diff options
Diffstat (limited to 'synapse/api')
| -rw-r--r-- | synapse/api/auth.py | 5 | ||||
| -rw-r--r-- | synapse/api/errors.py | 74 | ||||
| -rw-r--r-- | synapse/api/filtering.py | 298 |
3 files changed, 260 insertions, 117 deletions
diff --git a/synapse/api/auth.py b/synapse/api/auth.py index 03a215ab1b..9dbc7993df 100644 --- a/synapse/api/auth.py +++ b/synapse/api/auth.py @@ -23,7 +23,7 @@ from synapse import event_auth from synapse.api.constants import EventTypes, Membership, JoinRules from synapse.api.errors import AuthError, Codes from synapse.types import UserID -from synapse.util.logcontext import preserve_context_over_fn +from synapse.util import logcontext from synapse.util.metrics import Measure logger = logging.getLogger(__name__) @@ -209,8 +209,7 @@ class Auth(object): default=[""] )[0] if user and access_token and ip_addr: - preserve_context_over_fn( - self.store.insert_client_ip, + logcontext.preserve_fn(self.store.insert_client_ip)( user=user, access_token=access_token, ip=ip_addr, diff --git a/synapse/api/errors.py b/synapse/api/errors.py index 921c457738..6fbd5d6876 100644 --- a/synapse/api/errors.py +++ b/synapse/api/errors.py @@ -15,6 +15,7 @@ """Contains exceptions and error codes.""" +import json import logging logger = logging.getLogger(__name__) @@ -50,27 +51,35 @@ class Codes(object): class CodeMessageException(RuntimeError): - """An exception with integer code and message string attributes.""" + """An exception with integer code and message string attributes. + Attributes: + code (int): HTTP error code + msg (str): string describing the error + """ def __init__(self, code, msg): super(CodeMessageException, self).__init__("%d: %s" % (code, msg)) self.code = code self.msg = msg - self.response_code_message = None def error_dict(self): return cs_error(self.msg) class SynapseError(CodeMessageException): - """A base error which can be caught for all synapse events.""" + """A base exception type for matrix errors which have an errcode and error + message (as well as an HTTP status code). + + Attributes: + errcode (str): Matrix error code e.g 'M_FORBIDDEN' + """ def __init__(self, code, msg, errcode=Codes.UNKNOWN): """Constructs a synapse error. Args: code (int): The integer error code (an HTTP response code) msg (str): The human-readable error message. - err (str): The error code e.g 'M_FORBIDDEN' + errcode (str): The matrix error code e.g 'M_FORBIDDEN' """ super(SynapseError, self).__init__(code, msg) self.errcode = errcode @@ -81,6 +90,39 @@ class SynapseError(CodeMessageException): self.errcode, ) + @classmethod + def from_http_response_exception(cls, err): + """Make a SynapseError based on an HTTPResponseException + + This is useful when a proxied request has failed, and we need to + decide how to map the failure onto a matrix error to send back to the + client. + + An attempt is made to parse the body of the http response as a matrix + error. If that succeeds, the errcode and error message from the body + are used as the errcode and error message in the new synapse error. + + Otherwise, the errcode is set to M_UNKNOWN, and the error message is + set to the reason code from the HTTP response. + + Args: + err (HttpResponseException): + + Returns: + SynapseError: + """ + # 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(err.response) + except ValueError: + j = {} + errcode = j.get('errcode', Codes.UNKNOWN) + errmsg = j.get('error', err.msg) + + res = SynapseError(err.code, errmsg, errcode) + return res + class RegistrationError(SynapseError): """An error raised when a registration event fails.""" @@ -106,13 +148,11 @@ class UnrecognizedRequestError(SynapseError): class NotFoundError(SynapseError): """An error indicating we can't find the thing you asked for""" - def __init__(self, *args, **kwargs): - if "errcode" not in kwargs: - kwargs["errcode"] = Codes.NOT_FOUND + def __init__(self, msg="Not found", errcode=Codes.NOT_FOUND): super(NotFoundError, self).__init__( 404, - "Not found", - **kwargs + msg, + errcode=errcode ) @@ -173,7 +213,6 @@ class LimitExceededError(SynapseError): errcode=Codes.LIMIT_EXCEEDED): super(LimitExceededError, self).__init__(code, msg, errcode) self.retry_after_ms = retry_after_ms - self.response_code_message = "Too Many Requests" def error_dict(self): return cs_error( @@ -243,6 +282,19 @@ class FederationError(RuntimeError): class HttpResponseException(CodeMessageException): + """ + Represents an HTTP-level failure of an outbound request + + Attributes: + response (str): body of response + """ def __init__(self, code, msg, response): - self.response = response + """ + + Args: + code (int): HTTP status code + msg (str): reason phrase from HTTP response status line + response (str): body of response + """ super(HttpResponseException, self).__init__(code, msg) + self.response = response diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py index fb291d7fb9..83206348e5 100644 --- a/synapse/api/filtering.py +++ b/synapse/api/filtering.py @@ -13,11 +13,174 @@ # See the License for the specific language governing permissions and # limitations under the License. from synapse.api.errors import SynapseError +from synapse.storage.presence import UserPresenceState from synapse.types import UserID, RoomID - from twisted.internet import defer import ujson as json +import jsonschema +from jsonschema import FormatChecker + +FILTER_SCHEMA = { + "additionalProperties": False, + "type": "object", + "properties": { + "limit": { + "type": "number" + }, + "senders": { + "$ref": "#/definitions/user_id_array" + }, + "not_senders": { + "$ref": "#/definitions/user_id_array" + }, + # TODO: We don't limit event type values but we probably should... + # check types are valid event types + "types": { + "type": "array", + "items": { + "type": "string" + } + }, + "not_types": { + "type": "array", + "items": { + "type": "string" + } + } + } +} + +ROOM_FILTER_SCHEMA = { + "additionalProperties": False, + "type": "object", + "properties": { + "not_rooms": { + "$ref": "#/definitions/room_id_array" + }, + "rooms": { + "$ref": "#/definitions/room_id_array" + }, + "ephemeral": { + "$ref": "#/definitions/room_event_filter" + }, + "include_leave": { + "type": "boolean" + }, + "state": { + "$ref": "#/definitions/room_event_filter" + }, + "timeline": { + "$ref": "#/definitions/room_event_filter" + }, + "account_data": { + "$ref": "#/definitions/room_event_filter" + }, + } +} + +ROOM_EVENT_FILTER_SCHEMA = { + "additionalProperties": False, + "type": "object", + "properties": { + "limit": { + "type": "number" + }, + "senders": { + "$ref": "#/definitions/user_id_array" + }, + "not_senders": { + "$ref": "#/definitions/user_id_array" + }, + "types": { + "type": "array", + "items": { + "type": "string" + } + }, + "not_types": { + "type": "array", + "items": { + "type": "string" + } + }, + "rooms": { + "$ref": "#/definitions/room_id_array" + }, + "not_rooms": { + "$ref": "#/definitions/room_id_array" + }, + "contains_url": { + "type": "boolean" + } + } +} + +USER_ID_ARRAY_SCHEMA = { + "type": "array", + "items": { + "type": "string", + "format": "matrix_user_id" + } +} + +ROOM_ID_ARRAY_SCHEMA = { + "type": "array", + "items": { + "type": "string", + "format": "matrix_room_id" + } +} + +USER_FILTER_SCHEMA = { + "$schema": "http://json-schema.org/draft-04/schema#", + "description": "schema for a Sync filter", + "type": "object", + "definitions": { + "room_id_array": ROOM_ID_ARRAY_SCHEMA, + "user_id_array": USER_ID_ARRAY_SCHEMA, + "filter": FILTER_SCHEMA, + "room_filter": ROOM_FILTER_SCHEMA, + "room_event_filter": ROOM_EVENT_FILTER_SCHEMA + }, + "properties": { + "presence": { + "$ref": "#/definitions/filter" + }, + "account_data": { + "$ref": "#/definitions/filter" + }, + "room": { + "$ref": "#/definitions/room_filter" + }, + "event_format": { + "type": "string", + "enum": ["client", "federation"] + }, + "event_fields": { + "type": "array", + "items": { + "type": "string", + # Don't allow '\\' in event field filters. This makes matching + # events a lot easier as we can then use a negative lookbehind + # assertion to split '\.' If we allowed \\ then it would + # incorrectly split '\\.' See synapse.events.utils.serialize_event + "pattern": "^((?!\\\).)*$" + } + } + }, + "additionalProperties": False +} + + +@FormatChecker.cls_checks('matrix_room_id') +def matrix_room_id_validator(room_id_str): + return RoomID.from_string(room_id_str) + + +@FormatChecker.cls_checks('matrix_user_id') +def matrix_user_id_validator(user_id_str): + return UserID.from_string(user_id_str) class Filtering(object): @@ -52,98 +215,11 @@ class Filtering(object): # NB: Filters are the complete json blobs. "Definitions" are an # individual top-level key e.g. public_user_data. Filters are made of # many definitions. - - top_level_definitions = [ - "presence", "account_data" - ] - - room_level_definitions = [ - "state", "timeline", "ephemeral", "account_data" - ] - - for key in top_level_definitions: - if key in user_filter_json: - self._check_definition(user_filter_json[key]) - - if "room" in user_filter_json: - self._check_definition_room_lists(user_filter_json["room"]) - for key in room_level_definitions: - if key in user_filter_json["room"]: - self._check_definition(user_filter_json["room"][key]) - - if "event_fields" in user_filter_json: - if type(user_filter_json["event_fields"]) != list: - raise SynapseError(400, "event_fields must be a list of strings") - for field in user_filter_json["event_fields"]: - if not isinstance(field, basestring): - raise SynapseError(400, "Event field must be a string") - # Don't allow '\\' in event field filters. This makes matching - # events a lot easier as we can then use a negative lookbehind - # assertion to split '\.' If we allowed \\ then it would - # incorrectly split '\\.' See synapse.events.utils.serialize_event - if r'\\' in field: - raise SynapseError( - 400, r'The escape character \ cannot itself be escaped' - ) - - def _check_definition_room_lists(self, definition): - """Check that "rooms" and "not_rooms" are lists of room ids if they - are present - - Args: - definition(dict): The filter definition - Raises: - SynapseError: If there was a problem with this definition. - """ - # check rooms are valid room IDs - room_id_keys = ["rooms", "not_rooms"] - for key in room_id_keys: - if key in definition: - if type(definition[key]) != list: - raise SynapseError(400, "Expected %s to be a list." % key) - for room_id in definition[key]: - RoomID.from_string(room_id) - - def _check_definition(self, definition): - """Check if the provided definition is valid. - - This inspects not only the types but also the values to make sure they - make sense. - - Args: - definition(dict): The filter definition - Raises: - SynapseError: If there was a problem with this definition. - """ - # NB: Filters are the complete json blobs. "Definitions" are an - # individual top-level key e.g. public_user_data. Filters are made of - # many definitions. - if type(definition) != dict: - raise SynapseError( - 400, "Expected JSON object, not %s" % (definition,) - ) - - self._check_definition_room_lists(definition) - - # check senders are valid user IDs - user_id_keys = ["senders", "not_senders"] - for key in user_id_keys: - if key in definition: - if type(definition[key]) != list: - raise SynapseError(400, "Expected %s to be a list." % key) - for user_id in definition[key]: - UserID.from_string(user_id) - - # TODO: We don't limit event type values but we probably should... - # check types are valid event types - event_keys = ["types", "not_types"] - for key in event_keys: - if key in definition: - if type(definition[key]) != list: - raise SynapseError(400, "Expected %s to be a list." % key) - for event_type in definition[key]: - if not isinstance(event_type, basestring): - raise SynapseError(400, "Event type should be a string") + try: + jsonschema.validate(user_filter_json, USER_FILTER_SCHEMA, + format_checker=FormatChecker()) + except jsonschema.ValidationError as e: + raise SynapseError(400, e.message) class FilterCollection(object): @@ -253,19 +329,35 @@ class Filter(object): Returns: bool: True if the event matches """ - sender = event.get("sender", None) - if not sender: - # Presence events have their 'sender' in content.user_id - content = event.get("content") - # account_data has been allowed to have non-dict content, so check type first - if isinstance(content, dict): - sender = content.get("user_id") + # We usually get the full "events" as dictionaries coming through, + # except for presence which actually gets passed around as its own + # namedtuple type. + if isinstance(event, UserPresenceState): + sender = event.user_id + room_id = None + ev_type = "m.presence" + is_url = False + else: + sender = event.get("sender", None) + if not sender: + # Presence events had their 'sender' in content.user_id, but are + # now handled above. We don't know if anything else uses this + # form. TODO: Check this and probably remove it. + content = event.get("content") + # account_data has been allowed to have non-dict content, so + # check type first + if isinstance(content, dict): + sender = content.get("user_id") + + room_id = event.get("room_id", None) + ev_type = event.get("type", None) + is_url = "url" in event.get("content", {}) return self.check_fields( - event.get("room_id", None), + room_id, sender, - event.get("type", None), - "url" in event.get("content", {}) + ev_type, + is_url, ) def check_fields(self, room_id, sender, event_type, contains_url): |