Doesn't buy us much, but makes everything a bit more consistent.
Also removes the SQL trace driver, as it is unused and the output is
hard to read anyway.
Preparations to actually remove/replace `BaseDendrite`.
Quite a few changes:
- SyncAPI accepts an `fulltext.Indexer` interface (fulltext is removed
from `BaseDendrite`)
- Caches are removed from `BaseDendrite`
- Introduces a `Router` struct (likely to change)
- also fixes#2903
- Introduces a `sqlutil.ConnectionManager`, which should remove
`base.DatabaseConnection` later on
- probably more
This also fixes an issue regarding updates to relations for invalid
events, which could result in us retrying said event over and over
again, if we fail to unmarshal the event to
`gomatrixserverlib.RelationContent`, this was discovered by
`@sleroq:virto.community`
This PR changes the following:
- `StoreEvent` now only stores an event (and possibly prev event),
instead of also doing redactions
- Adds a `MaybeRedactEvent` (pulled out from `StoreEvent`), which should
be called after storing events
- a few other things
This removes most of the code used for polylith/API mode.
This removes the `/api` internal endpoints entirely.
Binary size change roughly 5%:
```
51437560 Feb 13 10:15 dendrite-monolith-server # old
48759008 Feb 13 10:15 dendrite-monolith-server # new
```
Should fix the following issues or make a lot less worse when using
Postgres:
The main issue behind #2911: The client gives up after a certain time,
causing a cascade of context errors, because the response couldn't be
built up fast enough. This mostly happens on accounts with many rooms,
due to the inefficient way we're getting recent events and current state
For #2777: The queries for getting the membership events for history
visibility were being executed for each room (I think 185?), resulting
in a whooping 2k queries for membership events. (Getting the
statesnapshot -> block nids -> actual wanted membership event)
Both should now be better by:
- Using a LATERAL join to get all recent events for all joined rooms in
one go (TODO: maybe do the same for room summary and current state etc)
- If we're lazy loading on initial syncs, we're now not getting the
whole current state, just to drop the majority of it because we're lazy
loading members - we add a filter to exclude membership events on the
first call to `CurrentState`.
- Using an optimized query to get the membership events needed to
calculate history visibility
---------
Co-authored-by: kegsay <kegan@matrix.org>
This adds Sytest and Complement coverage reporting to the nightly
scheduled CI runs.
Fixes a few API mode related issues as well, since we seemingly never
really ran them with Complement.
Also fixes a bug related to device list changes: When we pass in an
empty `newlyLeftRooms` slice, we got a list of all currently joined
rooms with the corresponding members. When we then got the
`newlyJoinedRooms`, we wouldn't update the `changed` slice, because we
already got the user from the `newlyLeftRooms` query. This is fixed by
simply ignoring empty `newlyLeftRooms`.
This adds a new admin endpoint `/_dendrite/admin/purgeRoom/{roomID}`. It
completely erases all database entries for a given room ID.
The roomserver will start by clearing all data for that room and then
will generate an output event to notify downstream components (i.e. the
sync API and federation API) to do the same.
It does not currently clear media and it is currently not implemented
for SQLite since it relies on SQL array operations right now.
Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
Co-authored-by: Till Faelligen <2353100+S7evinK@users.noreply.github.com>
Since #2849 there is no limit for the current state we fetch to
calculate history visibility. In large rooms this can cause us to fetch
thousands of membership events we don't really care about.
This now only gets the state event types and senders in our timeline,
which should significantly reduce the amount of events we fetch from the
database.
Also removes `MaxTopologicalPosition`, as it is an unnecessary DB call,
given we use the result in `topological_position < $1` calls.
Second part for guest access, this adds a `WithAllowGuests()` option to
`MakeAuthAPI`, allowing guests to access the specified endpoints.
Endpoints taken from the
[spec](https://spec.matrix.org/v1.4/client-server-api/#client-behaviour-14)
and by checking Synapse endpoints for `allow_guest=true`.
Proposed fix for issue:
https://github.com/matrix-org/dendrite/issues/2838
Suppose bob received invites to spaceA and spaceB.
When Bob joins spaceA, we add an OutputEvent event to retire the invite.
This sets the invite to "deleted" in the database. This makes sense.
The bug is in stream_invites.go. Triggered when bob received a new
invite for spaceB, and does a client sync.
In the block (line 76)
`for roomID := range retiredInvites
if _, ok := req.Response.Rooms.Invite[roomID]; ok {
continue
}
if _, ok := req.Response.Rooms.Join[roomID]; ok {
continue
}
...
`
Bob is not in either maps even though he had just accepted the invite
for spaceA. Consequently, the spaceA invite is treated as a retired
invite, and a membership Leave event is generated. What bob sees is that
after accepting the invite to spaceB, he lose access to spaceA.
### Pull Request Checklist
<!-- Please read
https://matrix-org.github.io/dendrite/development/contributing before
submitting your pull request -->
* [ ] I have added tests for PR _or_ I have justified why this PR
doesn't need tests.
* [x ] Pull request includes a [sign off below using a legally
identifiable
name](https://matrix-org.github.io/dendrite/development/contributing#sign-off)
_or_ I have already signed off privately
Signed-off-by: `Tak Wai Wong <tak@hntlabs.com>`
Co-authored-by: Neil Alexander <neilalexander@users.noreply.github.com>
This is apparently some incorrect behaviour that we built as a result of
a spec bug (matrix-org/matrix-spec#1314) where we were applying a filter
to the `"state"` section of the `/sync` response incorrectly. The client
then has no way to know that the state was limited.
This PR removes the state limiting, which probably also helps #2842.
This should stop state events disappearing down a gap where we'd try to
separate out the sections *before* applying history visibility instead
of after.
This may be a better approach than #2843 but I hope @tak-hntlabs will
shout if it isn't.
* Multiroom feature
* Run multiroom visibility expiration conditionally
Remove SQLite and go 1.18 for tests matrixes
* Remove sqlite from unit tests
* Fix linter errors
* Do not build with go1.18
* Do not run upgrade tests
* Fix dendrite workflow
* Add forgotten content and timestamp fields to multiroom in sync response
* Fix syncapi multiroom unit tests
* Review adjustments in queries and naming
* Remove no longer maintained linters from golangci-lint configuration
* Document sqlc code generation
If we're going backwards, we were selecting potentially thousands of
events, which in turn were fed to history visibility checks, resulting
in bad sync performance.
