Be able to test `/login/sso/redirect` in Complement

Spawning from
https://github.com/element-hq/sbg/pull/421#discussion_r1854926218 where
we have a proxy that intercepts responses to
`/_matrix/client/v3/login/sso/redirect(/{idpId})` in order to upgrade
them to use OAuth 2.0 Pushed Authorization Requests (PAR). We have some
Complement tests in that codebase that go over this flow and these
changes are required [in order for the URL's to line
up](https://github.com/element-hq/synapse/blob/d648c8ce3f4cbf61191b9f5302e405f7b0288677/synapse/rest/client/login.py#L652-L673).

MSC4115 has now completed FCP, so we can enable it by default and switch
to the stable identifier.

This is an alternative to
https://github.com/matrix-org/matrix-rust-sdk/issues/3340 where we don't
need to change our CI setup too much in the Rust SDK repository, and
still can test MSC3266.

Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com>

MSC3967 was updated recently to make it more robust to network failures:

> there is an existing cross-signing master key and it exactly matches
the cross-signing master key provided in the request body. If there are
any additional keys provided in the request (self signing key, user
signing key) they MUST also match the existing keys stored on the
server. In other words, the request contains no new keys. If there are
new keys, UIA MUST be performed.


https://github.com/matrix-org/matrix-spec-proposals/blob/hughns/device-signing-upload-uia/proposals/3967-device-signing-upload-uia.md#proposal

This covers the case where the 200 OK is lost in transit so the client
retries the upload, only to then get UIA'd.

Complement tests: https://github.com/matrix-org/complement/pull/713 -
passing example
https://github.com/element-hq/synapse/actions/runs/7976948122/job/21778795094?pr=16943#step:7:8820

### Pull Request Checklist

<!-- Please read
https://element-hq.github.io/synapse/latest/development/contributing_guide.html
before submitting your pull request -->

* [x] Pull request is based on the develop branch
* [x] Pull request includes a [changelog
file](https://element-hq.github.io/synapse/latest/development/contributing_guide.html#changelog).
The entry should:
- Be a short description of your change which makes sense to users.
"Fixed a bug that prevented receiving messages from other servers."
instead of "Moved X method from `EventStore` to `EventWorkerStore`.".
  - Use markdown where necessary, mostly for `code blocks`.
  - End with either a period (.) or an exclamation mark (!).
  - Start with a capital letter.
- Feel free to credit yourself, by adding a sentence "Contributed by
@github_username." or "Contributed by [Your Name]." to the end of the
entry.
* [x] [Code
style](https://element-hq.github.io/synapse/latest/code_style.html) is
correct
(run the
[linters](https://element-hq.github.io/synapse/latest/development/contributing_guide.html#run-the-linters)

)

---------

Co-authored-by: reivilibre <oliverw@matrix.org>

Prevent `start_for_complement.sh` from setting `START_POSTGRES` to `false` when it's already set (#16985)

I have a use case where I'd like the Synapse image to start up a
postgres instance that I can use, but don't want to force Synapse to use
postgres as well.

This commit prevents postgres from being started when it has already
been explicitly enabled elsewhere.

* complement: enable dirty runs

* Add changelog

* Set a low connpool limit when running in Complement

Dirty runs can cause many containers to be running concurrently,
which seems to easily exhaust resources on the host. The increased
speedup from dirty runs also seems to use more db connections on
workers, which are misconfigured currently to have
`SUM(workers * cp_max) > max_connections`, causing
```
FATAL:  sorry, too many clients already
```
which results in tests failing.

* Try p=2 concurrency to restrict slowness of servers which causes partial state join tests to flake

* Debug logging

* Only run flakey tests

* Only adjust connection pool limits in worker mode

* Move cp vars to somewhere where they get executed in CI

* Move cp values back to where they actually work

* Debug logging

* Try p=1 to see if this makes worker mode happier

* Remove debug logging

Remove experimental MSC2716 implementation to incrementally import history into existing rooms (#15748)

Context for why we're removing the implementation:

 - https://github.com/matrix-org/matrix-spec-proposals/pull/2716#issuecomment-1487441010
 - https://github.com/matrix-org/matrix-spec-proposals/pull/2716#issuecomment-1504262734

Anyone wanting to continue MSC2716, should also address these leftover tasks: https://github.com/matrix-org/synapse/issues/10737

Closes https://github.com/matrix-org/synapse/issues/10737 in the fact that it is not longer necessary to track those things.

Co-authored-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>

Signed-off-by: Jason Little <realtyem@gmail.com>

* Enable Complement tests for Faster Remote Room Joins on worker-mode

* (dangerous) Add an override to allow Complement to use FRRJ under workers

* Newsfile

Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>

* Fix race where we didn't send out replication notification

* MORE HACKS

* Fix get_un_partial_stated_rooms_token to take instance_name

* Fix bad merge

* Remove warning

* Correctly advance un_partial_stated_room_stream

* Fix merge

* Add another notify_replication

* Fixups

* Create a separate ReplicationNotifier

* Fix test

* Fix portdb

* Create a separate ReplicationNotifier

* Fix test

* Fix portdb

* Fix presence test

* Newsfile

* Apply suggestions from code review

* Update changelog.d/14752.misc

Co-authored-by: Erik Johnston <erik@matrix.org>

* lint

Signed-off-by: Olivier Wilkinson (reivilibre) <oliverw@matrix.org>
Co-authored-by: Erik Johnston <erik@matrix.org>

Serving partial join responses is no longer experimental. They will only be served under the stable identifier if the the undocumented config flag experimental.msc3706_enabled is set to true.

Synapse continues to request a partial join only if the undocumented config flag experimental.faster_joins is set to true; this setting remains present and unaffected.

Fix https://github.com/matrix-org/synapse/issues/14390

 - Client API: `/_matrix/client/unstable/org.matrix.msc3030/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>` -> `/_matrix/client/v1/rooms/<roomID>/timestamp_to_event?ts=<timestamp>&dir=<direction>`
 - Federation API: `/_matrix/federation/unstable/org.matrix.msc3030/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>` -> `/_matrix/federation/v1/timestamp_to_event/<roomID>?ts=<timestamp>&dir=<direction>`

Complement test changes: https://github.com/matrix-org/complement/pull/559

* Expose getting SYNAPSE_WORKER_TYPES from external, allowing override of workers requested.

* Add WORKER_TYPES variable option to complement.sh script that passes requested workers into start_for_complement.sh entrypoint.

* Update docs to reflect this new ability.

* Changelog

* Don't rely on soft wrapping to format long strings

Good idea dklimpel. Thanks for catching that.

Co-authored-by: Dirk Klimpel <5740567+dklimpel@users.noreply.github.com>

* Small nits just noticed in docs.

* Fixup new line in docs.

Co-authored-by: Dirk Klimpel <5740567+dklimpel@users.noreply.github.com>

Co-authored-by: Eric Eastwood <erice@element.io>

When building the docker images for complement testing, copy a preinstalled
complement over from a base image, rather than apt installing it. This avoids
network traffic and is much faster.

Improve startup times in Complement test runs against workers, particularly in CPU-constrained environments. (#13127)

Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>

If no database is configured explicitly, use sqlite.

This means that you don't have to pass any variables into the image.

Co-authored-by: Richard van der Hoff <1389908+richvdh@users.noreply.github.com>

When we join a room via the faster-joins mechanism, we end up with "partial
state" at some points on the event DAG. Many parts of the codebase need to
wait for the full state to load. So, we implement a mechanism to keep track of
which events have partial state, and wait for them to be fully-populated.

Fixes matrix-org/complement#330 (or it will, once we remove the old files).

It's not quite a lift-and-shift: I've also taken the opportunity to get rid of the custom CA that we used to use to sign the TLS certs, which has been superceded by the CA exposed by Complement.