This is a stepping stone before we can go full Rust everywhere. We're
providing a generic interface as we want database access to work in
Synapse and `synapse-rust-apps`. In `synapse-rust-apps`, we will use a
`tokio-postgres` based database connection pool so it's full Rust.
We want to avoid the situation where we have two database connection
pools (one for Python, one for Rust) as we've run into connection
exhaustion problems on Matrix.org before.
As an example of using it and sanity check for all this work (including
tests), I've also ported over the `/versions` handler to the Rust side
with database access. The `/versions` endpoint is the simplest endpoint
I could find that still had some database access. Hopefully the refactor
on `/versions` isn't that controversial as it's not really the point of
this PR. We can always remove it from this PR but it's just here as a
sanity check that all of this works.
### Why `runInteraction(...)`?
Using the same `runInteraction` pattern that we already have in Synapse
means we can port over existing Synapse code/endpoints without much
thought. But this pattern also makes sense because we want[^1]
transactions to have repeatable-read isolation (easy to think about,
less foot-guns). Having everything thappen in a function callback means
we can do retries for serialization/deadlock errors.
[^1]: To note: Ideally, we'd want the least isolation possible but the
problem is that there is no tooling to yell at you when your
queries/logic is wrong so repeatable-read isolation is a great balance.
> When an application receives this error message, it should abort the
current transaction and retry the whole transaction from the beginning.
The second time through, the transaction will see the
previously-committed change as part of its initial view of the database,
so there is no logical conflict in using the new version of the row as
the starting point for the new transaction's update.
>
> Note that only updating transactions might need to be retried;
read-only transactions will never have serialization conflicts.
>
> *--
https://www.postgresql.org/docs/current/transaction-iso.html#XACT-REPEATABLE-READ*
As a note, this strategy is less of an impedance mismatch (aligns more
closely) with Synapse so the glue code for the `python_db_pool` should
also be simpler.
### How does this interact with logcontext (`LoggingContext`)?
See [docs on log
contexts](https://github.com/element-hq/synapse/blob/4e9f7757f17ba81b8747b7f8f9646d17df145aa3/docs/log_contexts.md)
for more background.
We already support normal logging from Rust -> Python with `pyo3-log`
and `log` but as soon as we pass a thread boundary, everything is logged
against the `sentinel` log context. Normally, we want logs and CPU/DB
usage correlated with the request that spawned the work.
You can see how I took a stab at fixing this in
https://github.com/element-hq/synapse/pull/19846 by capturing the
logcontext in a Tokio task local and re-activating as necessary. For
example, in that PR, I reactivated the logcontext in
`run_python_awaitable(...)` which we use to call `runInteraction(...)`
from the Rust side which means all of the database usage is correlated
with the request as expected. It also means any `log:info!(...)` done in
`run_interaction(...)` is correlated correctly. But there needs to be a
better story for when you want to log everywhere else.
I haven't explored tracking CPU usage on the Rust side.
I've left all of this out of this PR as I think it will be better to
tackle this as a dedicated follow-up. For example, I'm thinking about
instead creating a new `LoggingContext` with the `parent_context` set to
the calling context and try to avoid needing to call
`set_current_context(...)` on the Python side where possible (like
tracking CPU).
### Testing strategy
Added some tests that exercise some `async` Rust handlers for the
`/versions` endpoint:
```
SYNAPSE_TEST_LOG_LEVEL=INFO poetry run trial tests.rest.client.test_versions.VersionsTestCase
```
Real-world:
1. `poetry run synapse_homeserver --config-path homeserver.yaml`
1. `GET http://localhost:8008/_matrix/client/versions`
Setting a profile field (including displayname / avatar_uri) now needs to happen on a profile updates stream writer. There are new "dispatch" methods to do the profile field set/delete, that handle pushing the update over replication, if required.
Profile field updates and profile updates stream writes are now in the same database transaction.
Change `/org.matrix.msc3814.v1/dehydrated_device/[device_id]/events` to
accept GET requests instead of POST.
The original version of
[MSC3814](https://github.com/matrix-org/matrix-spec-proposals/pull/3814)
said we should delete keys after returning them from this endpoint, but
it is being updated to say we should not delete them, and therefore the
appropriate verb is GET.
Synapse already doesn't delete anything, so we just need to change to a
GET with a `next_batch` query param. (Currently it is a POST with
`next_batch` in the JSON content.)
This code was initially written by @ara4n and Claude, but both he and I
have read it and think it makes sense. I am far from a Synapse expert,
so feel free to tell me it's all wrong and point me in the right
direction.
I don't know what system tests will be affected by this, but I guess we
will see when the CI runs (right?).
This is a change to an unstable endpoint so no need for notifications
about breaking changes or similar.
Part of https://github.com/element-hq/element-meta/issues/2704
### Pull Request Checklist
* [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:
* [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: Matthew Hodgson <matthew@matrix.org>
Follows: #19487
Part of: MSC4354 whose experimental feature tracking issue is #19409
This PR implements the Sliding Sync (MSC4186) extension described in
MSC4354, allowing sliding sync clients
to receive sticky events in a reliable way.
The logic is much the same as for oldschool sync (implementation in
#19487),
although in the sliding sync extension, the client can choose their own
limit
and must control their own pagination through an extra token in the
extension request/response bodies.
Note this does not yet send down existing sticky events in the
room when the room has been newly-joined.
This newly-discovered gap is tracked at #19662 and will be addressed for
both current sync and MSC4186 SSS soon.
---------
Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>
Co-authored-by: Eric Eastwood <erice@element.io>
I.e. `default_config("test")` and `default_config("test", False)` are so
opaque and hard to connect the dots with.
Spawning from trying to figure out what the `server_name` is set as for
our `HomeserverTestCase` in order to reference it in
https://github.com/element-hq/synapse/pull/19848#discussion_r3397455309
This is in prep for converting the event serialization to Rust.
This is a fairly mechanical port, except that we store the appservice ID
rather than the appservice object. This avoids us having to store a
`Py<..>` (or port the appservice object over).
This is based on https://github.com/element-hq/synapse/pull/18416, which
got reverted (#19614) due to it incorrectly rejecting to-device messages
to users with many devices (and thus breaking message sending).
Fix https://github.com/element-hq/synapse/issues/17035
A to-device message content looks like:
```jsonc
{
"@user:domain": {"device1": {...}, "device2": {...}},
...
}
```
The previous PR would split up into multiple EDUs, each with a subset of
the users. However, if one user's entry was too large it would not
further split it up and then error out.
The main change in this PR is to allow splitting up a single user into
multiple EDUs.
Other changes:
1. Rename to `SOFT_MAX_EDU_SIZE` to indicate that we sometimes send EDUs
with larger size than that, and its more a target than a hard limit.
2. Check early if any to-device message (to a specific device) is too
large to send, even if we're not going to send it over federation. This
ensures that we catch issues where clients try to send too large
to-device.
This still means that if a client send a large individual to-device
message it will fail, but I don't believe we ever send such large
to-device messages (normally they're in the range of a few KB).
---
I ended up changing the implementation a bunch to make it easy to reuse
the code to split up dictionaries. Instead of repeatedly splitting up
the EDU until each bit fits into the size, we instead record the size of
each entry in the dict and instead split up based on cumulative size.
This means we call `encode_canonical_json` on each entry rather than
once on the entire struct, but its not significantly slower to do so.
--
cc @MatMaul @MadLittleMods
---------
Co-authored-by: Mathieu Velten <matmaul@gmail.com>
Co-authored-by: mcalinghee <mcalinghee.dev@gmail.com>
Co-authored-by: Eric Eastwood <madlittlemods@gmail.com>
Ports the event class to Rust.
The main difference here are:
1. There is now a single event class
2. We now validate a lot more at event construction time than we
previously did (we basically checked nothing before). This required some
changes to the tests, including
https://github.com/matrix-org/sytest/pull/1423
Reviewable commit-by-commit.
### Overview of Event Rust structure
The format of the event struct in Rust is quite different than that in
Python.
The top-level looks like:
```rust
pub struct Event {
/// The parsed event JSON.
fields: FormattedEvent,
/// The event ID. For format v1 this is read directly from the JSON;
/// for v2+ it is computed from the canonical-JSON hash at
/// construction time and cached here.
event_id: Arc<str>,
/// Synapse-internal per-event state that lives outside the federated
/// JSON (e.g. outlier flag, soft-failure, stream positions).
#[pyo3(get)]
internal_metadata: EventInternalMetadata,
/// The room version this event was parsed for.
#[pyo3(get)]
room_version: &'static RoomVersion,
/// `None` for accepted events; otherwise a short reason set by auth
/// when the event was rejected.
rejected_reason: Option<Box<str>>,
}
```
which includes the actual parsed event in `FormattedEvent`, plus the
rest of the event metadata.
```rust
pub struct FormattedEvent<E = Arc<EventFormatEnum>> {
#[serde(default)]
pub signatures: Signatures,
#[serde(default)]
pub unsigned: Unsigned,
#[serde(flatten)]
pub specific_fields: E,
#[serde(flatten)]
pub common_fields: Arc<EventCommonFields>,
}
```
The struct is further split into the common fields, format specific
fields, plus the signatures and unsigned. We split out the signature and
unsigned fields as they are mutable, so when we clone the event we can
still share the common and specific fields and only copy signature and
unsigned.
The `specific_fields` are the fields that depend on the format version.
They can either be a specific format (e.g. `E = EventFormatV1`) or a
type-erased enum `EventFormatEnum` that is across all room versions:
```rust
pub enum EventFormatEnum {
V1(EventFormatV1),
V2V3(EventFormatV2V3),
V4(EventFormatV4),
VMSC4242(EventFormatVMSC4242),
}
```
For example:
```rust
/// Shared flat-list encoding of `auth_events` and `prev_events`, reused
/// by every format from v2/v3 onwards.
#[derive(Serialize, Deserialize)]
pub struct SimpleAuthPrevEvents {
pub auth_events: Vec<String>,
pub prev_events: Vec<String>,
}
/// Version-specific fields for room versions 3-10.
#[derive(Serialize, Deserialize)]
pub struct EventFormatV2V3 {
pub room_id: Box<str>,
#[serde(flatten)]
pub auth_prev_events: SimpleAuthPrevEvents,
}
```
### Dev notes
As discussed in
[`#element-backend-internal:matrix.org`](https://matrix.to/#/!SGNQGPGUwtcPBUotTL:matrix.org/$3gTjDO440GbAz57cXcCawwiyFLiD0crrarvS1uhzKOY?via=jki.re&via=element.io&via=matrix.org)
---------
Co-authored-by: Eric Eastwood <erice@element.io>
As per the spec, a room with m.room.name value that is absent, null or
empty should be treated as if there is no m.room.name event at all:
https://spec.matrix.org/v1.17/client-server-api/#mroomname
This fetches the full m.room.name event and checks the content.name
instead of only checking the existence of the m.room.name event. This
results in correctly sending heroes for those rooms.
Fixes: https://github.com/element-hq/synapse/issues/19447
Signed-off-by: Joe Groocock <me@frebib.net>
when an access token had a refresh token associated to it in the
database, deleting this refresh token (for example when deleting the
device using it) would cascade delete the access token, which wouldn't
be returned by the sql query that was supposed to delete it on its own,
and an empty array was passed to the cache invalidation function.
When we port the `Event` class to Rust, the constructor will check for
the existence of required fields. To support that, we tidy up the test
code where we construct fake events to add all the required fields.
There should be no behavioural changes.
Review commit-by-commit.
Based on #19708.
This is on the path to porting the entire event class to Rust, as
`event.content` will then return the new Rust class `JsonObject`.
This PR adds a pure Rust `JsonObject` class that is a `Mapping`
representing a json-style object. It uses `serde_json::Value` as its
in-memory representation and `pythonize` for conversion when a field is
looked up on the object.
I'm not thrilled with the name, but couldn't think of a better one.
This also adds `JsonObject` handling to the JSON serialisation functions
we use, as well as to the `freeze(..)` function.
Reviewable commit-by-commit.
When upgrading a room to v12, we accidentally ended up mutating the
content of the old power level. Since we cache events, this meant any
future usage of the old power level event would see the wrong content
(until it dropped from the cache).
This meant that the creator of the new room would not be able to perform
admin actions on the old room. Any federation requests for the event
would fail the hash checks, since the content had been changed.
All in all, quite a nasty bug.
This fixes the bug described in #19713 (and double-checked against the
SDK integration test, which now passes with this change). A sync
response must be returned immediately if a room subscription
configuration change caused a new non-empty response (checked with `if
response` in the code) to be produced.
Fixes#19713.
Fixes#18844.
---------
Co-authored-by: Erik Johnston <erik@matrix.org>
Currently synapse returns `M_FORBIDDEN` when trying to use the account
deactivation API, if the server admin disabled displayname changes. This
is undesirable, since it prevents GDPR erasure without admin
interaction. The admin API seems to work fine though. This also only
seems to affect the deactivate API, when the erase flag is true.
Relevant endpoint:
https://spec.matrix.org/latest/client-server-api/#post_matrixclientv3accountdeactivate
This change only removes the checked for condition that the displayname
and profile avatar are allowed to be changed per the configuration
setting. If a user is deleting themselves, why is that denied?
There did not seem to be a basic test for this endpoint that checks the
`erase` usage, so that was added as well as checking the above mentioned
behavior.
Closes: #19688
Part of: MSC4450 whose Experimental Feature tracking issue is #19691
Add an unstable, namespaced `idp_id` query parameter to `fallback/web` \
This allows clients to specify the identity provider they'd like to log
in with for SSO when they have multiple upstream IdPs associated with
their account.
Previously, Synapse would just pick one arbitrarily. But this was
undesirable as you may want to use a different one at that point in
time. When logging in, the user is able to choose when IdP they use -
during UIA (which uses fallback auth mechanism) they should be able to
do the same.
-----
Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>
Co-authored-by: Andrew Morgan <andrew@amorgan.xyz>
Co-authored-by: Eric Eastwood <madlittlemods@gmail.com>
When we return events to clients we need to annotate them with the
membership of the user at the time of the event, in the `unsigned`
section. We already check the membership at the event during the
visibility checks, and so we annotate events there. However, since this
a per-user field we end up having to clone the event in question.
Instead, let's add a `FilteredEvent` class that is returned by the
visibility checks, which allows returning the membership without editing
the event. This has three benefits:
1. Avoids the clones of the event.
2. Allows us to statically check that we have filtered events before
returning them to clients.
3. We no longer edit `unsigned` data after event deserialization, this
makes it easier to port the event class to Rust.
The last benefit is why we're doing this *now*, however IMV it shouldn't
affect whether we want this change or not.
Reviewable commit-by-commit
---------
Co-authored-by: Olivier 'reivilibre' <oliverw@element.io>
Follows: #19365
Part of: MSC4354 Sticky Events (experimental feature #19409)
This PR introduces a `spam_checker_spammy` flag, analogous to
`policy_server_spammy`, as an explicit flag
that an event was decided to be spammy by a spam-checker module.
The original Sticky Events PR (#18968) just reused
`policy_server_spammy`, but it didn't sit right with me
because we (at least appear to be experimenting with features that)
allow users to opt-in to seeing
`policy_server_spammy` events (presumably for moderation purposes).
Keeping these flags separate felt best, therefore.
As for why we need this flag: soon soft-failed status won't be
permanent, at least for sticky events.
The spam checker modules currently work by making events soft-failed.
We want to prevent spammy events from getting
reconsidered/un-soft-failed, so it seems like we need
a flag to track spam-checker spamminess *separately* from soft-failed.
Should be commit-by-commit friendly, but is also small.
---------
Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>
The spec says `device_keys` may be omitted, but not set to `null`.
This was temporarily allowed as a workaround for misbehaving clients
(see #19023), which have since been fixed.
Fixes#19030
Reverts element-hq/synapse#18416
Unfortunately, this causes failures on `/sendToDevice` endpoint in
normal circumstances. If a single user has, say, a hundred devices then
we easily go over the limit. This blocks message sending entirely in
encrypted rooms.
cc @MadLittleMods @MatMaul
This is a simplification so that `unsigned` only includes "simple"
values, to make it easier to port to Rust.
Reviewable commit-by-commit
Summary:
1. **Add `recheck` column to `redactions` table**
A new boolean `recheck` column (default true) is added to the
`redactions` table. This captures whether a redaction needs its sender
domain checked at read time — required for room v3+ where redactions are
accepted speculatively and later validated. When persisting a new
redaction, `recheck` is set directly from
`event.internal_metadata.need_to_check_redaction()`.
It's fine if initially we recheck all redactions, as it only results in
a little more CPU overhead (as we always pull out the redaction event
regardless).
2. **Backfill `recheck` via background update**
A background update (`redactions_recheck`) backfills the new column for
existing rows by reading `recheck_redaction` from each event's
`internal_metadata` JSON. This avoids loading full event objects by
reading `event_json` directly via a SQL JOIN.
3. **Don't fetch confirmed redaction events from the DB**
Previously, when loading events, Synapse recursively fetched all
redaction events regardless of whether they needed domain rechecking.
Now `_fetch_event_rows` reads the `recheck` column and splits redactions
into two lists:
- `unconfirmed_redactions` — need fetching and domain validation
- `confirmed_redactions` — already validated, applied directly without
fetching the event
This avoids unnecessary DB reads for the common case of
already-confirmed redactions.
4. **Move `redacted_because` population to `EventClientSerializer`**
Previously, `redacted_because` (the full redaction event object) was
stored in `event.unsigned` at DB fetch time, coupling storage-layer code
to client serialization concerns. This is removed from
`_maybe_redact_event_row` and moved into
`EventClientSerializer.serialize_event`, which fetches the redaction
event on demand. The storage layer now only sets
`unsigned["redacted_by"]` (the redaction event ID).
5. **Always use `EventClientSerializer`**
The standalone `serialize_event` function was made private
(`_serialize_event`). All external callers — `rest/client/room.py`,
`rest/admin/events.py, appservice/api.py`, and `tests` — were updated to
use `EventClientSerializer.serialize_event` / `serialize_events`,
ensuring
`redacted_because` is always populated correctly via the serializer.
6. **Batch-fetch redaction events in `serialize_events`**
`serialize_events` now collects all `redacted_by` IDs from the event
batch upfront and fetches them in a single `get_events` call, passing
the result as a `redaction_map` to each `serialize_event` call. This
reduces N individual DB round-trips to one when serializing a batch of
events that includes redacted events.
---------
Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Updates the error codes to match MSC2666 changes (user ID query param
validation + proper errcode for requesting rooms with self), added the
new `count` field, and stabilized the endpoint.
Fixes: #19540Fixes: #16290 (side effect of the proposed fix)
Closes: #12804 (side effect of the proposed fix)
Introduced in: https://github.com/matrix-org/synapse/pull/8932
---
This PR is a relatively simple simplification of the profile change on
deactivation that appears to remove multiple bugs.
This PR's **primary motivating fix** is #19540: when a user is
deactivated and erased, they would be kept in the user directory. This
bug appears to have been here since #8932 (previously
https://github.com/matrix-org/synapse/pull/8932) (v1.26.0).
The root cause of this bug is that after removing the user from the user
directory, we would immediately update their displayname and avatar to
empty strings (one at a time), which re-inserts
the user into the user directory.
With this PR, we now delete the entire `profiles` row upon user erasure,
which is cleaner (from a 'your database goes back to zero after
deactivating and erasing a user' point of view) and
only needs one database operation (instead of doing displayname then
avatar).
With this PR, we also no longer send the 2 (deferred) `m.room.member`
`join` events to every room to propagate the displayname and avatar_url
changes.
This is good for two reasons:
- the user is about to get parted from those rooms anyway, so this
reduces the number of state events sent per room from 3 to 1. (More
efficient for us in the moment and leaves less litter in the room DAG.)
- it is possible for the displayname/avatar update to be sent **after**
the user parting, which seems as though it could trigger the user to be
re-joined to a public room.
(With that said, although this sounds vaguely familiar in my lossy
memory, I can't find a ticket that actually describes this bug, so this
might be fictional. Edit: #16290 seems to describe this, although the
title is misleading.)
Additionally, as a side effect of the proposed fix (deleting the
`profiles` row), this PR also now deletes custom profile fields upon
user erasure, which is a new feature/bugfix (not sure which) in its own
right.
I do not see a ticket that corresponds to this feature gap, possibly
because custom profile fields are still a niche feature without
mainstream support (to the best of my knowledge).
Tests are included for the primary bugfix and for the cleanup of custom
profile fields.
### `set_displayname` module API change
This change includes a minor _technically_-breaking change to the module
API.
The change concerns `set_displayname` which is exposed to the module API
with a `deactivation: bool = False` flag, matching the internal handler
method it wraps.
I suspect that this is a mistake caused by overly-faithfully piping
through the args from the wrapped method (this Module API was introduced
in
https://github.com/matrix-org/synapse/pull/14629/changes#diff-0b449f6f95672437cf04f0b5512572b4a6a729d2759c438b7c206ea249619885R1592).
The linked PR did the same for `by_admin` originally before it was
changed.
The `deactivation` flag's only purpose is to be piped through to other
Module API callbacks when a module has registered to be notified about
profile changes.
My claim is that it makes no sense for the Module API to have this flag
because it is not the one doing the deactivation, thus it should never
be in a position to set this to `True`.
My proposed change keeps the flag (for function signature
compatibility), but turns it into a no-op (with a `ERROR` log when it's
set to True by the module).
The Module API callback notifying of the module-caused displayname
change will therefore now always have `deactivation = False`.
*Discussed in
[`#synapse-dev:matrix.org`](https://matrix.to/#/!i5D5LLct_DYG-4hQprLzrxdbZ580U9UB6AEgFnk6rZQ/$1f8N6G_EJUI_I_LvplnVAF2UFZTw_FzgsPfB6pbcPKk?via=element.io&via=matrix.org&via=beeper.com)*
---------
Signed-off-by: Olivier 'reivilibre <oliverw@matrix.org>
Jason Robinson