstanding_2) + self.assertEqual(len(outstanding_2), 17) + self.assertEqual(outstanding_2, server_names[25:-1]) + + # ACT: call _wake_destinations_needing_catchup + + # patch wake_destination to just count the destinations instead + woken = [] + + def wake_destination_track(destination): + woken.append(destination) + + self.hs.get_federation_sender().wake_destination = wake_destination_track + + # cancel the pre-existing timer for _wake_destinations_needing_catchup + # this is because we are calling it manually rather than waiting for it + # to be called automatically + self.hs.get_federation_sender()._catchup_after_startup_timer.cancel() + + self.get_success( + self.hs.get_federation_sender()._wake_destinations_needing_catchup(), by=5.0 + ) + + # ASSERT (_wake_destinations_needing_catchup): + # - all remotes are woken up, save for zzzerver + self.assertNotIn("zzzerver", woken) + # - all destinations are woken exactly once; they appear once in woken. + self.assertCountEqual(woken, server_names[:-1]) diff --git a/tests/federation/test_federation_sender.py b/tests/federation/test_federation_sender.py
index d1bd18da39..917762e6b6 100644 --- a/tests/federation/test_federation_sender.py +++ b/tests/federation/test_federation_sender.py
@@ -34,7 +34,7 @@ class FederationSenderReceiptsTestCases(HomeserverTestCase): def make_homeserver(self, reactor, clock): mock_state_handler = Mock(spec=["get_current_hosts_in_room"]) # Ensure a new Awaitable is created for each call. - mock_state_handler.get_current_hosts_in_room.side_effect = lambda room_Id: make_awaitable( + mock_state_handler.get_current_hosts_in_room.return_value = make_awaitable( ["test", "host2"] ) return self.setup_test_homeserver( @@ -47,13 +47,13 @@ class FederationSenderReceiptsTestCases(HomeserverTestCase): mock_send_transaction = ( self.hs.get_federation_transport_client().send_transaction ) - mock_send_transaction.return_value = defer.succeed({}) + mock_send_transaction.return_value = make_awaitable({}) sender = self.hs.get_federation_sender() receipt = ReadReceipt( "room_id", "m.read", "user_id", ["event_id"], {"ts": 1234} ) - self.successResultOf(sender.send_read_receipt(receipt)) + self.successResultOf(defer.ensureDeferred(sender.send_read_receipt(receipt))) self.pump() @@ -87,13 +87,13 @@ class FederationSenderReceiptsTestCases(HomeserverTestCase): mock_send_transaction = ( self.hs.get_federation_transport_client().send_transaction ) - mock_send_transaction.return_value = defer.succeed({}) + mock_send_transaction.return_value = make_awaitable({}) sender = self.hs.get_federation_sender() receipt = ReadReceipt( "room_id", "m.read", "user_id", ["event_id"], {"ts": 1234} ) - self.successResultOf(sender.send_read_receipt(receipt)) + self.successResultOf(defer.ensureDeferred(sender.send_read_receipt(receipt))) self.pump() @@ -125,7 +125,7 @@ class FederationSenderReceiptsTestCases(HomeserverTestCase): receipt = ReadReceipt( "room_id", "m.read", "user_id", ["other_id"], {"ts": 1234} ) - self.successResultOf(sender.send_read_receipt(receipt)) + self.successResultOf(defer.ensureDeferred(sender.send_read_receipt(receipt))) self.pump() mock_send_transaction.assert_not_called() diff --git a/tests/federation/test_federation_server.py b/tests/federation/test_federation_server.py
index 296dc887be..da933ecd75 100644 --- a/tests/federation/test_federation_server.py +++ b/tests/federation/test_federation_server.py
@@ -15,6 +15,8 @@ # limitations under the License. import logging +from parameterized import parameterized + from synapse.events import make_event_from_dict from synapse.federation.federation_server import server_matches_acl_event from synapse.rest import admin @@ -23,6 +25,37 @@ from synapse.rest.client.v1 import login, room from tests import unittest +class FederationServerTests(unittest.FederatingHomeserverTestCase): + + servlets = [ + admin.register_servlets, + room.register_servlets, + login.register_servlets, + ] + + @parameterized.expand([(b"",), (b"foo",), (b'{"limit": Infinity}',)]) + def test_bad_request(self, query_content): + """ + Querying with bad data returns a reasonable error code. + """ + u1 = self.register_user("u1", "pass") + u1_token = self.login("u1", "pass") + + room_1 = self.helper.create_room_as(u1, tok=u1_token) + self.inject_room_member(room_1, "@user:other.example.com", "join") + + "/get_missing_events/(?P<room_id>[^/]*)/?" + + request, channel = self.make_request( + "POST", + "/_matrix/federation/v1/get_missing_events/%s" % (room_1,), + query_content, + ) + self.render(request) + self.assertEquals(400, channel.code, channel.result) + self.assertEqual(channel.json_body["errcode"], "M_NOT_JSON") + + class ServerACLsTestCase(unittest.TestCase): def test_blacklisted_server(self): e = _create_acl_event({"allow": ["*"], "deny": ["evil.com"]}) diff --git a/tests/federation/transport/test_server.py b/tests/federation/transport/test_server.py
index 27d83bb7d9..72e22d655f 100644 --- a/tests/federation/transport/test_server.py +++ b/tests/federation/transport/test_server.py
@@ -26,7 +26,7 @@ from tests.unittest import override_config class RoomDirectoryFederationTests(unittest.HomeserverTestCase): def prepare(self, reactor, clock, homeserver): - class Authenticator(object): + class Authenticator: def authenticate_request(self, request, content): return defer.succeed("otherserver.nottld") diff --git a/tests/handlers/test_appservice.py b/tests/handlers/test_appservice.py
index ebabe9a7d6..2a0b7c1b56 100644 --- a/tests/handlers/test_appservice.py +++ b/tests/handlers/test_appservice.py
@@ -19,6 +19,7 @@ from twisted.internet import defer from synapse.handlers.appservice import ApplicationServicesHandler +from tests.test_utils import make_awaitable from tests.utils import MockClock from .. import unittest @@ -117,9 +118,9 @@ class AppServiceHandlerTestCase(unittest.TestCase): self._mkservice_alias(is_interested_in_alias=False), ] - self.mock_as_api.query_alias.return_value = defer.succeed(True) + self.mock_as_api.query_alias.return_value = make_awaitable(True) self.mock_store.get_app_services.return_value = services - self.mock_store.get_association_from_room_alias.return_value = defer.succeed( + self.mock_store.get_association_from_room_alias.return_value = make_awaitable( Mock(room_id=room_id, servers=servers) ) @@ -135,7 +136,7 @@ class AppServiceHandlerTestCase(unittest.TestCase): def _mkservice(self, is_interested): service = Mock() - service.is_interested.return_value = defer.succeed(is_interested) + service.is_interested.return_value = make_awaitable(is_interested) service.token = "mock_service_token" service.url = "mock_service_url" return service diff --git a/tests/handlers/test_auth.py b/tests/handlers/test_auth.py
index c01b04e1dc..97877c2e42 100644 --- a/tests/handlers/test_auth.py +++ b/tests/handlers/test_auth.py
@@ -24,10 +24,11 @@ from synapse.api.errors import ResourceLimitError from synapse.handlers.auth import AuthHandler from tests import unittest +from tests.test_utils import make_awaitable from tests.utils import setup_test_homeserver -class AuthHandlers(object): +class AuthHandlers: def __init__(self, hs): self.auth_handler = AuthHandler(hs) @@ -142,7 +143,7 @@ class AuthTestCase(unittest.TestCase): def test_mau_limits_exceeded_large(self): self.auth_blocking._limit_usage_by_mau = True self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.large_number_of_users) + return_value=make_awaitable(self.large_number_of_users) ) with self.assertRaises(ResourceLimitError): @@ -153,7 +154,7 @@ class AuthTestCase(unittest.TestCase): ) self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.large_number_of_users) + return_value=make_awaitable(self.large_number_of_users) ) with self.assertRaises(ResourceLimitError): yield defer.ensureDeferred( @@ -168,7 +169,7 @@ class AuthTestCase(unittest.TestCase): # If not in monthly active cohort self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.auth_blocking._max_mau_value) + return_value=make_awaitable(self.auth_blocking._max_mau_value) ) with self.assertRaises(ResourceLimitError): yield defer.ensureDeferred( @@ -178,7 +179,7 @@ class AuthTestCase(unittest.TestCase): ) self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.auth_blocking._max_mau_value) + return_value=make_awaitable(self.auth_blocking._max_mau_value) ) with self.assertRaises(ResourceLimitError): yield defer.ensureDeferred( @@ -188,10 +189,10 @@ class AuthTestCase(unittest.TestCase): ) # If in monthly active cohort self.hs.get_datastore().user_last_seen_monthly_active = Mock( - return_value=defer.succeed(self.hs.get_clock().time_msec()) + return_value=make_awaitable(self.hs.get_clock().time_msec()) ) self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.auth_blocking._max_mau_value) + return_value=make_awaitable(self.auth_blocking._max_mau_value) ) yield defer.ensureDeferred( self.auth_handler.get_access_token_for_user_id( @@ -199,10 +200,10 @@ class AuthTestCase(unittest.TestCase): ) ) self.hs.get_datastore().user_last_seen_monthly_active = Mock( - return_value=defer.succeed(self.hs.get_clock().time_msec()) + return_value=make_awaitable(self.hs.get_clock().time_msec()) ) self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.auth_blocking._max_mau_value) + return_value=make_awaitable(self.auth_blocking._max_mau_value) ) yield defer.ensureDeferred( self.auth_handler.validate_short_term_login_token_and_get_user_id( @@ -215,7 +216,7 @@ class AuthTestCase(unittest.TestCase): self.auth_blocking._limit_usage_by_mau = True self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.small_number_of_users) + return_value=make_awaitable(self.small_number_of_users) ) # Ensure does not raise exception yield defer.ensureDeferred( @@ -225,7 +226,7 @@ class AuthTestCase(unittest.TestCase): ) self.hs.get_datastore().get_monthly_active_count = Mock( - return_value=defer.succeed(self.small_number_of_users) + return_value=make_awaitable(self.small_number_of_users) ) yield defer.ensureDeferred( self.auth_handler.validate_short_term_login_token_and_get_user_id( diff --git a/tests/handlers/test_device.py b/tests/handlers/test_device.py
index 6aa322bf3a..969d44c787 100644 --- a/tests/handlers/test_device.py +++ b/tests/handlers/test_device.py
@@ -35,6 +35,17 @@ class DeviceTestCase(unittest.HomeserverTestCase): # These tests assume that it starts 1000 seconds in. self.reactor.advance(1000) + def test_device_is_created_with_invalid_name(self): + self.get_failure( + self.handler.check_device_registered( + user_id="@boris:foo", + device_id="foo", + initial_device_display_name="a" + * (synapse.handlers.device.MAX_DEVICE_DISPLAY_NAME_LEN + 1), + ), + synapse.api.errors.SynapseError, + ) + def test_device_is_created_if_doesnt_exist(self): res = self.get_success( self.handler.check_device_registered( diff --git a/tests/handlers/test_directory.py b/tests/handlers/test_directory.py
index 00bb776271..bc0c5aefdc 100644 --- a/tests/handlers/test_directory.py +++ b/tests/handlers/test_directory.py
@@ -16,8 +16,6 @@ from mock import Mock -from twisted.internet import defer - import synapse import synapse.api.errors from synapse.api.constants import EventTypes @@ -26,6 +24,7 @@ from synapse.rest.client.v1 import directory, login, room from synapse.types import RoomAlias, create_requester from tests import unittest +from tests.test_utils import make_awaitable class DirectoryTestCase(unittest.HomeserverTestCase): @@ -71,7 +70,7 @@ class DirectoryTestCase(unittest.HomeserverTestCase): self.assertEquals({"room_id": "!8765qwer:test", "servers": ["test"]}, result) def test_get_remote_association(self): - self.mock_federation.make_query.return_value = defer.succeed( + self.mock_federation.make_query.return_value = make_awaitable( {"room_id": "!8765qwer:test", "servers": ["test", "remote"]} ) diff --git a/tests/handlers/test_e2e_keys.py b/tests/handlers/test_e2e_keys.py
index 210ddcbb88..366dcfb670 100644 --- a/tests/handlers/test_e2e_keys.py +++ b/tests/handlers/test_e2e_keys.py
@@ -30,7 +30,7 @@ from tests import unittest, utils class E2eKeysHandlerTestCase(unittest.TestCase): def __init__(self, *args, **kwargs): - super(E2eKeysHandlerTestCase, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) self.hs = None # type: synapse.server.HomeServer self.handler = None # type: synapse.handlers.e2e_keys.E2eKeysHandler diff --git a/tests/handlers/test_e2e_room_keys.py b/tests/handlers/test_e2e_room_keys.py
index 3362050ce0..7adde9b9de 100644 --- a/tests/handlers/test_e2e_room_keys.py +++ b/tests/handlers/test_e2e_room_keys.py
@@ -47,7 +47,7 @@ room_keys = { class E2eRoomKeysHandlerTestCase(unittest.TestCase): def __init__(self, *args, **kwargs): - super(E2eRoomKeysHandlerTestCase, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) self.hs = None # type: synapse.server.HomeServer self.handler = None # type: synapse.handlers.e2e_keys.E2eRoomKeysHandler diff --git a/tests/handlers/test_oidc.py b/tests/handlers/test_oidc.py
index 1bb25ab684..d5087e58be 100644 --- a/tests/handlers/test_oidc.py +++ b/tests/handlers/test_oidc.py
@@ -21,7 +21,6 @@ from mock import Mock, patch import attr import pymacaroons -from twisted.internet import defer from twisted.python.failure import Failure from twisted.web._newclient import ResponseDone @@ -75,7 +74,24 @@ COMMON_CONFIG = { COOKIE_NAME = b"oidc_session" COOKIE_PATH = "/_synapse/oidc" -MockedMappingProvider = Mock(OidcMappingProvider) + +class TestMappingProvider(OidcMappingProvider): + @staticmethod + def parse_config(config): + return + + def get_remote_user_id(self, userinfo): + return userinfo["sub"] + + async def map_user_attributes(self, userinfo, token): + return {"localpart": userinfo["username"], "display_name": None} + + # Do not include get_extra_attributes to test backwards compatibility paths. + + +class TestMappingProviderExtra(TestMappingProvider): + async def get_extra_attributes(self, userinfo, token): + return {"phone": userinfo["phone"]} def simple_async_mock(return_value=None, raises=None): @@ -116,15 +132,19 @@ class OidcHandlerTestCase(HomeserverTestCase): config = self.default_config() config["public_baseurl"] = BASE_URL - oidc_config = config.get("oidc_config", {}) + oidc_config = {} oidc_config["enabled"] = True oidc_config["client_id"] = CLIENT_ID oidc_config["client_secret"] = CLIENT_SECRET oidc_config["issuer"] = ISSUER oidc_config["scopes"] = SCOPES oidc_config["user_mapping_provider"] = { - "module": __name__ + ".MockedMappingProvider" + "module": __name__ + ".TestMappingProvider", } + + # Update this config with what's in the default config so that + # override_config works as expected. + oidc_config.update(config.get("oidc_config", {})) config["oidc_config"] = oidc_config hs = self.setup_test_homeserver( @@ -155,11 +175,10 @@ class OidcHandlerTestCase(HomeserverTestCase): self.assertEqual(self.handler._client_auth.client_secret, CLIENT_SECRET) @override_config({"oidc_config": {"discover": True}}) - @defer.inlineCallbacks def test_discovery(self): """The handler should discover the endpoints from OIDC discovery document.""" # This would throw if some metadata were invalid - metadata = yield defer.ensureDeferred(self.handler.load_metadata()) + metadata = self.get_success(self.handler.load_metadata()) self.http_client.get_json.assert_called_once_with(WELL_KNOWN) self.assertEqual(metadata.issuer, ISSUER) @@ -171,43 +190,40 @@ class OidcHandlerTestCase(HomeserverTestCase): # subsequent calls should be cached self.http_client.reset_mock() - yield defer.ensureDeferred(self.handler.load_metadata()) + self.get_success(self.handler.load_metadata()) self.http_client.get_json.assert_not_called() @override_config({"oidc_config": COMMON_CONFIG}) - @defer.inlineCallbacks def test_no_discovery(self): """When discovery is disabled, it should not try to load from discovery document.""" - yield defer.ensureDeferred(self.handler.load_metadata()) + self.get_success(self.handler.load_metadata()) self.http_client.get_json.assert_not_called() @override_config({"oidc_config": COMMON_CONFIG}) - @defer.inlineCallbacks def test_load_jwks(self): """JWKS loading is done once (then cached) if used.""" - jwks = yield defer.ensureDeferred(self.handler.load_jwks()) + jwks = self.get_success(self.handler.load_jwks()) self.http_client.get_json.assert_called_once_with(JWKS_URI) self.assertEqual(jwks, {"keys": []}) # subsequent calls should be cached… self.http_client.reset_mock() - yield defer.ensureDeferred(self.handler.load_jwks()) + self.get_success(self.handler.load_jwks()) self.http_client.get_json.assert_not_called() # …unless forced self.http_client.reset_mock() - yield defer.ensureDeferred(self.handler.load_jwks(force=True)) + self.get_success(self.handler.load_jwks(force=True)) self.http_client.get_json.assert_called_once_with(JWKS_URI) # Throw if the JWKS uri is missing with self.metadata_edit({"jwks_uri": None}): - with self.assertRaises(RuntimeError): - yield defer.ensureDeferred(self.handler.load_jwks(force=True)) + self.get_failure(self.handler.load_jwks(force=True), RuntimeError) # Return empty key set if JWKS are not used self.handler._scopes = [] # not asking the openid scope self.http_client.get_json.reset_mock() - jwks = yield defer.ensureDeferred(self.handler.load_jwks(force=True)) + jwks = self.get_success(self.handler.load_jwks(force=True)) self.http_client.get_json.assert_not_called() self.assertEqual(jwks, {"keys": []}) @@ -289,11 +305,10 @@ class OidcHandlerTestCase(HomeserverTestCase): # This should not throw self.handler._validate_metadata() - @defer.inlineCallbacks def test_redirect_request(self): """The redirect request has the right arguments & generates a valid session cookie.""" req = Mock(spec=["addCookie"]) - url = yield defer.ensureDeferred( + url = self.get_success( self.handler.handle_redirect_request(req, b"http://client/redirect") ) url = urlparse(url) @@ -333,20 +348,18 @@ class OidcHandlerTestCase(HomeserverTestCase): self.assertEqual(params["nonce"], [nonce]) self.assertEqual(redirect, "http://client/redirect") - @defer.inlineCallbacks def test_callback_error(self): """Errors from the provider returned in the callback are displayed.""" self.handler._render_error = Mock() request = Mock(args={}) request.args[b"error"] = [b"invalid_client"] - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_client", "") request.args[b"error_description"] = [b"some description"] - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_client", "some description") - @defer.inlineCallbacks def test_callback(self): """Code callback works and display errors if something went wrong. @@ -367,39 +380,48 @@ class OidcHandlerTestCase(HomeserverTestCase): "sub": "foo", "preferred_username": "bar", } - user_id = UserID("foo", "domain.org") + user_id = "@foo:domain.org" self.handler._render_error = Mock(return_value=None) self.handler._exchange_code = simple_async_mock(return_value=token) self.handler._parse_id_token = simple_async_mock(return_value=userinfo) self.handler._fetch_userinfo = simple_async_mock(return_value=userinfo) self.handler._map_userinfo_to_user = simple_async_mock(return_value=user_id) self.handler._auth_handler.complete_sso_login = simple_async_mock() - request = Mock(spec=["args", "getCookie", "addCookie"]) + request = Mock( + spec=["args", "getCookie", "addCookie", "requestHeaders", "getClientIP"] + ) code = "code" state = "state" nonce = "nonce" client_redirect_url = "http://client/redirect" - session = self.handler._generate_oidc_session_token( + user_agent = "Browser" + ip_address = "10.0.0.1" + request.getCookie.return_value = self.handler._generate_oidc_session_token( state=state, nonce=nonce, client_redirect_url=client_redirect_url, ui_auth_session_id=None, ) - request.getCookie.return_value = session request.args = {} request.args[b"code"] = [code.encode("utf-8")] request.args[b"state"] = [state.encode("utf-8")] - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + request.requestHeaders = Mock(spec=["getRawHeaders"]) + request.requestHeaders.getRawHeaders.return_value = [user_agent.encode("ascii")] + request.getClientIP.return_value = ip_address + + self.get_success(self.handler.handle_oidc_callback(request)) self.handler._auth_handler.complete_sso_login.assert_called_once_with( - user_id, request, client_redirect_url, + user_id, request, client_redirect_url, {}, ) self.handler._exchange_code.assert_called_once_with(code) self.handler._parse_id_token.assert_called_once_with(token, nonce=nonce) - self.handler._map_userinfo_to_user.assert_called_once_with(userinfo, token) + self.handler._map_userinfo_to_user.assert_called_once_with( + userinfo, token, user_agent, ip_address + ) self.handler._fetch_userinfo.assert_not_called() self.handler._render_error.assert_not_called() @@ -407,13 +429,13 @@ class OidcHandlerTestCase(HomeserverTestCase): self.handler._map_userinfo_to_user = simple_async_mock( raises=MappingException() ) - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("mapping_error") self.handler._map_userinfo_to_user = simple_async_mock(return_value=user_id) # Handle ID token errors self.handler._parse_id_token = simple_async_mock(raises=Exception()) - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_token") self.handler._auth_handler.complete_sso_login.reset_mock() @@ -424,30 +446,31 @@ class OidcHandlerTestCase(HomeserverTestCase): # With userinfo fetching self.handler._scopes = [] # do not ask the "openid" scope - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.handler._auth_handler.complete_sso_login.assert_called_once_with( - user_id, request, client_redirect_url, + user_id, request, client_redirect_url, {}, ) self.handler._exchange_code.assert_called_once_with(code) self.handler._parse_id_token.assert_not_called() - self.handler._map_userinfo_to_user.assert_called_once_with(userinfo, token) + self.handler._map_userinfo_to_user.assert_called_once_with( + userinfo, token, user_agent, ip_address + ) self.handler._fetch_userinfo.assert_called_once_with(token) self.handler._render_error.assert_not_called() # Handle userinfo fetching error self.handler._fetch_userinfo = simple_async_mock(raises=Exception()) - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("fetch_error") # Handle code exchange failure self.handler._exchange_code = simple_async_mock( raises=OidcError("invalid_request") ) - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_request") - @defer.inlineCallbacks def test_callback_session(self): """The callback verifies the session presence and validity""" self.handler._render_error = Mock(return_value=None) @@ -456,20 +479,20 @@ class OidcHandlerTestCase(HomeserverTestCase): # Missing cookie request.args = {} request.getCookie.return_value = None - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("missing_session", "No session cookie found") # Missing session parameter request.args = {} request.getCookie.return_value = "session" - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_request", "State parameter is missing") # Invalid cookie request.args = {} request.args[b"state"] = [b"state"] request.getCookie.return_value = "session" - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_session") # Mismatching session @@ -482,18 +505,17 @@ class OidcHandlerTestCase(HomeserverTestCase): request.args = {} request.args[b"state"] = [b"mismatching state"] request.getCookie.return_value = session - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("mismatching_session") # Valid session request.args = {} request.args[b"state"] = [b"state"] request.getCookie.return_value = session - yield defer.ensureDeferred(self.handler.handle_oidc_callback(request)) + self.get_success(self.handler.handle_oidc_callback(request)) self.assertRenderedError("invalid_request") @override_config({"oidc_config": {"client_auth_method": "client_secret_post"}}) - @defer.inlineCallbacks def test_exchange_code(self): """Code exchange behaves correctly and handles various error scenarios.""" token = {"type": "bearer"} @@ -502,7 +524,7 @@ class OidcHandlerTestCase(HomeserverTestCase): return_value=FakeResponse(code=200, phrase=b"OK", body=token_json) ) code = "code" - ret = yield defer.ensureDeferred(self.handler._exchange_code(code)) + ret = self.get_success(self.handler._exchange_code(code)) kwargs = self.http_client.request.call_args[1] self.assertEqual(ret, token) @@ -524,10 +546,9 @@ class OidcHandlerTestCase(HomeserverTestCase): body=b'{"error": "foo", "error_description": "bar"}', ) ) - with self.assertRaises(OidcError) as exc: - yield defer.ensureDeferred(self.handler._exchange_code(code)) - self.assertEqual(exc.exception.error, "foo") - self.assertEqual(exc.exception.error_description, "bar") + exc = self.get_failure(self.handler._exchange_code(code), OidcError) + self.assertEqual(exc.value.error, "foo") + self.assertEqual(exc.value.error_description, "bar") # Internal server error with no JSON body self.http_client.request = simple_async_mock( @@ -535,9 +556,8 @@ class OidcHandlerTestCase(HomeserverTestCase): code=500, phrase=b"Internal Server Error", body=b"Not JSON", ) ) - with self.assertRaises(OidcError) as exc: - yield defer.ensureDeferred(self.handler._exchange_code(code)) - self.assertEqual(exc.exception.error, "server_error") + exc = self.get_failure(self.handler._exchange_code(code), OidcError) + self.assertEqual(exc.value.error, "server_error") # Internal server error with JSON body self.http_client.request = simple_async_mock( @@ -547,17 +567,16 @@ class OidcHandlerTestCase(HomeserverTestCase): body=b'{"error": "internal_server_error"}', ) ) - with self.assertRaises(OidcError) as exc: - yield defer.ensureDeferred(self.handler._exchange_code(code)) - self.assertEqual(exc.exception.error, "internal_server_error") + + exc = self.get_failure(self.handler._exchange_code(code), OidcError) + self.assertEqual(exc.value.error, "internal_server_error") # 4xx error without "error" field self.http_client.request = simple_async_mock( return_value=FakeResponse(code=400, phrase=b"Bad request", body=b"{}",) ) - with self.assertRaises(OidcError) as exc: - yield defer.ensureDeferred(self.handler._exchange_code(code)) - self.assertEqual(exc.exception.error, "server_error") + exc = self.get_failure(self.handler._exchange_code(code), OidcError) + self.assertEqual(exc.value.error, "server_error") # 2xx error with "error" field self.http_client.request = simple_async_mock( @@ -565,6 +584,121 @@ class OidcHandlerTestCase(HomeserverTestCase): code=200, phrase=b"OK", body=b'{"error": "some_error"}', ) ) - with self.assertRaises(OidcError) as exc: - yield defer.ensureDeferred(self.handler._exchange_code(code)) - self.assertEqual(exc.exception.error, "some_error") + exc = self.get_failure(self.handler._exchange_code(code), OidcError) + self.assertEqual(exc.value.error, "some_error") + + @override_config( + { + "oidc_config": { + "user_mapping_provider": { + "module": __name__ + ".TestMappingProviderExtra" + } + } + } + ) + def test_extra_attributes(self): + """ + Login while using a mapping provider that implements get_extra_attributes. + """ + token = { + "type": "bearer", + "id_token": "id_token", + "access_token": "access_token", + } + userinfo = { + "sub": "foo", + "phone": "1234567", + } + user_id = "@foo:domain.org" + self.handler._exchange_code = simple_async_mock(return_value=token) + self.handler._parse_id_token = simple_async_mock(return_value=userinfo) + self.handler._map_userinfo_to_user = simple_async_mock(return_value=user_id) + self.handler._auth_handler.complete_sso_login = simple_async_mock() + request = Mock( + spec=["args", "getCookie", "addCookie", "requestHeaders", "getClientIP"] + ) + + state = "state" + client_redirect_url = "http://client/redirect" + request.getCookie.return_value = self.handler._generate_oidc_session_token( + state=state, + nonce="nonce", + client_redirect_url=client_redirect_url, + ui_auth_session_id=None, + ) + + request.args = {} + request.args[b"code"] = [b"code"] + request.args[b"state"] = [state.encode("utf-8")] + + request.requestHeaders = Mock(spec=["getRawHeaders"]) + request.requestHeaders.getRawHeaders.return_value = [b"Browser"] + request.getClientIP.return_value = "10.0.0.1" + + self.get_success(self.handler.handle_oidc_callback(request)) + + self.handler._auth_handler.complete_sso_login.assert_called_once_with( + user_id, request, client_redirect_url, {"phone": "1234567"}, + ) + + def test_map_userinfo_to_user(self): + """Ensure that mapping the userinfo returned from a provider to an MXID works properly.""" + userinfo = { + "sub": "test_user", + "username": "test_user", + } + # The token doesn't matter with the default user mapping provider. + token = {} + mxid = self.get_success( + self.handler._map_userinfo_to_user( + userinfo, token, "user-agent", "10.10.10.10" + ) + ) + self.assertEqual(mxid, "@test_user:test") + + # Some providers return an integer ID. + userinfo = { + "sub": 1234, + "username": "test_user_2", + } + mxid = self.get_success( + self.handler._map_userinfo_to_user( + userinfo, token, "user-agent", "10.10.10.10" + ) + ) + self.assertEqual(mxid, "@test_user_2:test") + + # Test if the mxid is already taken + store = self.hs.get_datastore() + user3 = UserID.from_string("@test_user_3:test") + self.get_success( + store.register_user(user_id=user3.to_string(), password_hash=None) + ) + userinfo = {"sub": "test3", "username": "test_user_3"} + e = self.get_failure( + self.handler._map_userinfo_to_user( + userinfo, token, "user-agent", "10.10.10.10" + ), + MappingException, + ) + self.assertEqual(str(e.value), "mxid '@test_user_3:test' is already taken") + + @override_config({"oidc_config": {"allow_existing_users": True}}) + def test_map_userinfo_to_existing_user(self): + """Existing users can log in with OpenID Connect when allow_existing_users is True.""" + store = self.hs.get_datastore() + user4 = UserID.from_string("@test_user_4:test") + self.get_success( + store.register_user(user_id=user4.to_string(), password_hash=None) + ) + userinfo = { + "sub": "test4", + "username": "test_user_4", + } + token = {} + mxid = self.get_success( + self.handler._map_userinfo_to_user( + userinfo, token, "user-agent", "10.10.10.10" + ) + ) + self.assertEqual(mxid, "@test_user_4:test") diff --git a/tests/handlers/test_presence.py b/tests/handlers/test_presence.py
index 05ea40a7de..306dcfe944 100644 --- a/tests/handlers/test_presence.py +++ b/tests/handlers/test_presence.py
@@ -19,6 +19,7 @@ from mock import Mock, call from signedjson.key import generate_signing_key from synapse.api.constants import EventTypes, Membership, PresenceState +from synapse.api.presence import UserPresenceState from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.events.builder import EventBuilder from synapse.handlers.presence import ( @@ -32,7 +33,6 @@ from synapse.handlers.presence import ( handle_update, ) from synapse.rest.client.v1 import room -from synapse.storage.presence import UserPresenceState from synapse.types import UserID, get_domain_from_id from tests import unittest diff --git a/tests/handlers/test_profile.py b/tests/handlers/test_profile.py
index 42a236aa58..1cef10feff 100644 --- a/tests/handlers/test_profile.py +++ b/tests/handlers/test_profile.py
@@ -24,10 +24,11 @@ from synapse.handlers.profile import MasterProfileHandler from synapse.types import UserID from tests import unittest +from tests.test_utils import make_awaitable from tests.utils import setup_test_homeserver -class ProfileHandlers(object): +class ProfileHandlers: def __init__(self, hs): self.profile_handler = MasterProfileHandler(hs) @@ -63,12 +64,16 @@ class ProfileTestCase(unittest.TestCase): self.bob = UserID.from_string("@4567:test") self.alice = UserID.from_string("@alice:remote") + yield defer.ensureDeferred(self.store.create_profile(self.frank.localpart)) + self.handler = hs.get_profile_handler() self.hs = hs @defer.inlineCallbacks def test_get_my_name(self): - yield self.store.set_profile_displayname(self.frank.localpart, "Frank", 1) + yield defer.ensureDeferred( + self.store.set_profile_displayname(self.frank.localpart, "Frank", 1) + ) displayname = yield defer.ensureDeferred( self.handler.get_displayname(self.frank) @@ -101,7 +106,12 @@ class ProfileTestCase(unittest.TestCase): ) self.assertEquals( - (yield self.store.get_profile_displayname(self.frank.localpart)), "Frank", + ( + yield defer.ensureDeferred( + self.store.get_profile_displayname(self.frank.localpart) + ) + ), + "Frank", ) @defer.inlineCallbacks @@ -109,10 +119,17 @@ class ProfileTestCase(unittest.TestCase): self.hs.config.enable_set_displayname = False # Setting displayname for the first time is allowed - yield self.store.set_profile_displayname(self.frank.localpart, "Frank", 1) + yield defer.ensureDeferred( + self.store.set_profile_displayname(self.frank.localpart, "Frank", 1) + ) self.assertEquals( - (yield self.store.get_profile_displayname(self.frank.localpart)), "Frank", + ( + yield defer.ensureDeferred( + self.store.get_profile_displayname(self.frank.localpart) + ) + ), + "Frank", ) # Setting displayname a second time is forbidden @@ -136,7 +153,7 @@ class ProfileTestCase(unittest.TestCase): @defer.inlineCallbacks def test_get_other_name(self): - self.mock_federation.make_query.return_value = defer.succeed( + self.mock_federation.make_query.return_value = make_awaitable( {"displayname": "Alice"} ) @@ -154,7 +171,10 @@ class ProfileTestCase(unittest.TestCase): @defer.inlineCallbacks def test_incoming_fed_query(self): - yield self.store.set_profile_displayname("caroline", "Caroline", 1) + yield defer.ensureDeferred(self.store.create_profile("caroline")) + yield defer.ensureDeferred( + self.store.set_profile_displayname("caroline", "Caroline", 1) + ) response = yield defer.ensureDeferred( self.query_handlers["profile"]( @@ -166,8 +186,10 @@ class ProfileTestCase(unittest.TestCase): @defer.inlineCallbacks def test_get_my_avatar(self): - yield self.store.set_profile_avatar_url( - self.frank.localpart, "http://my.server/me.png", 1 + yield defer.ensureDeferred( + self.store.set_profile_avatar_url( + self.frank.localpart, "http://my.server/me.png", 1 + ) ) avatar_url = yield defer.ensureDeferred(self.handler.get_avatar_url(self.frank)) @@ -184,7 +206,11 @@ class ProfileTestCase(unittest.TestCase): ) self.assertEquals( - (yield self.store.get_profile_avatar_url(self.frank.localpart)), + ( + yield defer.ensureDeferred( + self.store.get_profile_avatar_url(self.frank.localpart) + ) + ), "http://my.server/pic.gif", ) @@ -198,7 +224,11 @@ class ProfileTestCase(unittest.TestCase): ) self.assertEquals( - (yield self.store.get_profile_avatar_url(self.frank.localpart)), + ( + yield defer.ensureDeferred( + self.store.get_profile_avatar_url(self.frank.localpart) + ) + ), "http://my.server/me.png", ) @@ -207,12 +237,18 @@ class ProfileTestCase(unittest.TestCase): self.hs.config.enable_set_avatar_url = False # Setting displayname for the first time is allowed - yield self.store.set_profile_avatar_url( - self.frank.localpart, "http://my.server/me.png", 1 + yield defer.ensureDeferred( + self.store.set_profile_avatar_url( + self.frank.localpart, "http://my.server/me.png", 1 + ) ) self.assertEquals( - (yield self.store.get_profile_avatar_url(self.frank.localpart)), + ( + yield defer.ensureDeferred( + self.store.get_profile_avatar_url(self.frank.localpart) + ) + ), "http://my.server/me.png", ) diff --git a/tests/handlers/test_register.py b/tests/handlers/test_register.py
index e951a62a6d..312c03c83d 100644 --- a/tests/handlers/test_register.py +++ b/tests/handlers/test_register.py
@@ -15,8 +15,7 @@ from mock import Mock -from twisted.internet import defer - +from synapse.api.auth import Auth from synapse.api.constants import UserTypes from synapse.api.errors import Codes, ResourceLimitError, SynapseError from synapse.handlers.register import RegistrationHandler @@ -25,15 +24,18 @@ from synapse.rest.client.v2_alpha.register import ( _map_email_to_displayname, register_servlets, ) +from synapse.spam_checker_api import RegistrationBehaviour from synapse.types import RoomAlias, UserID, create_requester from tests.server import FakeChannel +from tests.test_utils import make_awaitable from tests.unittest import override_config +from tests.utils import mock_getRawHeaders from .. import unittest -class RegistrationHandlers(object): +class RegistrationHandlers: def __init__(self, hs): self.registration_handler = RegistrationHandler(hs) @@ -108,7 +110,7 @@ class RegistrationTestCase(unittest.HomeserverTestCase): def test_get_or_create_user_mau_not_blocked(self): self.hs.config.limit_usage_by_mau = True self.store.count_monthly_users = Mock( - return_value=defer.succeed(self.hs.config.max_mau_value - 1) + return_value=make_awaitable(self.hs.config.max_mau_value - 1) ) # Ensure does not throw exception self.get_success(self.get_or_create_user(self.requester, "c", "User")) @@ -116,7 +118,7 @@ class RegistrationTestCase(unittest.HomeserverTestCase): def test_get_or_create_user_mau_blocked(self): self.hs.config.limit_usage_by_mau = True self.store.get_monthly_active_count = Mock( - return_value=defer.succeed(self.lots_of_users) + return_value=make_awaitable(self.lots_of_users) ) self.get_failure( self.get_or_create_user(self.requester, "b", "display_name"), @@ -124,7 +126,7 @@ class RegistrationTestCase(unittest.HomeserverTestCase): ) self.store.get_monthly_active_count = Mock( - return_value=defer.succeed(self.hs.config.max_mau_value) + return_value=make_awaitable(self.hs.config.max_mau_value) ) self.get_failure( self.get_or_create_user(self.requester, "b", "display_name"), @@ -134,14 +136,14 @@ class RegistrationTestCase(unittest.HomeserverTestCase): def test_register_mau_blocked(self): self.hs.config.limit_usage_by_mau = True self.store.get_monthly_active_count = Mock( - return_value=defer.succeed(self.lots_of_users) + return_value=make_awaitable(self.lots_of_users) ) self.get_failure( self.handler.register_user(localpart="local_part"), ResourceLimitError ) self.store.get_monthly_active_count = Mock( - return_value=defer.succeed(self.hs.config.max_mau_value) + return_value=make_awaitable(self.hs.config.max_mau_value) ) self.get_failure( self.handler.register_user(localpart="local_part"), ResourceLimitError @@ -197,7 +199,7 @@ class RegistrationTestCase(unittest.HomeserverTestCase): room_alias_str = "#room:test" self.hs.config.auto_join_rooms = [room_alias_str] - self.store.is_real_user = Mock(return_value=defer.succeed(False)) + self.store.is_real_user = Mock(return_value=make_awaitable(False)) user_id = self.get_success(self.handler.register_user(localpart="support")) rooms = self.get_success(self.store.get_rooms_for_user(user_id)) self.assertEqual(len(rooms), 0) @@ -209,8 +211,8 @@ class RegistrationTestCase(unittest.HomeserverTestCase): def test_auto_create_auto_join_rooms_when_user_is_the_first_real_user(self): room_alias_str = "#room:test" - self.store.count_real_users = Mock(return_value=defer.succeed(1)) - self.store.is_real_user = Mock(return_value=defer.succeed(True)) + self.store.count_real_users = Mock(return_value=make_awaitable(1)) + self.store.is_real_user = Mock(return_value=make_awaitable(True)) user_id = self.get_success(self.handler.register_user(localpart="real")) rooms = self.get_success(self.store.get_rooms_for_user(user_id)) directory_handler = self.hs.get_handlers().directory_handler @@ -224,8 +226,8 @@ class RegistrationTestCase(unittest.HomeserverTestCase): room_alias_str = "#room:test" self.hs.config.auto_join_rooms = [room_alias_str] - self.store.count_real_users = Mock(return_value=defer.succeed(2)) - self.store.is_real_user = Mock(return_value=defer.succeed(True)) + self.store.count_real_users = Mock(return_value=make_awaitable(2)) + self.store.is_real_user = Mock(return_value=make_awaitable(True)) user_id = self.get_success(self.handler.register_user(localpart="real")) rooms = self.get_success(self.store.get_rooms_for_user(user_id)) self.assertEqual(len(rooms), 0) @@ -484,6 +486,53 @@ class RegistrationTestCase(unittest.HomeserverTestCase): self.handler.register_user(localpart=invalid_user_id), SynapseError ) + def test_spam_checker_deny(self): + """A spam checker can deny registration, which results in an error.""" + + class DenyAll: + def check_registration_for_spam( + self, email_threepid, username, request_info + ): + return RegistrationBehaviour.DENY + + # Configure a spam checker that denies all users. + spam_checker = self.hs.get_spam_checker() + spam_checker.spam_checkers = [DenyAll()] + + self.get_failure(self.handler.register_user(localpart="user"), SynapseError) + + def test_spam_checker_shadow_ban(self): + """A spam checker can choose to shadow-ban a user, which allows registration to succeed.""" + + class BanAll: + def check_registration_for_spam( + self, email_threepid, username, request_info + ): + return RegistrationBehaviour.SHADOW_BAN + + # Configure a spam checker that denies all users. + spam_checker = self.hs.get_spam_checker() + spam_checker.spam_checkers = [BanAll()] + + user_id = self.get_success(self.handler.register_user(localpart="user")) + + # Get an access token. + token = self.macaroon_generator.generate_access_token(user_id) + self.get_success( + self.store.add_access_token_to_user( + user_id=user_id, token=token, device_id=None, valid_until_ms=None + ) + ) + + # Ensure the user was marked as shadow-banned. + request = Mock(args={}) + request.args[b"access_token"] = [token.encode("ascii")] + request.requestHeaders.getRawHeaders = mock_getRawHeaders() + auth = Auth(self.hs) + requester = self.get_success(auth.get_user_by_req(request)) + + self.assertTrue(requester.shadow_banned) + def test_email_to_displayname_mapping(self): """Test that custom emails are mapped to new user displaynames correctly""" self._check_mapping( @@ -527,16 +576,16 @@ class RegistrationTestCase(unittest.HomeserverTestCase): # Mock Synapse's threepid validator get_threepid_validation_session = Mock( - return_value=defer.succeed( + return_value=make_awaitable( {"medium": "email", "address": email, "validated_at": 0} ) ) self.store.get_threepid_validation_session = get_threepid_validation_session - delete_threepid_session = Mock(return_value=defer.succeed(None)) + delete_threepid_session = Mock(return_value=make_awaitable(None)) self.store.delete_threepid_session = delete_threepid_session # Mock Synapse's http json post method to check for the internal bind call - post_json_get_json = Mock(return_value=defer.succeed(None)) + post_json_get_json = Mock(return_value=make_awaitable(None)) self.hs.get_simple_http_client().post_json_get_json = post_json_get_json # Retrieve a UIA session ID diff --git a/tests/handlers/test_stats.py b/tests/handlers/test_stats.py
index 07092f026a..0229f58315 100644 --- a/tests/handlers/test_stats.py +++ b/tests/handlers/test_stats.py
@@ -15,7 +15,7 @@ from synapse.rest import admin from synapse.rest.client.v1 import login, room -from synapse.storage.data_stores.main import stats +from synapse.storage.databases.main import stats from tests import unittest @@ -48,16 +48,16 @@ class StatsRoomTests(unittest.HomeserverTestCase): Add the background updates we need to run. """ # Ugh, have to reset this flag - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", {"update_name": "populate_stats_prepare", "progress_json": "{}"}, ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_process_rooms", @@ -67,7 +67,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_process_users", @@ -77,7 +77,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_cleanup", @@ -87,8 +87,8 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) - def get_all_room_state(self): - return self.store.db.simple_select_list( + async def get_all_room_state(self): + return await self.store.db_pool.simple_select_list( "room_stats_state", None, retcols=("name", "topic", "canonical_alias") ) @@ -102,7 +102,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): end_ts = self.store.quantise_stats_time(self.reactor.seconds() * 1000) return self.get_success( - self.store.db.simple_select_one( + self.store.db_pool.simple_select_one( table + "_historical", {id_col: stat_id, end_ts: end_ts}, cols, @@ -115,10 +115,10 @@ class StatsRoomTests(unittest.HomeserverTestCase): self._add_background_updates() while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) def test_initial_room(self): @@ -152,10 +152,10 @@ class StatsRoomTests(unittest.HomeserverTestCase): self._add_background_updates() while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) r = self.get_success(self.get_all_room_state()) @@ -192,9 +192,9 @@ class StatsRoomTests(unittest.HomeserverTestCase): # the position that the deltas should begin at, once they take over. self.hs.config.stats_enabled = True self.handler.stats_enabled = True - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False self.get_success( - self.store.db.simple_update_one( + self.store.db_pool.simple_update_one( table="stats_incremental_position", keyvalues={}, updatevalues={"stream_id": 0}, @@ -202,17 +202,17 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", {"update_name": "populate_stats_prepare", "progress_json": "{}"}, ) ) while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) # Now, before the table is actually ingested, add some more events. @@ -223,13 +223,13 @@ class StatsRoomTests(unittest.HomeserverTestCase): # Now do the initial ingestion. self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", {"update_name": "populate_stats_process_rooms", "progress_json": "{}"}, ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_cleanup", @@ -239,12 +239,12 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) self.reactor.advance(86401) @@ -259,7 +259,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): # self.handler.notify_new_event() # We need to let the delta processor advance… - self.pump(10 * 60) + self.reactor.advance(10 * 60) # Get the slices! There should be two -- day 1, and day 2. r = self.get_success(self.store.get_statistics_for_subject("room", room_1, 0)) @@ -352,6 +352,37 @@ class StatsRoomTests(unittest.HomeserverTestCase): self.assertEqual(r1stats_post["total_events"] - r1stats_ante["total_events"], 1) + def test_updating_profile_information_does_not_increase_joined_members_count(self): + """ + Check that the joined_members count does not increase when a user changes their + profile information (which is done by sending another join membership event into + the room. + """ + self._perform_background_initial_update() + + # Create a user and room + u1 = self.register_user("u1", "pass") + u1token = self.login("u1", "pass") + r1 = self.helper.create_room_as(u1, tok=u1token) + + # Get the current room stats + r1stats_ante = self._get_current_stats("room", r1) + + # Send a profile update into the room + new_profile = {"displayname": "bob"} + self.helper.change_membership( + r1, u1, u1, "join", extra_data=new_profile, tok=u1token + ) + + # Get the new room stats + r1stats_post = self._get_current_stats("room", r1) + + # Ensure that the user count did not changed + self.assertEqual(r1stats_post["joined_members"], r1stats_ante["joined_members"]) + self.assertEqual( + r1stats_post["local_users_in_room"], r1stats_ante["local_users_in_room"] + ) + def test_send_state_event_nonoverwriting(self): """ When we send a non-overwriting state event, it increments total_events AND current_state_events @@ -675,15 +706,15 @@ class StatsRoomTests(unittest.HomeserverTestCase): # preparation stage of the initial background update # Ugh, have to reset this flag - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False self.get_success( - self.store.db.simple_delete( + self.store.db_pool.simple_delete( "room_stats_current", {"1": 1}, "test_delete_stats" ) ) self.get_success( - self.store.db.simple_delete( + self.store.db_pool.simple_delete( "user_stats_current", {"1": 1}, "test_delete_stats" ) ) @@ -695,9 +726,9 @@ class StatsRoomTests(unittest.HomeserverTestCase): # now do the background updates - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_process_rooms", @@ -707,7 +738,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_process_users", @@ -717,7 +748,7 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_stats_cleanup", @@ -728,10 +759,10 @@ class StatsRoomTests(unittest.HomeserverTestCase): ) while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) r1stats_complete = self._get_current_stats("room", r1) diff --git a/tests/handlers/test_typing.py b/tests/handlers/test_typing.py
index 5878f74175..3fec09ea8a 100644 --- a/tests/handlers/test_typing.py +++ b/tests/handlers/test_typing.py
@@ -21,9 +21,10 @@ from mock import ANY, Mock, call from twisted.internet import defer from synapse.api.errors import AuthError -from synapse.types import UserID +from synapse.types import UserID, create_requester from tests import unittest +from tests.test_utils import make_awaitable from tests.unittest import override_config from tests.utils import register_federation_servlets @@ -72,6 +73,7 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): "delivered_txn", "get_received_txn_response", "set_received_txn_response", + "get_destination_last_successful_stream_ordering", "get_destination_retry_timings", "get_devices_by_remote", "maybe_store_room_on_invite", @@ -79,6 +81,7 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): "get_user_directory_stream_pos", "get_current_state_deltas", "get_device_updates_by_remote", + "get_room_max_stream_ordering", ] ) @@ -115,10 +118,14 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): retry_timings_res ) - self.datastore.get_device_updates_by_remote.return_value = defer.succeed( + self.datastore.get_device_updates_by_remote.return_value = make_awaitable( (0, []) ) + self.datastore.get_destination_last_successful_stream_ordering.return_value = make_awaitable( + None + ) + def get_received_txn_response(*args): return defer.succeed(None) @@ -126,10 +133,10 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.room_members = [] - def check_user_in_room(room_id, user_id): + async def check_user_in_room(room_id, user_id): if user_id not in [u.to_string() for u in self.room_members]: raise AuthError(401, "User is not in the room") - return defer.succeed(None) + return None hs.get_auth().check_user_in_room = check_user_in_room @@ -143,19 +150,21 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.datastore.get_users_in_room = get_users_in_room - self.datastore.get_user_directory_stream_pos.return_value = ( + self.datastore.get_user_directory_stream_pos.side_effect = ( # we deliberately return a non-None stream pos to avoid doing an initial_spam - defer.succeed(1) + lambda: make_awaitable(1) ) self.datastore.get_current_state_deltas.return_value = (0, None) self.datastore.get_to_device_stream_token = lambda: 0 - self.datastore.get_new_device_msgs_for_remote = lambda *args, **kargs: defer.succeed( + self.datastore.get_new_device_msgs_for_remote = lambda *args, **kargs: make_awaitable( ([], 0) ) - self.datastore.delete_device_msgs_for_remote = lambda *args, **kargs: None - self.datastore.set_received_txn_response = lambda *args, **kwargs: defer.succeed( + self.datastore.delete_device_msgs_for_remote = lambda *args, **kargs: make_awaitable( + None + ) + self.datastore.set_received_txn_response = lambda *args, **kwargs: make_awaitable( None ) @@ -166,7 +175,10 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.get_success( self.handler.started_typing( - target_user=U_APPLE, auth_user=U_APPLE, room_id=ROOM_ID, timeout=20000 + target_user=U_APPLE, + requester=create_requester(U_APPLE), + room_id=ROOM_ID, + timeout=20000, ) ) @@ -193,7 +205,10 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.get_success( self.handler.started_typing( - target_user=U_APPLE, auth_user=U_APPLE, room_id=ROOM_ID, timeout=20000 + target_user=U_APPLE, + requester=create_requester(U_APPLE), + room_id=ROOM_ID, + timeout=20000, ) ) @@ -268,7 +283,9 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.get_success( self.handler.stopped_typing( - target_user=U_APPLE, auth_user=U_APPLE, room_id=ROOM_ID + target_user=U_APPLE, + requester=create_requester(U_APPLE), + room_id=ROOM_ID, ) ) @@ -308,7 +325,10 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.get_success( self.handler.started_typing( - target_user=U_APPLE, auth_user=U_APPLE, room_id=ROOM_ID, timeout=10000 + target_user=U_APPLE, + requester=create_requester(U_APPLE), + room_id=ROOM_ID, + timeout=10000, ) ) @@ -347,7 +367,10 @@ class TypingNotificationsTestCase(unittest.HomeserverTestCase): self.get_success( self.handler.started_typing( - target_user=U_APPLE, auth_user=U_APPLE, room_id=ROOM_ID, timeout=10000 + target_user=U_APPLE, + requester=create_requester(U_APPLE), + room_id=ROOM_ID, + timeout=10000, ) ) diff --git a/tests/handlers/test_user_directory.py b/tests/handlers/test_user_directory.py
index ddee8d9e3a..48f750d357 100644 --- a/tests/handlers/test_user_directory.py +++ b/tests/handlers/test_user_directory.py
@@ -238,7 +238,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): def test_spam_checker(self): """ - A user which fails to the spam checks will not appear in search results. + A user which fails the spam checks will not appear in search results. """ u1 = self.register_user("user1", "pass") u1_token = self.login(u1, "pass") @@ -269,7 +269,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): # Configure a spam checker that does not filter any users. spam_checker = self.hs.get_spam_checker() - class AllowAll(object): + class AllowAll: def check_username_for_spam(self, user_profile): # Allow all users. return False @@ -282,7 +282,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): self.assertEqual(len(s["results"]), 1) # Configure a spam checker that filters all users. - class BlockAll(object): + class BlockAll: def check_username_for_spam(self, user_profile): # All users are spammy. return True @@ -339,7 +339,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): def get_users_in_public_rooms(self): r = self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( "users_in_public_rooms", None, ("user_id", "room_id") ) ) @@ -350,7 +350,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): def get_users_who_share_private_rooms(self): return self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( "users_who_share_private_rooms", None, ["user_id", "other_user_id", "room_id"], @@ -362,10 +362,10 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): Add the background updates we need to run. """ # Ugh, have to reset this flag - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_user_directory_createtables", @@ -374,7 +374,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_user_directory_process_rooms", @@ -384,7 +384,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_user_directory_process_users", @@ -394,7 +394,7 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): ) ) self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( "background_updates", { "update_name": "populate_user_directory_cleanup", @@ -437,10 +437,10 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): self._add_background_updates() while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) shares_private = self.get_users_who_share_private_rooms() @@ -476,10 +476,10 @@ class UserDirectoryTestCase(unittest.HomeserverTestCase): self._add_background_updates() while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) shares_private = self.get_users_who_share_private_rooms() diff --git a/tests/http/__init__.py b/tests/http/__init__.py
index 2096ba3c91..3e5a856584 100644 --- a/tests/http/__init__.py +++ b/tests/http/__init__.py
@@ -133,7 +133,7 @@ def create_test_cert_file(sanlist): @implementer(IOpenSSLServerConnectionCreator) -class TestServerTLSConnectionFactory(object): +class TestServerTLSConnectionFactory: """An SSL connection creator which returns connections which present a certificate signed by our test CA.""" @@ -145,7 +145,7 @@ class TestServerTLSConnectionFactory(object): self._cert_file = create_test_cert_file(sanlist) def serverConnectionForTLS(self, tlsProtocol): - ctx = SSL.Context(SSL.TLSv1_METHOD) + ctx = SSL.Context(SSL.SSLv23_METHOD) ctx.use_certificate_file(self._cert_file) ctx.use_privatekey_file(get_test_key_file()) return Connection(ctx, None) diff --git a/tests/http/federation/test_matrix_federation_agent.py b/tests/http/federation/test_matrix_federation_agent.py
index db260d599e..c3f7a28dcc 100644 --- a/tests/http/federation/test_matrix_federation_agent.py +++ b/tests/http/federation/test_matrix_federation_agent.py
@@ -972,7 +972,9 @@ class MatrixFederationAgentTests(unittest.TestCase): def test_well_known_cache(self): self.reactor.lookups["testserv"] = "1.2.3.4" - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) # there should be an attempt to connect on port 443 for the .well-known clients = self.reactor.tcpClients @@ -995,7 +997,9 @@ class MatrixFederationAgentTests(unittest.TestCase): well_known_server.loseConnection() # repeat the request: it should hit the cache - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) r = self.successResultOf(fetch_d) self.assertEqual(r.delegated_server, b"target-server") @@ -1003,7 +1007,9 @@ class MatrixFederationAgentTests(unittest.TestCase): self.reactor.pump((1000.0,)) # now it should connect again - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) self.assertEqual(len(clients), 1) (host, port, client_factory, _timeout, _bindAddress) = clients.pop(0) @@ -1026,7 +1032,9 @@ class MatrixFederationAgentTests(unittest.TestCase): self.reactor.lookups["testserv"] = "1.2.3.4" - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) # there should be an attempt to connect on port 443 for the .well-known clients = self.reactor.tcpClients @@ -1052,7 +1060,9 @@ class MatrixFederationAgentTests(unittest.TestCase): # another lookup. self.reactor.pump((900.0,)) - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) # The resolver may retry a few times, so fonx all requests that come along attempts = 0 @@ -1082,7 +1092,9 @@ class MatrixFederationAgentTests(unittest.TestCase): self.reactor.pump((10000.0,)) # Repated the request, this time it should fail if the lookup fails. - fetch_d = self.well_known_resolver.get_well_known(b"testserv") + fetch_d = defer.ensureDeferred( + self.well_known_resolver.get_well_known(b"testserv") + ) clients = self.reactor.tcpClients (host, port, client_factory, _timeout, _bindAddress) = clients.pop(0) @@ -1252,7 +1264,7 @@ def _log_request(request): @implementer(IPolicyForHTTPS) -class TrustingTLSPolicyForHTTPS(object): +class TrustingTLSPolicyForHTTPS: """An IPolicyForHTTPS which checks that the certificate belongs to the right server, but doesn't check the certificate chain.""" diff --git a/tests/http/test_fedclient.py b/tests/http/test_fedclient.py
index fff4f0cbf4..212484a7fe 100644 --- a/tests/http/test_fedclient.py +++ b/tests/http/test_fedclient.py
@@ -16,6 +16,7 @@ from mock import Mock from netaddr import IPSet +from parameterized import parameterized from twisted.internet import defer from twisted.internet.defer import TimeoutError @@ -58,7 +59,9 @@ class FederationClientTests(HomeserverTestCase): @defer.inlineCallbacks def do_request(): with LoggingContext("one") as context: - fetch_d = self.cl.get_json("testserv:8008", "foo/bar") + fetch_d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar") + ) # Nothing happened yet self.assertNoResult(fetch_d) @@ -120,7 +123,9 @@ class FederationClientTests(HomeserverTestCase): """ If the DNS lookup returns an error, it will bubble up. """ - d = self.cl.get_json("testserv2:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred( + self.cl.get_json("testserv2:8008", "foo/bar", timeout=10000) + ) self.pump() f = self.failureResultOf(d) @@ -128,7 +133,9 @@ class FederationClientTests(HomeserverTestCase): self.assertIsInstance(f.value.inner_exception, DNSLookupError) def test_client_connection_refused(self): - d = self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + ) self.pump() @@ -154,7 +161,9 @@ class FederationClientTests(HomeserverTestCase): If the HTTP request is not connected and is timed out, it'll give a ConnectingCancelledError or TimeoutError. """ - d = self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + ) self.pump() @@ -184,7 +193,9 @@ class FederationClientTests(HomeserverTestCase): If the HTTP request is connected, but gets no response before being timed out, it'll give a ResponseNeverReceived. """ - d = self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar", timeout=10000) + ) self.pump() @@ -226,7 +237,7 @@ class FederationClientTests(HomeserverTestCase): # Try making a GET request to a blacklisted IPv4 address # ------------------------------------------------------ # Make the request - d = cl.get_json("internal:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred(cl.get_json("internal:8008", "foo/bar", timeout=10000)) # Nothing happened yet self.assertNoResult(d) @@ -244,7 +255,9 @@ class FederationClientTests(HomeserverTestCase): # Try making a POST request to a blacklisted IPv6 address # ------------------------------------------------------- # Make the request - d = cl.post_json("internalv6:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred( + cl.post_json("internalv6:8008", "foo/bar", timeout=10000) + ) # Nothing has happened yet self.assertNoResult(d) @@ -263,7 +276,7 @@ class FederationClientTests(HomeserverTestCase): # Try making a GET request to a non-blacklisted IPv4 address # ---------------------------------------------------------- # Make the request - d = cl.post_json("fine:8008", "foo/bar", timeout=10000) + d = defer.ensureDeferred(cl.post_json("fine:8008", "foo/bar", timeout=10000)) # Nothing has happened yet self.assertNoResult(d) @@ -286,7 +299,7 @@ class FederationClientTests(HomeserverTestCase): request = MatrixFederationRequest( method="GET", destination="testserv:8008", path="foo/bar" ) - d = self.cl._send_request(request, timeout=10000) + d = defer.ensureDeferred(self.cl._send_request(request, timeout=10000)) self.pump() @@ -305,12 +318,14 @@ class FederationClientTests(HomeserverTestCase): r = self.successResultOf(d) self.assertEqual(r.code, 200) - def test_client_headers_no_body(self): + @parameterized.expand(["get_json", "post_json", "delete_json", "put_json"]) + def test_timeout_reading_body(self, method_name: str): """ If the HTTP request is connected, but gets no response before being - timed out, it'll give a ResponseNeverReceived. + timed out, it'll give a RequestSendFailed with can_retry. """ - d = self.cl.post_json("testserv:8008", "foo/bar", timeout=10000) + method = getattr(self.cl, method_name) + d = defer.ensureDeferred(method("testserv:8008", "foo/bar", timeout=10000)) self.pump() @@ -334,7 +349,9 @@ class FederationClientTests(HomeserverTestCase): self.reactor.advance(10.5) f = self.failureResultOf(d) - self.assertIsInstance(f.value, TimeoutError) + self.assertIsInstance(f.value, RequestSendFailed) + self.assertTrue(f.value.can_retry) + self.assertIsInstance(f.value.inner_exception, defer.TimeoutError) def test_client_requires_trailing_slashes(self): """ @@ -342,7 +359,9 @@ class FederationClientTests(HomeserverTestCase): requiring a trailing slash. We need to retry the request with a trailing slash. Workaround for Synapse <= v0.99.3, explained in #3622. """ - d = self.cl.get_json("testserv:8008", "foo/bar", try_trailing_slash_on_400=True) + d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar", try_trailing_slash_on_400=True) + ) # Send the request self.pump() @@ -395,7 +414,9 @@ class FederationClientTests(HomeserverTestCase): See test_client_requires_trailing_slashes() for context. """ - d = self.cl.get_json("testserv:8008", "foo/bar", try_trailing_slash_on_400=True) + d = defer.ensureDeferred( + self.cl.get_json("testserv:8008", "foo/bar", try_trailing_slash_on_400=True) + ) # Send the request self.pump() @@ -432,7 +453,11 @@ class FederationClientTests(HomeserverTestCase): self.failureResultOf(d) def test_client_sends_body(self): - self.cl.post_json("testserv:8008", "foo/bar", timeout=10000, data={"a": "b"}) + defer.ensureDeferred( + self.cl.post_json( + "testserv:8008", "foo/bar", timeout=10000, data={"a": "b"} + ) + ) self.pump() @@ -453,7 +478,7 @@ class FederationClientTests(HomeserverTestCase): def test_closes_connection(self): """Check that the client closes unused HTTP connections""" - d = self.cl.get_json("testserv:8008", "foo/bar") + d = defer.ensureDeferred(self.cl.get_json("testserv:8008", "foo/bar")) self.pump() @@ -486,6 +511,53 @@ class FederationClientTests(HomeserverTestCase): self.assertFalse(conn.disconnecting) # wait for a while - self.pump(120) + self.reactor.advance(120) self.assertTrue(conn.disconnecting) + + @parameterized.expand([(b"",), (b"foo",), (b'{"a": Infinity}',)]) + def test_json_error(self, return_value): + """ + Test what happens if invalid JSON is returned from the remote endpoint. + """ + + test_d = defer.ensureDeferred(self.cl.get_json("testserv:8008", "foo/bar")) + + self.pump() + + # Nothing happened yet + self.assertNoResult(test_d) + + # Make sure treq is trying to connect + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 1) + (host, port, factory, _timeout, _bindAddress) = clients[0] + self.assertEqual(host, "1.2.3.4") + self.assertEqual(port, 8008) + + # complete the connection and wire it up to a fake transport + protocol = factory.buildProtocol(None) + transport = StringTransport() + protocol.makeConnection(transport) + + # that should have made it send the request to the transport + self.assertRegex(transport.value(), b"^GET /foo/bar") + self.assertRegex(transport.value(), b"Host: testserv:8008") + + # Deferred is still without a result + self.assertNoResult(test_d) + + # Send it the HTTP response + protocol.dataReceived( + b"HTTP/1.1 200 OK\r\n" + b"Server: Fake\r\n" + b"Content-Type: application/json\r\n" + b"Content-Length: %i\r\n" + b"\r\n" + b"%s" % (len(return_value), return_value) + ) + + self.pump() + + f = self.failureResultOf(test_d) + self.assertIsInstance(f.value, ValueError) diff --git a/tests/http/test_servlet.py b/tests/http/test_servlet.py new file mode 100644
index 0000000000..45089158ce --- /dev/null +++ b/tests/http/test_servlet.py
@@ -0,0 +1,80 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import json +from io import BytesIO + +from mock import Mock + +from synapse.api.errors import SynapseError +from synapse.http.servlet import ( + parse_json_object_from_request, + parse_json_value_from_request, +) + +from tests import unittest + + +def make_request(content): + """Make an object that acts enough like a request.""" + request = Mock(spec=["content"]) + + if isinstance(content, dict): + content = json.dumps(content).encode("utf8") + + request.content = BytesIO(content) + return request + + +class TestServletUtils(unittest.TestCase): + def test_parse_json_value(self): + """Basic tests for parse_json_value_from_request.""" + # Test round-tripping. + obj = {"foo": 1} + result = parse_json_value_from_request(make_request(obj)) + self.assertEqual(result, obj) + + # Results don't have to be objects. + result = parse_json_value_from_request(make_request(b'["foo"]')) + self.assertEqual(result, ["foo"]) + + # Test empty. + with self.assertRaises(SynapseError): + parse_json_value_from_request(make_request(b"")) + + result = parse_json_value_from_request(make_request(b""), allow_empty_body=True) + self.assertIsNone(result) + + # Invalid UTF-8. + with self.assertRaises(SynapseError): + parse_json_value_from_request(make_request(b"\xFF\x00")) + + # Invalid JSON. + with self.assertRaises(SynapseError): + parse_json_value_from_request(make_request(b"foo")) + + with self.assertRaises(SynapseError): + parse_json_value_from_request(make_request(b'{"foo": Infinity}')) + + def test_parse_json_object(self): + """Basic tests for parse_json_object_from_request.""" + # Test empty. + result = parse_json_object_from_request( + make_request(b""), allow_empty_body=True + ) + self.assertEqual(result, {}) + + # Test not an object + with self.assertRaises(SynapseError): + parse_json_object_from_request(make_request(b'["foo"]')) diff --git a/tests/http/test_simple_client.py b/tests/http/test_simple_client.py new file mode 100644
index 0000000000..a1cf0862d4 --- /dev/null +++ b/tests/http/test_simple_client.py
@@ -0,0 +1,180 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from mock import Mock + +from netaddr import IPSet + +from twisted.internet import defer +from twisted.internet.error import DNSLookupError + +from synapse.http import RequestTimedOutError +from synapse.http.client import SimpleHttpClient +from synapse.server import HomeServer + +from tests.unittest import HomeserverTestCase + + +class SimpleHttpClientTests(HomeserverTestCase): + def prepare(self, reactor, clock, hs: "HomeServer"): + # Add a DNS entry for a test server + self.reactor.lookups["testserv"] = "1.2.3.4" + + self.cl = hs.get_simple_http_client() + + def test_dns_error(self): + """ + If the DNS lookup returns an error, it will bubble up. + """ + d = defer.ensureDeferred(self.cl.get_json("http://testserv2:8008/foo/bar")) + self.pump() + + f = self.failureResultOf(d) + self.assertIsInstance(f.value, DNSLookupError) + + def test_client_connection_refused(self): + d = defer.ensureDeferred(self.cl.get_json("http://testserv:8008/foo/bar")) + + self.pump() + + # Nothing happened yet + self.assertNoResult(d) + + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 1) + (host, port, factory, _timeout, _bindAddress) = clients[0] + self.assertEqual(host, "1.2.3.4") + self.assertEqual(port, 8008) + e = Exception("go away") + factory.clientConnectionFailed(None, e) + self.pump(0.5) + + f = self.failureResultOf(d) + + self.assertIs(f.value, e) + + def test_client_never_connect(self): + """ + If the HTTP request is not connected and is timed out, it'll give a + ConnectingCancelledError or TimeoutError. + """ + d = defer.ensureDeferred(self.cl.get_json("http://testserv:8008/foo/bar")) + + self.pump() + + # Nothing happened yet + self.assertNoResult(d) + + # Make sure treq is trying to connect + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 1) + self.assertEqual(clients[0][0], "1.2.3.4") + self.assertEqual(clients[0][1], 8008) + + # Deferred is still without a result + self.assertNoResult(d) + + # Push by enough to time it out + self.reactor.advance(120) + f = self.failureResultOf(d) + + self.assertIsInstance(f.value, RequestTimedOutError) + + def test_client_connect_no_response(self): + """ + If the HTTP request is connected, but gets no response before being + timed out, it'll give a ResponseNeverReceived. + """ + d = defer.ensureDeferred(self.cl.get_json("http://testserv:8008/foo/bar")) + + self.pump() + + # Nothing happened yet + self.assertNoResult(d) + + # Make sure treq is trying to connect + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 1) + self.assertEqual(clients[0][0], "1.2.3.4") + self.assertEqual(clients[0][1], 8008) + + conn = Mock() + client = clients[0][2].buildProtocol(None) + client.makeConnection(conn) + + # Deferred is still without a result + self.assertNoResult(d) + + # Push by enough to time it out + self.reactor.advance(120) + f = self.failureResultOf(d) + + self.assertIsInstance(f.value, RequestTimedOutError) + + def test_client_ip_range_blacklist(self): + """Ensure that Synapse does not try to connect to blacklisted IPs""" + + # Add some DNS entries we'll blacklist + self.reactor.lookups["internal"] = "127.0.0.1" + self.reactor.lookups["internalv6"] = "fe80:0:0:0:0:8a2e:370:7337" + ip_blacklist = IPSet(["127.0.0.0/8", "fe80::/64"]) + + cl = SimpleHttpClient(self.hs, ip_blacklist=ip_blacklist) + + # Try making a GET request to a blacklisted IPv4 address + # ------------------------------------------------------ + # Make the request + d = defer.ensureDeferred(cl.get_json("http://internal:8008/foo/bar")) + self.pump(1) + + # Check that it was unable to resolve the address + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 0) + + self.failureResultOf(d, DNSLookupError) + + # Try making a POST request to a blacklisted IPv6 address + # ------------------------------------------------------- + # Make the request + d = defer.ensureDeferred( + cl.post_json_get_json("http://internalv6:8008/foo/bar", {}) + ) + + # Move the reactor forwards + self.pump(1) + + # Check that it was unable to resolve the address + clients = self.reactor.tcpClients + self.assertEqual(len(clients), 0) + + # Check that it was due to a blacklisted DNS lookup + self.failureResultOf(d, DNSLookupError) + + # Try making a GET request to a non-blacklisted IPv4 address + # ---------------------------------------------------------- + # Make the request + d = defer.ensureDeferred(cl.get_json("http://testserv:8008/foo/bar")) + + # Nothing has happened yet + self.assertNoResult(d) + + # Move the reactor forwards + self.pump(1) + + # Check that it was able to resolve the address + clients = self.reactor.tcpClients + self.assertNotEqual(len(clients), 0) + + # Connection will still fail as this IP address does not resolve to anything + self.failureResultOf(d, RequestTimedOutError) diff --git a/tests/logging/test_structured.py b/tests/logging/test_structured.py
index 451d05c0f0..d36f5f426c 100644 --- a/tests/logging/test_structured.py +++ b/tests/logging/test_structured.py
@@ -29,12 +29,12 @@ from synapse.logging.context import LoggingContext from tests.unittest import DEBUG, HomeserverTestCase -class FakeBeginner(object): +class FakeBeginner: def beginLoggingTo(self, observers, **kwargs): self.observers = observers -class StructuredLoggingTestBase(object): +class StructuredLoggingTestBase: """ Test base that registers a cleanup handler to reset the stdlib log handler to 'unset'. diff --git a/tests/module_api/test_api.py b/tests/module_api/test_api.py
index 9c778a0e45..ccbb82f6a3 100644 --- a/tests/module_api/test_api.py +++ b/tests/module_api/test_api.py
@@ -47,7 +47,7 @@ class ModuleApiTestCase(HomeserverTestCase): # Check that the new user exists with all provided attributes self.assertEqual(user_id, "@bob:test") self.assertTrue(access_token) - self.assertTrue(self.store.get_user_by_id(user_id)) + self.assertTrue(self.get_success(self.store.get_user_by_id(user_id))) # Check that the email was assigned emails = self.get_success(self.store.user_get_threepids(user_id)) diff --git a/tests/push/test_email.py b/tests/push/test_email.py
index 83032cc9ea..3224568640 100644 --- a/tests/push/test_email.py +++ b/tests/push/test_email.py
@@ -27,7 +27,7 @@ from tests.unittest import HomeserverTestCase @attr.s -class _User(object): +class _User: "Helper wrapper for user ID and access token" id = attr.ib() token = attr.ib() @@ -170,7 +170,7 @@ class EmailPusherTests(HomeserverTestCase): last_stream_ordering = pushers[0]["last_stream_ordering"] # Advance time a bit, so the pusher will register something has happened - self.pump(100) + self.pump(10) # It hasn't succeeded yet, so the stream ordering shouldn't have moved pushers = self.get_success( diff --git a/tests/replication/_base.py b/tests/replication/_base.py
index 06575ba0a6..ae60874ec3 100644 --- a/tests/replication/_base.py +++ b/tests/replication/_base.py
@@ -65,7 +65,7 @@ class BaseStreamTestCase(unittest.HomeserverTestCase): # Since we use sqlite in memory databases we need to make sure the # databases objects are the same. - self.worker_hs.get_datastore().db = hs.get_datastore().db + self.worker_hs.get_datastore().db_pool = hs.get_datastore().db_pool self.test_handler = self._build_replication_data_handler() self.worker_hs.replication_data_handler = self.test_handler @@ -198,7 +198,7 @@ class BaseMultiWorkerStreamTestCase(unittest.HomeserverTestCase): self.streamer = self.hs.get_replication_streamer() store = self.hs.get_datastore() - self.database = store.db + self.database_pool = store.db_pool self.reactor.lookups["testserv"] = "1.2.3.4" @@ -254,7 +254,7 @@ class BaseMultiWorkerStreamTestCase(unittest.HomeserverTestCase): ) store = worker_hs.get_datastore() - store.db._db_pool = self.database._db_pool + store.db_pool._db_pool = self.database_pool._db_pool repl_handler = ReplicationCommandHandler(worker_hs) client = ClientReplicationStreamProtocol( diff --git a/tests/replication/slave/storage/test_events.py b/tests/replication/slave/storage/test_events.py
index 1a88c7fb80..c0ee1cfbd6 100644 --- a/tests/replication/slave/storage/test_events.py +++ b/tests/replication/slave/storage/test_events.py
@@ -20,6 +20,7 @@ from synapse.events import FrozenEvent, _EventInternalMetadata, make_event_from_ from synapse.handlers.room import RoomEventSource from synapse.replication.slave.storage.events import SlavedEventStore from synapse.storage.roommember import RoomsForUser +from synapse.types import PersistedEventPosition from tests.server import FakeTransport @@ -58,7 +59,7 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): # Patch up the equality operator for events so that we can check # whether lists of events match using assertEquals self.unpatches = [patch__eq__(_EventInternalMetadata), patch__eq__(FrozenEvent)] - return super(SlavedEventStoreTestCase, self).setUp() + return super().setUp() def prepare(self, *args, **kwargs): super().prepare(*args, **kwargs) @@ -160,7 +161,7 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): self.check( "get_unread_event_push_actions_by_room_for_user", [ROOM_ID, USER_ID_2, event1.event_id], - {"highlight_count": 0, "notify_count": 0}, + {"highlight_count": 0, "unread_count": 0, "notify_count": 0}, ) self.persist( @@ -173,7 +174,7 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): self.check( "get_unread_event_push_actions_by_room_for_user", [ROOM_ID, USER_ID_2, event1.event_id], - {"highlight_count": 0, "notify_count": 1}, + {"highlight_count": 0, "unread_count": 0, "notify_count": 1}, ) self.persist( @@ -188,7 +189,7 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): self.check( "get_unread_event_push_actions_by_room_for_user", [ROOM_ID, USER_ID_2, event1.event_id], - {"highlight_count": 1, "notify_count": 2}, + {"highlight_count": 1, "unread_count": 0, "notify_count": 2}, ) def test_get_rooms_for_user_with_stream_ordering(self): @@ -204,10 +205,14 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): type="m.room.member", sender=USER_ID_2, key=USER_ID_2, membership="join" ) self.replicate() + + expected_pos = PersistedEventPosition( + "master", j2.internal_metadata.stream_ordering + ) self.check( "get_rooms_for_user_with_stream_ordering", (USER_ID_2,), - {(ROOM_ID, j2.internal_metadata.stream_ordering)}, + {(ROOM_ID, expected_pos)}, ) def test_get_rooms_for_user_with_stream_ordering_with_multi_event_persist(self): @@ -293,9 +298,10 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): # the membership change is only any use to us if the room is in the # joined_rooms list. if membership_changes: - self.assertEqual( - joined_rooms, {(ROOM_ID, j2.internal_metadata.stream_ordering)} + expected_pos = PersistedEventPosition( + "master", j2.internal_metadata.stream_ordering ) + self.assertEqual(joined_rooms, {(ROOM_ID, expected_pos)}) event_id = 0 @@ -366,7 +372,11 @@ class SlavedEventStoreTestCase(BaseSlavedStoreTestCase): state_handler = self.hs.get_state_handler() context = self.get_success(state_handler.compute_event_context(event)) - self.master_store.add_push_actions_to_staging( - event.event_id, {user_id: actions for user_id, actions in push_actions} + self.get_success( + self.master_store.add_push_actions_to_staging( + event.event_id, + {user_id: actions for user_id, actions in push_actions}, + False, + ) ) return event, context diff --git a/tests/replication/test_federation_sender_shard.py b/tests/replication/test_federation_sender_shard.py
index 8d4dbf232e..1d7edee5ba 100644 --- a/tests/replication/test_federation_sender_shard.py +++ b/tests/replication/test_federation_sender_shard.py
@@ -16,15 +16,14 @@ import logging from mock import Mock -from twisted.internet import defer - from synapse.api.constants import EventTypes, Membership from synapse.events.builder import EventBuilderFactory from synapse.rest.admin import register_servlets_for_client_rest_resource from synapse.rest.client.v1 import login, room -from synapse.types import UserID +from synapse.types import UserID, create_requester from tests.replication._base import BaseMultiWorkerStreamTestCase +from tests.test_utils import make_awaitable logger = logging.getLogger(__name__) @@ -46,7 +45,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): new event. """ mock_client = Mock(spec=["put_json"]) - mock_client.put_json.side_effect = lambda *_, **__: defer.succeed({}) + mock_client.put_json.return_value = make_awaitable({}) self.make_worker_hs( "synapse.app.federation_sender", @@ -74,7 +73,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): new events. """ mock_client1 = Mock(spec=["put_json"]) - mock_client1.put_json.side_effect = lambda *_, **__: defer.succeed({}) + mock_client1.put_json.return_value = make_awaitable({}) self.make_worker_hs( "synapse.app.federation_sender", { @@ -86,7 +85,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): ) mock_client2 = Mock(spec=["put_json"]) - mock_client2.put_json.side_effect = lambda *_, **__: defer.succeed({}) + mock_client2.put_json.return_value = make_awaitable({}) self.make_worker_hs( "synapse.app.federation_sender", { @@ -137,7 +136,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): new typing EDUs. """ mock_client1 = Mock(spec=["put_json"]) - mock_client1.put_json.side_effect = lambda *_, **__: defer.succeed({}) + mock_client1.put_json.return_value = make_awaitable({}) self.make_worker_hs( "synapse.app.federation_sender", { @@ -149,7 +148,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): ) mock_client2 = Mock(spec=["put_json"]) - mock_client2.put_json.side_effect = lambda *_, **__: defer.succeed({}) + mock_client2.put_json.return_value = make_awaitable({}) self.make_worker_hs( "synapse.app.federation_sender", { @@ -176,7 +175,7 @@ class FederationSenderTestCase(BaseMultiWorkerStreamTestCase): self.get_success( typing_handler.started_typing( target_user=UserID.from_string(user), - auth_user=UserID.from_string(user), + requester=create_requester(user), room_id=room, timeout=20000, ) diff --git a/tests/rest/admin/test_admin.py b/tests/rest/admin/test_admin.py
index b1a4decced..0f1144fe1e 100644 --- a/tests/rest/admin/test_admin.py +++ b/tests/rest/admin/test_admin.py
@@ -178,7 +178,7 @@ class QuarantineMediaTestCase(unittest.HomeserverTestCase): self.fetches = [] - def get_file(destination, path, output_stream, args=None, max_size=None): + async def get_file(destination, path, output_stream, args=None, max_size=None): """ Returns tuple[int,dict,str,int] of file length, response headers, absolute URI, and response code. @@ -192,7 +192,7 @@ class QuarantineMediaTestCase(unittest.HomeserverTestCase): d = Deferred() d.addCallback(write_to) self.fetches.append((d, destination, path, args)) - return make_deferred_yieldable(d) + return await make_deferred_yieldable(d) client = Mock() client.get_file = get_file diff --git a/tests/rest/admin/test_device.py b/tests/rest/admin/test_device.py
index faa7f381a9..92c9058887 100644 --- a/tests/rest/admin/test_device.py +++ b/tests/rest/admin/test_device.py
@@ -221,7 +221,7 @@ class DeviceRestTestCase(unittest.HomeserverTestCase): self.render(request) self.assertEqual(400, channel.code, msg=channel.json_body) - self.assertEqual(Codes.UNKNOWN, channel.json_body["errcode"]) + self.assertEqual(Codes.TOO_LARGE, channel.json_body["errcode"]) # Ensure the display name was not updated. request, channel = self.make_request( diff --git a/tests/rest/admin/test_event_reports.py b/tests/rest/admin/test_event_reports.py new file mode 100644
index 0000000000..bf79086f78 --- /dev/null +++ b/tests/rest/admin/test_event_reports.py
@@ -0,0 +1,382 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 Dirk Klimpel +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +import json + +import synapse.rest.admin +from synapse.api.errors import Codes +from synapse.rest.client.v1 import login, room +from synapse.rest.client.v2_alpha import report_event + +from tests import unittest + + +class EventReportsTestCase(unittest.HomeserverTestCase): + servlets = [ + synapse.rest.admin.register_servlets, + login.register_servlets, + room.register_servlets, + report_event.register_servlets, + ] + + def prepare(self, reactor, clock, hs): + self.store = hs.get_datastore() + + self.admin_user = self.register_user("admin", "pass", admin=True) + self.admin_user_tok = self.login("admin", "pass") + + self.other_user = self.register_user("user", "pass") + self.other_user_tok = self.login("user", "pass") + + self.room_id1 = self.helper.create_room_as( + self.other_user, tok=self.other_user_tok, is_public=True + ) + self.helper.join(self.room_id1, user=self.admin_user, tok=self.admin_user_tok) + + self.room_id2 = self.helper.create_room_as( + self.other_user, tok=self.other_user_tok, is_public=True + ) + self.helper.join(self.room_id2, user=self.admin_user, tok=self.admin_user_tok) + + # Two rooms and two users. Every user sends and reports every room event + for i in range(5): + self._create_event_and_report( + room_id=self.room_id1, user_tok=self.other_user_tok, + ) + for i in range(5): + self._create_event_and_report( + room_id=self.room_id2, user_tok=self.other_user_tok, + ) + for i in range(5): + self._create_event_and_report( + room_id=self.room_id1, user_tok=self.admin_user_tok, + ) + for i in range(5): + self._create_event_and_report( + room_id=self.room_id2, user_tok=self.admin_user_tok, + ) + + self.url = "/_synapse/admin/v1/event_reports" + + def test_requester_is_no_admin(self): + """ + If the user is not a server admin, an error 403 is returned. + """ + + request, channel = self.make_request( + "GET", self.url, access_token=self.other_user_tok, + ) + self.render(request) + + self.assertEqual(403, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.FORBIDDEN, channel.json_body["errcode"]) + + def test_default_success(self): + """ + Testing list of reported events + """ + + request, channel = self.make_request( + "GET", self.url, access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 20) + self.assertNotIn("next_token", channel.json_body) + self._check_fields(channel.json_body["event_reports"]) + + def test_limit(self): + """ + Testing list of reported events with limit + """ + + request, channel = self.make_request( + "GET", self.url + "?limit=5", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 5) + self.assertEqual(channel.json_body["next_token"], 5) + self._check_fields(channel.json_body["event_reports"]) + + def test_from(self): + """ + Testing list of reported events with a defined starting point (from) + """ + + request, channel = self.make_request( + "GET", self.url + "?from=5", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 15) + self.assertNotIn("next_token", channel.json_body) + self._check_fields(channel.json_body["event_reports"]) + + def test_limit_and_from(self): + """ + Testing list of reported events with a defined starting point and limit + """ + + request, channel = self.make_request( + "GET", self.url + "?from=5&limit=10", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(channel.json_body["next_token"], 15) + self.assertEqual(len(channel.json_body["event_reports"]), 10) + self._check_fields(channel.json_body["event_reports"]) + + def test_filter_room(self): + """ + Testing list of reported events with a filter of room + """ + + request, channel = self.make_request( + "GET", + self.url + "?room_id=%s" % self.room_id1, + access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 10) + self.assertEqual(len(channel.json_body["event_reports"]), 10) + self.assertNotIn("next_token", channel.json_body) + self._check_fields(channel.json_body["event_reports"]) + + for report in channel.json_body["event_reports"]: + self.assertEqual(report["room_id"], self.room_id1) + + def test_filter_user(self): + """ + Testing list of reported events with a filter of user + """ + + request, channel = self.make_request( + "GET", + self.url + "?user_id=%s" % self.other_user, + access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 10) + self.assertEqual(len(channel.json_body["event_reports"]), 10) + self.assertNotIn("next_token", channel.json_body) + self._check_fields(channel.json_body["event_reports"]) + + for report in channel.json_body["event_reports"]: + self.assertEqual(report["user_id"], self.other_user) + + def test_filter_user_and_room(self): + """ + Testing list of reported events with a filter of user and room + """ + + request, channel = self.make_request( + "GET", + self.url + "?user_id=%s&room_id=%s" % (self.other_user, self.room_id1), + access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 5) + self.assertEqual(len(channel.json_body["event_reports"]), 5) + self.assertNotIn("next_token", channel.json_body) + self._check_fields(channel.json_body["event_reports"]) + + for report in channel.json_body["event_reports"]: + self.assertEqual(report["user_id"], self.other_user) + self.assertEqual(report["room_id"], self.room_id1) + + def test_valid_search_order(self): + """ + Testing search order. Order by timestamps. + """ + + # fetch the most recent first, largest timestamp + request, channel = self.make_request( + "GET", self.url + "?dir=b", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 20) + report = 1 + while report < len(channel.json_body["event_reports"]): + self.assertGreaterEqual( + channel.json_body["event_reports"][report - 1]["received_ts"], + channel.json_body["event_reports"][report]["received_ts"], + ) + report += 1 + + # fetch the oldest first, smallest timestamp + request, channel = self.make_request( + "GET", self.url + "?dir=f", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 20) + report = 1 + while report < len(channel.json_body["event_reports"]): + self.assertLessEqual( + channel.json_body["event_reports"][report - 1]["received_ts"], + channel.json_body["event_reports"][report]["received_ts"], + ) + report += 1 + + def test_invalid_search_order(self): + """ + Testing that a invalid search order returns a 400 + """ + + request, channel = self.make_request( + "GET", self.url + "?dir=bar", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(400, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.INVALID_PARAM, channel.json_body["errcode"]) + self.assertEqual("Unknown direction: bar", channel.json_body["error"]) + + def test_limit_is_negative(self): + """ + Testing that a negative list parameter returns a 400 + """ + + request, channel = self.make_request( + "GET", self.url + "?limit=-5", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(400, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.INVALID_PARAM, channel.json_body["errcode"]) + + def test_from_is_negative(self): + """ + Testing that a negative from parameter returns a 400 + """ + + request, channel = self.make_request( + "GET", self.url + "?from=-5", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(400, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.INVALID_PARAM, channel.json_body["errcode"]) + + def test_next_token(self): + """ + Testing that `next_token` appears at the right place + """ + + # `next_token` does not appear + # Number of results is the number of entries + request, channel = self.make_request( + "GET", self.url + "?limit=20", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 20) + self.assertNotIn("next_token", channel.json_body) + + # `next_token` does not appear + # Number of max results is larger than the number of entries + request, channel = self.make_request( + "GET", self.url + "?limit=21", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 20) + self.assertNotIn("next_token", channel.json_body) + + # `next_token` does appear + # Number of max results is smaller than the number of entries + request, channel = self.make_request( + "GET", self.url + "?limit=19", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 19) + self.assertEqual(channel.json_body["next_token"], 19) + + # Check + # Set `from` to value of `next_token` for request remaining entries + # `next_token` does not appear + request, channel = self.make_request( + "GET", self.url + "?from=19", access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(channel.json_body["total"], 20) + self.assertEqual(len(channel.json_body["event_reports"]), 1) + self.assertNotIn("next_token", channel.json_body) + + def _create_event_and_report(self, room_id, user_tok): + """Create and report events + """ + resp = self.helper.send(room_id, tok=user_tok) + event_id = resp["event_id"] + + request, channel = self.make_request( + "POST", + "rooms/%s/report/%s" % (room_id, event_id), + json.dumps({"score": -100, "reason": "this makes me sad"}), + access_token=user_tok, + ) + self.render(request) + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + + def _check_fields(self, content): + """Checks that all attributes are present in a event report + """ + for c in content: + self.assertIn("id", c) + self.assertIn("received_ts", c) + self.assertIn("room_id", c) + self.assertIn("event_id", c) + self.assertIn("user_id", c) + self.assertIn("reason", c) + self.assertIn("content", c) + self.assertIn("sender", c) + self.assertIn("room_alias", c) + self.assertIn("event_json", c) + self.assertIn("score", c["content"]) + self.assertIn("reason", c["content"]) + self.assertIn("auth_events", c["event_json"]) + self.assertIn("type", c["event_json"]) + self.assertIn("room_id", c["event_json"]) + self.assertIn("sender", c["event_json"]) + self.assertIn("content", c["event_json"]) diff --git a/tests/rest/admin/test_room.py b/tests/rest/admin/test_room.py
index ba8552c29f..6dfc709dc5 100644 --- a/tests/rest/admin/test_room.py +++ b/tests/rest/admin/test_room.py
@@ -283,6 +283,23 @@ class DeleteRoomTestCase(unittest.HomeserverTestCase): self.assertEqual(400, int(channel.result["code"]), msg=channel.result["body"]) self.assertEqual(Codes.BAD_JSON, channel.json_body["errcode"]) + def test_purge_is_not_bool(self): + """ + If parameter `purge` is not boolean, return an error + """ + body = json.dumps({"purge": "NotBool"}) + + request, channel = self.make_request( + "POST", + self.url, + content=body.encode(encoding="utf_8"), + access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(400, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.BAD_JSON, channel.json_body["errcode"]) + def test_purge_room_and_block(self): """Test to purge a room and block it. Members will not be moved to a new room and will not receive a message. @@ -297,7 +314,7 @@ class DeleteRoomTestCase(unittest.HomeserverTestCase): # Assert one user in room self._is_member(room_id=self.room_id, user_id=self.other_user) - body = json.dumps({"block": True}) + body = json.dumps({"block": True, "purge": True}) request, channel = self.make_request( "POST", @@ -331,7 +348,7 @@ class DeleteRoomTestCase(unittest.HomeserverTestCase): # Assert one user in room self._is_member(room_id=self.room_id, user_id=self.other_user) - body = json.dumps({"block": False}) + body = json.dumps({"block": False, "purge": True}) request, channel = self.make_request( "POST", @@ -351,6 +368,42 @@ class DeleteRoomTestCase(unittest.HomeserverTestCase): self._is_blocked(self.room_id, expect=False) self._has_no_members(self.room_id) + def test_block_room_and_not_purge(self): + """Test to block a room without purging it. + Members will not be moved to a new room and will not receive a message. + The room will not be purged. + """ + # Test that room is not purged + with self.assertRaises(AssertionError): + self._is_purged(self.room_id) + + # Test that room is not blocked + self._is_blocked(self.room_id, expect=False) + + # Assert one user in room + self._is_member(room_id=self.room_id, user_id=self.other_user) + + body = json.dumps({"block": False, "purge": False}) + + request, channel = self.make_request( + "POST", + self.url.encode("ascii"), + content=body.encode(encoding="utf_8"), + access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(None, channel.json_body["new_room_id"]) + self.assertEqual(self.other_user, channel.json_body["kicked_users"][0]) + self.assertIn("failed_to_kick_users", channel.json_body) + self.assertIn("local_aliases", channel.json_body) + + with self.assertRaises(AssertionError): + self._is_purged(self.room_id) + self._is_blocked(self.room_id, expect=False) + self._has_no_members(self.room_id) + def test_shutdown_room_consent(self): """Test that we can shutdown rooms with local users who have not yet accepted the privacy policy. This used to fail when we tried to @@ -513,7 +566,7 @@ class DeleteRoomTestCase(unittest.HomeserverTestCase): "state_groups_state", ): count = self.get_success( - self.store.db.simple_select_one_onecol( + self.store.db_pool.simple_select_one_onecol( table=table, keyvalues={"room_id": room_id}, retcol="COUNT(*)", @@ -614,7 +667,7 @@ class PurgeRoomTestCase(unittest.HomeserverTestCase): "state_groups_state", ): count = self.get_success( - self.store.db.simple_select_one_onecol( + self.store.db_pool.simple_select_one_onecol( table=table, keyvalues={"room_id": room_id}, retcol="COUNT(*)", @@ -1121,6 +1174,8 @@ class RoomTestCase(unittest.HomeserverTestCase): self.assertIn("room_id", channel.json_body) self.assertIn("name", channel.json_body) + self.assertIn("topic", channel.json_body) + self.assertIn("avatar", channel.json_body) self.assertIn("canonical_alias", channel.json_body) self.assertIn("joined_members", channel.json_body) self.assertIn("joined_local_members", channel.json_body) diff --git a/tests/rest/admin/test_user.py b/tests/rest/admin/test_user.py
index f16eef15f7..98d0623734 100644 --- a/tests/rest/admin/test_user.py +++ b/tests/rest/admin/test_user.py
@@ -22,11 +22,12 @@ from mock import Mock import synapse.rest.admin from synapse.api.constants import UserTypes -from synapse.api.errors import HttpResponseException, ResourceLimitError -from synapse.rest.client.v1 import login +from synapse.api.errors import Codes, HttpResponseException, ResourceLimitError +from synapse.rest.client.v1 import login, room from synapse.rest.client.v2_alpha import sync from tests import unittest +from tests.test_utils import make_awaitable from tests.unittest import override_config @@ -335,7 +336,9 @@ class UserRegisterTestCase(unittest.HomeserverTestCase): store = self.hs.get_datastore() # Set monthly active users to the limit - store.get_monthly_active_count = Mock(return_value=self.hs.config.max_mau_value) + store.get_monthly_active_count = Mock( + return_value=make_awaitable(self.hs.config.max_mau_value) + ) # Check that the blocking of monthly active users is working as expected # The registration of a new user fails due to the limit self.get_failure( @@ -588,7 +591,7 @@ class UserRestTestCase(unittest.HomeserverTestCase): # Set monthly active users to the limit self.store.get_monthly_active_count = Mock( - return_value=self.hs.config.max_mau_value + return_value=make_awaitable(self.hs.config.max_mau_value) ) # Check that the blocking of monthly active users is working as expected # The registration of a new user fails due to the limit @@ -628,7 +631,7 @@ class UserRestTestCase(unittest.HomeserverTestCase): # Set monthly active users to the limit self.store.get_monthly_active_count = Mock( - return_value=self.hs.config.max_mau_value + return_value=make_awaitable(self.hs.config.max_mau_value) ) # Check that the blocking of monthly active users is working as expected # The registration of a new user fails due to the limit @@ -871,6 +874,10 @@ class UserRestTestCase(unittest.HomeserverTestCase): ) self.render(request) self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) + self._is_erased("@user:test", False) + d = self.store.mark_user_erased("@user:test") + self.assertIsNone(self.get_success(d)) + self._is_erased("@user:test", True) # Attempt to reactivate the user (without a password). request, channel = self.make_request( @@ -903,6 +910,7 @@ class UserRestTestCase(unittest.HomeserverTestCase): self.assertEqual(200, int(channel.result["code"]), msg=channel.result["body"]) self.assertEqual("@user:test", channel.json_body["name"]) self.assertEqual(False, channel.json_body["deactivated"]) + self._is_erased("@user:test", False) def test_set_user_as_admin(self): """ @@ -992,3 +1000,104 @@ class UserRestTestCase(unittest.HomeserverTestCase): # Ensure they're still alive self.assertEqual(0, channel.json_body["deactivated"]) + + def _is_erased(self, user_id, expect): + """Assert that the user is erased or not + """ + d = self.store.is_user_erased(user_id) + if expect: + self.assertTrue(self.get_success(d)) + else: + self.assertFalse(self.get_success(d)) + + +class UserMembershipRestTestCase(unittest.HomeserverTestCase): + + servlets = [ + synapse.rest.admin.register_servlets, + login.register_servlets, + sync.register_servlets, + room.register_servlets, + ] + + def prepare(self, reactor, clock, hs): + self.store = hs.get_datastore() + + self.admin_user = self.register_user("admin", "pass", admin=True) + self.admin_user_tok = self.login("admin", "pass") + + self.other_user = self.register_user("user", "pass") + self.url = "/_synapse/admin/v1/users/%s/joined_rooms" % urllib.parse.quote( + self.other_user + ) + + def test_no_auth(self): + """ + Try to list rooms of an user without authentication. + """ + request, channel = self.make_request("GET", self.url, b"{}") + self.render(request) + + self.assertEqual(401, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.MISSING_TOKEN, channel.json_body["errcode"]) + + def test_requester_is_no_admin(self): + """ + If the user is not a server admin, an error is returned. + """ + other_user_token = self.login("user", "pass") + + request, channel = self.make_request( + "GET", self.url, access_token=other_user_token, + ) + self.render(request) + + self.assertEqual(403, int(channel.result["code"]), msg=channel.result["body"]) + self.assertEqual(Codes.FORBIDDEN, channel.json_body["errcode"]) + + def test_user_does_not_exist(self): + """ + Tests that a lookup for a user that does not exist returns a 404 + """ + url = "/_synapse/admin/v1/users/@unknown_person:test/joined_rooms" + request, channel = self.make_request( + "GET", url, access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(404, channel.code, msg=channel.json_body) + self.assertEqual(Codes.NOT_FOUND, channel.json_body["errcode"]) + + def test_user_is_not_local(self): + """ + Tests that a lookup for a user that is not a local returns a 400 + """ + url = "/_synapse/admin/v1/users/@unknown_person:unknown_domain/joined_rooms" + + request, channel = self.make_request( + "GET", url, access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(400, channel.code, msg=channel.json_body) + self.assertEqual("Can only lookup local users", channel.json_body["error"]) + + def test_get_rooms(self): + """ + Tests that a normal lookup for rooms is successfully + """ + # Create rooms and join + other_user_tok = self.login("user", "pass") + number_rooms = 5 + for n in range(number_rooms): + self.helper.create_room_as(self.other_user, tok=other_user_tok) + + # Get rooms + request, channel = self.make_request( + "GET", self.url, access_token=self.admin_user_tok, + ) + self.render(request) + + self.assertEqual(200, channel.code, msg=channel.json_body) + self.assertEqual(number_rooms, channel.json_body["total"]) + self.assertEqual(number_rooms, len(channel.json_body["joined_rooms"])) diff --git a/tests/rest/client/test_retention.py b/tests/rest/client/test_retention.py
index cc264cf0b5..47c0d5634c 100644 --- a/tests/rest/client/test_retention.py +++ b/tests/rest/client/test_retention.py
@@ -46,50 +46,63 @@ class RetentionTestCase(unittest.HomeserverTestCase): } self.hs = self.setup_test_homeserver(config=config) + return self.hs def prepare(self, reactor, clock, homeserver): self.user_id = self.register_user("user", "password") self.token = self.login("user", "password") - def test_retention_state_event(self): - """Tests that the server configuration can limit the values a user can set to the - room's retention policy. + self.store = self.hs.get_datastore() + self.serializer = self.hs.get_event_client_serializer() + self.clock = self.hs.get_clock() + + def test_retention_event_purged_with_state_event(self): + """Tests that expired events are correctly purged when the room's retention policy + is defined by a state event. """ room_id = self.helper.create_room_as(self.user_id, tok=self.token) + # Set the room's retention period to 2 days. + lifetime = one_day_ms * 2 self.helper.send_state( room_id=room_id, event_type=EventTypes.Retention, - body={"max_lifetime": one_day_ms * 4}, + body={"max_lifetime": lifetime}, tok=self.token, - expect_code=400, ) + self._test_retention_event_purged(room_id, one_day_ms * 1.5) + + def test_retention_event_purged_with_state_event_outside_allowed(self): + """Tests that the server configuration can override the policy for a room when + running the purge jobs. + """ + room_id = self.helper.create_room_as(self.user_id, tok=self.token) + + # Set a max_lifetime higher than the maximum allowed value. self.helper.send_state( room_id=room_id, event_type=EventTypes.Retention, - body={"max_lifetime": one_hour_ms}, + body={"max_lifetime": one_day_ms * 4}, tok=self.token, - expect_code=400, ) - def test_retention_event_purged_with_state_event(self): - """Tests that expired events are correctly purged when the room's retention policy - is defined by a state event. - """ - room_id = self.helper.create_room_as(self.user_id, tok=self.token) + # Check that the event is purged after waiting for the maximum allowed duration + # instead of the one specified in the room's policy. + self._test_retention_event_purged(room_id, one_day_ms * 1.5) - # Set the room's retention period to 2 days. - lifetime = one_day_ms * 2 + # Set a max_lifetime lower than the minimum allowed value. self.helper.send_state( room_id=room_id, event_type=EventTypes.Retention, - body={"max_lifetime": lifetime}, + body={"max_lifetime": one_hour_ms}, tok=self.token, ) - self._test_retention_event_purged(room_id, one_day_ms * 1.5) + # Check that the event is purged after waiting for the minimum allowed duration + # instead of the one specified in the room's policy. + self._test_retention_event_purged(room_id, one_day_ms * 0.5) def test_retention_event_purged_without_state_event(self): """Tests that expired events are correctly purged when the room's retention policy @@ -141,11 +154,33 @@ class RetentionTestCase(unittest.HomeserverTestCase): # That event should be the second, not outdated event. self.assertEqual(filtered_events[0].event_id, valid_event_id, filtered_events) - def _test_retention_event_purged(self, room_id, increment): + def _test_retention_event_purged(self, room_id: str, increment: float): + """Run the following test scenario to test the message retention policy support: + + 1. Send event 1 + 2. Increment time by `increment` + 3. Send event 2 + 4. Increment time by `increment` + 5. Check that event 1 has been purged + 6. Check that event 2 has not been purged + 7. Check that state events that were sent before event 1 aren't purged. + The main reason for sending a second event is because currently Synapse won't + purge the latest message in a room because it would otherwise result in a lack of + forward extremities for this room. It's also a good thing to ensure the purge jobs + aren't too greedy and purge messages they shouldn't. + + Args: + room_id: The ID of the room to test retention in. + increment: The number of milliseconds to advance the clock each time. Must be + defined so that events in the room aren't purged if they are `increment` + old but are purged if they are `increment * 2` old. + """ # Get the create event to, later, check that we can still access it. message_handler = self.hs.get_message_handler() create_event = self.get_success( - message_handler.get_room_data(self.user_id, room_id, EventTypes.Create) + message_handler.get_room_data( + self.user_id, room_id, EventTypes.Create, state_key="" + ) ) # Send a first event to the room. This is the event we'll want to be purged at the @@ -155,7 +190,7 @@ class RetentionTestCase(unittest.HomeserverTestCase): expired_event_id = resp.get("event_id") # Check that we can retrieve the event. - expired_event = self.get_event(room_id, expired_event_id) + expired_event = self.get_event(expired_event_id) self.assertEqual( expired_event.get("content", {}).get("body"), "1", expired_event ) @@ -173,26 +208,31 @@ class RetentionTestCase(unittest.HomeserverTestCase): # one should still be kept. self.reactor.advance(increment / 1000) - # Check that the event has been purged from the database. - self.get_event(room_id, expired_event_id, expected_code=404) + # Check that the first event has been purged from the database, i.e. that we + # can't retrieve it anymore, because it has expired. + self.get_event(expired_event_id, expect_none=True) - # Check that the event that hasn't been purged can still be retrieved. - valid_event = self.get_event(room_id, valid_event_id) + # Check that the event that hasn't expired can still be retrieved. + valid_event = self.get_event(valid_event_id) self.assertEqual(valid_event.get("content", {}).get("body"), "2", valid_event) # Check that we can still access state events that were sent before the event that # has been purged. self.get_event(room_id, create_event.event_id) - def get_event(self, room_id, event_id, expected_code=200): - url = "/_matrix/client/r0/rooms/%s/event/%s" % (room_id, event_id) + def get_event(self, event_id, expect_none=False): + event = self.get_success(self.store.get_event(event_id, allow_none=True)) - request, channel = self.make_request("GET", url, access_token=self.token) - self.render(request) + if expect_none: + self.assertIsNone(event) + return {} - self.assertEqual(channel.code, expected_code, channel.result) + self.assertIsNotNone(event) - return channel.json_body + time_now = self.clock.time_msec() + serialized = self.get_success(self.serializer.serialize_event(event, time_now)) + + return serialized class RetentionNoDefaultPolicyTestCase(unittest.HomeserverTestCase): diff --git a/tests/rest/client/test_shadow_banned.py b/tests/rest/client/test_shadow_banned.py new file mode 100644
index 0000000000..dfe4bf7762 --- /dev/null +++ b/tests/rest/client/test_shadow_banned.py
@@ -0,0 +1,312 @@ +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +from mock import Mock, patch + +import synapse.rest.admin +from synapse.api.constants import EventTypes +from synapse.rest.client.v1 import directory, login, profile, room +from synapse.rest.client.v2_alpha import room_upgrade_rest_servlet + +from tests import unittest + + +class _ShadowBannedBase(unittest.HomeserverTestCase): + def prepare(self, reactor, clock, homeserver): + # Create two users, one of which is shadow-banned. + self.banned_user_id = self.register_user("banned", "test") + self.banned_access_token = self.login("banned", "test") + + self.store = self.hs.get_datastore() + + self.get_success( + self.store.db_pool.simple_update( + table="users", + keyvalues={"name": self.banned_user_id}, + updatevalues={"shadow_banned": True}, + desc="shadow_ban", + ) + ) + + self.other_user_id = self.register_user("otheruser", "pass") + self.other_access_token = self.login("otheruser", "pass") + + +# To avoid the tests timing out don't add a delay to "annoy the requester". +@patch("random.randint", new=lambda a, b: 0) +class RoomTestCase(_ShadowBannedBase): + servlets = [ + synapse.rest.admin.register_servlets_for_client_rest_resource, + directory.register_servlets, + login.register_servlets, + room.register_servlets, + room_upgrade_rest_servlet.register_servlets, + ] + + def test_invite(self): + """Invites from shadow-banned users don't actually get sent.""" + + # The create works fine. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + # Inviting the user completes successfully. + self.helper.invite( + room=room_id, + src=self.banned_user_id, + tok=self.banned_access_token, + targ=self.other_user_id, + ) + + # But the user wasn't actually invited. + invited_rooms = self.get_success( + self.store.get_invited_rooms_for_local_user(self.other_user_id) + ) + self.assertEqual(invited_rooms, []) + + def test_invite_3pid(self): + """Ensure that a 3PID invite does not attempt to contact the identity server.""" + identity_handler = self.hs.get_handlers().identity_handler + identity_handler.lookup_3pid = Mock( + side_effect=AssertionError("This should not get called") + ) + + # The create works fine. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + # Inviting the user completes successfully. + request, channel = self.make_request( + "POST", + "/rooms/%s/invite" % (room_id,), + {"id_server": "test", "medium": "email", "address": "test@test.test"}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code, channel.result) + + # This should have raised an error earlier, but double check this wasn't called. + identity_handler.lookup_3pid.assert_not_called() + + def test_create_room(self): + """Invitations during a room creation should be discarded, but the room still gets created.""" + # The room creation is successful. + request, channel = self.make_request( + "POST", + "/_matrix/client/r0/createRoom", + {"visibility": "public", "invite": [self.other_user_id]}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code, channel.result) + room_id = channel.json_body["room_id"] + + # But the user wasn't actually invited. + invited_rooms = self.get_success( + self.store.get_invited_rooms_for_local_user(self.other_user_id) + ) + self.assertEqual(invited_rooms, []) + + # Since a real room was created, the other user should be able to join it. + self.helper.join(room_id, self.other_user_id, tok=self.other_access_token) + + # Both users should be in the room. + users = self.get_success(self.store.get_users_in_room(room_id)) + self.assertCountEqual(users, ["@banned:test", "@otheruser:test"]) + + def test_message(self): + """Messages from shadow-banned users don't actually get sent.""" + + room_id = self.helper.create_room_as( + self.other_user_id, tok=self.other_access_token + ) + + # The user should be in the room. + self.helper.join(room_id, self.banned_user_id, tok=self.banned_access_token) + + # Sending a message should complete successfully. + result = self.helper.send_event( + room_id=room_id, + type=EventTypes.Message, + content={"msgtype": "m.text", "body": "with right label"}, + tok=self.banned_access_token, + ) + self.assertIn("event_id", result) + event_id = result["event_id"] + + latest_events = self.get_success( + self.store.get_latest_event_ids_in_room(room_id) + ) + self.assertNotIn(event_id, latest_events) + + def test_upgrade(self): + """A room upgrade should fail, but look like it succeeded.""" + + # The create works fine. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + request, channel = self.make_request( + "POST", + "/_matrix/client/r0/rooms/%s/upgrade" % (room_id,), + {"new_version": "6"}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code, channel.result) + # A new room_id should be returned. + self.assertIn("replacement_room", channel.json_body) + + new_room_id = channel.json_body["replacement_room"] + + # It doesn't really matter what API we use here, we just want to assert + # that the room doesn't exist. + summary = self.get_success(self.store.get_room_summary(new_room_id)) + # The summary should be empty since the room doesn't exist. + self.assertEqual(summary, {}) + + def test_typing(self): + """Typing notifications should not be propagated into the room.""" + # The create works fine. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + request, channel = self.make_request( + "PUT", + "/rooms/%s/typing/%s" % (room_id, self.banned_user_id), + {"typing": True, "timeout": 30000}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code) + + # There should be no typing events. + event_source = self.hs.get_event_sources().sources["typing"] + self.assertEquals(event_source.get_current_key(), 0) + + # The other user can join and send typing events. + self.helper.join(room_id, self.other_user_id, tok=self.other_access_token) + + request, channel = self.make_request( + "PUT", + "/rooms/%s/typing/%s" % (room_id, self.other_user_id), + {"typing": True, "timeout": 30000}, + access_token=self.other_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code) + + # These appear in the room. + self.assertEquals(event_source.get_current_key(), 1) + events = self.get_success( + event_source.get_new_events(from_key=0, room_ids=[room_id]) + ) + self.assertEquals( + events[0], + [ + { + "type": "m.typing", + "room_id": room_id, + "content": {"user_ids": [self.other_user_id]}, + } + ], + ) + + +# To avoid the tests timing out don't add a delay to "annoy the requester". +@patch("random.randint", new=lambda a, b: 0) +class ProfileTestCase(_ShadowBannedBase): + servlets = [ + synapse.rest.admin.register_servlets_for_client_rest_resource, + login.register_servlets, + profile.register_servlets, + room.register_servlets, + ] + + def test_displayname(self): + """Profile changes should succeed, but don't end up in a room.""" + original_display_name = "banned" + new_display_name = "new name" + + # Join a room. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + # The update should succeed. + request, channel = self.make_request( + "PUT", + "/_matrix/client/r0/profile/%s/displayname" % (self.banned_user_id,), + {"displayname": new_display_name}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code, channel.result) + self.assertEqual(channel.json_body, {}) + + # The user's display name should be updated. + request, channel = self.make_request( + "GET", "/profile/%s/displayname" % (self.banned_user_id,) + ) + self.render(request) + self.assertEqual(channel.code, 200, channel.result) + self.assertEqual(channel.json_body["displayname"], new_display_name) + + # But the display name in the room should not be. + message_handler = self.hs.get_message_handler() + event = self.get_success( + message_handler.get_room_data( + self.banned_user_id, room_id, "m.room.member", self.banned_user_id, + ) + ) + self.assertEqual( + event.content, {"membership": "join", "displayname": original_display_name} + ) + + def test_room_displayname(self): + """Changes to state events for a room should be processed, but not end up in the room.""" + original_display_name = "banned" + new_display_name = "new name" + + # Join a room. + room_id = self.helper.create_room_as( + self.banned_user_id, tok=self.banned_access_token + ) + + # The update should succeed. + request, channel = self.make_request( + "PUT", + "/_matrix/client/r0/rooms/%s/state/m.room.member/%s" + % (room_id, self.banned_user_id), + {"membership": "join", "displayname": new_display_name}, + access_token=self.banned_access_token, + ) + self.render(request) + self.assertEquals(200, channel.code, channel.result) + self.assertIn("event_id", channel.json_body) + + # The display name in the room should not be changed. + message_handler = self.hs.get_message_handler() + event = self.get_success( + message_handler.get_room_data( + self.banned_user_id, room_id, "m.room.member", self.banned_user_id, + ) + ) + self.assertEqual( + event.content, {"membership": "join", "displayname": original_display_name} + ) diff --git a/tests/rest/client/test_third_party_rules.py b/tests/rest/client/test_third_party_rules.py deleted file mode 100644
index d03e121664..0000000000 --- a/tests/rest/client/test_third_party_rules.py +++ /dev/null
@@ -1,170 +0,0 @@ -# -*- coding: utf-8 -*- -# Copyright 2019 The Matrix.org Foundation C.I.C. -# -# Licensed under the Apache License, Version 2.0 (the 'License'); -# you may not use this file except in compliance with the License. -# You may obtain a copy of the License at -# -# http://www.apache.org/licenses/LICENSE-2.0 -# -# Unless required by applicable law or agreed to in writing, software -# distributed under the License is distributed on an 'AS IS' BASIS, -# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. -# See the License for the specific language governing permissions and -# limitations under the License. -import threading -from typing import Dict - -from mock import Mock - -from synapse.events import EventBase -from synapse.module_api import ModuleApi -from synapse.rest import admin -from synapse.rest.client.v1 import login, room -from synapse.types import Requester, StateMap - -from tests import unittest - -thread_local = threading.local() - - -class ThirdPartyRulesTestModule: - def __init__(self, config: Dict, module_api: ModuleApi): - # keep a record of the "current" rules module, so that the test can patch - # it if desired. - thread_local.rules_module = self - self.module_api = module_api - - async def on_create_room( - self, requester: Requester, config: dict, is_requester_admin: bool - ): - return True - - async def check_event_allowed(self, event: EventBase, state: StateMap[EventBase]): - return True - - @staticmethod - def parse_config(config): - return config - - -def current_rules_module() -> ThirdPartyRulesTestModule: - return thread_local.rules_module - - -class ThirdPartyRulesTestCase(unittest.HomeserverTestCase): - servlets = [ - admin.register_servlets, - login.register_servlets, - room.register_servlets, - ] - - def default_config(self): - config = super().default_config() - config["third_party_event_rules"] = { - "module": __name__ + ".ThirdPartyRulesTestModule", - "config": {}, - } - return config - - def prepare(self, reactor, clock, homeserver): - # Create a user and room to play with during the tests - self.user_id = self.register_user("kermit", "monkey") - self.tok = self.login("kermit", "monkey") - - self.room_id = self.helper.create_room_as(self.user_id, tok=self.tok) - - def test_third_party_rules(self): - """Tests that a forbidden event is forbidden from being sent, but an allowed one - can be sent. - """ - # patch the rules module with a Mock which will return False for some event - # types - async def check(ev, state): - return ev.type != "foo.bar.forbidden" - - callback = Mock(spec=[], side_effect=check) - current_rules_module().check_event_allowed = callback - - request, channel = self.make_request( - "PUT", - "/_matrix/client/r0/rooms/%s/send/foo.bar.allowed/1" % self.room_id, - {}, - access_token=self.tok, - ) - self.render(request) - self.assertEquals(channel.result["code"], b"200", channel.result) - - callback.assert_called_once() - - # there should be various state events in the state arg: do some basic checks - state_arg = callback.call_args[0][1] - for k in (("m.room.create", ""), ("m.room.member", self.user_id)): - self.assertIn(k, state_arg) - ev = state_arg[k] - self.assertEqual(ev.type, k[0]) - self.assertEqual(ev.state_key, k[1]) - - request, channel = self.make_request( - "PUT", - "/_matrix/client/r0/rooms/%s/send/foo.bar.forbidden/1" % self.room_id, - {}, - access_token=self.tok, - ) - self.render(request) - self.assertEquals(channel.result["code"], b"403", channel.result) - - def test_modify_event(self): - """Tests that the module can successfully tweak an event before it is persisted. - """ - # first patch the event checker so that it will modify the event - async def check(ev: EventBase, state): - ev.content = {"x": "y"} - return True - - current_rules_module().check_event_allowed = check - - # now send the event - request, channel = self.make_request( - "PUT", - "/_matrix/client/r0/rooms/%s/send/modifyme/1" % self.room_id, - {"x": "x"}, - access_token=self.tok, - ) - self.render(request) - self.assertEqual(channel.result["code"], b"200", channel.result) - event_id = channel.json_body["event_id"] - - # ... and check that it got modified - request, channel = self.make_request( - "GET", - "/_matrix/client/r0/rooms/%s/event/%s" % (self.room_id, event_id), - access_token=self.tok, - ) - self.render(request) - self.assertEqual(channel.result["code"], b"200", channel.result) - ev = channel.json_body - self.assertEqual(ev["content"]["x"], "y") - - def test_send_event(self): - """Tests that the module can send an event into a room via the module api""" - content = { - "msgtype": "m.text", - "body": "Hello!", - } - event_dict = { - "room_id": self.room_id, - "type": "m.room.message", - "content": content, - "sender": self.user_id, - } - event = self.get_success( - current_rules_module().module_api.create_and_send_event_into_room( - event_dict - ) - ) # type: EventBase - - self.assertEquals(event.sender, self.user_id) - self.assertEquals(event.room_id, self.room_id) - self.assertEquals(event.type, "m.room.message") - self.assertEquals(event.content, content) diff --git a/tests/rest/client/third_party_rules.py b/tests/rest/client/third_party_rules.py new file mode 100644
index 0000000000..715e87de08 --- /dev/null +++ b/tests/rest/client/third_party_rules.py
@@ -0,0 +1,86 @@ +# -*- coding: utf-8 -*- +# Copyright 2019 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the 'License'); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an 'AS IS' BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +from synapse.rest import admin +from synapse.rest.client.v1 import login, room +from synapse.types import Requester + +from tests import unittest + + +class ThirdPartyRulesTestModule: + def __init__(self, config, *args, **kwargs): + pass + + async def on_create_room( + self, requester: Requester, config: dict, is_requester_admin: bool + ): + return True + + async def check_event_allowed(self, event, context): + if event.type == "foo.bar.forbidden": + return False + else: + return True + + @staticmethod + def parse_config(config): + return config + + +class ThirdPartyRulesTestCase(unittest.HomeserverTestCase): + servlets = [ + admin.register_servlets, + login.register_servlets, + room.register_servlets, + ] + + def make_homeserver(self, reactor, clock): + config = self.default_config() + config["third_party_event_rules"] = { + "module": "tests.rest.client.third_party_rules.ThirdPartyRulesTestModule", + "config": {}, + } + + self.hs = self.setup_test_homeserver(config=config) + return self.hs + + def prepare(self, reactor, clock, homeserver): + # Create a user and room to play with during the tests + self.user_id = self.register_user("kermit", "monkey") + self.tok = self.login("kermit", "monkey") + + self.room_id = self.helper.create_room_as(self.user_id, tok=self.tok) + + def test_third_party_rules(self): + """Tests that a forbidden event is forbidden from being sent, but an allowed one + can be sent. + """ + request, channel = self.make_request( + "PUT", + "/_matrix/client/r0/rooms/%s/send/foo.bar.allowed/1" % self.room_id, + {}, + access_token=self.tok, + ) + self.render(request) + self.assertEquals(channel.result["code"], b"200", channel.result) + + request, channel = self.make_request( + "PUT", + "/_matrix/client/r0/rooms/%s/send/foo.bar.forbidden/1" % self.room_id, + {}, + access_token=self.tok, + ) + self.render(request) + self.assertEquals(channel.result["code"], b"403", channel.result) diff --git a/tests/rest/client/v1/test_login.py b/tests/rest/client/v1/test_login.py
index db52725cfe..5d987a30c7 100644 --- a/tests/rest/client/v1/test_login.py +++ b/tests/rest/client/v1/test_login.py
@@ -7,8 +7,9 @@ from mock import Mock import jwt import synapse.rest.admin +from synapse.appservice import ApplicationService from synapse.rest.client.v1 import login, logout -from synapse.rest.client.v2_alpha import devices +from synapse.rest.client.v2_alpha import devices, register from synapse.rest.client.v2_alpha.account import WhoamiRestServlet from tests import unittest @@ -62,8 +63,7 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): "identifier": {"type": "m.id.user", "user": "kermit" + str(i)}, "password": "monkey", } - request_data = json.dumps(params) - request, channel = self.make_request(b"POST", LOGIN_URL, request_data) + request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) if i == 5: @@ -76,14 +76,13 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): # than 1min. self.assertTrue(retry_after_ms < 6000) - self.reactor.advance(retry_after_ms / 1000.0) + self.reactor.advance(retry_after_ms / 1000.0 + 1.0) params = { "type": "m.login.password", "identifier": {"type": "m.id.user", "user": "kermit" + str(i)}, "password": "monkey", } - request_data = json.dumps(params) request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) @@ -111,8 +110,7 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): "identifier": {"type": "m.id.user", "user": "kermit"}, "password": "monkey", } - request_data = json.dumps(params) - request, channel = self.make_request(b"POST", LOGIN_URL, request_data) + request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) if i == 5: @@ -132,7 +130,6 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): "identifier": {"type": "m.id.user", "user": "kermit"}, "password": "monkey", } - request_data = json.dumps(params) request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) @@ -160,8 +157,7 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): "identifier": {"type": "m.id.user", "user": "kermit"}, "password": "notamonkey", } - request_data = json.dumps(params) - request, channel = self.make_request(b"POST", LOGIN_URL, request_data) + request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) if i == 5: @@ -174,14 +170,13 @@ class LoginRestServletTestCase(unittest.HomeserverTestCase): # than 1min. self.assertTrue(retry_after_ms < 6000) - self.reactor.advance(retry_after_ms / 1000.0) + self.reactor.advance(retry_after_ms / 1000.0 + 1.0) params = { "type": "m.login.password", "identifier": {"type": "m.id.user", "user": "kermit"}, "password": "notamonkey", } - request_data = json.dumps(params) request, channel = self.make_request(b"POST", LOGIN_URL, params) self.render(request) @@ -754,3 +749,134 @@ class JWTPubKeyTestCase(unittest.HomeserverTestCase): channel.json_body["error"], "JWT validation failed: Signature verification failed", ) + + +AS_USER = "as_user_alice" + + +class AppserviceLoginRestServletTestCase(unittest.HomeserverTestCase): + servlets = [ + login.register_servlets, + register.register_servlets, + ] + + def register_as_user(self, username): + request, channel = self.make_request( + b"POST", + "/_matrix/client/r0/register?access_token=%s" % (self.service.token,), + {"username": username}, + ) + self.render(request) + + def make_homeserver(self, reactor, clock): + self.hs = self.setup_test_homeserver() + + self.service = ApplicationService( + id="unique_identifier", + token="some_token", + hostname="example.com", + sender="@asbot:example.com", + namespaces={ + ApplicationService.NS_USERS: [ + {"regex": r"@as_user.*", "exclusive": False} + ], + ApplicationService.NS_ROOMS: [], + ApplicationService.NS_ALIASES: [], + }, + ) + self.another_service = ApplicationService( + id="another__identifier", + token="another_token", + hostname="example.com", + sender="@as2bot:example.com", + namespaces={ + ApplicationService.NS_USERS: [ + {"regex": r"@as2_user.*", "exclusive": False} + ], + ApplicationService.NS_ROOMS: [], + ApplicationService.NS_ALIASES: [], + }, + ) + + self.hs.get_datastore().services_cache.append(self.service) + self.hs.get_datastore().services_cache.append(self.another_service) + return self.hs + + def test_login_appservice_user(self): + """Test that an appservice user can use /login + """ + self.register_as_user(AS_USER) + + params = { + "type": login.LoginRestServlet.APPSERVICE_TYPE, + "identifier": {"type": "m.id.user", "user": AS_USER}, + } + request, channel = self.make_request( + b"POST", LOGIN_URL, params, access_token=self.service.token + ) + + self.render(request) + self.assertEquals(channel.result["code"], b"200", channel.result) + + def test_login_appservice_user_bot(self): + """Test that the appservice bot can use /login + """ + self.register_as_user(AS_USER) + + params = { + "type": login.LoginRestServlet.APPSERVICE_TYPE, + "identifier": {"type": "m.id.user", "user": self.service.sender}, + } + request, channel = self.make_request( + b"POST", LOGIN_URL, params, access_token=self.service.token + ) + + self.render(request) + self.assertEquals(channel.result["code"], b"200", channel.result) + + def test_login_appservice_wrong_user(self): + """Test that non-as users cannot login with the as token + """ + self.register_as_user(AS_USER) + + params = { + "type": login.LoginRestServlet.APPSERVICE_TYPE, + "identifier": {"type": "m.id.user", "user": "fibble_wibble"}, + } + request, channel = self.make_request( + b"POST", LOGIN_URL, params, access_token=self.service.token + ) + + self.render(request) + self.assertEquals(channel.result["code"], b"403", channel.result) + + def test_login_appservice_wrong_as(self): + """Test that as users cannot login with wrong as token + """ + self.register_as_user(AS_USER) + + params = { + "type": login.LoginRestServlet.APPSERVICE_TYPE, + "identifier": {"type": "m.id.user", "user": AS_USER}, + } + request, channel = self.make_request( + b"POST", LOGIN_URL, params, access_token=self.another_service.token + ) + + self.render(request) + self.assertEquals(channel.result["code"], b"403", channel.result) + + def test_login_appservice_no_token(self): + """Test that users must provide a token when using the appservice + login method + """ + self.register_as_user(AS_USER) + + params = { + "type": login.LoginRestServlet.APPSERVICE_TYPE, + "identifier": {"type": "m.id.user", "user": AS_USER}, + } + request, channel = self.make_request(b"POST", LOGIN_URL, params) + + self.render(request) + self.assertEquals(channel.result["code"], b"401", channel.result) diff --git a/tests/rest/client/v1/test_profile.py b/tests/rest/client/v1/test_profile.py
index 8df58b4a63..ace0a3c08d 100644 --- a/tests/rest/client/v1/test_profile.py +++ b/tests/rest/client/v1/test_profile.py
@@ -70,8 +70,8 @@ class MockHandlerProfileTestCase(unittest.TestCase): profile_handler=self.mock_handler, ) - def _get_user_by_req(request=None, allow_guest=False): - return defer.succeed(synapse.types.create_requester(myid)) + async def _get_user_by_req(request=None, allow_guest=False): + return synapse.types.create_requester(myid) hs.get_auth().get_user_by_req = _get_user_by_req diff --git a/tests/rest/client/v1/test_push_rule_attrs.py b/tests/rest/client/v1/test_push_rule_attrs.py new file mode 100644
index 0000000000..081052f6a6 --- /dev/null +++ b/tests/rest/client/v1/test_push_rule_attrs.py
@@ -0,0 +1,448 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import synapse +from synapse.api.errors import Codes +from synapse.rest.client.v1 import login, push_rule, room + +from tests.unittest import HomeserverTestCase + + +class PushRuleAttributesTestCase(HomeserverTestCase): + servlets = [ + synapse.rest.admin.register_servlets_for_client_rest_resource, + room.register_servlets, + login.register_servlets, + push_rule.register_servlets, + ] + hijack_auth = False + + def test_enabled_on_creation(self): + """ + Tests the GET and PUT of push rules' `enabled` endpoints. + Tests that a rule is enabled upon creation, even though a rule with that + ruleId existed previously and was disabled. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # GET enabled for that new rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["enabled"], True) + + def test_enabled_on_recreation(self): + """ + Tests the GET and PUT of push rules' `enabled` endpoints. + Tests that a rule is enabled upon creation, even if a rule with that + ruleId existed previously and was disabled. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # disable the rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/enabled", + {"enabled": False}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # check rule disabled + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["enabled"], False) + + # DELETE the rule + request, channel = self.make_request( + "DELETE", "/pushrules/global/override/best.friend", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # GET enabled for that new rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["enabled"], True) + + def test_enabled_disable(self): + """ + Tests the GET and PUT of push rules' `enabled` endpoints. + Tests that a rule is disabled and enabled when we ask for it. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # disable the rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/enabled", + {"enabled": False}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # check rule disabled + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["enabled"], False) + + # re-enable the rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/enabled", + {"enabled": True}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # check rule enabled + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["enabled"], True) + + def test_enabled_404_when_get_non_existent(self): + """ + Tests that `enabled` gives 404 when the rule doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # check 404 for never-heard-of rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # GET enabled for that new rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # DELETE the rule + request, channel = self.make_request( + "DELETE", "/pushrules/global/override/best.friend", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # check 404 for deleted rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_enabled_404_when_get_non_existent_server_rule(self): + """ + Tests that `enabled` gives 404 when the server-default rule doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # check 404 for never-heard-of rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/.m.muahahaha/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_enabled_404_when_put_non_existent_rule(self): + """ + Tests that `enabled` gives 404 when we put to a rule that doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # enable & check 404 for never-heard-of rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/enabled", + {"enabled": True}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_enabled_404_when_put_non_existent_server_rule(self): + """ + Tests that `enabled` gives 404 when we put to a server-default rule that doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # enable & check 404 for never-heard-of rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/.m.muahahah/enabled", + {"enabled": True}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_actions_get(self): + """ + Tests that `actions` gives you what you expect on a fresh rule. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # GET actions for that new rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/actions", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual( + channel.json_body["actions"], ["notify", {"set_tweak": "highlight"}] + ) + + def test_actions_put(self): + """ + Tests that PUT on actions updates the value you'd get from GET. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # change the rule actions + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/actions", + {"actions": ["dont_notify"]}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # GET actions for that new rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/actions", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + self.assertEqual(channel.json_body["actions"], ["dont_notify"]) + + def test_actions_404_when_get_non_existent(self): + """ + Tests that `actions` gives 404 when the rule doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + body = { + "conditions": [ + {"kind": "event_match", "key": "sender", "pattern": "@user2:hs"} + ], + "actions": ["notify", {"set_tweak": "highlight"}], + } + + # check 404 for never-heard-of rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + # PUT a new rule + request, channel = self.make_request( + "PUT", "/pushrules/global/override/best.friend", body, access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # DELETE the rule + request, channel = self.make_request( + "DELETE", "/pushrules/global/override/best.friend", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 200) + + # check 404 for deleted rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/best.friend/enabled", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_actions_404_when_get_non_existent_server_rule(self): + """ + Tests that `actions` gives 404 when the server-default rule doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # check 404 for never-heard-of rule + request, channel = self.make_request( + "GET", "/pushrules/global/override/.m.muahahaha/actions", access_token=token + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_actions_404_when_put_non_existent_rule(self): + """ + Tests that `actions` gives 404 when putting to a rule that doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # enable & check 404 for never-heard-of rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/best.friend/actions", + {"actions": ["dont_notify"]}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) + + def test_actions_404_when_put_non_existent_server_rule(self): + """ + Tests that `actions` gives 404 when putting to a server-default rule that doesn't exist. + """ + self.register_user("user", "pass") + token = self.login("user", "pass") + + # enable & check 404 for never-heard-of rule + request, channel = self.make_request( + "PUT", + "/pushrules/global/override/.m.muahahah/actions", + {"actions": ["dont_notify"]}, + access_token=token, + ) + self.render(request) + self.assertEqual(channel.code, 404) + self.assertEqual(channel.json_body["errcode"], Codes.NOT_FOUND) diff --git a/tests/rest/client/v1/test_rooms.py b/tests/rest/client/v1/test_rooms.py
index 5ccda8b2bd..0d809d25d5 100644 --- a/tests/rest/client/v1/test_rooms.py +++ b/tests/rest/client/v1/test_rooms.py
@@ -23,14 +23,12 @@ from urllib import parse as urlparse from mock import Mock -from twisted.internet import defer - import synapse.rest.admin from synapse.api.constants import EventContentFields, EventTypes, Membership from synapse.handlers.pagination import PurgeStatus from synapse.rest.client.v1 import directory, login, profile, room from synapse.rest.client.v2_alpha import account -from synapse.types import JsonDict, RoomAlias +from synapse.types import JsonDict, RoomAlias, UserID from synapse.util.stringutils import random_string from tests import unittest @@ -51,8 +49,8 @@ class RoomBase(unittest.HomeserverTestCase): self.hs.get_federation_handler = Mock(return_value=Mock()) - def _insert_client_ip(*args, **kwargs): - return defer.succeed(None) + async def _insert_client_ip(*args, **kwargs): + return None self.hs.get_datastore().insert_client_ip = _insert_client_ip @@ -677,6 +675,92 @@ class RoomMemberStateTestCase(RoomBase): self.assertEquals(json.loads(content), channel.json_body) +class RoomJoinRatelimitTestCase(RoomBase): + user_id = "@sid1:red" + + servlets = [ + profile.register_servlets, + room.register_servlets, + ] + + @unittest.override_config( + {"rc_joins": {"local": {"per_second": 0.5, "burst_count": 3}}} + ) + def test_join_local_ratelimit(self): + """Tests that local joins are actually rate-limited.""" + for i in range(3): + self.helper.create_room_as(self.user_id) + + self.helper.create_room_as(self.user_id, expect_code=429) + + @unittest.override_config( + {"rc_joins": {"local": {"per_second": 0.5, "burst_count": 3}}} + ) + def test_join_local_ratelimit_profile_change(self): + """Tests that sending a profile update into all of the user's joined rooms isn't + rate-limited by the rate-limiter on joins.""" + + # Create and join as many rooms as the rate-limiting config allows in a second. + room_ids = [ + self.helper.create_room_as(self.user_id), + self.helper.create_room_as(self.user_id), + self.helper.create_room_as(self.user_id), + ] + # Let some time for the rate-limiter to forget about our multi-join. + self.reactor.advance(2) + # Add one to make sure we're joined to more rooms than the config allows us to + # join in a second. + room_ids.append(self.helper.create_room_as(self.user_id)) + + # Create a profile for the user, since it hasn't been done on registration. + store = self.hs.get_datastore() + self.get_success( + store.create_profile(UserID.from_string(self.user_id).localpart) + ) + + # Update the display name for the user. + path = "/_matrix/client/r0/profile/%s/displayname" % self.user_id + request, channel = self.make_request("PUT", path, {"displayname": "John Doe"}) + self.render(request) + self.assertEquals(channel.code, 200, channel.json_body) + + # Check that all the rooms have been sent a profile update into. + for room_id in room_ids: + path = "/_matrix/client/r0/rooms/%s/state/m.room.member/%s" % ( + room_id, + self.user_id, + ) + + request, channel = self.make_request("GET", path) + self.render(request) + self.assertEquals(channel.code, 200) + + self.assertIn("displayname", channel.json_body) + self.assertEquals(channel.json_body["displayname"], "John Doe") + + @unittest.override_config( + {"rc_joins": {"local": {"per_second": 0.5, "burst_count": 3}}} + ) + def test_join_local_ratelimit_idempotent(self): + """Tests that the room join endpoints remain idempotent despite rate-limiting + on room joins.""" + room_id = self.helper.create_room_as(self.user_id) + + # Let's test both paths to be sure. + paths_to_test = [ + "/_matrix/client/r0/rooms/%s/join", + "/_matrix/client/r0/join/%s", + ] + + for path in paths_to_test: + # Make sure we send more requests than the rate-limiting config would allow + # if all of these requests ended up joining the user to a room. + for i in range(4): + request, channel = self.make_request("POST", path % room_id, {}) + self.render(request) + self.assertEquals(channel.code, 200) + + class RoomMessagesTestCase(RoomBase): """ Tests /rooms/$room_id/messages/$user_id/$msg_id REST events. """ @@ -821,6 +905,7 @@ class RoomMessageListTestCase(RoomBase): first_token = self.get_success( store.get_topological_token_for_event(first_event_id) ) + first_token_str = self.get_success(first_token.to_string(store)) # Send a second message in the room, which won't be removed, and which we'll # use as the marker to purge events before. @@ -828,6 +913,7 @@ class RoomMessageListTestCase(RoomBase): second_token = self.get_success( store.get_topological_token_for_event(second_event_id) ) + second_token_str = self.get_success(second_token.to_string(store)) # Send a third event in the room to ensure we don't fall under any edge case # due to our marker being the latest forward extremity in the room. @@ -837,7 +923,11 @@ class RoomMessageListTestCase(RoomBase): request, channel = self.make_request( "GET", "/rooms/%s/messages?access_token=x&from=%s&dir=b&filter=%s" - % (self.room_id, second_token, json.dumps({"types": [EventTypes.Message]})), + % ( + self.room_id, + second_token_str, + json.dumps({"types": [EventTypes.Message]}), + ), ) self.render(request) self.assertEqual(channel.code, 200, channel.json_body) @@ -852,7 +942,7 @@ class RoomMessageListTestCase(RoomBase): pagination_handler._purge_history( purge_id=purge_id, room_id=self.room_id, - token=second_token, + token=second_token_str, delete_local_events=True, ) ) @@ -862,7 +952,11 @@ class RoomMessageListTestCase(RoomBase): request, channel = self.make_request( "GET", "/rooms/%s/messages?access_token=x&from=%s&dir=b&filter=%s" - % (self.room_id, second_token, json.dumps({"types": [EventTypes.Message]})), + % ( + self.room_id, + second_token_str, + json.dumps({"types": [EventTypes.Message]}), + ), ) self.render(request) self.assertEqual(channel.code, 200, channel.json_body) @@ -876,7 +970,11 @@ class RoomMessageListTestCase(RoomBase): request, channel = self.make_request( "GET", "/rooms/%s/messages?access_token=x&from=%s&dir=b&filter=%s" - % (self.room_id, first_token, json.dumps({"types": [EventTypes.Message]})), + % ( + self.room_id, + first_token_str, + json.dumps({"types": [EventTypes.Message]}), + ), ) self.render(request) self.assertEqual(channel.code, 200, channel.json_body) diff --git a/tests/rest/client/v1/test_typing.py b/tests/rest/client/v1/test_typing.py
index 18260bb90e..94d2bf2eb1 100644 --- a/tests/rest/client/v1/test_typing.py +++ b/tests/rest/client/v1/test_typing.py
@@ -46,7 +46,7 @@ class RoomTypingTestCase(unittest.HomeserverTestCase): hs.get_handlers().federation_handler = Mock() - def get_user_by_access_token(token=None, allow_guest=False): + async def get_user_by_access_token(token=None, allow_guest=False): return { "user": UserID.from_string(self.auth_user_id), "token_id": 1, @@ -55,8 +55,8 @@ class RoomTypingTestCase(unittest.HomeserverTestCase): hs.get_auth().get_user_by_access_token = get_user_by_access_token - def _insert_client_ip(*args, **kwargs): - return defer.succeed(None) + async def _insert_client_ip(*args, **kwargs): + return None hs.get_datastore().insert_client_ip = _insert_client_ip diff --git a/tests/rest/client/v1/utils.py b/tests/rest/client/v1/utils.py
index 22d734e763..afaf9f7b85 100644 --- a/tests/rest/client/v1/utils.py +++ b/tests/rest/client/v1/utils.py
@@ -30,7 +30,7 @@ from tests.server import make_request, render @attr.s -class RestHelper(object): +class RestHelper: """Contains extra helper functions to quickly and clearly perform a given REST action, which isn't the focus of the test. """ @@ -39,7 +39,9 @@ class RestHelper(object): resource = attr.ib() auth_user_id = attr.ib() - def create_room_as(self, room_creator=None, is_public=True, tok=None): + def create_room_as( + self, room_creator=None, is_public=True, tok=None, expect_code=200, + ): temp_id = self.auth_user_id self.auth_user_id = room_creator path = "/_matrix/client/r0/createRoom" @@ -54,9 +56,11 @@ class RestHelper(object): ) render(request, self.resource, self.hs.get_reactor()) - assert channel.result["code"] == b"200", channel.result + assert channel.result["code"] == b"%d" % expect_code, channel.result self.auth_user_id = temp_id - return channel.json_body["room_id"] + + if expect_code == 200: + return channel.json_body["room_id"] def invite(self, room=None, src=None, targ=None, expect_code=200, tok=None): self.change_membership( @@ -88,7 +92,28 @@ class RestHelper(object): expect_code=expect_code, ) - def change_membership(self, room, src, targ, membership, tok=None, expect_code=200): + def change_membership( + self, + room: str, + src: str, + targ: str, + membership: str, + extra_data: dict = {}, + tok: Optional[str] = None, + expect_code: int = 200, + ) -> None: + """ + Send a membership state event into a room. + + Args: + room: The ID of the room to send to + src: The mxid of the event sender + targ: The mxid of the event's target. The state key + membership: The type of membership event + extra_data: Extra information to include in the content of the event + tok: The user access token to use + expect_code: The expected HTTP response code + """ temp_id = self.auth_user_id self.auth_user_id = src @@ -97,6 +122,7 @@ class RestHelper(object): path = path + "?access_token=%s" % tok data = {"membership": membership} + data.update(extra_data) request, channel = make_request( self.hs.get_reactor(), "PUT", path, json.dumps(data).encode("utf8") diff --git a/tests/rest/client/v2_alpha/test_account.py b/tests/rest/client/v2_alpha/test_account.py
index 0a51aeff92..ae2cd67f35 100644 --- a/tests/rest/client/v2_alpha/test_account.py +++ b/tests/rest/client/v2_alpha/test_account.py
@@ -19,6 +19,7 @@ import os import re from email.parser import Parser from typing import Optional +from urllib.parse import urlencode import pkg_resources @@ -27,6 +28,7 @@ from synapse.api.constants import LoginType, Membership from synapse.api.errors import Codes from synapse.rest.client.v1 import login, room from synapse.rest.client.v2_alpha import account, register +from synapse.rest.synapse.client.password_reset import PasswordResetSubmitTokenResource from tests import unittest from tests.unittest import override_config @@ -70,6 +72,7 @@ class PasswordResetTestCase(unittest.HomeserverTestCase): def prepare(self, reactor, clock, hs): self.store = hs.get_datastore() + self.submit_token_resource = PasswordResetSubmitTokenResource(hs) def test_basic_password_reset(self): """Test basic password reset flow @@ -251,8 +254,32 @@ class PasswordResetTestCase(unittest.HomeserverTestCase): # Remove the host path = link.replace("https://example.com", "") + # Load the password reset confirmation page request, channel = self.make_request("GET", path, shorthand=False) - self.render(request) + request.render(self.submit_token_resource) + self.pump() + self.assertEquals(200, channel.code, channel.result) + + # Now POST to the same endpoint, mimicking the same behaviour as clicking the + # password reset confirm button + + # Send arguments as url-encoded form data, matching the template's behaviour + form_args = [] + for key, value_list in request.args.items(): + for value in value_list: + arg = (key, value) + form_args.append(arg) + + # Confirm the password reset + request, channel = self.make_request( + "POST", + path, + content=urlencode(form_args).encode("utf8"), + shorthand=False, + content_is_form=True, + ) + request.render(self.submit_token_resource) + self.pump() self.assertEquals(200, channel.code, channel.result) def _get_link_from_email(self): @@ -705,6 +732,12 @@ class ThreepidEmailRestTestCase(unittest.HomeserverTestCase): @override_config({"next_link_domain_whitelist": ["example.com", "example.org"]}) def test_next_link_domain_whitelist(self): """Tests next_link parameters must fit the whitelist if provided""" + + # Ensure not providing a next_link parameter still works + self._request_token( + "something@example.com", "some_secret", next_link=None, expect_code=200, + ) + self._request_token( "something@example.com", "some_secret", diff --git a/tests/rest/client/v2_alpha/test_filter.py b/tests/rest/client/v2_alpha/test_filter.py
index e0e9e94fbf..de00350580 100644 --- a/tests/rest/client/v2_alpha/test_filter.py +++ b/tests/rest/client/v2_alpha/test_filter.py
@@ -13,6 +13,8 @@ # See the License for the specific language governing permissions and # limitations under the License. +from twisted.internet import defer + from synapse.api.errors import Codes from synapse.rest.client.v2_alpha import filter @@ -73,8 +75,10 @@ class FilterTestCase(unittest.HomeserverTestCase): self.assertEquals(channel.json_body["errcode"], Codes.FORBIDDEN) def test_get_filter(self): - filter_id = self.filtering.add_user_filter( - user_localpart="apple", user_filter=self.EXAMPLE_FILTER + filter_id = defer.ensureDeferred( + self.filtering.add_user_filter( + user_localpart="apple", user_filter=self.EXAMPLE_FILTER + ) ) self.reactor.advance(1) filter_id = filter_id.result diff --git a/tests/rest/client/v2_alpha/test_register.py b/tests/rest/client/v2_alpha/test_register.py
index ceca4041e1..ecf697e5e0 100644 --- a/tests/rest/client/v2_alpha/test_register.py +++ b/tests/rest/client/v2_alpha/test_register.py
@@ -112,8 +112,8 @@ class RegisterRestServletTestCase(unittest.HomeserverTestCase): self.assertEquals(channel.result["code"], b"200", channel.result) self.assertDictContainsSubset(det_data, channel.json_body) + @override_config({"enable_registration": False}) def test_POST_disabled_registration(self): - self.hs.config.enable_registration = False request_data = json.dumps({"username": "kermit", "password": "monkey"}) self.auth_result = (None, {"username": "kermit", "password": "monkey"}, None) @@ -156,7 +156,7 @@ class RegisterRestServletTestCase(unittest.HomeserverTestCase): else: self.assertEquals(channel.result["code"], b"200", channel.result) - self.reactor.advance(retry_after_ms / 1000.0) + self.reactor.advance(retry_after_ms / 1000.0 + 1.0) request, channel = self.make_request(b"POST", self.url + b"?kind=guest", b"{}") self.render(request) @@ -182,7 +182,7 @@ class RegisterRestServletTestCase(unittest.HomeserverTestCase): else: self.assertEquals(channel.result["code"], b"200", channel.result) - self.reactor.advance(retry_after_ms / 1000.0) + self.reactor.advance(retry_after_ms / 1000.0 + 1.0) request, channel = self.make_request(b"POST", self.url + b"?kind=guest", b"{}") self.render(request) diff --git a/tests/rest/client/v2_alpha/test_shared_rooms.py b/tests/rest/client/v2_alpha/test_shared_rooms.py new file mode 100644
index 0000000000..5ae72fd008 --- /dev/null +++ b/tests/rest/client/v2_alpha/test_shared_rooms.py
@@ -0,0 +1,138 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 Half-Shot +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +import synapse.rest.admin +from synapse.rest.client.v1 import login, room +from synapse.rest.client.v2_alpha import shared_rooms + +from tests import unittest + + +class UserSharedRoomsTest(unittest.HomeserverTestCase): + """ + Tests the UserSharedRoomsServlet. + """ + + servlets = [ + login.register_servlets, + synapse.rest.admin.register_servlets_for_client_rest_resource, + room.register_servlets, + shared_rooms.register_servlets, + ] + + def make_homeserver(self, reactor, clock): + config = self.default_config() + config["update_user_directory"] = True + return self.setup_test_homeserver(config=config) + + def prepare(self, reactor, clock, hs): + self.store = hs.get_datastore() + self.handler = hs.get_user_directory_handler() + + def _get_shared_rooms(self, token, other_user): + request, channel = self.make_request( + "GET", + "/_matrix/client/unstable/uk.half-shot.msc2666/user/shared_rooms/%s" + % other_user, + access_token=token, + ) + self.render(request) + return request, channel + + def test_shared_room_list_public(self): + """ + A room should show up in the shared list of rooms between two users + if it is public. + """ + u1 = self.register_user("user1", "pass") + u1_token = self.login(u1, "pass") + u2 = self.register_user("user2", "pass") + u2_token = self.login(u2, "pass") + + room = self.helper.create_room_as(u1, is_public=True, tok=u1_token) + self.helper.invite(room, src=u1, targ=u2, tok=u1_token) + self.helper.join(room, user=u2, tok=u2_token) + + request, channel = self._get_shared_rooms(u1_token, u2) + self.assertEquals(200, channel.code, channel.result) + self.assertEquals(len(channel.json_body["joined"]), 1) + self.assertEquals(channel.json_body["joined"][0], room) + + def test_shared_room_list_private(self): + """ + A room should show up in the shared list of rooms between two users + if it is private. + """ + u1 = self.register_user("user1", "pass") + u1_token = self.login(u1, "pass") + u2 = self.register_user("user2", "pass") + u2_token = self.login(u2, "pass") + + room = self.helper.create_room_as(u1, is_public=False, tok=u1_token) + self.helper.invite(room, src=u1, targ=u2, tok=u1_token) + self.helper.join(room, user=u2, tok=u2_token) + + request, channel = self._get_shared_rooms(u1_token, u2) + self.assertEquals(200, channel.code, channel.result) + self.assertEquals(len(channel.json_body["joined"]), 1) + self.assertEquals(channel.json_body["joined"][0], room) + + def test_shared_room_list_mixed(self): + """ + The shared room list between two users should contain both public and private + rooms. + """ + u1 = self.register_user("user1", "pass") + u1_token = self.login(u1, "pass") + u2 = self.register_user("user2", "pass") + u2_token = self.login(u2, "pass") + + room_public = self.helper.create_room_as(u1, is_public=True, tok=u1_token) + room_private = self.helper.create_room_as(u2, is_public=False, tok=u2_token) + self.helper.invite(room_public, src=u1, targ=u2, tok=u1_token) + self.helper.invite(room_private, src=u2, targ=u1, tok=u2_token) + self.helper.join(room_public, user=u2, tok=u2_token) + self.helper.join(room_private, user=u1, tok=u1_token) + + request, channel = self._get_shared_rooms(u1_token, u2) + self.assertEquals(200, channel.code, channel.result) + self.assertEquals(len(channel.json_body["joined"]), 2) + self.assertTrue(room_public in channel.json_body["joined"]) + self.assertTrue(room_private in channel.json_body["joined"]) + + def test_shared_room_list_after_leave(self): + """ + A room should no longer be considered shared if the other + user has left it. + """ + u1 = self.register_user("user1", "pass") + u1_token = self.login(u1, "pass") + u2 = self.register_user("user2", "pass") + u2_token = self.login(u2, "pass") + + room = self.helper.create_room_as(u1, is_public=True, tok=u1_token) + self.helper.invite(room, src=u1, targ=u2, tok=u1_token) + self.helper.join(room, user=u2, tok=u2_token) + + # Assert user directory is not empty + request, channel = self._get_shared_rooms(u1_token, u2) + self.assertEquals(200, channel.code, channel.result) + self.assertEquals(len(channel.json_body["joined"]), 1) + self.assertEquals(channel.json_body["joined"][0], room) + + self.helper.leave(room, user=u1, tok=u1_token) + + request, channel = self._get_shared_rooms(u2_token, u1) + self.assertEquals(200, channel.code, channel.result) + self.assertEquals(len(channel.json_body["joined"]), 0) diff --git a/tests/rest/client/v2_alpha/test_sync.py b/tests/rest/client/v2_alpha/test_sync.py
index fa3a3ec1bd..a31e44c97e 100644 --- a/tests/rest/client/v2_alpha/test_sync.py +++ b/tests/rest/client/v2_alpha/test_sync.py
@@ -16,9 +16,9 @@ import json import synapse.rest.admin -from synapse.api.constants import EventContentFields, EventTypes +from synapse.api.constants import EventContentFields, EventTypes, RelationTypes from synapse.rest.client.v1 import login, room -from synapse.rest.client.v2_alpha import sync +from synapse.rest.client.v2_alpha import read_marker, sync from tests import unittest from tests.server import TimedOutException @@ -324,3 +324,156 @@ class SyncTypingTests(unittest.HomeserverTestCase): "GET", sync_url % (access_token, next_batch) ) self.assertRaises(TimedOutException, self.render, request) + + +class UnreadMessagesTestCase(unittest.HomeserverTestCase): + servlets = [ + synapse.rest.admin.register_servlets, + login.register_servlets, + read_marker.register_servlets, + room.register_servlets, + sync.register_servlets, + ] + + def prepare(self, reactor, clock, hs): + self.url = "/sync?since=%s" + self.next_batch = "s0" + + # Register the first user (used to check the unread counts). + self.user_id = self.register_user("kermit", "monkey") + self.tok = self.login("kermit", "monkey") + + # Create the room we'll check unread counts for. + self.room_id = self.helper.create_room_as(self.user_id, tok=self.tok) + + # Register the second user (used to send events to the room). + self.user2 = self.register_user("kermit2", "monkey") + self.tok2 = self.login("kermit2", "monkey") + + # Change the power levels of the room so that the second user can send state + # events. + self.helper.send_state( + self.room_id, + EventTypes.PowerLevels, + { + "users": {self.user_id: 100, self.user2: 100}, + "users_default": 0, + "events": { + "m.room.name": 50, + "m.room.power_levels": 100, + "m.room.history_visibility": 100, + "m.room.canonical_alias": 50, + "m.room.avatar": 50, + "m.room.tombstone": 100, + "m.room.server_acl": 100, + "m.room.encryption": 100, + }, + "events_default": 0, + "state_default": 50, + "ban": 50, + "kick": 50, + "redact": 50, + "invite": 0, + }, + tok=self.tok, + ) + + def test_unread_counts(self): + """Tests that /sync returns the right value for the unread count (MSC2654).""" + + # Check that our own messages don't increase the unread count. + self.helper.send(self.room_id, "hello", tok=self.tok) + self._check_unread_count(0) + + # Join the new user and check that this doesn't increase the unread count. + self.helper.join(room=self.room_id, user=self.user2, tok=self.tok2) + self._check_unread_count(0) + + # Check that the new user sending a message increases our unread count. + res = self.helper.send(self.room_id, "hello", tok=self.tok2) + self._check_unread_count(1) + + # Send a read receipt to tell the server we've read the latest event. + body = json.dumps({"m.read": res["event_id"]}).encode("utf8") + request, channel = self.make_request( + "POST", + "/rooms/%s/read_markers" % self.room_id, + body, + access_token=self.tok, + ) + self.render(request) + self.assertEqual(channel.code, 200, channel.json_body) + + # Check that the unread counter is back to 0. + self._check_unread_count(0) + + # Check that room name changes increase the unread counter. + self.helper.send_state( + self.room_id, "m.room.name", {"name": "my super room"}, tok=self.tok2, + ) + self._check_unread_count(1) + + # Check that room topic changes increase the unread counter. + self.helper.send_state( + self.room_id, "m.room.topic", {"topic": "welcome!!!"}, tok=self.tok2, + ) + self._check_unread_count(2) + + # Check that encrypted messages increase the unread counter. + self.helper.send_event(self.room_id, EventTypes.Encrypted, {}, tok=self.tok2) + self._check_unread_count(3) + + # Check that custom events with a body increase the unread counter. + self.helper.send_event( + self.room_id, "org.matrix.custom_type", {"body": "hello"}, tok=self.tok2, + ) + self._check_unread_count(4) + + # Check that edits don't increase the unread counter. + self.helper.send_event( + room_id=self.room_id, + type=EventTypes.Message, + content={ + "body": "hello", + "msgtype": "m.text", + "m.relates_to": {"rel_type": RelationTypes.REPLACE}, + }, + tok=self.tok2, + ) + self._check_unread_count(4) + + # Check that notices don't increase the unread counter. + self.helper.send_event( + room_id=self.room_id, + type=EventTypes.Message, + content={"body": "hello", "msgtype": "m.notice"}, + tok=self.tok2, + ) + self._check_unread_count(4) + + # Check that tombstone events changes increase the unread counter. + self.helper.send_state( + self.room_id, + EventTypes.Tombstone, + {"replacement_room": "!someroom:test"}, + tok=self.tok2, + ) + self._check_unread_count(5) + + def _check_unread_count(self, expected_count: True): + """Syncs and compares the unread count with the expected value.""" + + request, channel = self.make_request( + "GET", self.url % self.next_batch, access_token=self.tok, + ) + self.render(request) + + self.assertEqual(channel.code, 200, channel.json_body) + + room_entry = channel.json_body["rooms"]["join"][self.room_id] + self.assertEqual( + room_entry["org.matrix.msc2654.unread_count"], expected_count, room_entry, + ) + + # Store the next batch for the next request. + self.next_batch = channel.json_body["next_batch"] diff --git a/tests/rest/key/v2/test_remote_key_resource.py b/tests/rest/key/v2/test_remote_key_resource.py
index 99eb477149..6850c666be 100644 --- a/tests/rest/key/v2/test_remote_key_resource.py +++ b/tests/rest/key/v2/test_remote_key_resource.py
@@ -53,7 +53,7 @@ class BaseRemoteKeyResourceTestCase(unittest.HomeserverTestCase): Tell the mock http client to expect an outgoing GET request for the given key """ - def get_json(destination, path, ignore_backoff=False, **kwargs): + async def get_json(destination, path, ignore_backoff=False, **kwargs): self.assertTrue(ignore_backoff) self.assertEqual(destination, server_name) key_id = "%s:%s" % (signing_key.alg, signing_key.version) @@ -177,7 +177,7 @@ class EndToEndPerspectivesTests(BaseRemoteKeyResourceTestCase): # wire up outbound POST /key/v2/query requests from hs2 so that they # will be forwarded to hs1 - def post_json(destination, path, data): + async def post_json(destination, path, data): self.assertEqual(destination, self.hs.hostname) self.assertEqual( path, "/_matrix/key/v2/query", diff --git a/tests/rest/media/v1/test_media_storage.py b/tests/rest/media/v1/test_media_storage.py
index f4f3e56777..5f897d49cf 100644 --- a/tests/rest/media/v1/test_media_storage.py +++ b/tests/rest/media/v1/test_media_storage.py
@@ -120,12 +120,13 @@ class _TestImage: extension = attr.ib(type=bytes) expected_cropped = attr.ib(type=Optional[bytes]) expected_scaled = attr.ib(type=Optional[bytes]) + expected_found = attr.ib(default=True, type=bool) @parameterized_class( ("test_image",), [ - # smol png + # smoll png ( _TestImage( unhexlify( @@ -161,6 +162,8 @@ class _TestImage: None, ), ), + # an empty file + (_TestImage(b"", b"image/gif", b".gif", None, None, False,),), ], ) class MediaRepoTests(unittest.HomeserverTestCase): @@ -303,12 +306,16 @@ class MediaRepoTests(unittest.HomeserverTestCase): self.assertEqual(headers.getRawHeaders(b"Content-Disposition"), None) def test_thumbnail_crop(self): - self._test_thumbnail("crop", self.test_image.expected_cropped) + self._test_thumbnail( + "crop", self.test_image.expected_cropped, self.test_image.expected_found + ) def test_thumbnail_scale(self): - self._test_thumbnail("scale", self.test_image.expected_scaled) + self._test_thumbnail( + "scale", self.test_image.expected_scaled, self.test_image.expected_found + ) - def _test_thumbnail(self, method, expected_body): + def _test_thumbnail(self, method, expected_body, expected_found): params = "?width=32&height=32&method=" + method request, channel = self.make_request( "GET", self.media_id + params, shorthand=False @@ -325,11 +332,23 @@ class MediaRepoTests(unittest.HomeserverTestCase): ) self.pump() - self.assertEqual(channel.code, 200) - if expected_body is not None: + if expected_found: + self.assertEqual(channel.code, 200) + if expected_body is not None: + self.assertEqual( + channel.result["body"], expected_body, channel.result["body"] + ) + else: + # ensure that the result is at least some valid image + Image.open(BytesIO(channel.result["body"])) + else: + # A 404 with a JSON body. + self.assertEqual(channel.code, 404) self.assertEqual( - channel.result["body"], expected_body, channel.result["body"] + channel.json_body, + { + "errcode": "M_NOT_FOUND", + "error": "Not found [b'example.com', b'12345?width=32&height=32&method=%s']" + % method, + }, ) - else: - # ensure that the result is at least some valid image - Image.open(BytesIO(channel.result["body"])) diff --git a/tests/rest/media/v1/test_url_preview.py b/tests/rest/media/v1/test_url_preview.py
index 74765a582b..c00a7b9114 100644 --- a/tests/rest/media/v1/test_url_preview.py +++ b/tests/rest/media/v1/test_url_preview.py
@@ -32,7 +32,7 @@ from tests.server import FakeTransport @attr.s -class FakeResponse(object): +class FakeResponse: version = attr.ib() code = attr.ib() phrase = attr.ib() @@ -43,7 +43,7 @@ class FakeResponse(object): @property def request(self): @attr.s - class FakeTransport(object): + class FakeTransport: absoluteURI = self.absoluteURI return FakeTransport() @@ -111,7 +111,7 @@ class URLPreviewTests(unittest.HomeserverTestCase): self.lookups = {} - class Resolver(object): + class Resolver: def resolveHostName( _self, resolutionReceiver, diff --git a/tests/rest/test_health.py b/tests/rest/test_health.py new file mode 100644
index 0000000000..2d021f6565 --- /dev/null +++ b/tests/rest/test_health.py
@@ -0,0 +1,34 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + + +from synapse.rest.health import HealthResource + +from tests import unittest + + +class HealthCheckTests(unittest.HomeserverTestCase): + def setUp(self): + super().setUp() + + # replace the JsonResource with a HealthResource. + self.resource = HealthResource() + + def test_health(self): + request, channel = self.make_request("GET", "/health", shorthand=False) + self.render(request) + + self.assertEqual(request.code, 200) + self.assertEqual(channel.result["body"], b"OK") diff --git a/tests/rest/test_well_known.py b/tests/rest/test_well_known.py
index b090bb974c..dcd65c2a50 100644 --- a/tests/rest/test_well_known.py +++ b/tests/rest/test_well_known.py
@@ -21,7 +21,7 @@ from tests import unittest class WellKnownTests(unittest.HomeserverTestCase): def setUp(self): - super(WellKnownTests, self).setUp() + super().setUp() # replace the JsonResource with a WellKnownResource self.resource = WellKnownResource(self.hs) diff --git a/tests/server.py b/tests/server.py
index b6e0b14e78..b404ad4e2a 100644 --- a/tests/server.py +++ b/tests/server.py
@@ -1,6 +1,6 @@ import json import logging -from io import BytesIO +from io import SEEK_END, BytesIO import attr from zope.interface import implementer @@ -35,7 +35,7 @@ class TimedOutException(Exception): @attr.s -class FakeChannel(object): +class FakeChannel: """ A fake Twisted Web Channel (the part that interfaces with the wire). @@ -135,6 +135,7 @@ def make_request( request=SynapseRequest, shorthand=True, federation_auth_origin=None, + content_is_form=False, ): """ Make a web request using the given method and path, feed it the @@ -150,6 +151,8 @@ def make_request( with the usual REST API path, if it doesn't contain it. federation_auth_origin (bytes|None): if set to not-None, we will add a fake Authorization header pretenting to be the given server name. + content_is_form: Whether the content is URL encoded form data. Adds the + 'Content-Type': 'application/x-www-form-urlencoded' header. Returns: Tuple[synapse.http.site.SynapseRequest, channel] @@ -181,6 +184,8 @@ def make_request( req = request(channel) req.process = lambda: b"" req.content = BytesIO(content) + # Twisted expects to be at the end of the content when parsing the request. + req.content.seek(SEEK_END) req.postpath = list(map(unquote, path[1:].split(b"/"))) if access_token: @@ -195,7 +200,13 @@ def make_request( ) if content: - req.requestHeaders.addRawHeader(b"Content-Type", b"application/json") + if content_is_form: + req.requestHeaders.addRawHeader( + b"Content-Type", b"application/x-www-form-urlencoded" + ) + else: + # Assume the body is JSON + req.requestHeaders.addRawHeader(b"Content-Type", b"application/json") req.requestReceived(method, path, b"1.1") @@ -242,14 +253,14 @@ class ThreadedMemoryReactorClock(MemoryReactorClock): lookups = self.lookups = {} @implementer(IResolverSimple) - class FakeResolver(object): + class FakeResolver: def getHostByName(self, name, timeout=None): if name not in lookups: return fail(DNSLookupError("OH NO: unknown %s" % (name,))) return succeed(lookups[name]) self.nameResolver = SimpleResolverComplexifier(FakeResolver()) - super(ThreadedMemoryReactorClock, self).__init__() + super().__init__() def listenUDP(self, port, protocol, interface="", maxPacketSize=8196): p = udp.Port(port, protocol, interface, maxPacketSize, self) @@ -371,7 +382,7 @@ def get_clock(): @attr.s(cmp=False) -class FakeTransport(object): +class FakeTransport: """ A twisted.internet.interfaces.ITransport implementation which sends all its data straight into an IProtocol object: it exists to connect two IProtocols together. diff --git a/tests/server_notices/test_resource_limits_server_notices.py b/tests/server_notices/test_resource_limits_server_notices.py
index 99908edba3..6382b19dc3 100644 --- a/tests/server_notices/test_resource_limits_server_notices.py +++ b/tests/server_notices/test_resource_limits_server_notices.py
@@ -27,6 +27,7 @@ from synapse.server_notices.resource_limits_server_notices import ( ) from tests import unittest +from tests.test_utils import make_awaitable from tests.unittest import override_config from tests.utils import default_config @@ -66,7 +67,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): raise Exception("Failed to find reference to ResourceLimitsServerNotices") self._rlsn._store.user_last_seen_monthly_active = Mock( - return_value=defer.succeed(1000) + return_value=make_awaitable(1000) ) self._rlsn._server_notices_manager.send_notice = Mock( return_value=defer.succeed(Mock()) @@ -79,7 +80,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): return_value=defer.succeed("!something:localhost") ) self._rlsn._store.add_tag_to_room = Mock(return_value=defer.succeed(None)) - self._rlsn._store.get_tags_for_room = Mock(return_value=defer.succeed({})) + self._rlsn._store.get_tags_for_room = Mock(return_value=make_awaitable({})) @override_config({"hs_disabled": True}) def test_maybe_send_server_notice_disabled_hs(self): @@ -101,7 +102,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): type=EventTypes.Message, content={"msgtype": ServerNoticeMsgType} ) self._rlsn._store.get_events = Mock( - return_value=defer.succeed({"123": mock_event}) + return_value=make_awaitable({"123": mock_event}) ) self.get_success(self._rlsn.maybe_send_server_notice_to_user(self.user_id)) # Would be better to check the content, but once == remove blocking event @@ -119,7 +120,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): type=EventTypes.Message, content={"msgtype": ServerNoticeMsgType} ) self._rlsn._store.get_events = Mock( - return_value=defer.succeed({"123": mock_event}) + return_value=make_awaitable({"123": mock_event}) ) self.get_success(self._rlsn.maybe_send_server_notice_to_user(self.user_id)) @@ -155,7 +156,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): """ self._rlsn._auth.check_auth_blocking = Mock(return_value=defer.succeed(None)) self._rlsn._store.user_last_seen_monthly_active = Mock( - return_value=defer.succeed(None) + return_value=make_awaitable(None) ) self.get_success(self._rlsn.maybe_send_server_notice_to_user(self.user_id)) @@ -214,7 +215,7 @@ class TestResourceLimitsServerNotices(unittest.HomeserverTestCase): type=EventTypes.Message, content={"msgtype": ServerNoticeMsgType} ) self._rlsn._store.get_events = Mock( - return_value=defer.succeed({"123": mock_event}) + return_value=make_awaitable({"123": mock_event}) ) self.get_success(self._rlsn.maybe_send_server_notice_to_user(self.user_id)) @@ -258,10 +259,10 @@ class TestResourceLimitsServerNoticesWithRealRooms(unittest.HomeserverTestCase): self.user_id = "@user_id:test" def test_server_notice_only_sent_once(self): - self.store.get_monthly_active_count = Mock(return_value=1000) + self.store.get_monthly_active_count = Mock(return_value=make_awaitable(1000)) self.store.user_last_seen_monthly_active = Mock( - return_value=defer.succeed(1000) + return_value=make_awaitable(1000) ) # Call the function multiple times to ensure we only send the notice once @@ -275,7 +276,7 @@ class TestResourceLimitsServerNoticesWithRealRooms(unittest.HomeserverTestCase): self.server_notices_manager.get_or_create_notice_room_for_user(self.user_id) ) - token = self.get_success(self.event_source.get_current_token()) + token = self.event_source.get_current_token() events, _ = self.get_success( self.store.get_recent_events_for_room( room_id, limit=100, end_token=token.room_key diff --git a/tests/state/test_v2.py b/tests/state/test_v2.py
index f2955a9c69..ad9bbef9d2 100644 --- a/tests/state/test_v2.py +++ b/tests/state/test_v2.py
@@ -49,7 +49,7 @@ class FakeClock: return defer.succeed(None) -class FakeEvent(object): +class FakeEvent: """A fake event we use as a convenience. NOTE: Again as a convenience we use "node_ids" rather than event_ids to @@ -595,7 +595,7 @@ def pairwise(iterable): @attr.s -class TestStateResolutionStore(object): +class TestStateResolutionStore: event_map = attr.ib() def get_events(self, event_ids, allow_rejected=False): diff --git a/tests/storage/test__base.py b/tests/storage/test__base.py
index 5a50e4fdd4..f5afed017c 100644 --- a/tests/storage/test__base.py +++ b/tests/storage/test__base.py
@@ -99,7 +99,7 @@ class CacheTestCase(unittest.HomeserverTestCase): class CacheDecoratorTestCase(unittest.HomeserverTestCase): @defer.inlineCallbacks def test_passthrough(self): - class A(object): + class A: @cached() def func(self, key): return key @@ -113,7 +113,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): def test_hit(self): callcount = [0] - class A(object): + class A: @cached() def func(self, key): callcount[0] += 1 @@ -131,7 +131,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): def test_invalidate(self): callcount = [0] - class A(object): + class A: @cached() def func(self, key): callcount[0] += 1 @@ -149,7 +149,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): self.assertEquals(callcount[0], 2) def test_invalidate_missing(self): - class A(object): + class A: @cached() def func(self, key): return key @@ -160,7 +160,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): def test_max_entries(self): callcount = [0] - class A(object): + class A: @cached(max_entries=10) def func(self, key): callcount[0] += 1 @@ -187,7 +187,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): d = defer.succeed(123) - class A(object): + class A: @cached() def func(self, key): callcount[0] += 1 @@ -205,7 +205,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): callcount = [0] callcount2 = [0] - class A(object): + class A: @cached() def func(self, key): callcount[0] += 1 @@ -238,7 +238,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): callcount = [0] callcount2 = [0] - class A(object): + class A: @cached(max_entries=2) def func(self, key): callcount[0] += 1 @@ -275,7 +275,7 @@ class CacheDecoratorTestCase(unittest.HomeserverTestCase): callcount = [0] callcount2 = [0] - class A(object): + class A: @cached() def func(self, key): callcount[0] += 1 @@ -323,7 +323,7 @@ class UpsertManyTests(unittest.HomeserverTestCase): self.table_name = "table_" + hs.get_secrets().token_hex(6) self.get_success( - self.storage.db.runInteraction( + self.storage.db_pool.runInteraction( "create", lambda x, *a: x.execute(*a), "CREATE TABLE %s (id INTEGER, username TEXT, value TEXT)" @@ -331,7 +331,7 @@ class UpsertManyTests(unittest.HomeserverTestCase): ) ) self.get_success( - self.storage.db.runInteraction( + self.storage.db_pool.runInteraction( "index", lambda x, *a: x.execute(*a), "CREATE UNIQUE INDEX %sindex ON %s(id, username)" @@ -354,9 +354,9 @@ class UpsertManyTests(unittest.HomeserverTestCase): value_values = [["hello"], ["there"]] self.get_success( - self.storage.db.runInteraction( + self.storage.db_pool.runInteraction( "test", - self.storage.db.simple_upsert_many_txn, + self.storage.db_pool.simple_upsert_many_txn, self.table_name, key_names, key_values, @@ -367,7 +367,7 @@ class UpsertManyTests(unittest.HomeserverTestCase): # Check results are what we expect res = self.get_success( - self.storage.db.simple_select_list( + self.storage.db_pool.simple_select_list( self.table_name, None, ["id, username, value"] ) ) @@ -381,9 +381,9 @@ class UpsertManyTests(unittest.HomeserverTestCase): value_values = [["bleb"]] self.get_success( - self.storage.db.runInteraction( + self.storage.db_pool.runInteraction( "test", - self.storage.db.simple_upsert_many_txn, + self.storage.db_pool.simple_upsert_many_txn, self.table_name, key_names, key_values, @@ -394,7 +394,7 @@ class UpsertManyTests(unittest.HomeserverTestCase): # Check results are what we expect res = self.get_success( - self.storage.db.simple_select_list( + self.storage.db_pool.simple_select_list( self.table_name, None, ["id, username, value"] ) ) diff --git a/tests/storage/test_appservice.py b/tests/storage/test_appservice.py
index ef296e7dab..46f94914ff 100644 --- a/tests/storage/test_appservice.py +++ b/tests/storage/test_appservice.py
@@ -24,13 +24,14 @@ from twisted.internet import defer from synapse.appservice import ApplicationService, ApplicationServiceState from synapse.config._base import ConfigError -from synapse.storage.data_stores.main.appservice import ( +from synapse.storage.database import DatabasePool, make_conn +from synapse.storage.databases.main.appservice import ( ApplicationServiceStore, ApplicationServiceTransactionStore, ) -from synapse.storage.database import Database, make_conn from tests import unittest +from tests.test_utils import make_awaitable from tests.utils import setup_test_homeserver @@ -178,14 +179,14 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_get_appservice_state_none(self): service = Mock(id="999") - state = yield self.store.get_appservice_state(service) + state = yield defer.ensureDeferred(self.store.get_appservice_state(service)) self.assertEquals(None, state) @defer.inlineCallbacks def test_get_appservice_state_up(self): yield self._set_state(self.as_list[0]["id"], ApplicationServiceState.UP) service = Mock(id=self.as_list[0]["id"]) - state = yield self.store.get_appservice_state(service) + state = yield defer.ensureDeferred(self.store.get_appservice_state(service)) self.assertEquals(ApplicationServiceState.UP, state) @defer.inlineCallbacks @@ -194,20 +195,22 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): yield self._set_state(self.as_list[1]["id"], ApplicationServiceState.DOWN) yield self._set_state(self.as_list[2]["id"], ApplicationServiceState.DOWN) service = Mock(id=self.as_list[1]["id"]) - state = yield self.store.get_appservice_state(service) + state = yield defer.ensureDeferred(self.store.get_appservice_state(service)) self.assertEquals(ApplicationServiceState.DOWN, state) @defer.inlineCallbacks def test_get_appservices_by_state_none(self): - services = yield self.store.get_appservices_by_state( - ApplicationServiceState.DOWN + services = yield defer.ensureDeferred( + self.store.get_appservices_by_state(ApplicationServiceState.DOWN) ) self.assertEquals(0, len(services)) @defer.inlineCallbacks def test_set_appservices_state_down(self): service = Mock(id=self.as_list[1]["id"]) - yield self.store.set_appservice_state(service, ApplicationServiceState.DOWN) + yield defer.ensureDeferred( + self.store.set_appservice_state(service, ApplicationServiceState.DOWN) + ) rows = yield self.db_pool.runQuery( self.engine.convert_param_style( "SELECT as_id FROM application_services_state WHERE state=?" @@ -219,9 +222,15 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_set_appservices_state_multiple_up(self): service = Mock(id=self.as_list[1]["id"]) - yield self.store.set_appservice_state(service, ApplicationServiceState.UP) - yield self.store.set_appservice_state(service, ApplicationServiceState.DOWN) - yield self.store.set_appservice_state(service, ApplicationServiceState.UP) + yield defer.ensureDeferred( + self.store.set_appservice_state(service, ApplicationServiceState.UP) + ) + yield defer.ensureDeferred( + self.store.set_appservice_state(service, ApplicationServiceState.DOWN) + ) + yield defer.ensureDeferred( + self.store.set_appservice_state(service, ApplicationServiceState.UP) + ) rows = yield self.db_pool.runQuery( self.engine.convert_param_style( "SELECT as_id FROM application_services_state WHERE state=?" @@ -234,7 +243,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): def test_create_appservice_txn_first(self): service = Mock(id=self.as_list[0]["id"]) events = [Mock(event_id="e1"), Mock(event_id="e2")] - txn = yield self.store.create_appservice_txn(service, events) + txn = yield defer.ensureDeferred( + self.store.create_appservice_txn(service, events) + ) self.assertEquals(txn.id, 1) self.assertEquals(txn.events, events) self.assertEquals(txn.service, service) @@ -246,7 +257,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): yield self._set_last_txn(service.id, 9643) # AS is falling behind yield self._insert_txn(service.id, 9644, events) yield self._insert_txn(service.id, 9645, events) - txn = yield self.store.create_appservice_txn(service, events) + txn = yield defer.ensureDeferred( + self.store.create_appservice_txn(service, events) + ) self.assertEquals(txn.id, 9646) self.assertEquals(txn.events, events) self.assertEquals(txn.service, service) @@ -256,7 +269,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): service = Mock(id=self.as_list[0]["id"]) events = [Mock(event_id="e1"), Mock(event_id="e2")] yield self._set_last_txn(service.id, 9643) - txn = yield self.store.create_appservice_txn(service, events) + txn = yield defer.ensureDeferred( + self.store.create_appservice_txn(service, events) + ) self.assertEquals(txn.id, 9644) self.assertEquals(txn.events, events) self.assertEquals(txn.service, service) @@ -277,7 +292,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): yield self._insert_txn(self.as_list[2]["id"], 10, events) yield self._insert_txn(self.as_list[3]["id"], 9643, events) - txn = yield self.store.create_appservice_txn(service, events) + txn = yield defer.ensureDeferred( + self.store.create_appservice_txn(service, events) + ) self.assertEquals(txn.id, 9644) self.assertEquals(txn.events, events) self.assertEquals(txn.service, service) @@ -289,7 +306,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): txn_id = 1 yield self._insert_txn(service.id, txn_id, events) - yield self.store.complete_appservice_txn(txn_id=txn_id, service=service) + yield defer.ensureDeferred( + self.store.complete_appservice_txn(txn_id=txn_id, service=service) + ) res = yield self.db_pool.runQuery( self.engine.convert_param_style( @@ -315,7 +334,9 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): txn_id = 5 yield self._set_last_txn(service.id, 4) yield self._insert_txn(service.id, txn_id, events) - yield self.store.complete_appservice_txn(txn_id=txn_id, service=service) + yield defer.ensureDeferred( + self.store.complete_appservice_txn(txn_id=txn_id, service=service) + ) res = yield self.db_pool.runQuery( self.engine.convert_param_style( @@ -339,7 +360,7 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): def test_get_oldest_unsent_txn_none(self): service = Mock(id=self.as_list[0]["id"]) - txn = yield self.store.get_oldest_unsent_txn(service) + txn = yield defer.ensureDeferred(self.store.get_oldest_unsent_txn(service)) self.assertEquals(None, txn) @defer.inlineCallbacks @@ -349,14 +370,14 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): other_events = [Mock(event_id="e5"), Mock(event_id="e6")] # we aren't testing store._base stuff here, so mock this out - self.store.get_events_as_list = Mock(return_value=events) + self.store.get_events_as_list = Mock(return_value=make_awaitable(events)) yield self._insert_txn(self.as_list[1]["id"], 9, other_events) yield self._insert_txn(service.id, 10, events) yield self._insert_txn(service.id, 11, other_events) yield self._insert_txn(service.id, 12, other_events) - txn = yield self.store.get_oldest_unsent_txn(service) + txn = yield defer.ensureDeferred(self.store.get_oldest_unsent_txn(service)) self.assertEquals(service, txn.service) self.assertEquals(10, txn.id) self.assertEquals(events, txn.events) @@ -366,8 +387,8 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): yield self._set_state(self.as_list[0]["id"], ApplicationServiceState.DOWN) yield self._set_state(self.as_list[1]["id"], ApplicationServiceState.UP) - services = yield self.store.get_appservices_by_state( - ApplicationServiceState.DOWN + services = yield defer.ensureDeferred( + self.store.get_appservices_by_state(ApplicationServiceState.DOWN) ) self.assertEquals(1, len(services)) self.assertEquals(self.as_list[0]["id"], services[0].id) @@ -379,8 +400,8 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): yield self._set_state(self.as_list[2]["id"], ApplicationServiceState.DOWN) yield self._set_state(self.as_list[3]["id"], ApplicationServiceState.UP) - services = yield self.store.get_appservices_by_state( - ApplicationServiceState.DOWN + services = yield defer.ensureDeferred( + self.store.get_appservices_by_state(ApplicationServiceState.DOWN) ) self.assertEquals(2, len(services)) self.assertEquals( @@ -391,8 +412,8 @@ class ApplicationServiceTransactionStoreTestCase(unittest.TestCase): # required for ApplicationServiceTransactionStoreTestCase tests class TestTransactionStore(ApplicationServiceTransactionStore, ApplicationServiceStore): - def __init__(self, database: Database, db_conn, hs): - super(TestTransactionStore, self).__init__(database, db_conn, hs) + def __init__(self, database: DatabasePool, db_conn, hs): + super().__init__(database, db_conn, hs) class ApplicationServiceStoreConfigTestCase(unittest.TestCase): diff --git a/tests/storage/test_background_update.py b/tests/storage/test_background_update.py
index 940b166129..02aae1c13d 100644 --- a/tests/storage/test_background_update.py +++ b/tests/storage/test_background_update.py
@@ -1,7 +1,5 @@ from mock import Mock -from twisted.internet import defer - from synapse.storage.background_updates import BackgroundUpdater from tests import unittest @@ -9,7 +7,9 @@ from tests import unittest class BackgroundUpdateTestCase(unittest.HomeserverTestCase): def prepare(self, reactor, clock, homeserver): - self.updates = self.hs.get_datastore().db.updates # type: BackgroundUpdater + self.updates = ( + self.hs.get_datastore().db_pool.updates + ) # type: BackgroundUpdater # the base test class should have run the real bg updates for us self.assertTrue( self.get_success(self.updates.has_completed_background_updates()) @@ -29,18 +29,17 @@ class BackgroundUpdateTestCase(unittest.HomeserverTestCase): store = self.hs.get_datastore() self.get_success( - store.db.simple_insert( + store.db_pool.simple_insert( "background_updates", values={"update_name": "test_update", "progress_json": '{"my_key": 1}'}, ) ) # first step: make a bit of progress - @defer.inlineCallbacks - def update(progress, count): - yield self.clock.sleep((count * duration_ms) / 1000) + async def update(progress, count): + await self.clock.sleep((count * duration_ms) / 1000) progress = {"my_key": progress["my_key"] + 1} - yield store.db.runInteraction( + await store.db_pool.runInteraction( "update_progress", self.updates._background_update_progress_txn, "test_update", @@ -65,13 +64,12 @@ class BackgroundUpdateTestCase(unittest.HomeserverTestCase): # second step: complete the update # we should now get run with a much bigger number of items to update - @defer.inlineCallbacks - def update(progress, count): + async def update(progress, count): self.assertEqual(progress, {"my_key": 2}) self.assertAlmostEqual( count, target_background_update_duration_ms / duration_ms, places=0, ) - yield self.updates._end_background_update("test_update") + await self.updates._end_background_update("test_update") return count self.update_handler.side_effect = update diff --git a/tests/storage/test_base.py b/tests/storage/test_base.py
index b589506c60..eac7e4dcd2 100644 --- a/tests/storage/test_base.py +++ b/tests/storage/test_base.py
@@ -21,7 +21,7 @@ from mock import Mock from twisted.internet import defer from synapse.storage._base import SQLBaseStore -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool from synapse.storage.engines import create_engine from tests import unittest @@ -56,8 +56,9 @@ class SQLBaseStoreTestCase(unittest.TestCase): engine = create_engine(sqlite_config) fake_engine = Mock(wraps=engine) fake_engine.can_native_upsert = False + fake_engine.in_transaction.return_value = False - db = Database(Mock(), Mock(config=sqlite_config), fake_engine) + db = DatabasePool(Mock(), Mock(config=sqlite_config), fake_engine) db._db_pool = self.db_pool self.datastore = SQLBaseStore(db, None, hs) @@ -66,8 +67,10 @@ class SQLBaseStoreTestCase(unittest.TestCase): def test_insert_1col(self): self.mock_txn.rowcount = 1 - yield self.datastore.db.simple_insert( - table="tablename", values={"columname": "Value"} + yield defer.ensureDeferred( + self.datastore.db_pool.simple_insert( + table="tablename", values={"columname": "Value"} + ) ) self.mock_txn.execute.assert_called_with( @@ -78,10 +81,12 @@ class SQLBaseStoreTestCase(unittest.TestCase): def test_insert_3cols(self): self.mock_txn.rowcount = 1 - yield self.datastore.db.simple_insert( - table="tablename", - # Use OrderedDict() so we can assert on the SQL generated - values=OrderedDict([("colA", 1), ("colB", 2), ("colC", 3)]), + yield defer.ensureDeferred( + self.datastore.db_pool.simple_insert( + table="tablename", + # Use OrderedDict() so we can assert on the SQL generated + values=OrderedDict([("colA", 1), ("colB", 2), ("colC", 3)]), + ) ) self.mock_txn.execute.assert_called_with( @@ -93,8 +98,10 @@ class SQLBaseStoreTestCase(unittest.TestCase): self.mock_txn.rowcount = 1 self.mock_txn.__iter__ = Mock(return_value=iter([("Value",)])) - value = yield self.datastore.db.simple_select_one_onecol( - table="tablename", keyvalues={"keycol": "TheKey"}, retcol="retcol" + value = yield defer.ensureDeferred( + self.datastore.db_pool.simple_select_one_onecol( + table="tablename", keyvalues={"keycol": "TheKey"}, retcol="retcol" + ) ) self.assertEquals("Value", value) @@ -107,10 +114,12 @@ class SQLBaseStoreTestCase(unittest.TestCase): self.mock_txn.rowcount = 1 self.mock_txn.fetchone.return_value = (1, 2, 3) - ret = yield self.datastore.db.simple_select_one( - table="tablename", - keyvalues={"keycol": "TheKey"}, - retcols=["colA", "colB", "colC"], + ret = yield defer.ensureDeferred( + self.datastore.db_pool.simple_select_one( + table="tablename", + keyvalues={"keycol": "TheKey"}, + retcols=["colA", "colB", "colC"], + ) ) self.assertEquals({"colA": 1, "colB": 2, "colC": 3}, ret) @@ -123,11 +132,13 @@ class SQLBaseStoreTestCase(unittest.TestCase): self.mock_txn.rowcount = 0 self.mock_txn.fetchone.return_value = None - ret = yield self.datastore.db.simple_select_one( - table="tablename", - keyvalues={"keycol": "Not here"}, - retcols=["colA"], - allow_none=True, + ret = yield defer.ensureDeferred( + self.datastore.db_pool.simple_select_one( + table="tablename", + keyvalues={"keycol": "Not here"}, + retcols=["colA"], + allow_none=True, + ) ) self.assertFalse(ret) @@ -138,8 +149,10 @@ class SQLBaseStoreTestCase(unittest.TestCase): self.mock_txn.__iter__ = Mock(return_value=iter([(1,), (2,), (3,)])) self.mock_txn.description = (("colA", None, None, None, None, None, None),) - ret = yield self.datastore.db.simple_select_list( - table="tablename", keyvalues={"keycol": "A set"}, retcols=["colA"] + ret = yield defer.ensureDeferred( + self.datastore.db_pool.simple_select_list( + table="tablename", keyvalues={"keycol": "A set"}, retcols=["colA"] + ) ) self.assertEquals([{"colA": 1}, {"colA": 2}, {"colA": 3}], ret) @@ -151,10 +164,12 @@ class SQLBaseStoreTestCase(unittest.TestCase): def test_update_one_1col(self): self.mock_txn.rowcount = 1 - yield self.datastore.db.simple_update_one( - table="tablename", - keyvalues={"keycol": "TheKey"}, - updatevalues={"columnname": "New Value"}, + yield defer.ensureDeferred( + self.datastore.db_pool.simple_update_one( + table="tablename", + keyvalues={"keycol": "TheKey"}, + updatevalues={"columnname": "New Value"}, + ) ) self.mock_txn.execute.assert_called_with( @@ -166,10 +181,12 @@ class SQLBaseStoreTestCase(unittest.TestCase): def test_update_one_4cols(self): self.mock_txn.rowcount = 1 - yield self.datastore.db.simple_update_one( - table="tablename", - keyvalues=OrderedDict([("colA", 1), ("colB", 2)]), - updatevalues=OrderedDict([("colC", 3), ("colD", 4)]), + yield defer.ensureDeferred( + self.datastore.db_pool.simple_update_one( + table="tablename", + keyvalues=OrderedDict([("colA", 1), ("colB", 2)]), + updatevalues=OrderedDict([("colC", 3), ("colD", 4)]), + ) ) self.mock_txn.execute.assert_called_with( @@ -181,8 +198,10 @@ class SQLBaseStoreTestCase(unittest.TestCase): def test_delete_one(self): self.mock_txn.rowcount = 1 - yield self.datastore.db.simple_delete_one( - table="tablename", keyvalues={"keycol": "Go away"} + yield defer.ensureDeferred( + self.datastore.db_pool.simple_delete_one( + table="tablename", keyvalues={"keycol": "Go away"} + ) ) self.mock_txn.execute.assert_called_with( diff --git a/tests/storage/test_cleanup_extrems.py b/tests/storage/test_cleanup_extrems.py
index 43425c969a..080761d1d2 100644 --- a/tests/storage/test_cleanup_extrems.py +++ b/tests/storage/test_cleanup_extrems.py
@@ -38,7 +38,7 @@ class CleanupExtremBackgroundUpdateStoreTestCase(HomeserverTestCase): # Create a test user and room self.user = UserID("alice", "test") - self.requester = Requester(self.user, None, False, None, None) + self.requester = Requester(self.user, None, False, False, None, None) info, _ = self.get_success(self.room_creator.create_room(self.requester, {})) self.room_id = info["room_id"] @@ -47,12 +47,12 @@ class CleanupExtremBackgroundUpdateStoreTestCase(HomeserverTestCase): """ # Make sure we don't clash with in progress updates. self.assertTrue( - self.store.db.updates._all_done, "Background updates are still ongoing" + self.store.db_pool.updates._all_done, "Background updates are still ongoing" ) schema_path = os.path.join( prepare_database.dir_path, - "data_stores", + "databases", "main", "schema", "delta", @@ -64,19 +64,19 @@ class CleanupExtremBackgroundUpdateStoreTestCase(HomeserverTestCase): prepare_database.executescript(txn, schema_path) self.get_success( - self.store.db.runInteraction( + self.store.db_pool.runInteraction( "test_delete_forward_extremities", run_delta_file ) ) # Ugh, have to reset this flag - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) def test_soft_failed_extremities_handled_correctly(self): @@ -260,7 +260,7 @@ class CleanupExtremDummyEventsTestCase(HomeserverTestCase): # Create a test user and room self.user = UserID.from_string(self.register_user("user1", "password")) self.token1 = self.login("user1", "password") - self.requester = Requester(self.user, None, False, None, None) + self.requester = Requester(self.user, None, False, False, None, None) info, _ = self.get_success(self.room_creator.create_room(self.requester, {})) self.room_id = info["room_id"] self.event_creator = homeserver.get_event_creation_handler() @@ -271,7 +271,7 @@ class CleanupExtremDummyEventsTestCase(HomeserverTestCase): # Pump the reactor repeatedly so that the background updates have a # chance to run. - self.pump(10 * 60) + self.pump(20) latest_event_ids = self.get_success( self.store.get_latest_event_ids_in_room(self.room_id) @@ -353,6 +353,7 @@ class CleanupExtremDummyEventsTestCase(HomeserverTestCase): self.event_creator_handler._rooms_to_exclude_from_dummy_event_insertion[ "3" ] = 300000 + self.event_creator_handler._expire_rooms_to_exclude_from_dummy_event_insertion() # All entries within time frame self.assertEqual( @@ -362,7 +363,7 @@ class CleanupExtremDummyEventsTestCase(HomeserverTestCase): 3, ) # Oldest room to expire - self.pump(1) + self.pump(1.01) self.event_creator_handler._expire_rooms_to_exclude_from_dummy_event_insertion() self.assertEqual( len( diff --git a/tests/storage/test_client_ips.py b/tests/storage/test_client_ips.py
index 3b483bc7f0..755c70db31 100644 --- a/tests/storage/test_client_ips.py +++ b/tests/storage/test_client_ips.py
@@ -16,13 +16,12 @@ from mock import Mock -from twisted.internet import defer - import synapse.rest.admin from synapse.http.site import XForwardedForRequest from synapse.rest.client.v1 import login from tests import unittest +from tests.test_utils import make_awaitable from tests.unittest import override_config @@ -86,7 +85,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): self.pump(0) result = self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( table="user_ips", keyvalues={"user_id": user_id}, retcols=["access_token", "ip", "user_agent", "device_id", "last_seen"], @@ -117,7 +116,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): self.pump(0) result = self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( table="user_ips", keyvalues={"user_id": user_id}, retcols=["access_token", "ip", "user_agent", "device_id", "last_seen"], @@ -155,7 +154,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): user_id = "@user:server" self.store.get_monthly_active_count = Mock( - return_value=defer.succeed(lots_of_users) + return_value=make_awaitable(lots_of_users) ) self.get_success( self.store.insert_client_ip( @@ -204,10 +203,10 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): def test_devices_last_seen_bg_update(self): # First make sure we have completed all updates. while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) user_id = "@user:id" @@ -225,7 +224,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): # But clear the associated entry in devices table self.get_success( - self.store.db.simple_update( + self.store.db_pool.simple_update( table="devices", keyvalues={"user_id": user_id, "device_id": device_id}, updatevalues={"last_seen": None, "ip": None, "user_agent": None}, @@ -252,7 +251,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): # Register the background update to run again. self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( table="background_updates", values={ "update_name": "devices_last_seen", @@ -263,14 +262,14 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): ) # ... and tell the DataStore that it hasn't finished all updates yet - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False # Now let's actually drive the updates to completion while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) # We should now get the correct result again @@ -293,10 +292,10 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): def test_old_user_ips_pruned(self): # First make sure we have completed all updates. while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) user_id = "@user:id" @@ -315,7 +314,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): # We should see that in the DB result = self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( table="user_ips", keyvalues={"user_id": user_id}, retcols=["access_token", "ip", "user_agent", "device_id", "last_seen"], @@ -341,7 +340,7 @@ class ClientIpStoreTestCase(unittest.HomeserverTestCase): # We should get no results. result = self.get_success( - self.store.db.simple_select_list( + self.store.db_pool.simple_select_list( table="user_ips", keyvalues={"user_id": user_id}, retcols=["access_token", "ip", "user_agent", "device_id", "last_seen"], diff --git a/tests/storage/test_devices.py b/tests/storage/test_devices.py
index c2539b353a..ecb00f4e02 100644 --- a/tests/storage/test_devices.py +++ b/tests/storage/test_devices.py
@@ -23,7 +23,7 @@ import tests.utils class DeviceStoreTestCase(tests.unittest.TestCase): def __init__(self, *args, **kwargs): - super(DeviceStoreTestCase, self).__init__(*args, **kwargs) + super().__init__(*args, **kwargs) self.store = None # type: synapse.storage.DataStore @defer.inlineCallbacks @@ -34,9 +34,11 @@ class DeviceStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def test_store_new_device(self): - yield self.store.store_device("user_id", "device_id", "display_name") + yield defer.ensureDeferred( + self.store.store_device("user_id", "device_id", "display_name") + ) - res = yield self.store.get_device("user_id", "device_id") + res = yield defer.ensureDeferred(self.store.get_device("user_id", "device_id")) self.assertDictContainsSubset( { "user_id": "user_id", @@ -48,11 +50,17 @@ class DeviceStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def test_get_devices_by_user(self): - yield self.store.store_device("user_id", "device1", "display_name 1") - yield self.store.store_device("user_id", "device2", "display_name 2") - yield self.store.store_device("user_id2", "device3", "display_name 3") + yield defer.ensureDeferred( + self.store.store_device("user_id", "device1", "display_name 1") + ) + yield defer.ensureDeferred( + self.store.store_device("user_id", "device2", "display_name 2") + ) + yield defer.ensureDeferred( + self.store.store_device("user_id2", "device3", "display_name 3") + ) - res = yield self.store.get_devices_by_user("user_id") + res = yield defer.ensureDeferred(self.store.get_devices_by_user("user_id")) self.assertEqual(2, len(res.keys())) self.assertDictContainsSubset( { @@ -76,13 +84,13 @@ class DeviceStoreTestCase(tests.unittest.TestCase): device_ids = ["device_id1", "device_id2"] # Add two device updates with a single stream_id - yield self.store.add_device_change_to_streams( - "user_id", device_ids, ["somehost"] + yield defer.ensureDeferred( + self.store.add_device_change_to_streams("user_id", device_ids, ["somehost"]) ) # Get all device updates ever meant for this remote - now_stream_id, device_updates = yield self.store.get_device_updates_by_remote( - "somehost", -1, limit=100 + now_stream_id, device_updates = yield defer.ensureDeferred( + self.store.get_device_updates_by_remote("somehost", -1, limit=100) ) # Check original device_ids are contained within these updates @@ -99,29 +107,35 @@ class DeviceStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def test_update_device(self): - yield self.store.store_device("user_id", "device_id", "display_name 1") + yield defer.ensureDeferred( + self.store.store_device("user_id", "device_id", "display_name 1") + ) - res = yield self.store.get_device("user_id", "device_id") + res = yield defer.ensureDeferred(self.store.get_device("user_id", "device_id")) self.assertEqual("display_name 1", res["display_name"]) # do a no-op first - yield self.store.update_device("user_id", "device_id") - res = yield self.store.get_device("user_id", "device_id") + yield defer.ensureDeferred(self.store.update_device("user_id", "device_id")) + res = yield defer.ensureDeferred(self.store.get_device("user_id", "device_id")) self.assertEqual("display_name 1", res["display_name"]) # do the update - yield self.store.update_device( - "user_id", "device_id", new_display_name="display_name 2" + yield defer.ensureDeferred( + self.store.update_device( + "user_id", "device_id", new_display_name="display_name 2" + ) ) # check it worked - res = yield self.store.get_device("user_id", "device_id") + res = yield defer.ensureDeferred(self.store.get_device("user_id", "device_id")) self.assertEqual("display_name 2", res["display_name"]) @defer.inlineCallbacks def test_update_unknown_device(self): with self.assertRaises(synapse.api.errors.StoreError) as cm: - yield self.store.update_device( - "user_id", "unknown_device_id", new_display_name="display_name 2" + yield defer.ensureDeferred( + self.store.update_device( + "user_id", "unknown_device_id", new_display_name="display_name 2" + ) ) self.assertEqual(404, cm.exception.code) diff --git a/tests/storage/test_directory.py b/tests/storage/test_directory.py
index 4e128e1047..da93ca3980 100644 --- a/tests/storage/test_directory.py +++ b/tests/storage/test_directory.py
@@ -34,35 +34,53 @@ class DirectoryStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_room_to_alias(self): - yield self.store.create_room_alias_association( - room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + yield defer.ensureDeferred( + self.store.create_room_alias_association( + room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + ) ) self.assertEquals( ["#my-room:test"], - (yield self.store.get_aliases_for_room(self.room.to_string())), + ( + yield defer.ensureDeferred( + self.store.get_aliases_for_room(self.room.to_string()) + ) + ), ) @defer.inlineCallbacks def test_alias_to_room(self): - yield self.store.create_room_alias_association( - room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + yield defer.ensureDeferred( + self.store.create_room_alias_association( + room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + ) ) self.assertObjectHasAttributes( {"room_id": self.room.to_string(), "servers": ["test"]}, - (yield self.store.get_association_from_room_alias(self.alias)), + ( + yield defer.ensureDeferred( + self.store.get_association_from_room_alias(self.alias) + ) + ), ) @defer.inlineCallbacks def test_delete_alias(self): - yield self.store.create_room_alias_association( - room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + yield defer.ensureDeferred( + self.store.create_room_alias_association( + room_alias=self.alias, room_id=self.room.to_string(), servers=["test"] + ) ) - room_id = yield self.store.delete_room_alias(self.alias) + room_id = yield defer.ensureDeferred(self.store.delete_room_alias(self.alias)) self.assertEqual(self.room.to_string(), room_id) self.assertIsNone( - (yield self.store.get_association_from_room_alias(self.alias)) + ( + yield defer.ensureDeferred( + self.store.get_association_from_room_alias(self.alias) + ) + ) ) diff --git a/tests/storage/test_end_to_end_keys.py b/tests/storage/test_end_to_end_keys.py
index 398d546280..3fc4bb13b6 100644 --- a/tests/storage/test_end_to_end_keys.py +++ b/tests/storage/test_end_to_end_keys.py
@@ -30,11 +30,15 @@ class EndToEndKeyStoreTestCase(tests.unittest.TestCase): now = 1470174257070 json = {"key": "value"} - yield self.store.store_device("user", "device", None) + yield defer.ensureDeferred(self.store.store_device("user", "device", None)) - yield self.store.set_e2e_device_keys("user", "device", now, json) + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user", "device", now, json) + ) - res = yield self.store.get_e2e_device_keys((("user", "device"),)) + res = yield defer.ensureDeferred( + self.store.get_e2e_device_keys_for_cs_api((("user", "device"),)) + ) self.assertIn("user", res) self.assertIn("device", res["user"]) dev = res["user"]["device"] @@ -45,14 +49,18 @@ class EndToEndKeyStoreTestCase(tests.unittest.TestCase): now = 1470174257070 json = {"key": "value"} - yield self.store.store_device("user", "device", None) + yield defer.ensureDeferred(self.store.store_device("user", "device", None)) - changed = yield self.store.set_e2e_device_keys("user", "device", now, json) + changed = yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user", "device", now, json) + ) self.assertTrue(changed) # If we try to upload the same key then we should be told nothing # changed - changed = yield self.store.set_e2e_device_keys("user", "device", now, json) + changed = yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user", "device", now, json) + ) self.assertFalse(changed) @defer.inlineCallbacks @@ -60,10 +68,16 @@ class EndToEndKeyStoreTestCase(tests.unittest.TestCase): now = 1470174257070 json = {"key": "value"} - yield self.store.set_e2e_device_keys("user", "device", now, json) - yield self.store.store_device("user", "device", "display_name") + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user", "device", now, json) + ) + yield defer.ensureDeferred( + self.store.store_device("user", "device", "display_name") + ) - res = yield self.store.get_e2e_device_keys((("user", "device"),)) + res = yield defer.ensureDeferred( + self.store.get_e2e_device_keys_for_cs_api((("user", "device"),)) + ) self.assertIn("user", res) self.assertIn("device", res["user"]) dev = res["user"]["device"] @@ -75,18 +89,28 @@ class EndToEndKeyStoreTestCase(tests.unittest.TestCase): def test_multiple_devices(self): now = 1470174257070 - yield self.store.store_device("user1", "device1", None) - yield self.store.store_device("user1", "device2", None) - yield self.store.store_device("user2", "device1", None) - yield self.store.store_device("user2", "device2", None) + yield defer.ensureDeferred(self.store.store_device("user1", "device1", None)) + yield defer.ensureDeferred(self.store.store_device("user1", "device2", None)) + yield defer.ensureDeferred(self.store.store_device("user2", "device1", None)) + yield defer.ensureDeferred(self.store.store_device("user2", "device2", None)) - yield self.store.set_e2e_device_keys("user1", "device1", now, {"key": "json11"}) - yield self.store.set_e2e_device_keys("user1", "device2", now, {"key": "json12"}) - yield self.store.set_e2e_device_keys("user2", "device1", now, {"key": "json21"}) - yield self.store.set_e2e_device_keys("user2", "device2", now, {"key": "json22"}) + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user1", "device1", now, {"key": "json11"}) + ) + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user1", "device2", now, {"key": "json12"}) + ) + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user2", "device1", now, {"key": "json21"}) + ) + yield defer.ensureDeferred( + self.store.set_e2e_device_keys("user2", "device2", now, {"key": "json22"}) + ) - res = yield self.store.get_e2e_device_keys( - (("user1", "device1"), ("user2", "device2")) + res = yield defer.ensureDeferred( + self.store.get_e2e_device_keys_for_cs_api( + (("user1", "device1"), ("user2", "device2")) + ) ) self.assertIn("user1", res) self.assertIn("device1", res["user1"]) diff --git a/tests/storage/test_event_federation.py b/tests/storage/test_event_federation.py
index 3aeec0dc0f..d4c3b867e3 100644 --- a/tests/storage/test_event_federation.py +++ b/tests/storage/test_event_federation.py
@@ -56,7 +56,9 @@ class EventFederationWorkerStoreTestCase(tests.unittest.HomeserverTestCase): ) for i in range(0, 20): - self.get_success(self.store.db.runInteraction("insert", insert_event, i)) + self.get_success( + self.store.db_pool.runInteraction("insert", insert_event, i) + ) # this should get the last ten r = self.get_success(self.store.get_prev_events_for_room(room_id)) @@ -81,13 +83,13 @@ class EventFederationWorkerStoreTestCase(tests.unittest.HomeserverTestCase): for i in range(0, 20): self.get_success( - self.store.db.runInteraction("insert", insert_event, i, room1) + self.store.db_pool.runInteraction("insert", insert_event, i, room1) ) self.get_success( - self.store.db.runInteraction("insert", insert_event, i, room2) + self.store.db_pool.runInteraction("insert", insert_event, i, room2) ) self.get_success( - self.store.db.runInteraction("insert", insert_event, i, room3) + self.store.db_pool.runInteraction("insert", insert_event, i, room3) ) # Test simple case @@ -164,7 +166,7 @@ class EventFederationWorkerStoreTestCase(tests.unittest.HomeserverTestCase): depth = depth_map[event_id] - self.store.db.simple_insert_txn( + self.store.db_pool.simple_insert_txn( txn, table="events", values={ @@ -179,7 +181,7 @@ class EventFederationWorkerStoreTestCase(tests.unittest.HomeserverTestCase): }, ) - self.store.db.simple_insert_many_txn( + self.store.db_pool.simple_insert_many_txn( txn, table="event_auth", values=[ @@ -192,7 +194,7 @@ class EventFederationWorkerStoreTestCase(tests.unittest.HomeserverTestCase): for event_id in auth_graph: next_stream_ordering += 1 self.get_success( - self.store.db.runInteraction( + self.store.db_pool.runInteraction( "insert", insert_event, event_id, next_stream_ordering ) ) diff --git a/tests/storage/test_event_metrics.py b/tests/storage/test_event_metrics.py
index a7b85004e5..3957471f3f 100644 --- a/tests/storage/test_event_metrics.py +++ b/tests/storage/test_event_metrics.py
@@ -27,7 +27,7 @@ class ExtremStatisticsTestCase(HomeserverTestCase): room_creator = self.hs.get_room_creation_handler() user = UserID("alice", "test") - requester = Requester(user, None, False, None, None) + requester = Requester(user, None, False, False, None, None) # Real events, forward extremities events = [(3, 2), (6, 2), (4, 6)] @@ -52,14 +52,14 @@ class ExtremStatisticsTestCase(HomeserverTestCase): self.reactor.advance(60 * 60 * 1000) self.pump(1) - items = set( + items = list( filter( lambda x: b"synapse_forward_extremities_" in x, - generate_latest(REGISTRY).split(b"\n"), + generate_latest(REGISTRY, emit_help=False).split(b"\n"), ) ) - expected = { + expected = [ b'synapse_forward_extremities_bucket{le="1.0"} 0.0', b'synapse_forward_extremities_bucket{le="2.0"} 2.0', b'synapse_forward_extremities_bucket{le="3.0"} 2.0', @@ -72,9 +72,12 @@ class ExtremStatisticsTestCase(HomeserverTestCase): b'synapse_forward_extremities_bucket{le="100.0"} 3.0', b'synapse_forward_extremities_bucket{le="200.0"} 3.0', b'synapse_forward_extremities_bucket{le="500.0"} 3.0', - b'synapse_forward_extremities_bucket{le="+Inf"} 3.0', - b"synapse_forward_extremities_count 3.0", - b"synapse_forward_extremities_sum 10.0", - } - + # per https://docs.google.com/document/d/1KwV0mAXwwbvvifBvDKH_LU1YjyXE_wxCkHNoCGq1GX0/edit#heading=h.wghdjzzh72j9, + # "inf" is valid: "this includes variants such as inf" + b'synapse_forward_extremities_bucket{le="inf"} 3.0', + b"# TYPE synapse_forward_extremities_gcount gauge", + b"synapse_forward_extremities_gcount 3.0", + b"# TYPE synapse_forward_extremities_gsum gauge", + b"synapse_forward_extremities_gsum 10.0", + ] self.assertEqual(items, expected) diff --git a/tests/storage/test_event_push_actions.py b/tests/storage/test_event_push_actions.py
index b45bc9c115..c0595963dd 100644 --- a/tests/storage/test_event_push_actions.py +++ b/tests/storage/test_event_push_actions.py
@@ -39,14 +39,18 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def test_get_unread_push_actions_for_user_in_range_for_http(self): - yield self.store.get_unread_push_actions_for_user_in_range_for_http( - USER_ID, 0, 1000, 20 + yield defer.ensureDeferred( + self.store.get_unread_push_actions_for_user_in_range_for_http( + USER_ID, 0, 1000, 20 + ) ) @defer.inlineCallbacks def test_get_unread_push_actions_for_user_in_range_for_email(self): - yield self.store.get_unread_push_actions_for_user_in_range_for_email( - USER_ID, 0, 1000, 20 + yield defer.ensureDeferred( + self.store.get_unread_push_actions_for_user_in_range_for_email( + USER_ID, 0, 1000, 20 + ) ) @defer.inlineCallbacks @@ -56,12 +60,18 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def _assert_counts(noitf_count, highlight_count): - counts = yield self.store.db.runInteraction( - "", self.store._get_unread_counts_by_pos_txn, room_id, user_id, 0 + counts = yield defer.ensureDeferred( + self.store.db_pool.runInteraction( + "", self.store._get_unread_counts_by_pos_txn, room_id, user_id, 0 + ) ) self.assertEquals( counts, - {"notify_count": noitf_count, "highlight_count": highlight_count}, + { + "notify_count": noitf_count, + "unread_count": 0, # Unread counts are tested in the sync tests. + "highlight_count": highlight_count, + }, ) @defer.inlineCallbacks @@ -72,28 +82,36 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): event.internal_metadata.stream_ordering = stream event.depth = stream - yield self.store.add_push_actions_to_staging( - event.event_id, {user_id: action} + yield defer.ensureDeferred( + self.store.add_push_actions_to_staging( + event.event_id, {user_id: action}, False, + ) ) - yield self.store.db.runInteraction( - "", - self.persist_events_store._set_push_actions_for_event_and_users_txn, - [(event, None)], - [(event, None)], + yield defer.ensureDeferred( + self.store.db_pool.runInteraction( + "", + self.persist_events_store._set_push_actions_for_event_and_users_txn, + [(event, None)], + [(event, None)], + ) ) def _rotate(stream): - return self.store.db.runInteraction( - "", self.store._rotate_notifs_before_txn, stream + return defer.ensureDeferred( + self.store.db_pool.runInteraction( + "", self.store._rotate_notifs_before_txn, stream + ) ) def _mark_read(stream, depth): - return self.store.db.runInteraction( - "", - self.store._remove_old_push_actions_before_txn, - room_id, - user_id, - stream, + return defer.ensureDeferred( + self.store.db_pool.runInteraction( + "", + self.store._remove_old_push_actions_before_txn, + room_id, + user_id, + stream, + ) ) yield _assert_counts(0, 0) @@ -117,8 +135,10 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): yield _inject_actions(6, PlAIN_NOTIF) yield _rotate(7) - yield self.store.db.simple_delete( - table="event_push_actions", keyvalues={"1": 1}, desc="" + yield defer.ensureDeferred( + self.store.db_pool.simple_delete( + table="event_push_actions", keyvalues={"1": 1}, desc="" + ) ) yield _assert_counts(1, 0) @@ -136,33 +156,43 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): @defer.inlineCallbacks def test_find_first_stream_ordering_after_ts(self): def add_event(so, ts): - return self.store.db.simple_insert( - "events", - { - "stream_ordering": so, - "received_ts": ts, - "event_id": "event%i" % so, - "type": "", - "room_id": "", - "content": "", - "processed": True, - "outlier": False, - "topological_ordering": 0, - "depth": 0, - }, + return defer.ensureDeferred( + self.store.db_pool.simple_insert( + "events", + { + "stream_ordering": so, + "received_ts": ts, + "event_id": "event%i" % so, + "type": "", + "room_id": "", + "content": "", + "processed": True, + "outlier": False, + "topological_ordering": 0, + "depth": 0, + }, + ) ) # start with the base case where there are no events in the table - r = yield self.store.find_first_stream_ordering_after_ts(11) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(11) + ) self.assertEqual(r, 0) # now with one event yield add_event(2, 10) - r = yield self.store.find_first_stream_ordering_after_ts(9) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(9) + ) self.assertEqual(r, 2) - r = yield self.store.find_first_stream_ordering_after_ts(10) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(10) + ) self.assertEqual(r, 2) - r = yield self.store.find_first_stream_ordering_after_ts(11) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(11) + ) self.assertEqual(r, 3) # add a bunch of dummy events to the events table @@ -175,25 +205,37 @@ class EventPushActionsStoreTestCase(tests.unittest.TestCase): ): yield add_event(stream_ordering, ts) - r = yield self.store.find_first_stream_ordering_after_ts(110) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(110) + ) self.assertEqual(r, 3, "First event after 110ms should be 3, was %i" % r) # 4 and 5 are both after 120: we want 4 rather than 5 - r = yield self.store.find_first_stream_ordering_after_ts(120) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(120) + ) self.assertEqual(r, 4, "First event after 120ms should be 4, was %i" % r) - r = yield self.store.find_first_stream_ordering_after_ts(129) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(129) + ) self.assertEqual(r, 10, "First event after 129ms should be 10, was %i" % r) # check we can get the last event - r = yield self.store.find_first_stream_ordering_after_ts(140) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(140) + ) self.assertEqual(r, 20, "First event after 14ms should be 20, was %i" % r) # off the end - r = yield self.store.find_first_stream_ordering_after_ts(160) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(160) + ) self.assertEqual(r, 21) # check we can find an event at ordering zero yield add_event(0, 5) - r = yield self.store.find_first_stream_ordering_after_ts(1) + r = yield defer.ensureDeferred( + self.store.find_first_stream_ordering_after_ts(1) + ) self.assertEqual(r, 0) diff --git a/tests/storage/test_id_generators.py b/tests/storage/test_id_generators.py
index 55e9ecf264..392b08832b 100644 --- a/tests/storage/test_id_generators.py +++ b/tests/storage/test_id_generators.py
@@ -12,9 +12,8 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - - -from synapse.storage.database import Database +from synapse.storage.database import DatabasePool +from synapse.storage.engines import IncorrectDatabaseSetup from synapse.storage.util.id_generators import MultiWriterIdGenerator from tests.unittest import HomeserverTestCase @@ -27,9 +26,9 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): def prepare(self, reactor, clock, hs): self.store = hs.get_datastore() - self.db = self.store.db # type: Database + self.db_pool = self.store.db_pool # type: DatabasePool - self.get_success(self.db.runInteraction("_setup_db", self._setup_db)) + self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db)) def _setup_db(self, txn): txn.execute("CREATE SEQUENCE foobar_seq") @@ -43,29 +42,64 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): """ ) - def _create_id_generator(self, instance_name="master") -> MultiWriterIdGenerator: + def _create_id_generator( + self, instance_name="master", writers=["master"] + ) -> MultiWriterIdGenerator: def _create(conn): return MultiWriterIdGenerator( conn, - self.db, + self.db_pool, + stream_name="test_stream", instance_name=instance_name, table="foobar", instance_column="instance_name", id_column="stream_id", sequence_name="foobar_seq", + writers=writers, ) - return self.get_success(self.db.runWithConnection(_create)) + return self.get_success_or_raise(self.db_pool.runWithConnection(_create)) def _insert_rows(self, instance_name: str, number: int): + """Insert N rows as the given instance, inserting with stream IDs pulled + from the postgres sequence. + """ + def _insert(txn): for _ in range(number): txn.execute( "INSERT INTO foobar VALUES (nextval('foobar_seq'), ?)", (instance_name,), ) + txn.execute( + """ + INSERT INTO stream_positions VALUES ('test_stream', ?, lastval()) + ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = lastval() + """, + (instance_name,), + ) + + self.get_success(self.db_pool.runInteraction("_insert_rows", _insert)) - self.get_success(self.db.runInteraction("test_single_instance", _insert)) + def _insert_row_with_id(self, instance_name: str, stream_id: int): + """Insert one row as the given instance with given stream_id, updating + the postgres sequence position to match. + """ + + def _insert(txn): + txn.execute( + "INSERT INTO foobar VALUES (?, ?)", (stream_id, instance_name,), + ) + txn.execute("SELECT setval('foobar_seq', ?)", (stream_id,)) + txn.execute( + """ + INSERT INTO stream_positions VALUES ('test_stream', ?, ?) + ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = ? + """, + (instance_name, stream_id, stream_id), + ) + + self.get_success(self.db_pool.runInteraction("_insert_row_with_id", _insert)) def test_empty(self): """Test an ID generator against an empty database gives sensible @@ -88,22 +122,72 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): id_gen = self._create_id_generator() self.assertEqual(id_gen.get_positions(), {"master": 7}) - self.assertEqual(id_gen.get_current_token("master"), 7) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) # Try allocating a new ID gen and check that we only see position # advanced after we leave the context manager. async def _get_next_async(): - with await id_gen.get_next() as stream_id: + async with id_gen.get_next() as stream_id: self.assertEqual(stream_id, 8) self.assertEqual(id_gen.get_positions(), {"master": 7}) - self.assertEqual(id_gen.get_current_token("master"), 7) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) self.get_success(_get_next_async()) self.assertEqual(id_gen.get_positions(), {"master": 8}) - self.assertEqual(id_gen.get_current_token("master"), 8) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 8) + + def test_out_of_order_finish(self): + """Test that IDs persisted out of order are correctly handled + """ + + # Prefill table with 7 rows written by 'master' + self._insert_rows("master", 7) + + id_gen = self._create_id_generator() + + self.assertEqual(id_gen.get_positions(), {"master": 7}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) + + ctx1 = self.get_success(id_gen.get_next()) + ctx2 = self.get_success(id_gen.get_next()) + ctx3 = self.get_success(id_gen.get_next()) + ctx4 = self.get_success(id_gen.get_next()) + + s1 = self.get_success(ctx1.__aenter__()) + s2 = self.get_success(ctx2.__aenter__()) + s3 = self.get_success(ctx3.__aenter__()) + s4 = self.get_success(ctx4.__aenter__()) + + self.assertEqual(s1, 8) + self.assertEqual(s2, 9) + self.assertEqual(s3, 10) + self.assertEqual(s4, 11) + + self.assertEqual(id_gen.get_positions(), {"master": 7}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) + + self.get_success(ctx2.__aexit__(None, None, None)) + + self.assertEqual(id_gen.get_positions(), {"master": 7}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) + + self.get_success(ctx1.__aexit__(None, None, None)) + + self.assertEqual(id_gen.get_positions(), {"master": 9}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 9) + + self.get_success(ctx4.__aexit__(None, None, None)) + + self.assertEqual(id_gen.get_positions(), {"master": 9}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 9) + + self.get_success(ctx3.__aexit__(None, None, None)) + + self.assertEqual(id_gen.get_positions(), {"master": 11}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 11) def test_multi_instance(self): """Test that reads and writes from multiple processes are handled @@ -112,18 +196,18 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): self._insert_rows("first", 3) self._insert_rows("second", 4) - first_id_gen = self._create_id_generator("first") - second_id_gen = self._create_id_generator("second") + first_id_gen = self._create_id_generator("first", writers=["first", "second"]) + second_id_gen = self._create_id_generator("second", writers=["first", "second"]) self.assertEqual(first_id_gen.get_positions(), {"first": 3, "second": 7}) - self.assertEqual(first_id_gen.get_current_token("first"), 3) - self.assertEqual(first_id_gen.get_current_token("second"), 7) + self.assertEqual(first_id_gen.get_current_token_for_writer("first"), 3) + self.assertEqual(first_id_gen.get_current_token_for_writer("second"), 7) # Try allocating a new ID gen and check that we only see position # advanced after we leave the context manager. async def _get_next_async(): - with await first_id_gen.get_next() as stream_id: + async with first_id_gen.get_next() as stream_id: self.assertEqual(stream_id, 8) self.assertEqual( @@ -141,7 +225,7 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): # stream ID async def _get_next_async(): - with await second_id_gen.get_next() as stream_id: + async with second_id_gen.get_next() as stream_id: self.assertEqual(stream_id, 9) self.assertEqual( @@ -166,7 +250,7 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): id_gen = self._create_id_generator() self.assertEqual(id_gen.get_positions(), {"master": 7}) - self.assertEqual(id_gen.get_current_token("master"), 7) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) # Try allocating a new ID gen and check that we only see position # advanced after we leave the context manager. @@ -176,9 +260,315 @@ class MultiWriterIdGeneratorTestCase(HomeserverTestCase): self.assertEqual(stream_id, 8) self.assertEqual(id_gen.get_positions(), {"master": 7}) - self.assertEqual(id_gen.get_current_token("master"), 7) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) - self.get_success(self.db.runInteraction("test", _get_next_txn)) + self.get_success(self.db_pool.runInteraction("test", _get_next_txn)) self.assertEqual(id_gen.get_positions(), {"master": 8}) - self.assertEqual(id_gen.get_current_token("master"), 8) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 8) + + def test_get_persisted_upto_position(self): + """Test that `get_persisted_upto_position` correctly tracks updates to + positions. + """ + + # The following tests are a bit cheeky in that we notify about new + # positions via `advance` without *actually* advancing the postgres + # sequence. + + self._insert_row_with_id("first", 3) + self._insert_row_with_id("second", 5) + + id_gen = self._create_id_generator("first", writers=["first", "second"]) + + self.assertEqual(id_gen.get_positions(), {"first": 3, "second": 5}) + + # Min is 3 and there is a gap between 5, so we expect it to be 3. + self.assertEqual(id_gen.get_persisted_upto_position(), 3) + + # We advance "first" straight to 6. Min is now 5 but there is no gap so + # we expect it to be 6 + id_gen.advance("first", 6) + self.assertEqual(id_gen.get_persisted_upto_position(), 6) + + # No gap, so we expect 7. + id_gen.advance("second", 7) + self.assertEqual(id_gen.get_persisted_upto_position(), 7) + + # We haven't seen 8 yet, so we expect 7 still. + id_gen.advance("second", 9) + self.assertEqual(id_gen.get_persisted_upto_position(), 7) + + # Now that we've seen 7, 8 and 9 we can got straight to 9. + id_gen.advance("first", 8) + self.assertEqual(id_gen.get_persisted_upto_position(), 9) + + # Jump forward with gaps. The minimum is 11, even though we haven't seen + # 10 we know that everything before 11 must be persisted. + id_gen.advance("first", 11) + id_gen.advance("second", 15) + self.assertEqual(id_gen.get_persisted_upto_position(), 11) + + def test_get_persisted_upto_position_get_next(self): + """Test that `get_persisted_upto_position` correctly tracks updates to + positions when `get_next` is called. + """ + + self._insert_row_with_id("first", 3) + self._insert_row_with_id("second", 5) + + id_gen = self._create_id_generator("first", writers=["first", "second"]) + + self.assertEqual(id_gen.get_positions(), {"first": 3, "second": 5}) + + self.assertEqual(id_gen.get_persisted_upto_position(), 3) + + async def _get_next_async(): + async with id_gen.get_next() as stream_id: + self.assertEqual(stream_id, 6) + self.assertEqual(id_gen.get_persisted_upto_position(), 3) + + self.get_success(_get_next_async()) + + self.assertEqual(id_gen.get_persisted_upto_position(), 6) + + # We assume that so long as `get_next` does correctly advance the + # `persisted_upto_position` in this case, then it will be correct in the + # other cases that are tested above (since they'll hit the same code). + + def test_restart_during_out_of_order_persistence(self): + """Test that restarting a process while another process is writing out + of order updates are handled correctly. + """ + + # Prefill table with 7 rows written by 'master' + self._insert_rows("master", 7) + + id_gen = self._create_id_generator() + + self.assertEqual(id_gen.get_positions(), {"master": 7}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) + + # Persist two rows at once + ctx1 = self.get_success(id_gen.get_next()) + ctx2 = self.get_success(id_gen.get_next()) + + s1 = self.get_success(ctx1.__aenter__()) + s2 = self.get_success(ctx2.__aenter__()) + + self.assertEqual(s1, 8) + self.assertEqual(s2, 9) + + self.assertEqual(id_gen.get_positions(), {"master": 7}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), 7) + + # We finish persisting the second row before restart + self.get_success(ctx2.__aexit__(None, None, None)) + + # We simulate a restart of another worker by just creating a new ID gen. + id_gen_worker = self._create_id_generator("worker") + + # Restarted worker should not see the second persisted row + self.assertEqual(id_gen_worker.get_positions(), {"master": 7}) + self.assertEqual(id_gen_worker.get_current_token_for_writer("master"), 7) + + # Now if we persist the first row then both instances should jump ahead + # correctly. + self.get_success(ctx1.__aexit__(None, None, None)) + + self.assertEqual(id_gen.get_positions(), {"master": 9}) + id_gen_worker.advance("master", 9) + self.assertEqual(id_gen_worker.get_positions(), {"master": 9}) + + def test_writer_config_change(self): + """Test that changing the writer config correctly works. + """ + + self._insert_row_with_id("first", 3) + self._insert_row_with_id("second", 5) + + # Initial config has two writers + id_gen = self._create_id_generator("first", writers=["first", "second"]) + self.assertEqual(id_gen.get_persisted_upto_position(), 3) + self.assertEqual(id_gen.get_current_token_for_writer("first"), 3) + self.assertEqual(id_gen.get_current_token_for_writer("second"), 5) + + # New config removes one of the configs. Note that if the writer is + # removed from config we assume that it has been shut down and has + # finished persisting, hence why the persisted upto position is 5. + id_gen_2 = self._create_id_generator("second", writers=["second"]) + self.assertEqual(id_gen_2.get_persisted_upto_position(), 5) + self.assertEqual(id_gen_2.get_current_token_for_writer("second"), 5) + + # This config points to a single, previously unused writer. + id_gen_3 = self._create_id_generator("third", writers=["third"]) + self.assertEqual(id_gen_3.get_persisted_upto_position(), 5) + + # For new writers we assume their initial position to be the current + # persisted up to position. This stops Synapse from doing a full table + # scan when a new writer comes along. + self.assertEqual(id_gen_3.get_current_token_for_writer("third"), 5) + + id_gen_4 = self._create_id_generator("fourth", writers=["third"]) + self.assertEqual(id_gen_4.get_current_token_for_writer("third"), 5) + + # Check that we get a sane next stream ID with this new config. + + async def _get_next_async(): + async with id_gen_3.get_next() as stream_id: + self.assertEqual(stream_id, 6) + + self.get_success(_get_next_async()) + self.assertEqual(id_gen_3.get_persisted_upto_position(), 6) + + # If we add back the old "first" then we shouldn't see the persisted up + # to position revert back to 3. + id_gen_5 = self._create_id_generator("five", writers=["first", "third"]) + self.assertEqual(id_gen_5.get_persisted_upto_position(), 6) + self.assertEqual(id_gen_5.get_current_token_for_writer("first"), 6) + self.assertEqual(id_gen_5.get_current_token_for_writer("third"), 6) + + def test_sequence_consistency(self): + """Test that we error out if the table and sequence diverges. + """ + + # Prefill with some rows + self._insert_row_with_id("master", 3) + + # Now we add a row *without* updating the stream ID + def _insert(txn): + txn.execute("INSERT INTO foobar VALUES (26, 'master')") + + self.get_success(self.db_pool.runInteraction("_insert", _insert)) + + # Creating the ID gen should error + with self.assertRaises(IncorrectDatabaseSetup): + self._create_id_generator("first") + + +class BackwardsMultiWriterIdGeneratorTestCase(HomeserverTestCase): + """Tests MultiWriterIdGenerator that produce *negative* stream IDs. + """ + + if not USE_POSTGRES_FOR_TESTS: + skip = "Requires Postgres" + + def prepare(self, reactor, clock, hs): + self.store = hs.get_datastore() + self.db_pool = self.store.db_pool # type: DatabasePool + + self.get_success(self.db_pool.runInteraction("_setup_db", self._setup_db)) + + def _setup_db(self, txn): + txn.execute("CREATE SEQUENCE foobar_seq") + txn.execute( + """ + CREATE TABLE foobar ( + stream_id BIGINT NOT NULL, + instance_name TEXT NOT NULL, + data TEXT + ); + """ + ) + + def _create_id_generator( + self, instance_name="master", writers=["master"] + ) -> MultiWriterIdGenerator: + def _create(conn): + return MultiWriterIdGenerator( + conn, + self.db_pool, + stream_name="test_stream", + instance_name=instance_name, + table="foobar", + instance_column="instance_name", + id_column="stream_id", + sequence_name="foobar_seq", + writers=writers, + positive=False, + ) + + return self.get_success(self.db_pool.runWithConnection(_create)) + + def _insert_row(self, instance_name: str, stream_id: int): + """Insert one row as the given instance with given stream_id. + """ + + def _insert(txn): + txn.execute( + "INSERT INTO foobar VALUES (?, ?)", (stream_id, instance_name,), + ) + txn.execute( + """ + INSERT INTO stream_positions VALUES ('test_stream', ?, ?) + ON CONFLICT (stream_name, instance_name) DO UPDATE SET stream_id = ? + """, + (instance_name, -stream_id, -stream_id), + ) + + self.get_success(self.db_pool.runInteraction("_insert_row", _insert)) + + def test_single_instance(self): + """Test that reads and writes from a single process are handled + correctly. + """ + id_gen = self._create_id_generator() + + async def _get_next_async(): + async with id_gen.get_next() as stream_id: + self._insert_row("master", stream_id) + + self.get_success(_get_next_async()) + + self.assertEqual(id_gen.get_positions(), {"master": -1}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), -1) + self.assertEqual(id_gen.get_persisted_upto_position(), -1) + + async def _get_next_async2(): + async with id_gen.get_next_mult(3) as stream_ids: + for stream_id in stream_ids: + self._insert_row("master", stream_id) + + self.get_success(_get_next_async2()) + + self.assertEqual(id_gen.get_positions(), {"master": -4}) + self.assertEqual(id_gen.get_current_token_for_writer("master"), -4) + self.assertEqual(id_gen.get_persisted_upto_position(), -4) + + # Test loading from DB by creating a second ID gen + second_id_gen = self._create_id_generator() + + self.assertEqual(second_id_gen.get_positions(), {"master": -4}) + self.assertEqual(second_id_gen.get_current_token_for_writer("master"), -4) + self.assertEqual(second_id_gen.get_persisted_upto_position(), -4) + + def test_multiple_instance(self): + """Tests that having multiple instances that get advanced over + federation works corretly. + """ + id_gen_1 = self._create_id_generator("first", writers=["first", "second"]) + id_gen_2 = self._create_id_generator("second", writers=["first", "second"]) + + async def _get_next_async(): + async with id_gen_1.get_next() as stream_id: + self._insert_row("first", stream_id) + id_gen_2.advance("first", stream_id) + + self.get_success(_get_next_async()) + + self.assertEqual(id_gen_1.get_positions(), {"first": -1}) + self.assertEqual(id_gen_2.get_positions(), {"first": -1}) + self.assertEqual(id_gen_1.get_persisted_upto_position(), -1) + self.assertEqual(id_gen_2.get_persisted_upto_position(), -1) + + async def _get_next_async2(): + async with id_gen_2.get_next() as stream_id: + self._insert_row("second", stream_id) + id_gen_1.advance("second", stream_id) + + self.get_success(_get_next_async2()) + + self.assertEqual(id_gen_1.get_positions(), {"first": -1, "second": -2}) + self.assertEqual(id_gen_2.get_positions(), {"first": -1, "second": -2}) + self.assertEqual(id_gen_1.get_persisted_upto_position(), -2) + self.assertEqual(id_gen_2.get_persisted_upto_position(), -2) diff --git a/tests/storage/test_main.py b/tests/storage/test_main.py
index 0155ffd04e..fe37d2ed5a 100644 --- a/tests/storage/test_main.py +++ b/tests/storage/test_main.py
@@ -34,14 +34,16 @@ class DataStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_get_users_paginate(self): - yield self.store.register_user(self.user.to_string(), "pass") - yield self.store.create_profile(self.user.localpart) - yield self.store.set_profile_displayname( - self.user.localpart, self.displayname, 1 + yield defer.ensureDeferred( + self.store.register_user(self.user.to_string(), "pass") + ) + yield defer.ensureDeferred(self.store.create_profile(self.user.localpart)) + yield defer.ensureDeferred( + self.store.set_profile_displayname(self.user.localpart, self.displayname, 1) ) - users, total = yield self.store.get_users_paginate( - 0, 10, name="bc", guests=False + users, total = yield defer.ensureDeferred( + self.store.get_users_paginate(0, 10, name="bc", guests=False) ) self.assertEquals(1, total) diff --git a/tests/storage/test_monthly_active_users.py b/tests/storage/test_monthly_active_users.py
index 9c04e92577..8d97b6d4cd 100644 --- a/tests/storage/test_monthly_active_users.py +++ b/tests/storage/test_monthly_active_users.py
@@ -19,6 +19,7 @@ from twisted.internet import defer from synapse.api.constants import UserTypes from tests import unittest +from tests.test_utils import make_awaitable from tests.unittest import default_config, override_config FORTY_DAYS = 40 * 24 * 60 * 60 @@ -78,7 +79,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): # XXX why are we doing this here? this function is only run at startup # so it is odd to re-run it here. self.get_success( - self.store.db.runInteraction( + self.store.db_pool.runInteraction( "initialise", self.store._initialise_reserved_users, threepids ) ) @@ -136,6 +137,21 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): count = self.get_success(self.store.get_monthly_active_count()) self.assertEqual(count, 1) + def test_appservice_user_not_counted_in_mau(self): + self.get_success( + self.store.register_user( + user_id="@appservice_user:server", appservice_id="wibble" + ) + ) + count = self.get_success(self.store.get_monthly_active_count()) + self.assertEqual(count, 0) + + d = self.store.upsert_monthly_active_user("@appservice_user:server") + self.get_success(d) + + count = self.get_success(self.store.get_monthly_active_count()) + self.assertEqual(count, 0) + def test_user_last_seen_monthly_active(self): user_id1 = "@user1:server" user_id2 = "@user2:server" @@ -204,7 +220,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): self.store.user_add_threepid(user, "email", email, now, now) ) - d = self.store.db.runInteraction( + d = self.store.db_pool.runInteraction( "initialise", self.store._initialise_reserved_users, threepids ) self.get_success(d) @@ -230,7 +246,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): ) self.get_success(d) - self.store.upsert_monthly_active_user = Mock() + self.store.upsert_monthly_active_user = Mock(return_value=make_awaitable(None)) d = self.store.populate_monthly_active_users(user_id) self.get_success(d) @@ -238,7 +254,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): self.store.upsert_monthly_active_user.assert_not_called() def test_populate_monthly_users_should_update(self): - self.store.upsert_monthly_active_user = Mock() + self.store.upsert_monthly_active_user = Mock(return_value=make_awaitable(None)) self.store.is_trial_user = Mock(return_value=defer.succeed(False)) @@ -251,7 +267,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): self.store.upsert_monthly_active_user.assert_called_once() def test_populate_monthly_users_should_not_update(self): - self.store.upsert_monthly_active_user = Mock() + self.store.upsert_monthly_active_user = Mock(return_value=make_awaitable(None)) self.store.is_trial_user = Mock(return_value=defer.succeed(False)) self.store.user_last_seen_monthly_active = Mock( @@ -280,7 +296,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): ] self.hs.config.mau_limits_reserved_threepids = threepids - d = self.store.db.runInteraction( + d = self.store.db_pool.runInteraction( "initialise", self.store._initialise_reserved_users, threepids ) self.get_success(d) @@ -293,8 +309,12 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): self.get_success(self.store.register_user(user_id=user2, password_hash=None)) now = int(self.hs.get_clock().time_msec()) - self.store.user_add_threepid(user1, "email", user1_email, now, now) - self.store.user_add_threepid(user2, "email", user2_email, now, now) + self.get_success( + self.store.user_add_threepid(user1, "email", user1_email, now, now) + ) + self.get_success( + self.store.user_add_threepid(user2, "email", user2_email, now, now) + ) users = self.get_success(self.store.get_registered_reserved_users()) self.assertEqual(len(users), len(threepids)) @@ -333,7 +353,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): @override_config({"limit_usage_by_mau": False, "mau_stats_only": False}) def test_no_users_when_not_tracking(self): - self.store.upsert_monthly_active_user = Mock() + self.store.upsert_monthly_active_user = Mock(return_value=make_awaitable(None)) self.get_success(self.store.populate_monthly_active_users("@user:sever")) @@ -378,7 +398,7 @@ class MonthlyActiveUsersTestCase(unittest.HomeserverTestCase): self.get_success(self.store.upsert_monthly_active_user(appservice2_user1)) count = self.get_success(self.store.get_monthly_active_count()) - self.assertEqual(count, 4) + self.assertEqual(count, 1) d = self.store.get_monthly_active_count_by_service() result = self.get_success(d) diff --git a/tests/storage/test_profile.py b/tests/storage/test_profile.py
index 7458a37e54..7a38022e71 100644 --- a/tests/storage/test_profile.py +++ b/tests/storage/test_profile.py
@@ -33,19 +33,36 @@ class ProfileStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_displayname(self): - yield self.store.set_profile_displayname(self.u_frank.localpart, "Frank", 1) + yield defer.ensureDeferred(self.store.create_profile(self.u_frank.localpart)) + + yield defer.ensureDeferred( + self.store.set_profile_displayname(self.u_frank.localpart, "Frank", 1) + ) self.assertEquals( - "Frank", (yield self.store.get_profile_displayname(self.u_frank.localpart)) + "Frank", + ( + yield defer.ensureDeferred( + self.store.get_profile_displayname(self.u_frank.localpart) + ) + ), ) @defer.inlineCallbacks def test_avatar_url(self): - yield self.store.set_profile_avatar_url( - self.u_frank.localpart, "http://my.site/here", 1 + yield defer.ensureDeferred(self.store.create_profile(self.u_frank.localpart)) + + yield defer.ensureDeferred( + self.store.set_profile_avatar_url( + self.u_frank.localpart, "http://my.site/here", 1 + ) ) self.assertEquals( "http://my.site/here", - (yield self.store.get_profile_avatar_url(self.u_frank.localpart)), + ( + yield defer.ensureDeferred( + self.store.get_profile_avatar_url(self.u_frank.localpart) + ) + ), ) diff --git a/tests/storage/test_purge.py b/tests/storage/test_purge.py
index b9fafaa1a6..cc1f3c53c5 100644 --- a/tests/storage/test_purge.py +++ b/tests/storage/test_purge.py
@@ -13,6 +13,9 @@ # See the License for the specific language governing permissions and # limitations under the License. +from twisted.internet import defer + +from synapse.api.errors import NotFoundError from synapse.rest.client.v1 import room from tests.unittest import HomeserverTestCase @@ -44,28 +47,22 @@ class PurgeTests(HomeserverTestCase): storage = self.hs.get_storage() # Get the topological token - event = store.get_topological_token_for_event(last["event_id"]) - self.pump() - event = self.successResultOf(event) + token = self.get_success( + store.get_topological_token_for_event(last["event_id"]) + ) + token_str = self.get_success(token.to_string(self.hs.get_datastore())) # Purge everything before this topological token - purge = storage.purge_events.purge_history(self.room_id, event, True) - self.pump() - self.assertEqual(self.successResultOf(purge), None) - - # Try and get the events - get_first = store.get_event(first["event_id"]) - get_second = store.get_event(second["event_id"]) - get_third = store.get_event(third["event_id"]) - get_last = store.get_event(last["event_id"]) - self.pump() + self.get_success( + storage.purge_events.purge_history(self.room_id, token_str, True) + ) # 1-3 should fail and last will succeed, meaning that 1-3 are deleted # and last is not. - self.failureResultOf(get_first) - self.failureResultOf(get_second) - self.failureResultOf(get_third) - self.successResultOf(get_last) + self.get_failure(store.get_event(first["event_id"]), NotFoundError) + self.get_failure(store.get_event(second["event_id"]), NotFoundError) + self.get_failure(store.get_event(third["event_id"]), NotFoundError) + self.get_success(store.get_event(last["event_id"])) def test_purge_wont_delete_extrems(self): """ @@ -80,28 +77,19 @@ class PurgeTests(HomeserverTestCase): storage = self.hs.get_datastore() # Set the topological token higher than it should be - event = storage.get_topological_token_for_event(last["event_id"]) - self.pump() - event = self.successResultOf(event) - event = "t{}-{}".format( - *list(map(lambda x: x + 1, map(int, event[1:].split("-")))) + token = self.get_success( + storage.get_topological_token_for_event(last["event_id"]) ) + event = "t{}-{}".format(token.topological + 1, token.stream + 1) # Purge everything before this topological token - purge = storage.purge_history(self.room_id, event, True) + purge = defer.ensureDeferred(storage.purge_history(self.room_id, event, True)) self.pump() f = self.failureResultOf(purge) self.assertIn("greater than forward", f.value.args[0]) # Try and get the events - get_first = storage.get_event(first["event_id"]) - get_second = storage.get_event(second["event_id"]) - get_third = storage.get_event(third["event_id"]) - get_last = storage.get_event(last["event_id"]) - self.pump() - - # Nothing is deleted. - self.successResultOf(get_first) - self.successResultOf(get_second) - self.successResultOf(get_third) - self.successResultOf(get_last) + self.get_success(storage.get_event(first["event_id"])) + self.get_success(storage.get_event(second["event_id"])) + self.get_success(storage.get_event(third["event_id"])) + self.get_success(storage.get_event(last["event_id"])) diff --git a/tests/storage/test_redaction.py b/tests/storage/test_redaction.py
index db3667dc43..1ea35d60c1 100644 --- a/tests/storage/test_redaction.py +++ b/tests/storage/test_redaction.py
@@ -237,7 +237,9 @@ class RedactionTestCase(unittest.HomeserverTestCase): @defer.inlineCallbacks def build(self, prev_event_ids): - built_event = yield self._base_builder.build(prev_event_ids) + built_event = yield defer.ensureDeferred( + self._base_builder.build(prev_event_ids) + ) built_event._event_id = self._event_id built_event._dict["event_id"] = self._event_id @@ -249,6 +251,10 @@ class RedactionTestCase(unittest.HomeserverTestCase): def room_id(self): return self._base_builder.room_id + @property + def type(self): + return self._base_builder.type + event_1, context_1 = self.get_success( self.event_creation_handler.create_new_client_event( EventIdManglingBuilder( @@ -341,7 +347,7 @@ class RedactionTestCase(unittest.HomeserverTestCase): ) event_json = self.get_success( - self.store.db.simple_select_one_onecol( + self.store.db_pool.simple_select_one_onecol( table="event_json", keyvalues={"event_id": msg_event.event_id}, retcol="json", @@ -359,7 +365,7 @@ class RedactionTestCase(unittest.HomeserverTestCase): self.reactor.advance(60 * 60 * 2) event_json = self.get_success( - self.store.db.simple_select_one_onecol( + self.store.db_pool.simple_select_one_onecol( table="event_json", keyvalues={"event_id": msg_event.event_id}, retcol="json", diff --git a/tests/storage/test_registration.py b/tests/storage/test_registration.py
index 71a40a0a49..6b582771fe 100644 --- a/tests/storage/test_registration.py +++ b/tests/storage/test_registration.py
@@ -17,6 +17,7 @@ from twisted.internet import defer from synapse.api.constants import UserTypes +from synapse.api.errors import ThreepidValidationError from tests import unittest from tests.utils import setup_test_homeserver @@ -36,7 +37,7 @@ class RegistrationStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_register(self): - yield self.store.register_user(self.user_id, self.pwhash) + yield defer.ensureDeferred(self.store.register_user(self.user_id, self.pwhash)) self.assertEquals( { @@ -52,17 +53,21 @@ class RegistrationStoreTestCase(unittest.TestCase): "user_type": None, "deactivated": 0, }, - (yield self.store.get_user_by_id(self.user_id)), + (yield defer.ensureDeferred(self.store.get_user_by_id(self.user_id))), ) @defer.inlineCallbacks def test_add_tokens(self): - yield self.store.register_user(self.user_id, self.pwhash) - yield self.store.add_access_token_to_user( - self.user_id, self.tokens[1], self.device_id, valid_until_ms=None + yield defer.ensureDeferred(self.store.register_user(self.user_id, self.pwhash)) + yield defer.ensureDeferred( + self.store.add_access_token_to_user( + self.user_id, self.tokens[1], self.device_id, valid_until_ms=None + ) ) - result = yield self.store.get_user_by_access_token(self.tokens[1]) + result = yield defer.ensureDeferred( + self.store.get_user_by_access_token(self.tokens[1]) + ) self.assertDictContainsSubset( {"name": self.user_id, "device_id": self.device_id}, result @@ -73,31 +78,41 @@ class RegistrationStoreTestCase(unittest.TestCase): @defer.inlineCallbacks def test_user_delete_access_tokens(self): # add some tokens - yield self.store.register_user(self.user_id, self.pwhash) - yield self.store.add_access_token_to_user( - self.user_id, self.tokens[0], device_id=None, valid_until_ms=None + yield defer.ensureDeferred(self.store.register_user(self.user_id, self.pwhash)) + yield defer.ensureDeferred( + self.store.add_access_token_to_user( + self.user_id, self.tokens[0], device_id=None, valid_until_ms=None + ) ) - yield self.store.add_access_token_to_user( - self.user_id, self.tokens[1], self.device_id, valid_until_ms=None + yield defer.ensureDeferred( + self.store.add_access_token_to_user( + self.user_id, self.tokens[1], self.device_id, valid_until_ms=None + ) ) # now delete some - yield self.store.user_delete_access_tokens( - self.user_id, device_id=self.device_id + yield defer.ensureDeferred( + self.store.user_delete_access_tokens(self.user_id, device_id=self.device_id) ) # check they were deleted - user = yield self.store.get_user_by_access_token(self.tokens[1]) + user = yield defer.ensureDeferred( + self.store.get_user_by_access_token(self.tokens[1]) + ) self.assertIsNone(user, "access token was not deleted by device_id") # check the one not associated with the device was not deleted - user = yield self.store.get_user_by_access_token(self.tokens[0]) + user = yield defer.ensureDeferred( + self.store.get_user_by_access_token(self.tokens[0]) + ) self.assertEqual(self.user_id, user["name"]) # now delete the rest - yield self.store.user_delete_access_tokens(self.user_id) + yield defer.ensureDeferred(self.store.user_delete_access_tokens(self.user_id)) - user = yield self.store.get_user_by_access_token(self.tokens[0]) + user = yield defer.ensureDeferred( + self.store.get_user_by_access_token(self.tokens[0]) + ) self.assertIsNone(user, "access token was not deleted without device_id") @defer.inlineCallbacks @@ -105,14 +120,48 @@ class RegistrationStoreTestCase(unittest.TestCase): TEST_USER = "@test:test" SUPPORT_USER = "@support:test" - res = yield self.store.is_support_user(None) + res = yield defer.ensureDeferred(self.store.is_support_user(None)) self.assertFalse(res) - yield self.store.register_user(user_id=TEST_USER, password_hash=None) - res = yield self.store.is_support_user(TEST_USER) + yield defer.ensureDeferred( + self.store.register_user(user_id=TEST_USER, password_hash=None) + ) + res = yield defer.ensureDeferred(self.store.is_support_user(TEST_USER)) self.assertFalse(res) - yield self.store.register_user( - user_id=SUPPORT_USER, password_hash=None, user_type=UserTypes.SUPPORT + yield defer.ensureDeferred( + self.store.register_user( + user_id=SUPPORT_USER, password_hash=None, user_type=UserTypes.SUPPORT + ) ) - res = yield self.store.is_support_user(SUPPORT_USER) + res = yield defer.ensureDeferred(self.store.is_support_user(SUPPORT_USER)) self.assertTrue(res) + + @defer.inlineCallbacks + def test_3pid_inhibit_invalid_validation_session_error(self): + """Tests that enabling the configuration option to inhibit 3PID errors on + /requestToken also inhibits validation errors caused by an unknown session ID. + """ + + # Check that, with the config setting set to false (the default value), a + # validation error is caused by the unknown session ID. + try: + yield defer.ensureDeferred( + self.store.validate_threepid_session( + "fake_sid", "fake_client_secret", "fake_token", 0, + ) + ) + except ThreepidValidationError as e: + self.assertEquals(e.msg, "Unknown session_id", e) + + # Set the config setting to true. + self.store._ignore_unknown_session_error = True + + # Check that now the validation error is caused by the token not matching. + try: + yield defer.ensureDeferred( + self.store.validate_threepid_session( + "fake_sid", "fake_client_secret", "fake_token", 0, + ) + ) + except ThreepidValidationError as e: + self.assertEquals(e.msg, "Validation token not found or has expired", e) diff --git a/tests/storage/test_room.py b/tests/storage/test_room.py
index 1d77b4a2d6..bc8400f240 100644 --- a/tests/storage/test_room.py +++ b/tests/storage/test_room.py
@@ -37,11 +37,13 @@ class RoomStoreTestCase(unittest.TestCase): self.alias = RoomAlias.from_string("#a-room-name:test") self.u_creator = UserID.from_string("@creator:test") - yield self.store.store_room( - self.room.to_string(), - room_creator_user_id=self.u_creator.to_string(), - is_public=True, - room_version=RoomVersions.V1, + yield defer.ensureDeferred( + self.store.store_room( + self.room.to_string(), + room_creator_user_id=self.u_creator.to_string(), + is_public=True, + room_version=RoomVersions.V1, + ) ) @defer.inlineCallbacks @@ -52,12 +54,14 @@ class RoomStoreTestCase(unittest.TestCase): "creator": self.u_creator.to_string(), "is_public": True, }, - (yield self.store.get_room(self.room.to_string())), + (yield defer.ensureDeferred(self.store.get_room(self.room.to_string()))), ) @defer.inlineCallbacks def test_get_room_unknown_room(self): - self.assertIsNone((yield self.store.get_room("!uknown:test")),) + self.assertIsNone( + (yield defer.ensureDeferred(self.store.get_room("!uknown:test"))) + ) @defer.inlineCallbacks def test_get_room_with_stats(self): @@ -67,12 +71,22 @@ class RoomStoreTestCase(unittest.TestCase): "creator": self.u_creator.to_string(), "public": True, }, - (yield self.store.get_room_with_stats(self.room.to_string())), + ( + yield defer.ensureDeferred( + self.store.get_room_with_stats(self.room.to_string()) + ) + ), ) @defer.inlineCallbacks def test_get_room_with_stats_unknown_room(self): - self.assertIsNone((yield self.store.get_room_with_stats("!uknown:test")),) + self.assertIsNone( + ( + yield defer.ensureDeferred( + self.store.get_room_with_stats("!uknown:test") + ) + ), + ) class RoomEventsStoreTestCase(unittest.TestCase): @@ -88,17 +102,21 @@ class RoomEventsStoreTestCase(unittest.TestCase): self.room = RoomID.from_string("!abcde:test") - yield self.store.store_room( - self.room.to_string(), - room_creator_user_id="@creator:text", - is_public=True, - room_version=RoomVersions.V1, + yield defer.ensureDeferred( + self.store.store_room( + self.room.to_string(), + room_creator_user_id="@creator:text", + is_public=True, + room_version=RoomVersions.V1, + ) ) @defer.inlineCallbacks def inject_room_event(self, **kwargs): - yield self.storage.persistence.persist_event( - self.event_factory.create_event(room_id=self.room.to_string(), **kwargs) + yield defer.ensureDeferred( + self.storage.persistence.persist_event( + self.event_factory.create_event(room_id=self.room.to_string(), **kwargs) + ) ) @defer.inlineCallbacks diff --git a/tests/storage/test_roommember.py b/tests/storage/test_roommember.py
index f282921538..12ccc1f53e 100644 --- a/tests/storage/test_roommember.py +++ b/tests/storage/test_roommember.py
@@ -87,7 +87,7 @@ class RoomMemberStoreTestCase(unittest.HomeserverTestCase): self.inject_room_member(self.room, self.u_bob, Membership.JOIN) self.inject_room_member(self.room, self.u_charlie.to_string(), Membership.JOIN) - self.pump(20) + self.pump() self.assertTrue("_known_servers_count" not in self.store.__dict__.keys()) @@ -101,7 +101,7 @@ class RoomMemberStoreTestCase(unittest.HomeserverTestCase): # Initialises to 1 -- itself self.assertEqual(self.store._known_servers_count, 1) - self.pump(20) + self.pump() # No rooms have been joined, so technically the SQL returns 0, but it # will still say it knows about itself. @@ -111,7 +111,7 @@ class RoomMemberStoreTestCase(unittest.HomeserverTestCase): self.inject_room_member(self.room, self.u_bob, Membership.JOIN) self.inject_room_member(self.room, self.u_charlie.to_string(), Membership.JOIN) - self.pump(20) + self.pump(1) # It now knows about Charlie's server. self.assertEqual(self.store._known_servers_count, 2) @@ -179,20 +179,20 @@ class CurrentStateMembershipUpdateTestCase(unittest.HomeserverTestCase): def test_can_rerun_update(self): # First make sure we have completed all updates. while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) # Now let's create a room, which will insert a membership user = UserID("alice", "test") - requester = Requester(user, None, False, None, None) + requester = Requester(user, None, False, False, None, None) self.get_success(self.room_creator.create_room(requester, {})) # Register the background update to run again. self.get_success( - self.store.db.simple_insert( + self.store.db_pool.simple_insert( table="background_updates", values={ "update_name": "current_state_events_membership", @@ -203,12 +203,12 @@ class CurrentStateMembershipUpdateTestCase(unittest.HomeserverTestCase): ) # ... and tell the DataStore that it hasn't finished all updates yet - self.store.db.updates._all_done = False + self.store.db_pool.updates._all_done = False # Now let's actually drive the updates to completion while not self.get_success( - self.store.db.updates.has_completed_background_updates() + self.store.db_pool.updates.has_completed_background_updates() ): self.get_success( - self.store.db.updates.do_next_background_update(100), by=0.1 + self.store.db_pool.updates.do_next_background_update(100), by=0.1 ) diff --git a/tests/storage/test_state.py b/tests/storage/test_state.py
index a0e133cd4a..8bd12fa847 100644 --- a/tests/storage/test_state.py +++ b/tests/storage/test_state.py
@@ -44,11 +44,13 @@ class StateStoreTestCase(tests.unittest.TestCase): self.room = RoomID.from_string("!abc123:test") - yield self.store.store_room( - self.room.to_string(), - room_creator_user_id="@creator:text", - is_public=True, - room_version=RoomVersions.V1, + yield defer.ensureDeferred( + self.store.store_room( + self.room.to_string(), + room_creator_user_id="@creator:text", + is_public=True, + room_version=RoomVersions.V1, + ) ) @defer.inlineCallbacks @@ -68,7 +70,9 @@ class StateStoreTestCase(tests.unittest.TestCase): self.event_creation_handler.create_new_client_event(builder) ) - yield self.storage.persistence.persist_event(event, context) + yield defer.ensureDeferred( + self.storage.persistence.persist_event(event, context) + ) return event @@ -87,8 +91,8 @@ class StateStoreTestCase(tests.unittest.TestCase): self.room, self.u_alice, EventTypes.Name, "", {"name": "test room"} ) - state_group_map = yield self.storage.state.get_state_groups_ids( - self.room, [e2.event_id] + state_group_map = yield defer.ensureDeferred( + self.storage.state.get_state_groups_ids(self.room, [e2.event_id]) ) self.assertEqual(len(state_group_map), 1) state_map = list(state_group_map.values())[0] @@ -106,8 +110,8 @@ class StateStoreTestCase(tests.unittest.TestCase): self.room, self.u_alice, EventTypes.Name, "", {"name": "test room"} ) - state_group_map = yield self.storage.state.get_state_groups( - self.room, [e2.event_id] + state_group_map = yield defer.ensureDeferred( + self.storage.state.get_state_groups(self.room, [e2.event_id]) ) self.assertEqual(len(state_group_map), 1) state_list = list(state_group_map.values())[0] @@ -148,7 +152,9 @@ class StateStoreTestCase(tests.unittest.TestCase): ) # check we get the full state as of the final event - state = yield self.storage.state.get_state_for_event(e5.event_id) + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event(e5.event_id) + ) self.assertIsNotNone(e4) @@ -164,22 +170,28 @@ class StateStoreTestCase(tests.unittest.TestCase): ) # check we can filter to the m.room.name event (with a '' state key) - state = yield self.storage.state.get_state_for_event( - e5.event_id, StateFilter.from_types([(EventTypes.Name, "")]) + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event( + e5.event_id, StateFilter.from_types([(EventTypes.Name, "")]) + ) ) self.assertStateMapEqual({(e2.type, e2.state_key): e2}, state) # check we can filter to the m.room.name event (with a wildcard None state key) - state = yield self.storage.state.get_state_for_event( - e5.event_id, StateFilter.from_types([(EventTypes.Name, None)]) + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event( + e5.event_id, StateFilter.from_types([(EventTypes.Name, None)]) + ) ) self.assertStateMapEqual({(e2.type, e2.state_key): e2}, state) # check we can grab the m.room.member events (with a wildcard None state key) - state = yield self.storage.state.get_state_for_event( - e5.event_id, StateFilter.from_types([(EventTypes.Member, None)]) + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event( + e5.event_id, StateFilter.from_types([(EventTypes.Member, None)]) + ) ) self.assertStateMapEqual( @@ -188,12 +200,14 @@ class StateStoreTestCase(tests.unittest.TestCase): # check we can grab a specific room member without filtering out the # other event types - state = yield self.storage.state.get_state_for_event( - e5.event_id, - state_filter=StateFilter( - types={EventTypes.Member: {self.u_alice.to_string()}}, - include_others=True, - ), + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event( + e5.event_id, + state_filter=StateFilter( + types={EventTypes.Member: {self.u_alice.to_string()}}, + include_others=True, + ), + ) ) self.assertStateMapEqual( @@ -206,11 +220,13 @@ class StateStoreTestCase(tests.unittest.TestCase): ) # check that we can grab everything except members - state = yield self.storage.state.get_state_for_event( - e5.event_id, - state_filter=StateFilter( - types={EventTypes.Member: set()}, include_others=True - ), + state = yield defer.ensureDeferred( + self.storage.state.get_state_for_event( + e5.event_id, + state_filter=StateFilter( + types={EventTypes.Member: set()}, include_others=True + ), + ) ) self.assertStateMapEqual( @@ -222,8 +238,8 @@ class StateStoreTestCase(tests.unittest.TestCase): ####################################################### room_id = self.room.to_string() - group_ids = yield self.storage.state.get_state_groups_ids( - room_id, [e5.event_id] + group_ids = yield defer.ensureDeferred( + self.storage.state.get_state_groups_ids(room_id, [e5.event_id]) ) group = list(group_ids.keys())[0] diff --git a/tests/storage/test_user_directory.py b/tests/storage/test_user_directory.py
index 6a545d2eb0..738e912468 100644 --- a/tests/storage/test_user_directory.py +++ b/tests/storage/test_user_directory.py
@@ -31,16 +31,24 @@ class UserDirectoryStoreTestCase(unittest.TestCase): # alice and bob are both in !room_id. bobby is not but shares # a homeserver with alice. - yield self.store.update_profile_in_user_dir(ALICE, "alice", None) - yield self.store.update_profile_in_user_dir(BOB, "bob", None) - yield self.store.update_profile_in_user_dir(BOBBY, "bobby", None) - yield self.store.add_users_in_public_rooms("!room:id", (ALICE, BOB)) + yield defer.ensureDeferred( + self.store.update_profile_in_user_dir(ALICE, "alice", None) + ) + yield defer.ensureDeferred( + self.store.update_profile_in_user_dir(BOB, "bob", None) + ) + yield defer.ensureDeferred( + self.store.update_profile_in_user_dir(BOBBY, "bobby", None) + ) + yield defer.ensureDeferred( + self.store.add_users_in_public_rooms("!room:id", (ALICE, BOB)) + ) @defer.inlineCallbacks def test_search_user_dir(self): # normally when alice searches the directory she should just find # bob because bobby doesn't share a room with her. - r = yield self.store.search_user_dir(ALICE, "bob", 10) + r = yield defer.ensureDeferred(self.store.search_user_dir(ALICE, "bob", 10)) self.assertFalse(r["limited"]) self.assertEqual(1, len(r["results"])) self.assertDictEqual( @@ -51,7 +59,7 @@ class UserDirectoryStoreTestCase(unittest.TestCase): def test_search_user_dir_all_users(self): self.hs.config.user_directory_search_all_users = True try: - r = yield self.store.search_user_dir(ALICE, "bob", 10) + r = yield defer.ensureDeferred(self.store.search_user_dir(ALICE, "bob", 10)) self.assertFalse(r["limited"]) self.assertEqual(2, len(r["results"])) self.assertDictEqual( diff --git a/tests/test_federation.py b/tests/test_federation.py
index 87a16d7d7a..27a7fc9ed7 100644 --- a/tests/test_federation.py +++ b/tests/test_federation.py
@@ -1,7 +1,23 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 The Matrix.org Foundation C.I.C. +# +# Licensed under the Apache License, Version 2.0 (the "License"); +# you may not use this file except in compliance with the License. +# You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + from mock import Mock -from twisted.internet.defer import ensureDeferred, maybeDeferred, succeed +from twisted.internet.defer import succeed +from synapse.api.errors import FederationError from synapse.events import make_event_from_dict from synapse.logging.context import LoggingContext from synapse.types import Requester, UserID @@ -10,6 +26,7 @@ from synapse.util.retryutils import NotRetryingDestination from tests import unittest from tests.server import ThreadedMemoryReactorClock, setup_test_homeserver +from tests.test_utils import make_awaitable class MessageAcceptTests(unittest.HomeserverTestCase): @@ -26,24 +43,19 @@ class MessageAcceptTests(unittest.HomeserverTestCase): ) user_id = UserID("us", "test") - our_user = Requester(user_id, None, False, None, None) + our_user = Requester(user_id, None, False, False, None, None) room_creator = self.homeserver.get_room_creation_handler() - room_deferred = ensureDeferred( + self.room_id = self.get_success( room_creator.create_room( our_user, room_creator._presets_dict["public_chat"], ratelimit=False ) - ) - self.reactor.advance(0.1) - self.room_id = self.successResultOf(room_deferred)[0]["room_id"] + )[0]["room_id"] self.store = self.homeserver.get_datastore() # Figure out what the most recent event is - most_recent = self.successResultOf( - maybeDeferred( - self.homeserver.get_datastore().get_latest_event_ids_in_room, - self.room_id, - ) + most_recent = self.get_success( + self.homeserver.get_datastore().get_latest_event_ids_in_room(self.room_id) )[0] join_event = make_event_from_dict( @@ -73,19 +85,18 @@ class MessageAcceptTests(unittest.HomeserverTestCase): ) # Send the join, it should return None (which is not an error) - d = ensureDeferred( - self.handler.on_receive_pdu( - "test.serv", join_event, sent_to_us_directly=True - ) + self.assertEqual( + self.get_success( + self.handler.on_receive_pdu( + "test.serv", join_event, sent_to_us_directly=True + ) + ), + None, ) - self.reactor.advance(1) - self.assertEqual(self.successResultOf(d), None) # Make sure we actually joined the room self.assertEqual( - self.successResultOf( - maybeDeferred(self.store.get_latest_event_ids_in_room, self.room_id) - )[0], + self.get_success(self.store.get_latest_event_ids_in_room(self.room_id))[0], "$join:test.serv", ) @@ -95,7 +106,7 @@ class MessageAcceptTests(unittest.HomeserverTestCase): prev_events that said event references. """ - def post_json(destination, path, data, headers=None, timeout=0): + async def post_json(destination, path, data, headers=None, timeout=0): # If it asks us for new missing events, give them NOTHING if path.startswith("/_matrix/federation/v1/get_missing_events/"): return {"events": []} @@ -103,8 +114,8 @@ class MessageAcceptTests(unittest.HomeserverTestCase): self.http_client.post_json = post_json # Figure out what the most recent event is - most_recent = self.successResultOf( - maybeDeferred(self.store.get_latest_event_ids_in_room, self.room_id) + most_recent = self.get_success( + self.store.get_latest_event_ids_in_room(self.room_id) )[0] # Now lie about an event @@ -124,17 +135,14 @@ class MessageAcceptTests(unittest.HomeserverTestCase): ) with LoggingContext(request="lying_event"): - d = ensureDeferred( + failure = self.get_failure( self.handler.on_receive_pdu( "test.serv", lying_event, sent_to_us_directly=True - ) + ), + FederationError, ) - # Step the reactor, so the database fetches come back - self.reactor.advance(1) - # on_receive_pdu should throw an error - failure = self.failureResultOf(d) self.assertEqual( failure.value.args[0], ( @@ -144,8 +152,8 @@ class MessageAcceptTests(unittest.HomeserverTestCase): ) # Make sure the invalid event isn't there - extrem = maybeDeferred(self.store.get_latest_event_ids_in_room, self.room_id) - self.assertEqual(self.successResultOf(extrem)[0], "$join:test.serv") + extrem = self.get_success(self.store.get_latest_event_ids_in_room(self.room_id)) + self.assertEqual(extrem[0], "$join:test.serv") def test_retry_device_list_resync(self): """Tests that device lists are marked as stale if they couldn't be synced, and @@ -173,7 +181,7 @@ class MessageAcceptTests(unittest.HomeserverTestCase): # Register a mock on the store so that the incoming update doesn't fail because # we don't share a room with the user. store = self.homeserver.get_datastore() - store.get_rooms_for_user = Mock(return_value=succeed(["!someroom:test"])) + store.get_rooms_for_user = Mock(return_value=make_awaitable(["!someroom:test"])) # Manually inject a fake device list update. We need this update to include at # least one prev_id so that the user's device list will need to be retried. diff --git a/tests/test_server.py b/tests/test_server.py
index 073b2362cc..655c918a15 100644 --- a/tests/test_server.py +++ b/tests/test_server.py
@@ -157,6 +157,28 @@ class JsonResourceTests(unittest.TestCase): self.assertEqual(channel.json_body["error"], "Unrecognized request") self.assertEqual(channel.json_body["errcode"], "M_UNRECOGNIZED") + def test_head_request(self): + """ + JsonResource.handler_for_request gives correctly decoded URL args to + the callback, while Twisted will give the raw bytes of URL query + arguments. + """ + + def _callback(request, **kwargs): + return 200, {"result": True} + + res = JsonResource(self.homeserver) + res.register_paths( + "GET", [re.compile("^/_matrix/foo$")], _callback, "test_servlet", + ) + + # The path was registered as GET, but this is a HEAD request. + request, channel = make_request(self.reactor, b"HEAD", b"/_matrix/foo") + render(request, res, self.reactor) + + self.assertEqual(channel.result["code"], b"200") + self.assertNotIn("body", channel.result) + class OptionsResourceTests(unittest.TestCase): def setUp(self): @@ -255,7 +277,7 @@ class WrapHtmlRequestHandlerTests(unittest.TestCase): self.reactor = ThreadedMemoryReactorClock() def test_good_response(self): - def callback(request): + async def callback(request): request.write(b"response") request.finish() @@ -275,7 +297,7 @@ class WrapHtmlRequestHandlerTests(unittest.TestCase): with the right location. """ - def callback(request, **kwargs): + async def callback(request, **kwargs): raise RedirectException(b"/look/an/eagle", 301) res = WrapHtmlRequestHandlerTests.TestResource() @@ -295,7 +317,7 @@ class WrapHtmlRequestHandlerTests(unittest.TestCase): returned too """ - def callback(request, **kwargs): + async def callback(request, **kwargs): e = RedirectException(b"/no/over/there", 304) e.cookies.append(b"session=yespls") raise e @@ -312,3 +334,19 @@ class WrapHtmlRequestHandlerTests(unittest.TestCase): self.assertEqual(location_headers, [b"/no/over/there"]) cookies_headers = [v for k, v in headers if k == b"Set-Cookie"] self.assertEqual(cookies_headers, [b"session=yespls"]) + + def test_head_request(self): + """A head request should work by being turned into a GET request.""" + + async def callback(request): + request.write(b"response") + request.finish() + + res = WrapHtmlRequestHandlerTests.TestResource() + res.callback = callback + + request, channel = make_request(self.reactor, b"HEAD", b"/path") + render(request, res, self.reactor) + + self.assertEqual(channel.result["code"], b"200") + self.assertNotIn("body", channel.result) diff --git a/tests/test_state.py b/tests/test_state.py
index 4858e8fc59..80b0ccbc40 100644 --- a/tests/test_state.py +++ b/tests/test_state.py
@@ -71,7 +71,7 @@ def create_event( return event -class StateGroupStore(object): +class StateGroupStore: def __init__(self): self._event_to_state_group = {} self._group_to_state = {} @@ -80,16 +80,16 @@ class StateGroupStore(object): self._next_group = 1 - def get_state_groups_ids(self, room_id, event_ids): + async def get_state_groups_ids(self, room_id, event_ids): groups = {} for event_id in event_ids: group = self._event_to_state_group.get(event_id) if group: groups[group] = self._group_to_state[group] - return defer.succeed(groups) + return groups - def store_state_group( + async def store_state_group( self, event_id, room_id, prev_group, delta_ids, current_state_ids ): state_group = self._next_group @@ -97,19 +97,17 @@ class StateGroupStore(object): self._group_to_state[state_group] = dict(current_state_ids) - return defer.succeed(state_group) + return state_group - def get_events(self, event_ids, **kwargs): - return defer.succeed( - { - e_id: self._event_id_to_event[e_id] - for e_id in event_ids - if e_id in self._event_id_to_event - } - ) + async def get_events(self, event_ids, **kwargs): + return { + e_id: self._event_id_to_event[e_id] + for e_id in event_ids + if e_id in self._event_id_to_event + } - def get_state_group_delta(self, name): - return defer.succeed((None, None)) + async def get_state_group_delta(self, name): + return (None, None) def register_events(self, events): for e in events: @@ -121,17 +119,17 @@ class StateGroupStore(object): def register_event_id_state_group(self, event_id, state_group): self._event_to_state_group[event_id] = state_group - def get_room_version_id(self, room_id): - return defer.succeed(RoomVersions.V1.identifier) + async def get_room_version_id(self, room_id): + return RoomVersions.V1.identifier class DictObj(dict): def __init__(self, **kwargs): - super(DictObj, self).__init__(kwargs) + super().__init__(kwargs) self.__dict__ = self -class Graph(object): +class Graph: def __init__(self, nodes, edges): events = {} clobbered = set(events.keys()) @@ -213,7 +211,7 @@ class StateTestCase(unittest.TestCase): ctx_c = context_store["C"] ctx_d = context_store["D"] - prev_state_ids = yield ctx_d.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(ctx_d.get_prev_state_ids()) self.assertEqual(2, len(prev_state_ids)) self.assertEqual(ctx_c.state_group, ctx_d.state_group_before_event) @@ -259,7 +257,7 @@ class StateTestCase(unittest.TestCase): ctx_c = context_store["C"] ctx_d = context_store["D"] - prev_state_ids = yield ctx_d.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(ctx_d.get_prev_state_ids()) self.assertSetEqual({"START", "A", "C"}, set(prev_state_ids.values())) self.assertEqual(ctx_c.state_group, ctx_d.state_group_before_event) @@ -318,7 +316,7 @@ class StateTestCase(unittest.TestCase): ctx_c = context_store["C"] ctx_e = context_store["E"] - prev_state_ids = yield ctx_e.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(ctx_e.get_prev_state_ids()) self.assertSetEqual({"START", "A", "B", "C"}, set(prev_state_ids.values())) self.assertEqual(ctx_c.state_group, ctx_e.state_group_before_event) self.assertEqual(ctx_e.state_group_before_event, ctx_e.state_group) @@ -393,7 +391,7 @@ class StateTestCase(unittest.TestCase): ctx_b = context_store["B"] ctx_d = context_store["D"] - prev_state_ids = yield ctx_d.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(ctx_d.get_prev_state_ids()) self.assertSetEqual({"A1", "A2", "A3", "A5", "B"}, set(prev_state_ids.values())) self.assertEqual(ctx_b.state_group, ctx_d.state_group_before_event) @@ -425,7 +423,7 @@ class StateTestCase(unittest.TestCase): self.state.compute_event_context(event, old_state=old_state) ) - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(context.get_prev_state_ids()) self.assertCountEqual((e.event_id for e in old_state), prev_state_ids.values()) current_state_ids = yield defer.ensureDeferred(context.get_current_state_ids()) @@ -450,7 +448,7 @@ class StateTestCase(unittest.TestCase): self.state.compute_event_context(event, old_state=old_state) ) - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(context.get_prev_state_ids()) self.assertCountEqual((e.event_id for e in old_state), prev_state_ids.values()) current_state_ids = yield defer.ensureDeferred(context.get_current_state_ids()) @@ -476,12 +474,14 @@ class StateTestCase(unittest.TestCase): create_event(type="test2", state_key=""), ] - group_name = yield self.store.store_state_group( - prev_event_id, - event.room_id, - None, - None, - {(e.type, e.state_key): e.event_id for e in old_state}, + group_name = yield defer.ensureDeferred( + self.store.store_state_group( + prev_event_id, + event.room_id, + None, + None, + {(e.type, e.state_key): e.event_id for e in old_state}, + ) ) self.store.register_event_id_state_group(prev_event_id, group_name) @@ -508,18 +508,20 @@ class StateTestCase(unittest.TestCase): create_event(type="test2", state_key=""), ] - group_name = yield self.store.store_state_group( - prev_event_id, - event.room_id, - None, - None, - {(e.type, e.state_key): e.event_id for e in old_state}, + group_name = yield defer.ensureDeferred( + self.store.store_state_group( + prev_event_id, + event.room_id, + None, + None, + {(e.type, e.state_key): e.event_id for e in old_state}, + ) ) self.store.register_event_id_state_group(prev_event_id, group_name) context = yield defer.ensureDeferred(self.state.compute_event_context(event)) - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = yield defer.ensureDeferred(context.get_prev_state_ids()) self.assertEqual({e.event_id for e in old_state}, set(prev_state_ids.values())) @@ -691,21 +693,25 @@ class StateTestCase(unittest.TestCase): def _get_context( self, event, prev_event_id_1, old_state_1, prev_event_id_2, old_state_2 ): - sg1 = yield self.store.store_state_group( - prev_event_id_1, - event.room_id, - None, - None, - {(e.type, e.state_key): e.event_id for e in old_state_1}, + sg1 = yield defer.ensureDeferred( + self.store.store_state_group( + prev_event_id_1, + event.room_id, + None, + None, + {(e.type, e.state_key): e.event_id for e in old_state_1}, + ) ) self.store.register_event_id_state_group(prev_event_id_1, sg1) - sg2 = yield self.store.store_state_group( - prev_event_id_2, - event.room_id, - None, - None, - {(e.type, e.state_key): e.event_id for e in old_state_2}, + sg2 = yield defer.ensureDeferred( + self.store.store_state_group( + prev_event_id_2, + event.room_id, + None, + None, + {(e.type, e.state_key): e.event_id for e in old_state_2}, + ) ) self.store.register_event_id_state_group(prev_event_id_2, sg2) diff --git a/tests/test_utils/__init__.py b/tests/test_utils/__init__.py
index 508aeba078..a298cc0fd3 100644 --- a/tests/test_utils/__init__.py +++ b/tests/test_utils/__init__.py
@@ -17,6 +17,7 @@ """ Utilities for running the unit tests """ +from asyncio import Future from typing import Any, Awaitable, TypeVar TV = TypeVar("TV") @@ -38,6 +39,12 @@ def get_awaitable_result(awaitable: Awaitable[TV]) -> TV: raise Exception("awaitable has not yet completed") -async def make_awaitable(result: Any): - """Create an awaitable that just returns a result.""" - return result +def make_awaitable(result: Any) -> Awaitable[Any]: + """ + Makes an awaitable, suitable for mocking an `async` function. + This uses Futures as they can be awaited multiple times so can be returned + to multiple callers. + """ + future = Future() # type: ignore + future.set_result(result) + return future diff --git a/tests/test_utils/event_injection.py b/tests/test_utils/event_injection.py
index 8522c6fc09..e93aa84405 100644 --- a/tests/test_utils/event_injection.py +++ b/tests/test_utils/event_injection.py
@@ -13,14 +13,13 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. -from typing import Optional, Tuple +from typing import List, Optional, Tuple import synapse.server from synapse.api.constants import EventTypes from synapse.api.room_versions import KNOWN_ROOM_VERSIONS from synapse.events import EventBase from synapse.events.snapshot import EventContext -from synapse.types import Collection """ Utility functions for poking events into the storage of the server under test. @@ -58,7 +57,7 @@ async def inject_member_event( async def inject_event( hs: synapse.server.HomeServer, room_version: Optional[str] = None, - prev_event_ids: Optional[Collection[str]] = None, + prev_event_ids: Optional[List[str]] = None, **kwargs ) -> EventBase: """Inject a generic event into a room @@ -72,7 +71,10 @@ async def inject_event( """ event, context = await create_event(hs, room_version, prev_event_ids, **kwargs) - await hs.get_storage().persistence.persist_event(event, context) + persistence = hs.get_storage().persistence + assert persistence is not None + + await persistence.persist_event(event, context) return event @@ -80,7 +82,7 @@ async def inject_event( async def create_event( hs: synapse.server.HomeServer, room_version: Optional[str] = None, - prev_event_ids: Optional[Collection[str]] = None, + prev_event_ids: Optional[List[str]] = None, **kwargs ) -> Tuple[EventBase, EventContext]: if room_version is None: diff --git a/tests/test_utils/logging_setup.py b/tests/test_utils/logging_setup.py
index 2d96b0fa8d..fdfb840b62 100644 --- a/tests/test_utils/logging_setup.py +++ b/tests/test_utils/logging_setup.py
@@ -29,8 +29,7 @@ class ToTwistedHandler(logging.Handler): log_entry = self.format(record) log_level = record.levelname.lower().replace("warning", "warn") self.tx_log.emit( - twisted.logger.LogLevel.levelWithName(log_level), - log_entry.replace("{", r"(").replace("}", r")"), + twisted.logger.LogLevel.levelWithName(log_level), "{entry}", entry=log_entry ) diff --git a/tests/test_visibility.py b/tests/test_visibility.py
index b371efc0df..510b630114 100644 --- a/tests/test_visibility.py +++ b/tests/test_visibility.py
@@ -37,10 +37,9 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): self.hs = yield setup_test_homeserver(self.addCleanup) self.event_creation_handler = self.hs.get_event_creation_handler() self.event_builder_factory = self.hs.get_event_builder_factory() - self.store = self.hs.get_datastore() self.storage = self.hs.get_storage() - yield create_room(self.hs, TEST_ROOM_ID, "@someone:ROOM") + yield defer.ensureDeferred(create_room(self.hs, TEST_ROOM_ID, "@someone:ROOM")) @defer.inlineCallbacks def test_filtering(self): @@ -64,8 +63,8 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): evt = yield self.inject_room_member(user, extra_content={"a": "b"}) events_to_filter.append(evt) - filtered = yield filter_events_for_server( - self.storage, "test_server", events_to_filter + filtered = yield defer.ensureDeferred( + filter_events_for_server(self.storage, "test_server", events_to_filter) ) # the result should be 5 redacted events, and 5 unredacted events. @@ -99,11 +98,13 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): events_to_filter.append(evt) # the erasey user gets erased - yield self.hs.get_datastore().mark_user_erased("@erased:local_hs") + yield defer.ensureDeferred( + self.hs.get_datastore().mark_user_erased("@erased:local_hs") + ) # ... and the filtering happens. - filtered = yield filter_events_for_server( - self.storage, "test_server", events_to_filter + filtered = yield defer.ensureDeferred( + filter_events_for_server(self.storage, "test_server", events_to_filter) ) for i in range(0, len(events_to_filter)): @@ -140,7 +141,9 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): event, context = yield defer.ensureDeferred( self.event_creation_handler.create_new_client_event(builder) ) - yield self.storage.persistence.persist_event(event, context) + yield defer.ensureDeferred( + self.storage.persistence.persist_event(event, context) + ) return event @defer.inlineCallbacks @@ -162,7 +165,9 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): self.event_creation_handler.create_new_client_event(builder) ) - yield self.storage.persistence.persist_event(event, context) + yield defer.ensureDeferred( + self.storage.persistence.persist_event(event, context) + ) return event @defer.inlineCallbacks @@ -183,7 +188,9 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): self.event_creation_handler.create_new_client_event(builder) ) - yield self.storage.persistence.persist_event(event, context) + yield defer.ensureDeferred( + self.storage.persistence.persist_event(event, context) + ) return event @defer.inlineCallbacks @@ -265,8 +272,8 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): storage.main = test_store storage.state = test_store - filtered = yield filter_events_for_server( - test_store, "test_server", events_to_filter + filtered = yield defer.ensureDeferred( + filter_events_for_server(test_store, "test_server", events_to_filter) ) logger.info("Filtering took %f seconds", time.time() - start) @@ -287,7 +294,7 @@ class FilterEventsForServerTestCase(tests.unittest.TestCase): test_large_room.skip = "Disabled by default because it's slow" -class _TestStore(object): +class _TestStore: """Implements a few methods of the DataStore, so that we can test filter_events_for_server diff --git a/tests/unittest.py b/tests/unittest.py
index 68d2586efd..e654c0442d 100644 --- a/tests/unittest.py +++ b/tests/unittest.py
@@ -14,7 +14,6 @@ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. - import gc import hashlib import hmac @@ -23,11 +22,12 @@ import logging import time from typing import Optional, Tuple, Type, TypeVar, Union -from mock import Mock +from mock import Mock, patch from canonicaljson import json from twisted.internet.defer import Deferred, ensureDeferred, succeed +from twisted.python.failure import Failure from twisted.python.threadpool import ThreadPool from twisted.trial import unittest @@ -92,7 +92,7 @@ class TestCase(unittest.TestCase): root logger's logging level while that test (case|method) runs.""" def __init__(self, methodName, *args, **kwargs): - super(TestCase, self).__init__(methodName, *args, **kwargs) + super().__init__(methodName, *args, **kwargs) method = getattr(self, methodName) @@ -169,6 +169,19 @@ def INFO(target): return target +def logcontext_clean(target): + """A decorator which marks the TestCase or method as 'logcontext_clean' + + ... ie, any logcontext errors should cause a test failure + """ + + def logcontext_error(msg): + raise AssertionError("logcontext error: %s" % (msg)) + + patcher = patch("synapse.logging.context.logcontext_error", new=logcontext_error) + return patcher(target) + + class HomeserverTestCase(TestCase): """ A base TestCase that reduces boilerplate for HomeServer-using test cases. @@ -241,20 +254,20 @@ class HomeserverTestCase(TestCase): if hasattr(self, "user_id"): if self.hijack_auth: - def get_user_by_access_token(token=None, allow_guest=False): - return succeed( - { - "user": UserID.from_string(self.helper.auth_user_id), - "token_id": 1, - "is_guest": False, - } - ) - - def get_user_by_req(request, allow_guest=False, rights="access"): - return succeed( - create_requester( - UserID.from_string(self.helper.auth_user_id), 1, False, None - ) + async def get_user_by_access_token(token=None, allow_guest=False): + return { + "user": UserID.from_string(self.helper.auth_user_id), + "token_id": 1, + "is_guest": False, + } + + async def get_user_by_req(request, allow_guest=False, rights="access"): + return create_requester( + UserID.from_string(self.helper.auth_user_id), + 1, + False, + False, + None, ) self.hs.get_auth().get_user_by_req = get_user_by_req @@ -353,6 +366,7 @@ class HomeserverTestCase(TestCase): request: Type[T] = SynapseRequest, shorthand: bool = True, federation_auth_origin: str = None, + content_is_form: bool = False, ) -> Tuple[T, FakeChannel]: """ Create a SynapseRequest at the path using the method and containing the @@ -368,6 +382,8 @@ class HomeserverTestCase(TestCase): with the usual REST API path, if it doesn't contain it. federation_auth_origin (bytes|None): if set to not-None, we will add a fake Authorization header pretenting to be the given server name. + content_is_form: Whether the content is URL encoded form data. Adds the + 'Content-Type': 'application/x-www-form-urlencoded' header. Returns: Tuple[synapse.http.site.SynapseRequest, channel] @@ -384,6 +400,7 @@ class HomeserverTestCase(TestCase): request, shorthand, federation_auth_origin, + content_is_form, ) def render(self, request): @@ -422,8 +439,8 @@ class HomeserverTestCase(TestCase): async def run_bg_updates(): with LoggingContext("run_bg_updates", request="run_bg_updates-1"): - while not await stor.db.updates.has_completed_background_updates(): - await stor.db.updates.do_next_background_update(1) + while not await stor.db_pool.updates.has_completed_background_updates(): + await stor.db_pool.updates.do_next_background_update(1) hs = setup_test_homeserver(self.addCleanup, *args, **kwargs) stor = hs.get_datastore() @@ -459,6 +476,35 @@ class HomeserverTestCase(TestCase): self.pump() return self.failureResultOf(d, exc) + def get_success_or_raise(self, d, by=0.0): + """Drive deferred to completion and return result or raise exception + on failure. + """ + + if inspect.isawaitable(d): + deferred = ensureDeferred(d) + if not isinstance(deferred, Deferred): + return d + + results = [] # type: list + deferred.addBoth(results.append) + + self.pump(by=by) + + if not results: + self.fail( + "Success result expected on {!r}, found no result instead".format( + deferred + ) + ) + + result = results[0] + + if isinstance(result, Failure): + result.raiseException() + + return result + def register_user(self, username, password, admin=False): """ Register a user. Requires the Admin API be registered. @@ -544,7 +590,7 @@ class HomeserverTestCase(TestCase): """ event_creator = self.hs.get_event_creation_handler() secrets = self.hs.get_secrets() - requester = Requester(user, None, False, None, None) + requester = Requester(user, None, False, False, None, None) event, context = self.get_success( event_creator.create_event( @@ -571,7 +617,7 @@ class HomeserverTestCase(TestCase): Add the given event as an extremity to the room. """ self.get_success( - self.hs.get_datastore().db.simple_insert( + self.hs.get_datastore().db_pool.simple_insert( table="event_forward_extremities", values={"room_id": room_id, "event_id": event_id}, desc="test_add_extremity", @@ -614,7 +660,7 @@ class FederatingHomeserverTestCase(HomeserverTestCase): """ def prepare(self, reactor, clock, homeserver): - class Authenticator(object): + class Authenticator: def authenticate_request(self, request, content): return succeed("other.example.com") diff --git a/tests/util/caches/test_descriptors.py b/tests/util/caches/test_descriptors.py
index 4d2b9e0d64..677e925477 100644 --- a/tests/util/caches/test_descriptors.py +++ b/tests/util/caches/test_descriptors.py
@@ -88,7 +88,7 @@ class CacheTestCase(unittest.TestCase): class DescriptorTestCase(unittest.TestCase): @defer.inlineCallbacks def test_cache(self): - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -122,7 +122,7 @@ class DescriptorTestCase(unittest.TestCase): def test_cache_num_args(self): """Only the first num_args arguments should matter to the cache""" - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -156,7 +156,7 @@ class DescriptorTestCase(unittest.TestCase): """If the wrapped function throws synchronously, things should continue to work """ - class Cls(object): + class Cls: @cached() def fn(self, arg1): raise SynapseError(100, "mai spoon iz too big!!1") @@ -180,7 +180,7 @@ class DescriptorTestCase(unittest.TestCase): complete_lookup = defer.Deferred() - class Cls(object): + class Cls: @descriptors.cached() def fn(self, arg1): @defer.inlineCallbacks @@ -223,7 +223,7 @@ class DescriptorTestCase(unittest.TestCase): """Check that the cache sets and restores logcontexts correctly when the lookup function throws an exception""" - class Cls(object): + class Cls: @descriptors.cached() def fn(self, arg1): @defer.inlineCallbacks @@ -263,7 +263,7 @@ class DescriptorTestCase(unittest.TestCase): @defer.inlineCallbacks def test_cache_default_args(self): - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -300,7 +300,7 @@ class DescriptorTestCase(unittest.TestCase): obj.mock.assert_not_called() def test_cache_iterable(self): - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -336,7 +336,7 @@ class DescriptorTestCase(unittest.TestCase): """If the wrapped function throws synchronously, things should continue to work """ - class Cls(object): + class Cls: @descriptors.cached(iterable=True) def fn(self, arg1): raise SynapseError(100, "mai spoon iz too big!!1") @@ -358,7 +358,7 @@ class DescriptorTestCase(unittest.TestCase): class CachedListDescriptorTestCase(unittest.TestCase): @defer.inlineCallbacks def test_cache(self): - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -366,11 +366,11 @@ class CachedListDescriptorTestCase(unittest.TestCase): def fn(self, arg1, arg2): pass - @descriptors.cachedList("fn", "args1", inlineCallbacks=True) - def list_fn(self, args1, arg2): + @descriptors.cachedList("fn", "args1") + async def list_fn(self, args1, arg2): assert current_context().request == "c1" # we want this to behave like an asynchronous function - yield run_on_reactor() + await run_on_reactor() assert current_context().request == "c1" return self.mock(args1, arg2) @@ -408,7 +408,7 @@ class CachedListDescriptorTestCase(unittest.TestCase): def test_invalidate(self): """Make sure that invalidation callbacks are called.""" - class Cls(object): + class Cls: def __init__(self): self.mock = mock.Mock() @@ -416,10 +416,10 @@ class CachedListDescriptorTestCase(unittest.TestCase): def fn(self, arg1, arg2): pass - @descriptors.cachedList("fn", "args1", inlineCallbacks=True) - def list_fn(self, args1, arg2): + @descriptors.cachedList("fn", "args1") + async def list_fn(self, args1, arg2): # we want this to behave like an asynchronous function - yield run_on_reactor() + await run_on_reactor() return self.mock(args1, arg2) obj = Cls() diff --git a/tests/util/test_file_consumer.py b/tests/util/test_file_consumer.py
index 8d6627ec33..2012263184 100644 --- a/tests/util/test_file_consumer.py +++ b/tests/util/test_file_consumer.py
@@ -112,7 +112,7 @@ class FileConsumerTests(unittest.TestCase): self.assertTrue(string_file.closed) -class DummyPullProducer(object): +class DummyPullProducer: def __init__(self): self.consumer = None self.deferred = defer.Deferred() @@ -134,7 +134,7 @@ class DummyPullProducer(object): return d -class BlockingStringWrite(object): +class BlockingStringWrite: def __init__(self): self.buffer = "" self.closed = False diff --git a/tests/util/test_retryutils.py b/tests/util/test_retryutils.py
index 9e348694ad..5f46ed0cef 100644 --- a/tests/util/test_retryutils.py +++ b/tests/util/test_retryutils.py
@@ -26,9 +26,7 @@ class RetryLimiterTestCase(HomeserverTestCase): def test_new_destination(self): """A happy-path case with a new destination and a successful operation""" store = self.hs.get_datastore() - d = get_retry_limiter("test_dest", self.clock, store) - self.pump() - limiter = self.successResultOf(d) + limiter = self.get_success(get_retry_limiter("test_dest", self.clock, store)) # advance the clock a bit before making the request self.pump(1) @@ -36,18 +34,14 @@ class RetryLimiterTestCase(HomeserverTestCase): with limiter: pass - d = store.get_destination_retry_timings("test_dest") - self.pump() - new_timings = self.successResultOf(d) + new_timings = self.get_success(store.get_destination_retry_timings("test_dest")) self.assertIsNone(new_timings) def test_limiter(self): """General test case which walks through the process of a failing request""" store = self.hs.get_datastore() - d = get_retry_limiter("test_dest", self.clock, store) - self.pump() - limiter = self.successResultOf(d) + limiter = self.get_success(get_retry_limiter("test_dest", self.clock, store)) self.pump(1) try: @@ -58,29 +52,22 @@ class RetryLimiterTestCase(HomeserverTestCase): except AssertionError: pass - # wait for the update to land - self.pump() - - d = store.get_destination_retry_timings("test_dest") - self.pump() - new_timings = self.successResultOf(d) + new_timings = self.get_success(store.get_destination_retry_timings("test_dest")) self.assertEqual(new_timings["failure_ts"], failure_ts) self.assertEqual(new_timings["retry_last_ts"], failure_ts) self.assertEqual(new_timings["retry_interval"], MIN_RETRY_INTERVAL) # now if we try again we should get a failure - d = get_retry_limiter("test_dest", self.clock, store) - self.pump() - self.failureResultOf(d, NotRetryingDestination) + self.get_failure( + get_retry_limiter("test_dest", self.clock, store), NotRetryingDestination + ) # # advance the clock and try again # self.pump(MIN_RETRY_INTERVAL) - d = get_retry_limiter("test_dest", self.clock, store) - self.pump() - limiter = self.successResultOf(d) + limiter = self.get_success(get_retry_limiter("test_dest", self.clock, store)) self.pump(1) try: @@ -91,12 +78,7 @@ class RetryLimiterTestCase(HomeserverTestCase): except AssertionError: pass - # wait for the update to land - self.pump() - - d = store.get_destination_retry_timings("test_dest") - self.pump() - new_timings = self.successResultOf(d) + new_timings = self.get_success(store.get_destination_retry_timings("test_dest")) self.assertEqual(new_timings["failure_ts"], failure_ts) self.assertEqual(new_timings["retry_last_ts"], retry_ts) self.assertGreaterEqual( @@ -109,10 +91,8 @@ class RetryLimiterTestCase(HomeserverTestCase): # # one more go, with success # - self.pump(MIN_RETRY_INTERVAL * RETRY_MULTIPLIER * 2.0) - d = get_retry_limiter("test_dest", self.clock, store) - self.pump() - limiter = self.successResultOf(d) + self.reactor.advance(MIN_RETRY_INTERVAL * RETRY_MULTIPLIER * 2.0) + limiter = self.get_success(get_retry_limiter("test_dest", self.clock, store)) self.pump(1) with limiter: @@ -121,7 +101,5 @@ class RetryLimiterTestCase(HomeserverTestCase): # wait for the update to land self.pump() - d = store.get_destination_retry_timings("test_dest") - self.pump() - new_timings = self.successResultOf(d) + new_timings = self.get_success(store.get_destination_retry_timings("test_dest")) self.assertIsNone(new_timings) diff --git a/tests/util/test_rwlock.py b/tests/util/test_rwlock.py
index bd32e2cee7..d3dea3b52a 100644 --- a/tests/util/test_rwlock.py +++ b/tests/util/test_rwlock.py
@@ -13,6 +13,7 @@ # See the License for the specific language governing permissions and # limitations under the License. +from twisted.internet import defer from synapse.util.async_helpers import ReadWriteLock @@ -43,6 +44,7 @@ class ReadWriteLockTestCase(unittest.TestCase): rwlock.read(key), # 5 rwlock.write(key), # 6 ] + ds = [defer.ensureDeferred(d) for d in ds] self._assert_called_before_not_after(ds, 2) @@ -73,12 +75,12 @@ class ReadWriteLockTestCase(unittest.TestCase): with ds[6].result: pass - d = rwlock.write(key) + d = defer.ensureDeferred(rwlock.write(key)) self.assertTrue(d.called) with d.result: pass - d = rwlock.read(key) + d = defer.ensureDeferred(rwlock.read(key)) self.assertTrue(d.called) with d.result: pass diff --git a/tests/util/test_stringutils.py b/tests/util/test_stringutils.py
index 4f4da29a98..8491f7cc83 100644 --- a/tests/util/test_stringutils.py +++ b/tests/util/test_stringutils.py
@@ -28,9 +28,6 @@ class StringUtilsTestCase(unittest.TestCase): "_--something==_", "...--==-18913", "8Dj2odd-e9asd.cd==_--ddas-secret-", - # We temporarily allow : characters: https://github.com/matrix-org/synapse/issues/6766 - # To be removed in a future release - "SECRET:1234567890", ] bad = [ diff --git a/tests/utils.py b/tests/utils.py
index 8eb6fdd98e..867cf88977 100644 --- a/tests/utils.py +++ b/tests/utils.py
@@ -154,6 +154,10 @@ def default_config(name, parse=False): "account": {"per_second": 10000, "burst_count": 10000}, "failed_attempts": {"per_second": 10000, "burst_count": 10000}, }, + "rc_joins": { + "local": {"per_second": 10000, "burst_count": 10000}, + "remote": {"per_second": 10000, "burst_count": 10000}, + }, "saml2_enabled": False, "public_baseurl": None, "default_identity_server": None, @@ -470,7 +474,7 @@ class MockHttpResource(HttpServer): self.callbacks.append((method, path_pattern, callback)) -class MockKey(object): +class MockKey: alg = "mock_alg" version = "mock_version" signature = b"\x9a\x87$" @@ -489,7 +493,7 @@ class MockKey(object): return b"<fake_encoded_key>" -class MockClock(object): +class MockClock: now = 1000 def __init__(self): @@ -566,7 +570,7 @@ def _format_call(args, kwargs): ) -class DeferredMockCallable(object): +class DeferredMockCallable: """A callable instance that stores a set of pending call expectations and return values for them. It allows a unit test to assert that the given set of function calls are eventually made, by awaiting on them to be called. @@ -640,14 +644,8 @@ class DeferredMockCallable(object): ) -@defer.inlineCallbacks -def create_room(hs, room_id, creator_id): +async def create_room(hs, room_id: str, creator_id: str): """Creates and persist a creation event for the given room - - Args: - hs - room_id (str) - creator_id (str) """ persistence_store = hs.get_storage().persistence @@ -655,7 +653,7 @@ def create_room(hs, room_id, creator_id): event_builder_factory = hs.get_event_builder_factory() event_creation_handler = hs.get_event_creation_handler() - yield store.store_room( + await store.store_room( room_id=room_id, room_creator_user_id=creator_id, is_public=False, @@ -673,8 +671,6 @@ def create_room(hs, room_id, creator_id): }, ) - event, context = yield defer.ensureDeferred( - event_creation_handler.create_new_client_event(builder) - ) + event, context = await event_creation_handler.create_new_client_event(builder) - yield persistence_store.persist_event(event, context) + await persistence_store.persist_event(event, context) diff --git a/tox.ini b/tox.ini
index ea41f1c1de..0a2d14aec4 100644 --- a/tox.ini +++ b/tox.ini
@@ -2,13 +2,12 @@ envlist = packaging, py35, py36, py37, py38, check_codestyle, check_isort [base] +extras = test deps = - mock python-subunit junitxml coverage coverage-enable-subprocess - parameterized # cyptography 2.2 requires setuptools >= 18.5 # @@ -36,7 +35,7 @@ setenv = [testenv] deps = {[base]deps} -extras = all +extras = all, test whitelist_externals = sh @@ -84,7 +83,6 @@ deps = # Old automat version for Twisted Automat == 0.3.0 - mock lxml coverage coverage-enable-subprocess @@ -97,7 +95,7 @@ commands = /bin/sh -c 'python -m synapse.python_dependencies | sed -e "s/>=/==/g" -e "s/psycopg2==2.6//" -e "s/pyopenssl==16.0.0/pyopenssl==17.0.0/" | xargs -d"\n" pip install' # Install Synapse itself. This won't update any libraries. - pip install -e . + pip install -e ".[test]" {envbindir}/coverage run "{envbindir}/trial" {env:TRIAL_FLAGS:} {posargs:tests} {env:TOXSUFFIX:} @@ -118,20 +116,14 @@ commands = check-manifest [testenv:check_codestyle] -skip_install = True -deps = - flake8==3.8.3 - flake8-comprehensions - # We pin so that our tests don't start failing on new releases of black. - black==19.10b0 +extras = lint commands = python -m black --check --diff . /bin/sh -c "flake8 synapse tests scripts scripts-dev contrib synctl {env:PEP8SUFFIX:}" {toxinidir}/scripts-dev/config-lint.sh [testenv:check_isort] -skip_install = True -deps = isort==5.0.3 +extras = lint commands = /bin/sh -c "isort -c --df --sp setup.cfg synapse tests scripts-dev scripts" [testenv:check-newsfragment] @@ -170,50 +162,10 @@ commands= skip_install = True deps = {[base]deps} - mypy==0.750 + mypy==0.782 mypy-zope -env = - MYPYPATH = stubs/ extras = all -commands = mypy \ - synapse/api \ - synapse/appservice \ - synapse/config \ - synapse/event_auth.py \ - synapse/events/spamcheck.py \ - synapse/federation \ - synapse/handlers/auth.py \ - synapse/handlers/cas_handler.py \ - synapse/handlers/directory.py \ - synapse/handlers/federation.py \ - synapse/handlers/identity.py \ - synapse/handlers/oidc_handler.py \ - synapse/handlers/presence.py \ - synapse/handlers/room_member.py \ - synapse/handlers/room_member_worker.py \ - synapse/handlers/saml_handler.py \ - synapse/handlers/sync.py \ - synapse/handlers/ui_auth \ - synapse/http/server.py \ - synapse/http/site.py \ - synapse/logging/ \ - synapse/metrics \ - synapse/module_api \ - synapse/push/pusherpool.py \ - synapse/push/push_rule_evaluator.py \ - synapse/replication \ - synapse/rest \ - synapse/spam_checker_api \ - synapse/storage/data_stores/main/ui_auth.py \ - synapse/storage/database.py \ - synapse/storage/engines \ - synapse/storage/util \ - synapse/streams \ - synapse/util/caches/stream_change_cache.py \ - tests/replication \ - tests/test_utils \ - tests/rest/client/v2_alpha/test_auth.py \ - tests/util/test_stream_change_cache.py +commands = mypy # To find all folders that pass mypy you run: #