2016-01-07 05:26:29 +01:00
|
|
|
# Copyright 2014-2016 OpenMarket Ltd
|
2021-11-23 12:43:56 +01:00
|
|
|
# Copyright 2021 The Matrix.org Foundation C.I.C.
|
2014-12-03 17:07:21 +01:00
|
|
|
#
|
|
|
|
# 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.
|
2020-07-20 19:33:04 +02:00
|
|
|
import collections.abc
|
2018-07-09 08:09:20 +02:00
|
|
|
import re
|
2022-01-26 14:27:04 +01:00
|
|
|
from typing import (
|
|
|
|
TYPE_CHECKING,
|
|
|
|
Any,
|
|
|
|
Callable,
|
|
|
|
Dict,
|
|
|
|
Iterable,
|
|
|
|
List,
|
|
|
|
Mapping,
|
2022-05-07 14:37:29 +02:00
|
|
|
MutableMapping,
|
2022-01-26 14:27:04 +01:00
|
|
|
Optional,
|
|
|
|
Union,
|
|
|
|
)
|
2018-07-09 08:09:20 +02:00
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
import attr
|
2016-11-21 18:52:45 +01:00
|
|
|
|
2021-08-10 03:23:31 +02:00
|
|
|
from synapse.api.constants import EventContentFields, EventTypes, RelationTypes
|
2020-05-14 19:24:01 +02:00
|
|
|
from synapse.api.errors import Codes, SynapseError
|
2020-03-09 13:58:25 +01:00
|
|
|
from synapse.api.room_versions import RoomVersion
|
2021-10-13 13:24:07 +02:00
|
|
|
from synapse.types import JsonDict
|
2021-03-17 17:51:55 +01:00
|
|
|
from synapse.util.frozenutils import unfreeze
|
2016-11-21 18:42:16 +01:00
|
|
|
|
2018-07-09 08:09:20 +02:00
|
|
|
from . import EventBase
|
2018-04-15 21:43:35 +02:00
|
|
|
|
2022-01-26 14:27:04 +01:00
|
|
|
if TYPE_CHECKING:
|
2022-03-18 18:49:32 +01:00
|
|
|
from synapse.handlers.relations import BundledAggregations
|
2022-01-26 14:27:04 +01:00
|
|
|
|
|
|
|
|
2016-11-21 18:42:16 +01:00
|
|
|
# Split strings on "." but not "\." This uses a negative lookbehind assertion for '\'
|
|
|
|
# (?<!stuff) matches if the current position in the string is not preceded
|
|
|
|
# by a match for 'stuff'.
|
|
|
|
# TODO: This is fast, but fails to handle "foo\\.bar" which should be treated as
|
|
|
|
# the literal fields "foo\" and "bar" but will instead be treated as "foo\\.bar"
|
2019-06-20 11:32:02 +02:00
|
|
|
SPLIT_FIELD_REGEX = re.compile(r"(?<!\\)\.")
|
2016-11-21 18:42:16 +01:00
|
|
|
|
2022-03-29 12:41:19 +02:00
|
|
|
CANONICALJSON_MAX_INT = (2**53) - 1
|
2021-08-26 18:07:58 +02:00
|
|
|
CANONICALJSON_MIN_INT = -CANONICALJSON_MAX_INT
|
|
|
|
|
2014-12-03 17:07:21 +01:00
|
|
|
|
2020-03-05 16:46:44 +01:00
|
|
|
def prune_event(event: EventBase) -> EventBase:
|
2021-02-16 23:32:34 +01:00
|
|
|
"""Returns a pruned version of the given event, which removes all keys we
|
2014-12-03 17:07:21 +01:00
|
|
|
don't know about or think could potentially be dodgy.
|
|
|
|
|
|
|
|
This is used when we "redact" an event. We want to remove all fields that
|
|
|
|
the user has specified, but we do want to keep necessary information like
|
|
|
|
type, state_key etc.
|
2019-01-28 17:42:10 +01:00
|
|
|
"""
|
2020-03-09 13:58:25 +01:00
|
|
|
pruned_event_dict = prune_event_dict(event.room_version, event.get_dict())
|
2019-01-28 17:42:10 +01:00
|
|
|
|
2020-03-05 16:46:44 +01:00
|
|
|
from . import make_event_from_dict
|
2019-06-20 11:32:02 +02:00
|
|
|
|
2020-03-05 16:46:44 +01:00
|
|
|
pruned_event = make_event_from_dict(
|
|
|
|
pruned_event_dict, event.room_version, event.internal_metadata.get_dict()
|
2019-01-28 17:42:10 +01:00
|
|
|
)
|
|
|
|
|
2020-10-05 15:43:14 +02:00
|
|
|
# copy the internal fields
|
|
|
|
pruned_event.internal_metadata.stream_ordering = (
|
|
|
|
event.internal_metadata.stream_ordering
|
|
|
|
)
|
|
|
|
|
2021-03-17 13:33:18 +01:00
|
|
|
pruned_event.internal_metadata.outlier = event.internal_metadata.outlier
|
|
|
|
|
2019-07-18 15:41:42 +02:00
|
|
|
# Mark the event as redacted
|
|
|
|
pruned_event.internal_metadata.redacted = True
|
|
|
|
|
|
|
|
return pruned_event
|
|
|
|
|
2019-01-28 17:42:10 +01:00
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def prune_event_dict(room_version: RoomVersion, event_dict: JsonDict) -> JsonDict:
|
2019-01-28 17:42:10 +01:00
|
|
|
"""Redacts the event_dict in the same way as `prune_event`, except it
|
|
|
|
operates on dicts rather than event objects
|
|
|
|
|
|
|
|
Returns:
|
2020-03-09 13:58:25 +01:00
|
|
|
A copy of the pruned event dict
|
2014-12-03 17:07:21 +01:00
|
|
|
"""
|
|
|
|
|
|
|
|
allowed_keys = [
|
|
|
|
"event_id",
|
|
|
|
"sender",
|
|
|
|
"room_id",
|
|
|
|
"hashes",
|
|
|
|
"signatures",
|
|
|
|
"content",
|
|
|
|
"type",
|
|
|
|
"state_key",
|
|
|
|
"depth",
|
|
|
|
"prev_events",
|
|
|
|
"auth_events",
|
|
|
|
"origin",
|
|
|
|
"origin_server_ts",
|
|
|
|
]
|
|
|
|
|
2021-01-05 13:41:48 +01:00
|
|
|
# Room versions from before MSC2176 had additional allowed keys.
|
|
|
|
if not room_version.msc2176_redaction_rules:
|
|
|
|
allowed_keys.extend(["prev_state", "membership"])
|
|
|
|
|
2019-01-28 17:42:10 +01:00
|
|
|
event_type = event_dict["type"]
|
2015-01-29 17:50:23 +01:00
|
|
|
|
2014-12-03 17:07:21 +01:00
|
|
|
new_content = {}
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def add_fields(*fields: str) -> None:
|
2014-12-03 17:07:21 +01:00
|
|
|
for field in fields:
|
2019-01-28 17:42:10 +01:00
|
|
|
if field in event_dict["content"]:
|
2015-01-29 17:50:23 +01:00
|
|
|
new_content[field] = event_dict["content"][field]
|
2014-12-03 17:07:21 +01:00
|
|
|
|
|
|
|
if event_type == EventTypes.Member:
|
|
|
|
add_fields("membership")
|
2021-09-03 18:51:15 +02:00
|
|
|
if room_version.msc3375_redaction_rules:
|
2021-09-30 17:13:59 +02:00
|
|
|
add_fields(EventContentFields.AUTHORISING_USER)
|
2014-12-03 17:07:21 +01:00
|
|
|
elif event_type == EventTypes.Create:
|
2021-01-05 13:41:48 +01:00
|
|
|
# MSC2176 rules state that create events cannot be redacted.
|
|
|
|
if room_version.msc2176_redaction_rules:
|
|
|
|
return event_dict
|
|
|
|
|
2014-12-03 17:07:21 +01:00
|
|
|
add_fields("creator")
|
|
|
|
elif event_type == EventTypes.JoinRules:
|
|
|
|
add_fields("join_rule")
|
2021-07-28 13:03:01 +02:00
|
|
|
if room_version.msc3083_join_rules:
|
|
|
|
add_fields("allow")
|
2014-12-03 17:07:21 +01:00
|
|
|
elif event_type == EventTypes.PowerLevels:
|
|
|
|
add_fields(
|
|
|
|
"users",
|
|
|
|
"users_default",
|
|
|
|
"events",
|
|
|
|
"events_default",
|
|
|
|
"state_default",
|
|
|
|
"ban",
|
|
|
|
"kick",
|
|
|
|
"redact",
|
|
|
|
)
|
2021-01-05 13:41:48 +01:00
|
|
|
|
|
|
|
if room_version.msc2176_redaction_rules:
|
|
|
|
add_fields("invite")
|
|
|
|
|
2021-07-28 17:46:37 +02:00
|
|
|
if room_version.msc2716_historical:
|
|
|
|
add_fields("historical")
|
|
|
|
|
2020-03-09 13:58:25 +01:00
|
|
|
elif event_type == EventTypes.Aliases and room_version.special_case_aliases_auth:
|
2014-12-03 17:07:21 +01:00
|
|
|
add_fields("aliases")
|
2015-07-03 11:31:17 +02:00
|
|
|
elif event_type == EventTypes.RoomHistoryVisibility:
|
2015-07-06 14:05:52 +02:00
|
|
|
add_fields("history_visibility")
|
2021-01-05 13:41:48 +01:00
|
|
|
elif event_type == EventTypes.Redaction and room_version.msc2176_redaction_rules:
|
|
|
|
add_fields("redacts")
|
2021-08-10 03:23:31 +02:00
|
|
|
elif room_version.msc2716_redactions and event_type == EventTypes.MSC2716_INSERTION:
|
2021-09-21 22:06:28 +02:00
|
|
|
add_fields(EventContentFields.MSC2716_NEXT_BATCH_ID)
|
|
|
|
elif room_version.msc2716_redactions and event_type == EventTypes.MSC2716_BATCH:
|
|
|
|
add_fields(EventContentFields.MSC2716_BATCH_ID)
|
2021-08-10 03:23:31 +02:00
|
|
|
elif room_version.msc2716_redactions and event_type == EventTypes.MSC2716_MARKER:
|
2022-08-19 22:37:01 +02:00
|
|
|
add_fields(EventContentFields.MSC2716_INSERTION_EVENT_REFERENCE)
|
2014-12-03 17:07:21 +01:00
|
|
|
|
2019-06-20 11:32:02 +02:00
|
|
|
allowed_fields = {k: v for k, v in event_dict.items() if k in allowed_keys}
|
2014-12-03 17:07:21 +01:00
|
|
|
|
|
|
|
allowed_fields["content"] = new_content
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
unsigned: JsonDict = {}
|
2019-01-28 17:42:10 +01:00
|
|
|
allowed_fields["unsigned"] = unsigned
|
2014-12-11 14:25:19 +01:00
|
|
|
|
2019-01-28 17:42:10 +01:00
|
|
|
event_unsigned = event_dict.get("unsigned", {})
|
2014-12-11 14:25:19 +01:00
|
|
|
|
2019-01-28 17:42:10 +01:00
|
|
|
if "age_ts" in event_unsigned:
|
|
|
|
unsigned["age_ts"] = event_unsigned["age_ts"]
|
|
|
|
if "replaces_state" in event_unsigned:
|
|
|
|
unsigned["replaces_state"] = event_unsigned["replaces_state"]
|
|
|
|
|
|
|
|
return allowed_fields
|
2014-12-05 17:20:48 +01:00
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def _copy_field(src: JsonDict, dst: JsonDict, field: List[str]) -> None:
|
2016-11-21 18:42:16 +01:00
|
|
|
"""Copy the field in 'src' to 'dst'.
|
|
|
|
|
|
|
|
For example, if src={"foo":{"bar":5}} and dst={}, and field=["foo","bar"]
|
|
|
|
then dst={"foo":{"bar":5}}.
|
|
|
|
|
|
|
|
Args:
|
2021-10-13 13:24:07 +02:00
|
|
|
src: The dict to read from.
|
|
|
|
dst: The dict to modify.
|
|
|
|
field: List of keys to drill down to in 'src'.
|
2016-11-21 18:42:16 +01:00
|
|
|
"""
|
|
|
|
if len(field) == 0: # this should be impossible
|
|
|
|
return
|
|
|
|
if len(field) == 1: # common case e.g. 'origin_server_ts'
|
|
|
|
if field[0] in src:
|
|
|
|
dst[field[0]] = src[field[0]]
|
|
|
|
return
|
|
|
|
|
|
|
|
# Else is a nested field e.g. 'content.body'
|
|
|
|
# Pop the last field as that's the key to move across and we need the
|
|
|
|
# parent dict in order to access the data. Drill down to the right dict.
|
|
|
|
key_to_move = field.pop(-1)
|
|
|
|
sub_dict = src
|
|
|
|
for sub_field in field: # e.g. sub_field => "content"
|
2022-04-27 15:00:07 +02:00
|
|
|
if sub_field in sub_dict and isinstance(
|
|
|
|
sub_dict[sub_field], collections.abc.Mapping
|
|
|
|
):
|
2016-11-21 18:42:16 +01:00
|
|
|
sub_dict = sub_dict[sub_field]
|
|
|
|
else:
|
|
|
|
return
|
|
|
|
|
|
|
|
if key_to_move not in sub_dict:
|
|
|
|
return
|
|
|
|
|
|
|
|
# Insert the key into the output dictionary, creating nested objects
|
|
|
|
# as required. We couldn't do this any earlier or else we'd need to delete
|
|
|
|
# the empty objects if the key didn't exist.
|
|
|
|
sub_out_dict = dst
|
|
|
|
for sub_field in field:
|
2016-11-22 14:42:11 +01:00
|
|
|
sub_out_dict = sub_out_dict.setdefault(sub_field, {})
|
2016-11-21 18:42:16 +01:00
|
|
|
sub_out_dict[key_to_move] = sub_dict[key_to_move]
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def only_fields(dictionary: JsonDict, fields: List[str]) -> JsonDict:
|
2016-11-21 18:42:16 +01:00
|
|
|
"""Return a new dict with only the fields in 'dictionary' which are present
|
|
|
|
in 'fields'.
|
|
|
|
|
|
|
|
If there are no event fields specified then all fields are included.
|
2020-10-23 18:38:40 +02:00
|
|
|
The entries may include '.' characters to indicate sub-fields.
|
2016-11-21 18:42:16 +01:00
|
|
|
So ['content.body'] will include the 'body' field of the 'content' object.
|
|
|
|
A literal '.' character in a field name may be escaped using a '\'.
|
|
|
|
|
|
|
|
Args:
|
2021-10-13 13:24:07 +02:00
|
|
|
dictionary: The dictionary to read from.
|
|
|
|
fields: A list of fields to copy over. Only shallow refs are
|
2016-11-21 18:42:16 +01:00
|
|
|
taken.
|
|
|
|
Returns:
|
2021-10-13 13:24:07 +02:00
|
|
|
A new dictionary with only the given fields. If fields was empty,
|
2016-11-21 18:42:16 +01:00
|
|
|
the same dictionary is returned.
|
|
|
|
"""
|
|
|
|
if len(fields) == 0:
|
|
|
|
return dictionary
|
|
|
|
|
|
|
|
# for each field, convert it:
|
|
|
|
# ["content.body.thing\.with\.dots"] => [["content", "body", "thing\.with\.dots"]]
|
|
|
|
split_fields = [SPLIT_FIELD_REGEX.split(f) for f in fields]
|
|
|
|
|
|
|
|
# for each element of the output array of arrays:
|
2016-11-22 14:42:11 +01:00
|
|
|
# remove escaping so we can use the right key names.
|
|
|
|
split_fields[:] = [
|
2019-06-20 11:32:02 +02:00
|
|
|
[f.replace(r"\.", r".") for f in field_array] for field_array in split_fields
|
2016-11-22 14:42:11 +01:00
|
|
|
]
|
2016-11-21 18:42:16 +01:00
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
output: JsonDict = {}
|
2016-11-21 18:42:16 +01:00
|
|
|
for field_array in split_fields:
|
|
|
|
_copy_field(dictionary, output, field_array)
|
|
|
|
return output
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def format_event_raw(d: JsonDict) -> JsonDict:
|
2015-01-29 03:34:35 +01:00
|
|
|
return d
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def format_event_for_client_v1(d: JsonDict) -> JsonDict:
|
2015-11-30 18:46:35 +01:00
|
|
|
d = format_event_for_client_v2(d)
|
|
|
|
|
2015-12-01 12:14:48 +01:00
|
|
|
sender = d.get("sender")
|
|
|
|
if sender is not None:
|
|
|
|
d["user_id"] = sender
|
2015-01-29 03:34:35 +01:00
|
|
|
|
2015-11-30 18:46:35 +01:00
|
|
|
copy_keys = (
|
2019-06-20 11:32:02 +02:00
|
|
|
"age",
|
|
|
|
"redacted_because",
|
|
|
|
"replaces_state",
|
|
|
|
"prev_content",
|
2015-09-10 15:25:54 +02:00
|
|
|
"invite_room_state",
|
2021-06-09 20:39:51 +02:00
|
|
|
"knock_room_state",
|
2015-09-10 15:25:54 +02:00
|
|
|
)
|
2015-11-30 18:46:35 +01:00
|
|
|
for key in copy_keys:
|
2015-01-29 03:34:35 +01:00
|
|
|
if key in d["unsigned"]:
|
|
|
|
d[key] = d["unsigned"][key]
|
|
|
|
|
|
|
|
return d
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def format_event_for_client_v2(d: JsonDict) -> JsonDict:
|
2015-01-29 03:34:35 +01:00
|
|
|
drop_keys = (
|
2019-06-20 11:32:02 +02:00
|
|
|
"auth_events",
|
|
|
|
"prev_events",
|
|
|
|
"hashes",
|
|
|
|
"signatures",
|
|
|
|
"depth",
|
|
|
|
"origin",
|
|
|
|
"prev_state",
|
2015-01-29 03:34:35 +01:00
|
|
|
)
|
|
|
|
for key in drop_keys:
|
|
|
|
d.pop(key, None)
|
|
|
|
return d
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def format_event_for_client_v2_without_room_id(d: JsonDict) -> JsonDict:
|
2015-01-29 03:34:35 +01:00
|
|
|
d = format_event_for_client_v2(d)
|
|
|
|
d.pop("room_id", None)
|
|
|
|
return d
|
|
|
|
|
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
@attr.s(slots=True, frozen=True, auto_attribs=True)
|
|
|
|
class SerializeEventConfig:
|
|
|
|
as_client_event: bool = True
|
|
|
|
# Function to convert from federation format to client format
|
|
|
|
event_format: Callable[[JsonDict], JsonDict] = format_event_for_client_v1
|
|
|
|
# ID of the user's auth token - used for namespacing of transaction IDs
|
|
|
|
token_id: Optional[int] = None
|
|
|
|
# List of event fields to include. If empty, all fields will be returned.
|
|
|
|
only_event_fields: Optional[List[str]] = None
|
|
|
|
# Some events can have stripped room state stored in the `unsigned` field.
|
|
|
|
# This is required for invite and knock functionality. If this option is
|
|
|
|
# False, that state will be removed from the event before it is returned.
|
|
|
|
# Otherwise, it will be kept.
|
|
|
|
include_stripped_room_state: bool = False
|
|
|
|
|
|
|
|
|
|
|
|
_DEFAULT_SERIALIZE_EVENT_CONFIG = SerializeEventConfig()
|
|
|
|
|
|
|
|
|
2019-06-20 11:32:02 +02:00
|
|
|
def serialize_event(
|
2021-10-13 13:24:07 +02:00
|
|
|
e: Union[JsonDict, EventBase],
|
|
|
|
time_now_ms: int,
|
2021-12-06 16:51:15 +01:00
|
|
|
*,
|
2022-03-03 16:43:06 +01:00
|
|
|
config: SerializeEventConfig = _DEFAULT_SERIALIZE_EVENT_CONFIG,
|
2021-10-13 13:24:07 +02:00
|
|
|
) -> JsonDict:
|
2017-04-26 17:18:08 +02:00
|
|
|
"""Serialize event for clients
|
|
|
|
|
|
|
|
Args:
|
2021-10-13 13:24:07 +02:00
|
|
|
e
|
|
|
|
time_now_ms
|
2022-03-03 16:43:06 +01:00
|
|
|
config: Event serialization config
|
2017-04-26 17:18:08 +02:00
|
|
|
|
|
|
|
Returns:
|
2021-10-13 13:24:07 +02:00
|
|
|
The serialized event dictionary.
|
2017-04-26 17:18:08 +02:00
|
|
|
"""
|
2019-01-29 18:26:24 +01:00
|
|
|
|
2014-12-05 17:20:48 +01:00
|
|
|
# FIXME(erikj): To handle the case of presence events and the like
|
|
|
|
if not isinstance(e, EventBase):
|
|
|
|
return e
|
|
|
|
|
2015-01-26 17:11:28 +01:00
|
|
|
time_now_ms = int(time_now_ms)
|
|
|
|
|
2014-12-05 17:20:48 +01:00
|
|
|
# Should this strip out None's?
|
|
|
|
d = {k: v for k, v in e.get_dict().items()}
|
2015-01-08 15:27:04 +01:00
|
|
|
|
2019-01-29 18:26:24 +01:00
|
|
|
d["event_id"] = e.event_id
|
|
|
|
|
2014-12-05 17:20:48 +01:00
|
|
|
if "age_ts" in d["unsigned"]:
|
2015-01-29 03:34:35 +01:00
|
|
|
d["unsigned"]["age"] = time_now_ms - d["unsigned"]["age_ts"]
|
2015-01-29 03:45:33 +01:00
|
|
|
del d["unsigned"]["age_ts"]
|
2014-12-08 10:08:26 +01:00
|
|
|
|
2014-12-11 14:25:19 +01:00
|
|
|
if "redacted_because" in e.unsigned:
|
2015-01-29 03:34:35 +01:00
|
|
|
d["unsigned"]["redacted_because"] = serialize_event(
|
2022-03-03 16:43:06 +01:00
|
|
|
e.unsigned["redacted_because"], time_now_ms, config=config
|
2014-12-11 14:25:19 +01:00
|
|
|
)
|
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
if config.token_id is not None:
|
|
|
|
if config.token_id == getattr(e.internal_metadata, "token_id", None):
|
2015-01-29 03:45:33 +01:00
|
|
|
txn_id = getattr(e.internal_metadata, "txn_id", None)
|
2015-01-29 03:34:35 +01:00
|
|
|
if txn_id is not None:
|
|
|
|
d["unsigned"]["transaction_id"] = txn_id
|
2014-12-11 14:25:19 +01:00
|
|
|
|
2021-06-09 20:39:51 +02:00
|
|
|
# invite_room_state and knock_room_state are a list of stripped room state events
|
|
|
|
# that are meant to provide metadata about a room to an invitee/knocker. They are
|
|
|
|
# intended to only be included in specific circumstances, such as down sync, and
|
|
|
|
# should not be included in any other case.
|
2022-03-03 16:43:06 +01:00
|
|
|
if not config.include_stripped_room_state:
|
2017-04-26 17:23:30 +02:00
|
|
|
d["unsigned"].pop("invite_room_state", None)
|
2021-06-09 20:39:51 +02:00
|
|
|
d["unsigned"].pop("knock_room_state", None)
|
2017-04-26 17:23:30 +02:00
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
if config.as_client_event:
|
|
|
|
d = config.event_format(d)
|
2016-11-21 18:42:16 +01:00
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
only_event_fields = config.only_event_fields
|
2016-11-22 14:42:11 +01:00
|
|
|
if only_event_fields:
|
2019-06-20 11:32:02 +02:00
|
|
|
if not isinstance(only_event_fields, list) or not all(
|
2020-06-16 14:51:47 +02:00
|
|
|
isinstance(f, str) for f in only_event_fields
|
2019-06-20 11:32:02 +02:00
|
|
|
):
|
2016-11-22 14:42:11 +01:00
|
|
|
raise TypeError("only_event_fields must be a list of strings")
|
2016-11-22 10:59:27 +01:00
|
|
|
d = only_fields(d, only_event_fields)
|
2016-11-21 18:42:16 +01:00
|
|
|
|
|
|
|
return d
|
2019-05-09 14:21:57 +02:00
|
|
|
|
|
|
|
|
2020-09-04 12:54:56 +02:00
|
|
|
class EventClientSerializer:
|
2019-05-09 14:21:57 +02:00
|
|
|
"""Serializes events that are to be sent to clients.
|
|
|
|
|
|
|
|
This is used for bundling extra information with any events to be sent to
|
|
|
|
clients.
|
|
|
|
"""
|
|
|
|
|
2022-01-07 15:10:46 +01:00
|
|
|
def serialize_event(
|
2021-10-13 13:24:07 +02:00
|
|
|
self,
|
|
|
|
event: Union[JsonDict, EventBase],
|
|
|
|
time_now: int,
|
2021-12-06 16:51:15 +01:00
|
|
|
*,
|
2022-03-03 16:43:06 +01:00
|
|
|
config: SerializeEventConfig = _DEFAULT_SERIALIZE_EVENT_CONFIG,
|
2022-01-26 14:27:04 +01:00
|
|
|
bundle_aggregations: Optional[Dict[str, "BundledAggregations"]] = None,
|
2022-04-19 17:42:19 +02:00
|
|
|
apply_edits: bool = True,
|
2021-10-13 13:24:07 +02:00
|
|
|
) -> JsonDict:
|
2019-05-09 14:21:57 +02:00
|
|
|
"""Serializes a single event.
|
|
|
|
|
|
|
|
Args:
|
2021-11-23 12:43:56 +01:00
|
|
|
event: The event being serialized.
|
2021-10-13 13:24:07 +02:00
|
|
|
time_now: The current time in milliseconds
|
2022-03-03 16:43:06 +01:00
|
|
|
config: Event serialization config
|
2022-04-19 17:42:19 +02:00
|
|
|
bundle_aggregations: A map from event_id to the aggregations to be bundled
|
|
|
|
into the event.
|
|
|
|
apply_edits: Whether the content of the event should be modified to reflect
|
|
|
|
any replacement in `bundle_aggregations[<event_id>].replace`.
|
2019-05-09 14:21:57 +02:00
|
|
|
Returns:
|
2021-10-13 13:24:07 +02:00
|
|
|
The serialized event
|
2019-05-09 14:21:57 +02:00
|
|
|
"""
|
2019-05-14 17:59:21 +02:00
|
|
|
# To handle the case of presence events and the like
|
|
|
|
if not isinstance(event, EventBase):
|
2019-07-23 15:00:55 +02:00
|
|
|
return event
|
2019-05-14 17:59:21 +02:00
|
|
|
|
2022-03-03 16:43:06 +01:00
|
|
|
serialized_event = serialize_event(event, time_now, config=config)
|
2019-05-14 17:59:21 +02:00
|
|
|
|
2021-12-06 16:51:15 +01:00
|
|
|
# Check if there are any bundled aggregations to include with the event.
|
2022-01-07 15:10:46 +01:00
|
|
|
if bundle_aggregations:
|
2022-05-04 14:38:18 +02:00
|
|
|
if event.event_id in bundle_aggregations:
|
2022-01-21 11:31:31 +01:00
|
|
|
self._inject_bundled_aggregations(
|
2022-01-07 15:10:46 +01:00
|
|
|
event,
|
|
|
|
time_now,
|
2022-03-03 16:43:06 +01:00
|
|
|
config,
|
2022-05-04 14:38:18 +02:00
|
|
|
bundle_aggregations,
|
2022-01-07 15:10:46 +01:00
|
|
|
serialized_event,
|
2022-04-19 17:42:19 +02:00
|
|
|
apply_edits=apply_edits,
|
2022-01-07 15:10:46 +01:00
|
|
|
)
|
2021-10-21 20:39:16 +02:00
|
|
|
|
2019-07-23 15:00:55 +02:00
|
|
|
return serialized_event
|
2019-05-09 14:21:57 +02:00
|
|
|
|
2022-02-15 14:26:57 +01:00
|
|
|
def _apply_edit(
|
|
|
|
self, orig_event: EventBase, serialized_event: JsonDict, edit: EventBase
|
|
|
|
) -> None:
|
|
|
|
"""Replace the content, preserving existing relations of the serialized event.
|
|
|
|
|
|
|
|
Args:
|
|
|
|
orig_event: The original event.
|
|
|
|
serialized_event: The original event, serialized. This is modified.
|
|
|
|
edit: The event which edits the above.
|
|
|
|
"""
|
|
|
|
|
|
|
|
# Ensure we take copies of the edit content, otherwise we risk modifying
|
|
|
|
# the original event.
|
|
|
|
edit_content = edit.content.copy()
|
|
|
|
|
|
|
|
# Unfreeze the event content if necessary, so that we may modify it below
|
|
|
|
edit_content = unfreeze(edit_content)
|
|
|
|
serialized_event["content"] = edit_content.get("m.new_content", {})
|
|
|
|
|
|
|
|
# Check for existing relations
|
|
|
|
relates_to = orig_event.content.get("m.relates_to")
|
|
|
|
if relates_to:
|
|
|
|
# Keep the relations, ensuring we use a dict copy of the original
|
|
|
|
serialized_event["content"]["m.relates_to"] = relates_to.copy()
|
|
|
|
else:
|
|
|
|
serialized_event["content"].pop("m.relates_to", None)
|
|
|
|
|
2022-01-21 11:31:31 +01:00
|
|
|
def _inject_bundled_aggregations(
|
2022-01-07 15:10:46 +01:00
|
|
|
self,
|
|
|
|
event: EventBase,
|
|
|
|
time_now: int,
|
2022-03-03 16:43:06 +01:00
|
|
|
config: SerializeEventConfig,
|
2022-05-04 14:38:18 +02:00
|
|
|
bundled_aggregations: Dict[str, "BundledAggregations"],
|
2022-01-07 15:10:46 +01:00
|
|
|
serialized_event: JsonDict,
|
2022-04-19 17:42:19 +02:00
|
|
|
apply_edits: bool,
|
2021-11-23 12:43:56 +01:00
|
|
|
) -> None:
|
2021-12-06 16:51:15 +01:00
|
|
|
"""Potentially injects bundled aggregations into the unsigned portion of the serialized event.
|
2021-11-23 12:43:56 +01:00
|
|
|
|
|
|
|
Args:
|
|
|
|
event: The event being serialized.
|
|
|
|
time_now: The current time in milliseconds
|
2022-04-25 14:25:56 +02:00
|
|
|
config: Event serialization config
|
2022-05-04 14:38:18 +02:00
|
|
|
bundled_aggregations: Bundled aggregations to be injected.
|
|
|
|
A map from event_id to aggregation data. Must contain at least an
|
|
|
|
entry for `event`.
|
|
|
|
|
|
|
|
While serializing the bundled aggregations this map may be searched
|
|
|
|
again for additional events in a recursive manner.
|
2021-11-23 12:43:56 +01:00
|
|
|
serialized_event: The serialized event which may be modified.
|
2022-04-19 17:42:19 +02:00
|
|
|
apply_edits: Whether the content of the event should be modified to reflect
|
|
|
|
any replacement in `aggregations.replace`.
|
2021-11-23 12:43:56 +01:00
|
|
|
"""
|
2022-05-04 14:38:18 +02:00
|
|
|
|
|
|
|
# We have already checked that aggregations exist for this event.
|
|
|
|
event_aggregations = bundled_aggregations[event.event_id]
|
|
|
|
|
|
|
|
# The JSON dictionary to be added under the unsigned property of the event
|
|
|
|
# being serialized.
|
2022-01-26 14:27:04 +01:00
|
|
|
serialized_aggregations = {}
|
|
|
|
|
2022-05-04 14:38:18 +02:00
|
|
|
if event_aggregations.annotations:
|
|
|
|
serialized_aggregations[
|
|
|
|
RelationTypes.ANNOTATION
|
|
|
|
] = event_aggregations.annotations
|
2022-01-26 14:27:04 +01:00
|
|
|
|
2022-05-04 14:38:18 +02:00
|
|
|
if event_aggregations.references:
|
|
|
|
serialized_aggregations[
|
|
|
|
RelationTypes.REFERENCE
|
|
|
|
] = event_aggregations.references
|
2021-11-23 12:43:56 +01:00
|
|
|
|
2022-05-04 14:38:18 +02:00
|
|
|
if event_aggregations.replace:
|
2022-04-19 17:42:19 +02:00
|
|
|
# If there is an edit, optionally apply it to the event.
|
2022-05-04 14:38:18 +02:00
|
|
|
edit = event_aggregations.replace
|
2022-04-19 17:42:19 +02:00
|
|
|
if apply_edits:
|
|
|
|
self._apply_edit(event, serialized_event, edit)
|
2021-11-23 12:43:56 +01:00
|
|
|
|
2022-02-15 14:26:57 +01:00
|
|
|
# Include information about it in the relations dict.
|
2022-01-26 14:27:04 +01:00
|
|
|
serialized_aggregations[RelationTypes.REPLACE] = {
|
2021-11-23 12:43:56 +01:00
|
|
|
"event_id": edit.event_id,
|
|
|
|
"origin_server_ts": edit.origin_server_ts,
|
|
|
|
"sender": edit.sender,
|
|
|
|
}
|
|
|
|
|
2022-05-04 14:38:18 +02:00
|
|
|
# Include any threaded replies to this event.
|
|
|
|
if event_aggregations.thread:
|
|
|
|
thread = event_aggregations.thread
|
2022-02-15 14:26:57 +01:00
|
|
|
|
2022-05-04 14:38:18 +02:00
|
|
|
serialized_latest_event = self.serialize_event(
|
|
|
|
thread.latest_event,
|
|
|
|
time_now,
|
|
|
|
config=config,
|
|
|
|
bundle_aggregations=bundled_aggregations,
|
2022-02-15 14:26:57 +01:00
|
|
|
)
|
|
|
|
|
2022-03-10 16:36:13 +01:00
|
|
|
thread_summary = {
|
2022-02-15 14:26:57 +01:00
|
|
|
"latest_event": serialized_latest_event,
|
|
|
|
"count": thread.count,
|
|
|
|
"current_user_participated": thread.current_user_participated,
|
2022-01-26 14:27:04 +01:00
|
|
|
}
|
2022-03-10 16:36:13 +01:00
|
|
|
serialized_aggregations[RelationTypes.THREAD] = thread_summary
|
2021-11-23 12:43:56 +01:00
|
|
|
|
2022-01-07 15:10:46 +01:00
|
|
|
# Include the bundled aggregations in the event.
|
2022-01-26 14:27:04 +01:00
|
|
|
if serialized_aggregations:
|
2022-03-16 17:17:39 +01:00
|
|
|
# There is likely already an "unsigned" field, but a filter might
|
|
|
|
# have stripped it off (via the event_fields option). The server is
|
|
|
|
# allowed to return additional fields, so add it back.
|
|
|
|
serialized_event.setdefault("unsigned", {}).setdefault(
|
|
|
|
"m.relations", {}
|
|
|
|
).update(serialized_aggregations)
|
2022-01-07 15:10:46 +01:00
|
|
|
|
|
|
|
def serialize_events(
|
2022-03-03 16:43:06 +01:00
|
|
|
self,
|
|
|
|
events: Iterable[Union[JsonDict, EventBase]],
|
|
|
|
time_now: int,
|
|
|
|
*,
|
|
|
|
config: SerializeEventConfig = _DEFAULT_SERIALIZE_EVENT_CONFIG,
|
|
|
|
bundle_aggregations: Optional[Dict[str, "BundledAggregations"]] = None,
|
2021-10-13 13:24:07 +02:00
|
|
|
) -> List[JsonDict]:
|
2019-05-09 14:21:57 +02:00
|
|
|
"""Serializes multiple events.
|
|
|
|
|
|
|
|
Args:
|
2021-10-13 13:24:07 +02:00
|
|
|
event
|
|
|
|
time_now: The current time in milliseconds
|
2022-03-03 16:43:06 +01:00
|
|
|
config: Event serialization config
|
|
|
|
bundle_aggregations: Whether to include the bundled aggregations for this
|
|
|
|
event. Only applies to non-state events. (State events never include
|
|
|
|
bundled aggregations.)
|
2019-05-09 14:21:57 +02:00
|
|
|
|
|
|
|
Returns:
|
2021-10-13 13:24:07 +02:00
|
|
|
The list of serialized events
|
2019-05-09 14:21:57 +02:00
|
|
|
"""
|
2022-01-07 15:10:46 +01:00
|
|
|
return [
|
2022-03-03 16:43:06 +01:00
|
|
|
self.serialize_event(
|
|
|
|
event,
|
|
|
|
time_now,
|
|
|
|
config=config,
|
|
|
|
bundle_aggregations=bundle_aggregations,
|
|
|
|
)
|
|
|
|
for event in events
|
2022-01-07 15:10:46 +01:00
|
|
|
]
|
2020-01-28 12:02:55 +01:00
|
|
|
|
|
|
|
|
2022-05-07 14:37:29 +02:00
|
|
|
_PowerLevel = Union[str, int]
|
|
|
|
|
|
|
|
|
|
|
|
def copy_and_fixup_power_levels_contents(
|
|
|
|
old_power_levels: Mapping[str, Union[_PowerLevel, Mapping[str, _PowerLevel]]]
|
2021-10-13 13:24:07 +02:00
|
|
|
) -> Dict[str, Union[int, Dict[str, int]]]:
|
2022-05-07 14:37:29 +02:00
|
|
|
"""Copy the content of a power_levels event, unfreezing frozendicts along the way.
|
|
|
|
|
|
|
|
We accept as input power level values which are strings, provided they represent an
|
|
|
|
integer, e.g. `"`100"` instead of 100. Such strings are converted to integers
|
|
|
|
in the returned dictionary (hence "fixup" in the function name).
|
|
|
|
|
|
|
|
Note that future room versions will outlaw such stringy power levels (see
|
|
|
|
https://github.com/matrix-org/matrix-spec/issues/853).
|
2020-01-28 12:02:55 +01:00
|
|
|
|
|
|
|
Raises:
|
|
|
|
TypeError if the input does not look like a valid power levels event content
|
|
|
|
"""
|
2020-07-20 19:33:04 +02:00
|
|
|
if not isinstance(old_power_levels, collections.abc.Mapping):
|
2020-01-28 12:02:55 +01:00
|
|
|
raise TypeError("Not a valid power-levels content: %r" % (old_power_levels,))
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
power_levels: Dict[str, Union[int, Dict[str, int]]] = {}
|
2020-01-28 12:02:55 +01:00
|
|
|
|
2022-05-07 14:37:29 +02:00
|
|
|
for k, v in old_power_levels.items():
|
2020-07-20 19:33:04 +02:00
|
|
|
if isinstance(v, collections.abc.Mapping):
|
2021-10-13 13:24:07 +02:00
|
|
|
h: Dict[str, int] = {}
|
|
|
|
power_levels[k] = h
|
2020-01-28 12:02:55 +01:00
|
|
|
for k1, v1 in v.items():
|
2022-05-07 14:37:29 +02:00
|
|
|
_copy_power_level_value_as_integer(v1, h, k1)
|
2020-01-28 12:02:55 +01:00
|
|
|
|
2022-05-07 14:37:29 +02:00
|
|
|
else:
|
|
|
|
_copy_power_level_value_as_integer(v, power_levels, k)
|
2020-01-28 12:02:55 +01:00
|
|
|
|
|
|
|
return power_levels
|
2020-05-14 19:24:01 +02:00
|
|
|
|
|
|
|
|
2022-05-07 14:37:29 +02:00
|
|
|
def _copy_power_level_value_as_integer(
|
|
|
|
old_value: object,
|
|
|
|
power_levels: MutableMapping[str, Any],
|
|
|
|
key: str,
|
|
|
|
) -> None:
|
|
|
|
"""Set `power_levels[key]` to the integer represented by `old_value`.
|
|
|
|
|
|
|
|
:raises TypeError: if `old_value` is not an integer, nor a base-10 string
|
|
|
|
representation of an integer.
|
|
|
|
"""
|
|
|
|
if isinstance(old_value, int):
|
|
|
|
power_levels[key] = old_value
|
|
|
|
return
|
|
|
|
|
|
|
|
if isinstance(old_value, str):
|
|
|
|
try:
|
|
|
|
parsed_value = int(old_value, base=10)
|
|
|
|
except ValueError:
|
|
|
|
# Fall through to the final TypeError.
|
|
|
|
pass
|
|
|
|
else:
|
|
|
|
power_levels[key] = parsed_value
|
|
|
|
return
|
|
|
|
|
|
|
|
raise TypeError(f"Invalid power_levels value for {key}: {old_value}")
|
|
|
|
|
|
|
|
|
2021-10-13 13:24:07 +02:00
|
|
|
def validate_canonicaljson(value: Any) -> None:
|
2020-05-14 19:24:01 +02:00
|
|
|
"""
|
|
|
|
Ensure that the JSON object is valid according to the rules of canonical JSON.
|
|
|
|
|
|
|
|
See the appendix section 3.1: Canonical JSON.
|
|
|
|
|
|
|
|
This rejects JSON that has:
|
|
|
|
* An integer outside the range of [-2 ^ 53 + 1, 2 ^ 53 - 1]
|
|
|
|
* Floats
|
|
|
|
* NaN, Infinity, -Infinity
|
|
|
|
"""
|
|
|
|
if isinstance(value, int):
|
2021-08-26 18:07:58 +02:00
|
|
|
if value < CANONICALJSON_MIN_INT or CANONICALJSON_MAX_INT < value:
|
2020-05-14 19:24:01 +02:00
|
|
|
raise SynapseError(400, "JSON integer out of range", Codes.BAD_JSON)
|
|
|
|
|
|
|
|
elif isinstance(value, float):
|
|
|
|
# Note that Infinity, -Infinity, and NaN are also considered floats.
|
|
|
|
raise SynapseError(400, "Bad JSON value: float", Codes.BAD_JSON)
|
|
|
|
|
2022-04-27 15:00:07 +02:00
|
|
|
elif isinstance(value, collections.abc.Mapping):
|
2020-05-14 19:24:01 +02:00
|
|
|
for v in value.values():
|
|
|
|
validate_canonicaljson(v)
|
|
|
|
|
|
|
|
elif isinstance(value, (list, tuple)):
|
|
|
|
for i in value:
|
|
|
|
validate_canonicaljson(i)
|
|
|
|
|
|
|
|
elif not isinstance(value, (bool, str)) and value is not None:
|
|
|
|
# Other potential JSON values (bool, None, str) are safe.
|
|
|
|
raise SynapseError(400, "Unknown JSON value", Codes.BAD_JSON)
|