Merge branch 'develop' into rav/enforce_report_api

1 files changed, 182 insertions, 77 deletions

diff --git a/synapse/handlers/federation.py b/synapse/handlers/federation.py
index 87c0615820..d3ecebd29f 100644
--- a/synapse/handlers/federation.py
+++ b/synapse/handlers/federation.py
@@ -20,38 +20,42 @@ import itertools
 import logging
 import sys
 
-from signedjson.key import decode_verify_key_bytes
-from signedjson.sign import verify_signed_json
 import six
-from six.moves import http_client
 from six import iteritems
-from twisted.internet import defer
+from six.moves import http_client
+
+from signedjson.key import decode_verify_key_bytes
+from signedjson.sign import verify_signed_json
 from unpaddedbase64 import decode_base64
 
-from ._base import BaseHandler
+from twisted.internet import defer
 
+from synapse.api.constants import EventTypes, Membership, RejectedReason
 from synapse.api.errors import (
-    AuthError, FederationError, StoreError, CodeMessageException, SynapseError,
+    AuthError,
+    CodeMessageException,
     FederationDeniedError,
+    FederationError,
+    StoreError,
+    SynapseError,
 )
-from synapse.api.constants import EventTypes, Membership, RejectedReason
-from synapse.events.validator import EventValidator
-from synapse.util import unwrapFirstError, logcontext
-from synapse.util.metrics import measure_func
-from synapse.util.logutils import log_function
-from synapse.util.async import run_on_reactor, Linearizer
-from synapse.util.frozenutils import unfreeze
 from synapse.crypto.event_signing import (
-    compute_event_signature, add_hashes_and_signatures,
+    add_hashes_and_signatures,
+    compute_event_signature,
 )
-from synapse.types import UserID, get_domain_from_id
-
 from synapse.events.utils import prune_event
-
-from synapse.util.retryutils import NotRetryingDestination
-
+from synapse.events.validator import EventValidator
+from synapse.state import resolve_events_with_factory
+from synapse.types import UserID, get_domain_from_id
+from synapse.util import logcontext, unwrapFirstError
+from synapse.util.async import Linearizer
 from synapse.util.distributor import user_joined_room
+from synapse.util.frozenutils import unfreeze
+from synapse.util.logutils import log_function
+from synapse.util.metrics import measure_func
+from synapse.util.retryutils import NotRetryingDestination
 
+from ._base import BaseHandler
 
 logger = logging.getLogger(__name__)
 
@@ -90,7 +94,9 @@ class FederationHandler(BaseHandler):
 
     @defer.inlineCallbacks
     @log_function
-    def on_receive_pdu(self, origin, pdu, get_missing=True):
+    def on_receive_pdu(
+            self, origin, pdu, get_missing=True, sent_to_us_directly=False,
+    ):
         """ Process a PDU received via a federation /send/ transaction, or
         via backfill of missing prev_events
 
@@ -104,8 +110,10 @@ class FederationHandler(BaseHandler):
         """
 
         # We reprocess pdus when we have seen them only as outliers
-        existing = yield self.get_persisted_pdu(
-            origin, pdu.event_id, do_auth=False
+        existing = yield self.store.get_event(
+            pdu.event_id,
+            allow_none=True,
+            allow_rejected=True,
         )
 
         # FIXME: Currently we fetch an event again when we already have it
@@ -162,14 +170,11 @@ class FederationHandler(BaseHandler):
                     "Ignoring PDU %s for room %s from %s as we've left the room!",
                     pdu.event_id, pdu.room_id, origin,
                 )
-                return
+                defer.returnValue(None)
 
         state = None
-
         auth_chain = []
 
-        fetch_state = False
-
         # Get missing pdus if necessary.
         if not pdu.internal_metadata.is_outlier():
             # We only backfill backwards to the min depth.
@@ -224,26 +229,60 @@ class FederationHandler(BaseHandler):
                         list(prevs - seen)[:5],
                     )
 
-            if prevs - seen:
-                logger.info(
-                    "Still missing %d events for room %r: %r...",
-                    len(prevs - seen), pdu.room_id, list(prevs - seen)[:5]
+            if sent_to_us_directly and prevs - seen:
+                # If they have sent it to us directly, and the server
+                # isn't telling us about the auth events that it's
+                # made a message referencing, we explode
+                raise FederationError(
+                    "ERROR",
+                    403,
+                    (
+                        "Your server isn't divulging details about prev_events "
+                        "referenced in this event."
+                    ),
+                    affected=pdu.event_id,
                 )
-                fetch_state = True
+            elif prevs - seen:
+                # Calculate the state of the previous events, and
+                # de-conflict them to find the current state.
+                state_groups = []
+                auth_chains = set()
+                try:
+                    # Get the state of the events we know about
+                    ours = yield self.store.get_state_groups(pdu.room_id, list(seen))
+                    state_groups.append(ours)
+
+                    # Ask the remote server for the states we don't
+                    # know about
+                    for p in prevs - seen:
+                        state, got_auth_chain = (
+                            yield self.replication_layer.get_state_for_room(
+                                origin, pdu.room_id, p
+                            )
+                        )
+                        auth_chains.update(got_auth_chain)
+                        state_group = {(x.type, x.state_key): x.event_id for x in state}
+                        state_groups.append(state_group)
+
+                    # Resolve any conflicting state
+                    def fetch(ev_ids):
+                        return self.store.get_events(
+                            ev_ids, get_prev_content=False, check_redacted=False
+                        )
 
-        if fetch_state:
-            # We need to get the state at this event, since we haven't
-            # processed all the prev events.
-            logger.debug(
-                "_handle_new_pdu getting state for %s",
-                pdu.room_id
-            )
-            try:
-                state, auth_chain = yield self.replication_layer.get_state_for_room(
-                    origin, pdu.room_id, pdu.event_id,
-                )
-            except Exception:
-                logger.exception("Failed to get state for event: %s", pdu.event_id)
+                    state_map = yield resolve_events_with_factory(
+                        state_groups, {pdu.event_id: pdu}, fetch
+                    )
+
+                    state = (yield self.store.get_events(state_map.values())).values()
+                    auth_chain = list(auth_chains)
+                except Exception:
+                    raise FederationError(
+                        "ERROR",
+                        403,
+                        "We can't get valid state history.",
+                        affected=pdu.event_id,
+                    )
 
         yield self._process_received_pdu(
             origin,
@@ -321,11 +360,17 @@ class FederationHandler(BaseHandler):
 
         for e in missing_events:
             logger.info("Handling found event %s", e.event_id)
-            yield self.on_receive_pdu(
-                origin,
-                e,
-                get_missing=False
-            )
+            try:
+                yield self.on_receive_pdu(
+                    origin,
+                    e,
+                    get_missing=False
+                )
+            except FederationError as e:
+                if e.code == 403:
+                    logger.warn("Event %s failed history check.")
+                else:
+                    raise
 
     @log_function
     @defer.inlineCallbacks
@@ -459,6 +504,47 @@ class FederationHandler(BaseHandler):
     @measure_func("_filter_events_for_server")
     @defer.inlineCallbacks
     def _filter_events_for_server(self, server_name, room_id, events):
+        """Filter the given events for the given server, redacting those the
+        server can't see.
+
+        Assumes the server is currently in the room.
+
+        Returns
+            list[FrozenEvent]
+        """
+        # First lets check to see if all the events have a history visibility
+        # of "shared" or "world_readable". If thats the case then we don't
+        # need to check membership (as we know the server is in the room).
+        event_to_state_ids = yield self.store.get_state_ids_for_events(
+            frozenset(e.event_id for e in events),
+            types=(
+                (EventTypes.RoomHistoryVisibility, ""),
+            )
+        )
+
+        visibility_ids = set()
+        for sids in event_to_state_ids.itervalues():
+            hist = sids.get((EventTypes.RoomHistoryVisibility, ""))
+            if hist:
+                visibility_ids.add(hist)
+
+        # If we failed to find any history visibility events then the default
+        # is "shared" visiblity.
+        if not visibility_ids:
+            defer.returnValue(events)
+
+        event_map = yield self.store.get_events(visibility_ids)
+        all_open = all(
+            e.content.get("history_visibility") in (None, "shared", "world_readable")
+            for e in event_map.itervalues()
+        )
+
+        if all_open:
+            defer.returnValue(events)
+
+        # Ok, so we're dealing with events that have non-trivial visibility
+        # rules, so we need to also get the memberships of the room.
+
         event_to_state_ids = yield self.store.get_state_ids_for_events(
             frozenset(e.event_id for e in events),
             types=(
@@ -480,8 +566,8 @@ class FederationHandler(BaseHandler):
         # to get all state ids that we're interested in.
         event_map = yield self.store.get_events([
             e_id
-            for key_to_eid in event_to_state_ids.itervalues()
-            for key, e_id in key_to_eid.iteritems()
+            for key_to_eid in list(event_to_state_ids.values())
+            for key, e_id in key_to_eid.items()
             if key[0] != EventTypes.Member or check_match(key[1])
         ])
 
@@ -494,7 +580,20 @@ class FederationHandler(BaseHandler):
             for e_id, key_to_eid in event_to_state_ids.iteritems()
         }
 
+        erased_senders = yield self.store.are_users_erased(
+            e.sender for e in events,
+        )
+
         def redact_disallowed(event, state):
+            # if the sender has been gdpr17ed, always return a redacted
+            # copy of the event.
+            if erased_senders[event.sender]:
+                logger.info(
+                    "Sender of %s has been erased, redacting",
+                    event.event_id,
+                )
+                return prune_event(event)
+
             if not state:
                 return event
 
@@ -1149,13 +1248,13 @@ class FederationHandler(BaseHandler):
                 user = UserID.from_string(event.state_key)
                 yield user_joined_room(self.distributor, user, event.room_id)
 
-        state_ids = context.prev_state_ids.values()
+        state_ids = list(context.prev_state_ids.values())
         auth_chain = yield self.store.get_auth_chain(state_ids)
 
-        state = yield self.store.get_events(context.prev_state_ids.values())
+        state = yield self.store.get_events(list(context.prev_state_ids.values()))
 
         defer.returnValue({
-            "state": state.values(),
+            "state": list(state.values()),
             "auth_chain": auth_chain,
         })
 
@@ -1382,8 +1481,6 @@ class FederationHandler(BaseHandler):
     def get_state_for_pdu(self, room_id, event_id):
         """Returns the state at the event. i.e. not including said event.
         """
-        yield run_on_reactor()
-
         state_groups = yield self.store.get_state_groups(
             room_id, [event_id]
         )
@@ -1405,7 +1502,7 @@ class FederationHandler(BaseHandler):
                 else:
                     del results[(event.type, event.state_key)]
 
-            res = results.values()
+            res = list(results.values())
             for event in res:
                 # We sign these again because there was a bug where we
                 # incorrectly signed things the first time round
@@ -1426,8 +1523,6 @@ class FederationHandler(BaseHandler):
     def get_state_ids_for_pdu(self, room_id, event_id):
         """Returns the state at the event. i.e. not including said event.
         """
-        yield run_on_reactor()
-
         state_groups = yield self.store.get_state_groups_ids(
             room_id, [event_id]
         )
@@ -1446,7 +1541,7 @@ class FederationHandler(BaseHandler):
                 else:
                     results.pop((event.type, event.state_key), None)
 
-            defer.returnValue(results.values())
+            defer.returnValue(list(results.values()))
         else:
             defer.returnValue([])
 
@@ -1469,11 +1564,20 @@ class FederationHandler(BaseHandler):
 
     @defer.inlineCallbacks
     @log_function
-    def get_persisted_pdu(self, origin, event_id, do_auth=True):
-        """ Get a PDU from the database with given origin and id.
+    def get_persisted_pdu(self, origin, event_id):
+        """Get an event from the database for the given server.
+
+        Args:
+            origin [str]: hostname of server which is requesting the event; we
+               will check that the server is allowed to see it.
+            event_id [str]: id of the event being requested
 
         Returns:
-            Deferred: Results in a `Pdu`.
+            Deferred[EventBase|None]: None if we know nothing about the event;
+                otherwise the (possibly-redacted) event.
+
+        Raises:
+            AuthError if the server is not currently in the room
         """
         event = yield self.store.get_event(
             event_id,
@@ -1494,20 +1598,17 @@ class FederationHandler(BaseHandler):
                     )
                 )
 
-            if do_auth:
-                in_room = yield self.auth.check_host_in_room(
-                    event.room_id,
-                    origin
-                )
-                if not in_room:
-                    raise AuthError(403, "Host not in room.")
-
-                events = yield self._filter_events_for_server(
-                    origin, event.room_id, [event]
-                )
-
-                event = events[0]
+            in_room = yield self.auth.check_host_in_room(
+                event.room_id,
+                origin
+            )
+            if not in_room:
+                raise AuthError(403, "Host not in room.")
 
+            events = yield self._filter_events_for_server(
+                origin, event.room_id, [event]
+            )
+            event = events[0]
             defer.returnValue(event)
         else:
             defer.returnValue(None)
@@ -1795,6 +1896,10 @@ class FederationHandler(BaseHandler):
             min_depth=min_depth,
         )
 
+        missing_events = yield self._filter_events_for_server(
+            origin, room_id, missing_events,
+        )
+
         defer.returnValue(missing_events)
 
     @defer.inlineCallbacks
@@ -1915,7 +2020,7 @@ class FederationHandler(BaseHandler):
                 })
 
                 new_state = self.state_handler.resolve_events(
-                    [local_view.values(), remote_view.values()],
+                    [list(local_view.values()), list(remote_view.values())],
                     event
                 )