OpenCloud
/
MatrixSynapse
mirror of https://github.com/matrix-org/synapse
1
0
Fork 0
Code Issues Releases Wiki Activity

Allow non-member state sent in room batch to resolve for historic events (MSC2716) (#12329) 

Part of https://github.com/matrix-org/synapse/issues/12110

Complement test: https://github.com/matrix-org/complement/pull/354

Previously, they didn't resolve because async `filter_events_for_client`
removes all outlier state except for out-of-band membership.

And fundamentally, we have the state at these events so they shouldn't be marked as outliers.
pull/12369/head
Eric Eastwood 2022-04-04 14:54:50 -05:00 committed by GitHub
parent b446c99ac9
commit 9633eb2162
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 13 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
changelog.d/12329.bugfix Normal file
View File

@ -0,0 +1 @@
Fix non-member state events not resolving for historical events when used in [MSC2716](https://github.com/matrix-org/matrix-spec-proposals/pull/2716) `/batch_send` `state_events_at_start`.

38
synapse/handlers/room_batch.py
View File

@ -156,8 +156,8 @@ class RoomBatchHandler:
) -> List[str]:
"""Takes all `state_events_at_start` event dictionaries and creates/persists
them in a floating state event chain which don't resolve into the current room
state. They are floating because they reference no prev_events and are marked
as outliers which disconnects them from the normal DAG.
state. They are floating because they reference no prev_events which disconnects
them from the normal DAG.
Args:
state_events_at_start:
@ -213,31 +213,23 @@ class RoomBatchHandler:
room_id=room_id,
action=membership,
content=event_dict["content"],
# Mark as an outlier to disconnect it from the normal DAG
# and not show up between batches of history.
outlier=True,
historical=True,
# Only the first event in the state chain should be floating.
# The rest should hang off each other in a chain.
allow_no_prev_events=index == 0,
prev_event_ids=prev_event_ids_for_state_chain,
# Since each state event is marked as an outlier, the
# `EventContext.for_outlier()` won't have any `state_ids`
# set and therefore can't derive any state even though the
# prev_events are set. Also since the first event in the
# state chain is floating with no `prev_events`, it can't
# derive state from anywhere automatically. So we need to
# set some state explicitly.
# The first event in the state chain is floating with no
# `prev_events` which means it can't derive state from
# anywhere automatically. So we need to set some state
# explicitly.
#
# Make sure to use a copy of this list because we modify it
# later in the loop here. Otherwise it will be the same
# reference and also update in the event when we append later.
# reference and also update in the event when we append
# later.
state_event_ids=state_event_ids.copy(),
)
else:
# TODO: Add some complement tests that adds state that is not member joins
# and will use this code path. Maybe we only want to support join state events
# and can get rid of this `else`?
(
event,
_,
@ -246,21 +238,15 @@ class RoomBatchHandler:
state_event["sender"], app_service_requester.app_service
),
event_dict,
# Mark as an outlier to disconnect it from the normal DAG
# and not show up between batches of history.
outlier=True,
historical=True,
# Only the first event in the state chain should be floating.
# The rest should hang off each other in a chain.
allow_no_prev_events=index == 0,
prev_event_ids=prev_event_ids_for_state_chain,
# Since each state event is marked as an outlier, the
# `EventContext.for_outlier()` won't have any `state_ids`
# set and therefore can't derive any state even though the
# prev_events are set. Also since the first event in the
# state chain is floating with no `prev_events`, it can't
# derive state from anywhere automatically. So we need to
# set some state explicitly.
# The first event in the state chain is floating with no
# `prev_events` which means it can't derive state from
# anywhere automatically. So we need to set some state
# explicitly.
#
# Make sure to use a copy of this list because we modify it
# later in the loop here. Otherwise it will be the same