newsfile

There is a bug with the `StreamChangeCache` where it would incorrectly
return that all entities had changed if asked for entities changed
*since* the earliest stream position.

Note that for streams we use the inequalities: `$min_stream_id <
stream_id <= $max_stream_id`, i.e. when we ask the stream change cache
for all things that have changed since `$stream_id` we don't care for
events that happened *at* `$stream_id`.

Specifically: `_earliest_known_stream_pos` is the position at which we
know that we'll have entries for all changes since that point, we can
use the cache for any stream IDs that equal
`_earliest_known_stream_pos`.

`_earliest_known_stream_pos` is set in three places:
- On startup we set it either to:
  - the current maximum stream ID, with not prefilled values; or
  - the minimum of the latest N values we pulled from the DB
- When we evict items from the bottom, we set it to the stream ID of the
evicted items.

This was changed in https://github.com/matrix-org/synapse/pull/14435,
but I think we were overly conservative there.

---------

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

.github/workflows/fix_lint.yaml

@@ -29,17 +29,13 @@ jobs:
         with:
           install-project: "false"
 
-      - name: Import order (isort)
+      - name: Run ruff check
         continue-on-error: true
-        run: poetry run isort .
+        run: poetry run ruff check --fix .
 
-      - name: Code style (black)
+      - name: Run ruff format
         continue-on-error: true
-        run: poetry run black .
-
-      - name: Semantic checks (ruff)
-        continue-on-error: true
-        run: poetry run ruff --fix .
+        run: poetry run ruff format --quiet .
 
       - run: cargo clippy --all-features --fix -- -D warnings
         continue-on-error: true
@@ -49,4 +45,4 @@ jobs:
 
       - uses: stefanzweifel/git-auto-commit-action@v5
         with:
-            commit_message: "Attempt to fix linting"
+          commit_message: "Attempt to fix linting"

.github/workflows/tests.yml

@@ -131,15 +131,11 @@ jobs:
         with:
           install-project: "false"
 
-      - name: Import order (isort)
-        run: poetry run isort --check --diff .
+      - name: Run ruff check
+        run: poetry run ruff check --output-format=github .
 
-      - name: Code style (black)
-        run: poetry run black --check --diff .
-
-      - name: Semantic checks (ruff)
-        # --quiet suppresses the update check.
-        run: poetry run ruff check --quiet .
+      - name: Run ruff format
+        run: poetry run ruff format --check .
 
   lint-mypy:
     runs-on: ubuntu-latest

CHANGES.md

@@ -1,3 +1,215 @@
+# Synapse 1.116.0 (2024-10-01)
+
+No significant changes since 1.116.0rc2.
+
+
+
+
+# Synapse 1.116.0rc2 (2024-09-26)
+
+### Features
+
+- Add implementation of restricting who can overwrite a state event as proposed by [MSC3757](https://github.com/matrix-org/matrix-spec-proposals/pull/3757). ([\#17513](https://github.com/element-hq/synapse/issues/17513))
+
+
+
+
+# Synapse 1.116.0rc1 (2024-09-25)
+
+### Features
+
+- Add initial implementation of delayed events as proposed by [MSC4140](https://github.com/matrix-org/matrix-spec-proposals/pull/4140). ([\#17326](https://github.com/element-hq/synapse/issues/17326))
+- Add an asynchronous Admin API endpoint [to redact all a user's events](https://element-hq.github.io/synapse/v1.116/admin_api/user_admin_api.html#redact-all-the-events-of-a-user),
+  and [an endpoint to check on the status of that redaction task](https://element-hq.github.io/synapse/v1.116/admin_api/user_admin_api.html#check-the-status-of-a-redaction-process). ([\#17506](https://github.com/element-hq/synapse/issues/17506))
+- Add support for the `tags` and `not_tags` filters for [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync. ([\#17662](https://github.com/element-hq/synapse/issues/17662))
+- Guests can use the new media endpoints to download media, as described by [MSC4189](https://github.com/matrix-org/matrix-spec-proposals/pull/4189). ([\#17675](https://github.com/element-hq/synapse/issues/17675))
+- Add config option `turn_shared_secret_path`. ([\#17690](https://github.com/element-hq/synapse/issues/17690))
+- Return room tags in [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync account data extension. ([\#17707](https://github.com/element-hq/synapse/issues/17707))
+
+### Bugfixes
+
+- Make sure we get up-to-date state information when using the new [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync tables to derive room membership. ([\#17692](https://github.com/element-hq/synapse/issues/17692))
+- Fix bug where room account data would not correctly be sent down [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync for old rooms. ([\#17695](https://github.com/element-hq/synapse/issues/17695))
+- Fix a bug in [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync which could prevent /sync from working for certain user accounts. ([\#17727](https://github.com/element-hq/synapse/issues/17727), [\#17733](https://github.com/element-hq/synapse/issues/17733))
+- Ignore invites from ignored users in Sliding Sync. ([\#17729](https://github.com/element-hq/synapse/issues/17729))
+- Fix bug in [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync where the server would incorrectly return a negative bump stamp, which caused Element X apps to stop syncing. ([\#17748](https://github.com/element-hq/synapse/issues/17748))
+
+### Internal Changes
+
+- Import pydantic objects from the `_pydantic_compat` module.
+  This allows `check_pydantic_models.py` to mock those pydantic objects
+  only in the synapse module, and not interfere with pydantic objects in
+  external dependencies. ([\#17667](https://github.com/element-hq/synapse/issues/17667))
+- Use [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync tables as a bulk shortcut for getting the max `event_stream_ordering` of rooms. ([\#17693](https://github.com/element-hq/synapse/issues/17693))
+- Speed up [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) sliding sync requests a bit where there are many room changes. ([\#17696](https://github.com/element-hq/synapse/issues/17696))
+- Refactor [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) sliding sync filter unit tests so the sliding sync API has better test coverage. ([\#17703](https://github.com/element-hq/synapse/issues/17703))
+- Fetch `bump_stamp`s more efficiently in [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync. ([\#17723](https://github.com/element-hq/synapse/issues/17723))
+- Shortcut for checking if certain background updates have completed (utilized in [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync). ([\#17724](https://github.com/element-hq/synapse/issues/17724))
+- More efficiently fetch rooms for [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync. ([\#17725](https://github.com/element-hq/synapse/issues/17725))
+- Fix `_bulk_get_max_event_pos` being inefficient. ([\#17728](https://github.com/element-hq/synapse/issues/17728))
+- Add cache to `get_tags_for_room(...)`. ([\#17730](https://github.com/element-hq/synapse/issues/17730))
+- Small performance improvement in speeding up [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) Sliding Sync. ([\#17731](https://github.com/element-hq/synapse/issues/17731))
+- Minor speed up of initial [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) sliding sync requests. ([\#17734](https://github.com/element-hq/synapse/issues/17734))
+- Remove usage of the deprecated `cgi` module, deprecated in Python 3.11 and removed in Python 3.13. ([\#17741](https://github.com/element-hq/synapse/issues/17741))
+- Fix typing of a variable that is not `Unknown` anymore after updating `treq`. ([\#17744](https://github.com/element-hq/synapse/issues/17744))
+
+
+
+### Updates to locked dependencies
+
+* Bump anyhow from 1.0.86 to 1.0.89. ([\#17685](https://github.com/element-hq/synapse/issues/17685), [\#17716](https://github.com/element-hq/synapse/issues/17716))
+* Bump bytes from 1.7.1 to 1.7.2. ([\#17743](https://github.com/element-hq/synapse/issues/17743))
+* Bump cryptography from 43.0.0 to 43.0.1. ([\#17689](https://github.com/element-hq/synapse/issues/17689))
+* Bump idna from 3.8 to 3.10. ([\#17758](https://github.com/element-hq/synapse/issues/17758))
+* Bump msgpack from 1.0.8 to 1.1.0. ([\#17759](https://github.com/element-hq/synapse/issues/17759))
+* Bump phonenumbers from 8.13.44 to 8.13.45. ([\#17762](https://github.com/element-hq/synapse/issues/17762))
+* Bump prometheus-client from 0.20.0 to 0.21.0. ([\#17746](https://github.com/element-hq/synapse/issues/17746))
+* Bump pyasn1 from 0.6.0 to 0.6.1. ([\#17714](https://github.com/element-hq/synapse/issues/17714))
+* Bump pyasn1-modules from 0.4.0 to 0.4.1. ([\#17747](https://github.com/element-hq/synapse/issues/17747))
+* Bump pydantic from 2.8.2 to 2.9.2. ([\#17756](https://github.com/element-hq/synapse/issues/17756))
+* Bump python-multipart from 0.0.9 to 0.0.10. ([\#17745](https://github.com/element-hq/synapse/issues/17745))
+* Bump ruff from 0.6.4 to 0.6.7. ([\#17715](https://github.com/element-hq/synapse/issues/17715), [\#17760](https://github.com/element-hq/synapse/issues/17760))
+* Bump sentry-sdk from 2.13.0 to 2.14.0. ([\#17712](https://github.com/element-hq/synapse/issues/17712))
+* Bump serde from 1.0.209 to 1.0.210. ([\#17686](https://github.com/element-hq/synapse/issues/17686))
+* Bump serde_json from 1.0.127 to 1.0.128. ([\#17687](https://github.com/element-hq/synapse/issues/17687))
+* Bump treq from 23.11.0 to 24.9.1. ([\#17744](https://github.com/element-hq/synapse/issues/17744))
+* Bump types-pyyaml from 6.0.12.20240808 to 6.0.12.20240917. ([\#17755](https://github.com/element-hq/synapse/issues/17755))
+* Bump types-requests from 2.32.0.20240712 to 2.32.0.20240914. ([\#17713](https://github.com/element-hq/synapse/issues/17713))
+* Bump types-setuptools from 74.1.0.20240907 to 75.1.0.20240917. ([\#17757](https://github.com/element-hq/synapse/issues/17757))
+
+# Synapse 1.115.0 (2024-09-17)
+
+No significant changes since 1.115.0rc2.
+
+
+
+
+# Synapse 1.115.0rc2 (2024-09-12)
+
+### Internal Changes
+
+- Pre-populate room data used in experimental [MSC3575](https://github.com/matrix-org/matrix-spec-proposals/pull/3575) Sliding Sync `/sync` endpoint for quick filtering/sorting. ([\#17652](https://github.com/element-hq/synapse/issues/17652))
+- Speed up sliding sync by reducing amount of data pulled out of the database for large rooms. ([\#17683](https://github.com/element-hq/synapse/issues/17683))
+
+
+
+
+# Synapse 1.115.0rc1 (2024-09-10)
+
+### Features
+
+- Improve cross-signing upload when using [MSC3861](https://github.com/matrix-org/matrix-spec-proposals/pull/3861) to use a custom UIA flow stage, with web fallback support. ([\#17509](https://github.com/element-hq/synapse/issues/17509))
+
+### Bugfixes
+
+- Return `400 M_BAD_JSON` upon attempting to complete various room actions with a non-local user ID and unknown room ID, rather than an internal server error. ([\#17607](https://github.com/element-hq/synapse/issues/17607))
+- Fix authenticated media responses using a wrong limit when following redirects over federation. ([\#17626](https://github.com/element-hq/synapse/issues/17626))
+- Fix bug where we returned the wrong `bump_stamp` for invites in sliding sync response, causing incorrect ordering of invites in the room list. ([\#17674](https://github.com/element-hq/synapse/issues/17674))
+
+### Improved Documentation
+
+- Clarify that the admin api resource is only loaded on the main process and not workers. ([\#17590](https://github.com/element-hq/synapse/issues/17590))
+- Fixed typo in `saml2_config` config [example](https://element-hq.github.io/synapse/latest/usage/configuration/config_documentation.html#saml2_config). ([\#17594](https://github.com/element-hq/synapse/issues/17594))
+
+### Deprecations and Removals
+
+- Stabilise [MSC4156](https://github.com/matrix-org/matrix-spec-proposals/pull/4156) by removing the `msc4156_enabled` config setting and defaulting it to `true`. ([\#17650](https://github.com/element-hq/synapse/issues/17650))
+
+### Internal Changes
+
+- Update [MSC3861](https://github.com/matrix-org/matrix-spec-proposals/pull/3861) implementation: load the issuer and account management URLs from OIDC discovery. ([\#17407](https://github.com/element-hq/synapse/issues/17407))
+- Pre-populate room data used in experimental [MSC3575](https://github.com/matrix-org/matrix-spec-proposals/pull/3575) Sliding Sync `/sync` endpoint for quick filtering/sorting. ([\#17512](https://github.com/element-hq/synapse/issues/17512), [\#17632](https://github.com/element-hq/synapse/issues/17632), [\#17633](https://github.com/element-hq/synapse/issues/17633), [\#17634](https://github.com/element-hq/synapse/issues/17634), [\#17635](https://github.com/element-hq/synapse/issues/17635), [\#17636](https://github.com/element-hq/synapse/issues/17636), [\#17641](https://github.com/element-hq/synapse/issues/17641), [\#17654](https://github.com/element-hq/synapse/issues/17654), [\#17673](https://github.com/element-hq/synapse/issues/17673))
+- Store sliding sync per-connection state in the database. ([\#17599](https://github.com/element-hq/synapse/issues/17599), [\#17631](https://github.com/element-hq/synapse/issues/17631))
+- Make the sliding sync `PerConnectionState` class immutable. ([\#17600](https://github.com/element-hq/synapse/issues/17600))
+- Replace `isort` and `black` with `ruff`. ([\#17620](https://github.com/element-hq/synapse/issues/17620), [\#17643](https://github.com/element-hq/synapse/issues/17643))
+- Sliding Sync: Split up `get_room_membership_for_user_at_to_token`. ([\#17629](https://github.com/element-hq/synapse/issues/17629))
+- Use new database tables for sliding sync. ([\#17630](https://github.com/element-hq/synapse/issues/17630), [\#17649](https://github.com/element-hq/synapse/issues/17649))
+- Prevent duplicate tags being added to Sliding Sync traces. ([\#17655](https://github.com/element-hq/synapse/issues/17655))
+- Get `bump_stamp` from [new sliding sync tables](https://github.com/element-hq/synapse/pull/17512) which should be faster. ([\#17658](https://github.com/element-hq/synapse/issues/17658))
+- Speed up incremental Sliding Sync requests by avoiding extra work. ([\#17665](https://github.com/element-hq/synapse/issues/17665))
+- Small performance improvement in speeding up sliding sync. ([\#17666](https://github.com/element-hq/synapse/issues/17666), [\#17670](https://github.com/element-hq/synapse/issues/17670), [\#17672](https://github.com/element-hq/synapse/issues/17672))
+- Speed up sliding sync by reducing number of database calls. ([\#17684](https://github.com/element-hq/synapse/issues/17684))
+- Speed up sync by pulling out fewer events from the database. ([\#17688](https://github.com/element-hq/synapse/issues/17688))
+
+
+
+### Updates to locked dependencies
+
+* Bump authlib from 1.3.1 to 1.3.2. ([\#17679](https://github.com/element-hq/synapse/issues/17679))
+* Bump idna from 3.7 to 3.8. ([\#17682](https://github.com/element-hq/synapse/issues/17682))
+* Bump ruff from 0.6.2 to 0.6.4. ([\#17680](https://github.com/element-hq/synapse/issues/17680))
+* Bump towncrier from 24.7.1 to 24.8.0. ([\#17645](https://github.com/element-hq/synapse/issues/17645))
+* Bump twisted from 24.7.0rc1 to 24.7.0. ([\#17647](https://github.com/element-hq/synapse/issues/17647))
+* Bump types-pillow from 10.2.0.20240520 to 10.2.0.20240822. ([\#17644](https://github.com/element-hq/synapse/issues/17644))
+* Bump types-psycopg2 from 2.9.21.20240417 to 2.9.21.20240819. ([\#17646](https://github.com/element-hq/synapse/issues/17646))
+* Bump types-setuptools from 71.1.0.20240818 to 74.1.0.20240907. ([\#17681](https://github.com/element-hq/synapse/issues/17681))
+
+# Synapse 1.114.0 (2024-09-02)
+
+This release enables support for
+[MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186) —
+Simplified Sliding Sync. This allows using the upcoming releases of the Element
+X mobile apps without having to run a Sliding Sync Proxy.
+
+
+### Features
+
+- Enable native sliding sync support ([MSC3575](https://github.com/matrix-org/matrix-spec-proposals/pull/3575) and [MSC4186](https://github.com/matrix-org/matrix-spec-proposals/pull/4186)) by default. ([\#17648](https://github.com/element-hq/synapse/issues/17648))
+
+
+
+
+# Synapse 1.114.0rc3 (2024-08-30)
+
+### Bugfixes
+
+- Fix regression in v1.114.0rc2 that caused workers to fail to start. ([\#17626](https://github.com/element-hq/synapse/issues/17626))
+
+
+
+
+# Synapse 1.114.0rc2 (2024-08-30)
+
+### Features
+
+- Improve cross-signing upload when using [MSC3861](https://github.com/matrix-org/matrix-spec-proposals/pull/3861) to use a custom UIA flow stage, with web fallback support. ([\#17509](https://github.com/element-hq/synapse/issues/17509))
+- Make `hash_password` script accept password input from stdin. ([\#17608](https://github.com/element-hq/synapse/issues/17608))
+
+### Bugfixes
+
+- Fix hierarchy returning 403 when room is accessible through federation. Contributed by Krishan (@kfiven). ([\#17194](https://github.com/element-hq/synapse/issues/17194))
+- Fix content-length on federation `/thumbnail` responses. ([\#17532](https://github.com/element-hq/synapse/issues/17532))
+- Fix authenticated media responses using a wrong limit when following redirects over federation. ([\#17543](https://github.com/element-hq/synapse/issues/17543))
+
+### Internal Changes
+
+- MSC3861: load the issuer and account management URLs from OIDC discovery. ([\#17407](https://github.com/element-hq/synapse/issues/17407))
+- Refactor sliding sync class into multiple files. ([\#17595](https://github.com/element-hq/synapse/issues/17595))
+- Store sliding sync per-connection state in the database. ([\#17599](https://github.com/element-hq/synapse/issues/17599))
+- Make the sliding sync `PerConnectionState` class immutable. ([\#17600](https://github.com/element-hq/synapse/issues/17600))
+- Add support to `@tag_args` for standalone functions. ([\#17604](https://github.com/element-hq/synapse/issues/17604))
+- Speed up incremental syncs in sliding sync by adding some more caching. ([\#17606](https://github.com/element-hq/synapse/issues/17606))
+- Always return the user's own read receipts in sliding sync. ([\#17617](https://github.com/element-hq/synapse/issues/17617))
+- Replace `isort` and `black` with `ruff`. ([\#17620](https://github.com/element-hq/synapse/issues/17620))
+- Refactor sliding sync code to move room list logic out into a separate class. ([\#17622](https://github.com/element-hq/synapse/issues/17622))
+
+
+
+### Updates to locked dependencies
+
+* Bump attrs from 23.2.0 to 24.2.0. ([\#17609](https://github.com/element-hq/synapse/issues/17609))
+* Bump cryptography from 42.0.8 to 43.0.0. ([\#17584](https://github.com/element-hq/synapse/issues/17584))
+* Bump phonenumbers from 8.13.43 to 8.13.44. ([\#17610](https://github.com/element-hq/synapse/issues/17610))
+* Bump pygithub from 2.3.0 to 2.4.0. ([\#17612](https://github.com/element-hq/synapse/issues/17612))
+* Bump pyyaml from 6.0.1 to 6.0.2. ([\#17611](https://github.com/element-hq/synapse/issues/17611))
+* Bump sentry-sdk from 2.12.0 to 2.13.0. ([\#17585](https://github.com/element-hq/synapse/issues/17585))
+* Bump serde from 1.0.206 to 1.0.208. ([\#17581](https://github.com/element-hq/synapse/issues/17581))
+* Bump serde from 1.0.208 to 1.0.209. ([\#17613](https://github.com/element-hq/synapse/issues/17613))
+* Bump serde_json from 1.0.124 to 1.0.125. ([\#17582](https://github.com/element-hq/synapse/issues/17582))
+* Bump serde_json from 1.0.125 to 1.0.127. ([\#17614](https://github.com/element-hq/synapse/issues/17614))
+* Bump types-jsonschema from 4.23.0.20240712 to 4.23.0.20240813. ([\#17583](https://github.com/element-hq/synapse/issues/17583))
+* Bump types-setuptools from 71.1.0.20240726 to 71.1.0.20240818. ([\#17586](https://github.com/element-hq/synapse/issues/17586))
+
 # Synapse 1.114.0rc1 (2024-08-20)
 
 ### Features

Cargo.lock

@@ -13,9 +13,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.86"
+version = "1.0.89"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b3d1d046238990b9cf5bcde22a3fb3584ee5cf65fb2765f454ed428c7a0063da"
+checksum = "86fdf8605db99b54d3cd748a44c6d04df638eb5dafb219b135d0149bd0db01f6"
 
 [[package]]
 name = "arc-swap"
@@ -67,9 +67,9 @@ checksum = "79296716171880943b8470b5f8d03aa55eb2e645a4874bdbb28adb49162e012c"
 
 [[package]]
 name = "bytes"
-version = "1.7.1"
+version = "1.7.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8318a53db07bb3f8dca91a600466bdb3f2eaadeedfdbcf02e1accbad9271ba50"
+checksum = "428d9aa8fbc0670b7b8d6030a7fadd0f86151cae55e4dbbece15f3780a3dfaf3"
 
 [[package]]
 name = "cfg-if"
@@ -444,9 +444,9 @@ dependencies = [
 
 [[package]]
 name = "regex"
-version = "1.10.6"
+version = "1.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4219d74c6b67a3654a9fbebc4b419e22126d13d2f3c4a07ee0cb61ff79a79619"
+checksum = "38200e5ee88914975b69f657f0801b6f6dccafd44fd9326302a4aaeecfacb1d8"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -456,9 +456,9 @@ dependencies = [
 
 [[package]]
 name = "regex-automata"
-version = "0.4.6"
+version = "0.4.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "86b83b8b9847f9bf95ef68afb0b8e6cdb80f498442f5179a29fad448fcc1eaea"
+checksum = "368758f23274712b504848e9d5a6f010445cc8b87a7cdb4d7cbee666c1288da3"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -467,9 +467,9 @@ dependencies = [
 
 [[package]]
 name = "regex-syntax"
-version = "0.8.3"
+version = "0.8.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "adad44e29e4c806119491a7f06f03de4d1af22c3a680dd47f1e6e179439d1f56"
+checksum = "2b15c43186be67a4fd63bee50d0303afffcef381492ebe2c5d87f324e1b8815c"
 
 [[package]]
 name = "ryu"
@@ -485,18 +485,18 @@ checksum = "94143f37725109f92c262ed2cf5e59bce7498c01bcc1502d7b9afe439a4e9f49"
 
 [[package]]
 name = "serde"
-version = "1.0.208"
+version = "1.0.210"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cff085d2cb684faa248efb494c39b68e522822ac0de72ccf08109abde717cfb2"
+checksum = "c8e3592472072e6e22e0a54d5904d9febf8508f65fb8552499a1abc7d1078c3a"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.208"
+version = "1.0.210"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "24008e81ff7613ed8e5ba0cfaf24e2c2f1e5b8a0495711e44fcd4882fca62bcf"
+checksum = "243902eda00fad750862fc144cea25caca5e20d615af0a81bee94ca738f1df1f"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -505,9 +505,9 @@ dependencies = [
 
 [[package]]
 name = "serde_json"
-version = "1.0.125"
+version = "1.0.128"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "83c8e735a073ccf5be70aa8066aa984eaf2fa000db6c8d0100ae605b366d31ed"
+checksum = "6ff5456707a1de34e7e37f2a6fd3d3f808c318259cbd01ab6377795054b483d8"
 dependencies = [
  "itoa",
  "memchr",

README.rst

@@ -158,7 +158,7 @@ it:
 
 We **strongly** recommend using a CAPTCHA, particularly if your homeserver is exposed to
 the public internet. Without it, anyone can freely register accounts on your homeserver.
-This can be exploited by attackers to create spambots targetting the rest of the Matrix
+This can be exploited by attackers to create spambots targeting the rest of the Matrix
 federation.
 
 Your new user name will be formed partly from the ``server_name``, and partly

changelog.d/17512.misc

@@ -1 +0,0 @@
-Pre-populate room data used in experimental [MSC3575](https://github.com/matrix-org/matrix-spec-proposals/pull/3575) Sliding Sync `/sync` endpoint for quick filtering/sorting.

changelog.d/17595.misc

@@ -1 +0,0 @@
-Refactor sliding sync class into multiple files.

changelog.d/17599.misc

@@ -1 +0,0 @@
-Store sliding sync per-connection state in the database.

changelog.d/17600.misc

@@ -1 +0,0 @@
-Make the sliding sync `PerConnectionState` class immutable.

changelog.d/17749.doc

@@ -0,0 +1 @@
+Remove spurious "TODO UPDATE ALL THIS" note in the Debian installation docs.

changelog.d/17751.misc

@@ -0,0 +1 @@
+Minor performance increase for large accounts using sliding sync.

changelog.d/17765.misc

@@ -0,0 +1 @@
+Increase performance of the notifier when there are many syncing users.

changelog.d/17766.misc

@@ -0,0 +1 @@
+Increase performance of the notifier when there are many syncing users.

changelog.d/17767.misc

@@ -0,0 +1 @@
+Fix performance of streams that don't change often.

changelog.d/17768.misc

@@ -0,0 +1 @@
+Improve performance of sliding sync connections that do not ask for any rooms.

changelog.d/17771.misc

@@ -0,0 +1 @@
+Reduce overhead of sliding sync E2EE loops.

changelog.d/17779.bugfix

@@ -0,0 +1 @@
+Fix a rare bug introduced in v1.29.0 where invalidating a user's access token from a worker could raise an error.

contrib/cmdclient/console.py

@@ -21,7 +21,8 @@
 #
 #
 
-""" Starts a synapse client console. """
+"""Starts a synapse client console."""
+
 import argparse
 import binascii
 import cmd

contrib/graph/graph.py

@@ -20,8 +20,8 @@
 #
 
 import argparse
-import cgi
 import datetime
+import html
 import json
 import urllib.request
 from typing import List
@@ -85,7 +85,7 @@ def make_graph(pdus: List[dict], filename_prefix: str) -> None:
             "name": name,
             "type": pdu.get("pdu_type"),
             "state_key": pdu.get("state_key"),
-            "content": cgi.escape(json.dumps(pdu.get("content")), quote=True),
+            "content": html.escape(json.dumps(pdu.get("content")), quote=True),
             "time": t,
             "depth": pdu.get("depth"),
         }

debian/changelog

@@ -1,3 +1,57 @@
+matrix-synapse-py3 (1.116.0) stable; urgency=medium
+
+  * New Synapse release 1.116.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 01 Oct 2024 11:14:07 +0100
+
+matrix-synapse-py3 (1.116.0~rc2) stable; urgency=medium
+
+  * New synapse release 1.116.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Thu, 26 Sep 2024 13:28:43 +0000
+
+matrix-synapse-py3 (1.116.0~rc1) stable; urgency=medium
+
+  * New synapse release 1.116.0rc1.
+
+ -- Synapse Packaging team <packages@matrix.org>  Wed, 25 Sep 2024 09:34:07 +0000
+
+matrix-synapse-py3 (1.115.0) stable; urgency=medium
+
+  * New Synapse release 1.115.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 17 Sep 2024 14:32:10 +0100
+
+matrix-synapse-py3 (1.115.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.115.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Thu, 12 Sep 2024 11:10:15 +0100
+
+matrix-synapse-py3 (1.115.0~rc1) stable; urgency=medium
+
+  * New Synapse release 1.115.0rc1.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 10 Sep 2024 08:39:09 -0600
+
+matrix-synapse-py3 (1.114.0) stable; urgency=medium
+
+  * New Synapse release 1.114.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Mon, 02 Sep 2024 15:14:53 +0100
+
+matrix-synapse-py3 (1.114.0~rc3) stable; urgency=medium
+
+  * New Synapse release 1.114.0rc3.
+
+ -- Synapse Packaging team <packages@matrix.org>  Fri, 30 Aug 2024 16:38:05 +0100
+
+matrix-synapse-py3 (1.114.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.114.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Fri, 30 Aug 2024 15:35:13 +0100
+
 matrix-synapse-py3 (1.114.0~rc1) stable; urgency=medium
 
   * New synapse release 1.114.0rc1.

debian/hash_password.1

@@ -1,10 +1,13 @@
-.\" generated with Ronn-NG/v0.8.0
-.\" http://github.com/apjanke/ronn-ng/tree/0.8.0
-.TH "HASH_PASSWORD" "1" "July 2021" "" ""
+.\" generated with Ronn-NG/v0.10.1
+.\" http://github.com/apjanke/ronn-ng/tree/0.10.1
+.TH "HASH_PASSWORD" "1" "August 2024" ""
 .SH "NAME"
 \fBhash_password\fR \- Calculate the hash of a new password, so that passwords can be reset
 .SH "SYNOPSIS"
-\fBhash_password\fR [\fB\-p\fR|\fB\-\-password\fR [password]] [\fB\-c\fR|\fB\-\-config\fR \fIfile\fR]
+.TS
+allbox;
+\fBhash_password\fR [\fB\-p\fR	\fB\-\-password\fR [password]] [\fB\-c\fR	\fB\-\-config\fR \fIfile\fR]
+.TE
 .SH "DESCRIPTION"
 \fBhash_password\fR calculates the hash of a supplied password using bcrypt\.
 .P
@@ -20,7 +23,7 @@ bcrypt_rounds: 17 password_config: pepper: "random hashing pepper"
 .SH "OPTIONS"
 .TP
 \fB\-p\fR, \fB\-\-password\fR
-Read the password form the command line if [password] is supplied\. If not, prompt the user and read the password form the \fBSTDIN\fR\. It is not recommended to type the password on the command line directly\. Use the STDIN instead\.
+Read the password form the command line if [password] is supplied, or from \fBSTDIN\fR\. If not, prompt the user and read the password from the tty prompt\. It is not recommended to type the password on the command line directly\. Use the STDIN instead\.
 .TP
 \fB\-c\fR, \fB\-\-config\fR
 Read the supplied YAML \fIfile\fR containing the options \fBbcrypt_rounds\fR and the \fBpassword_config\fR section containing the \fBpepper\fR value\.
@@ -33,7 +36,17 @@ $2b$12$VJNqWQYfsWTEwcELfoSi4Oa8eA17movHqqi8\.X8fWFpum7SxZ9MFe
 .fi
 .IP "" 0
 .P
-Hash from the STDIN:
+Hash from the stdin:
+.IP "" 4
+.nf
+$ cat password_file | hash_password
+Password:
+Confirm password:
+$2b$12$AszlvfmJl2esnyhmn8m/kuR2tdXgROWtWxnX\.rcuAbM8ErLoUhybG
+.fi
+.IP "" 0
+.P
+Hash from the prompt:
 .IP "" 4
 .nf
 $ hash_password
@@ -53,6 +66,6 @@ $2b$12$CwI\.wBNr\.w3kmiUlV3T5s\.GT2wH7uebDCovDrCOh18dFedlANK99O
 .fi
 .IP "" 0
 .SH "COPYRIGHT"
-This man page was written by Rahul De <\fI\%mailto:rahulde@swecha\.net\fR> for Debian GNU/Linux distribution\.
+This man page was written by Rahul De «rahulde@swecha\.net» for Debian GNU/Linux distribution\.
 .SH "SEE ALSO"
 synctl(1), synapse_port_db(1), register_new_matrix_user(1), synapse_review_recent_signups(1)

debian/hash_password.1.html

@@ -0,0 +1,182 @@
+<!DOCTYPE html>
+<html>
+<head>
+  <meta http-equiv='content-type' content='text/html;charset=utf-8'>
+  <meta name='generator' content='Ronn-NG/v0.10.1 (http://github.com/apjanke/ronn-ng/tree/0.10.1)'>
+  <title>hash_password(1) - Calculate the hash of a new password, so that passwords can be reset</title>
+  <style type='text/css' media='all'>
+  /* style: man */
+  body#manpage {margin:0}
+  .mp {max-width:100ex;padding:0 9ex 1ex 4ex}
+  .mp p,.mp pre,.mp ul,.mp ol,.mp dl {margin:0 0 20px 0}
+  .mp h2 {margin:10px 0 0 0}
+  .mp > p,.mp > pre,.mp > ul,.mp > ol,.mp > dl {margin-left:8ex}
+  .mp h3 {margin:0 0 0 4ex}
+  .mp dt {margin:0;clear:left}
+  .mp dt.flush {float:left;width:8ex}
+  .mp dd {margin:0 0 0 9ex}
+  .mp h1,.mp h2,.mp h3,.mp h4 {clear:left}
+  .mp pre {margin-bottom:20px}
+  .mp pre+h2,.mp pre+h3 {margin-top:22px}
+  .mp h2+pre,.mp h3+pre {margin-top:5px}
+  .mp img {display:block;margin:auto}
+  .mp h1.man-title {display:none}
+  .mp,.mp code,.mp pre,.mp tt,.mp kbd,.mp samp,.mp h3,.mp h4 {font-family:monospace;font-size:14px;line-height:1.42857142857143}
+  .mp h2 {font-size:16px;line-height:1.25}
+  .mp h1 {font-size:20px;line-height:2}
+  .mp {text-align:justify;background:#fff}
+  .mp,.mp code,.mp pre,.mp pre code,.mp tt,.mp kbd,.mp samp {color:#131211}
+  .mp h1,.mp h2,.mp h3,.mp h4 {color:#030201}
+  .mp u {text-decoration:underline}
+  .mp code,.mp strong,.mp b {font-weight:bold;color:#131211}
+  .mp em,.mp var {font-style:italic;color:#232221;text-decoration:none}
+  .mp a,.mp a:link,.mp a:hover,.mp a code,.mp a pre,.mp a tt,.mp a kbd,.mp a samp {color:#0000ff}
+  .mp b.man-ref {font-weight:normal;color:#434241}
+  .mp pre {padding:0 4ex}
+  .mp pre code {font-weight:normal;color:#434241}
+  .mp h2+pre,h3+pre {padding-left:0}
+  ol.man-decor,ol.man-decor li {margin:3px 0 10px 0;padding:0;float:left;width:33%;list-style-type:none;text-transform:uppercase;color:#999;letter-spacing:1px}
+  ol.man-decor {width:100%}
+  ol.man-decor li.tl {text-align:left}
+  ol.man-decor li.tc {text-align:center;letter-spacing:4px}
+  ol.man-decor li.tr {text-align:right;float:right}
+  </style>
+</head>
+<!--
+  The following styles are deprecated and will be removed at some point:
+  div#man, div#man ol.man, div#man ol.head, div#man ol.man.
+
+  The .man-page, .man-decor, .man-head, .man-foot, .man-title, and
+  .man-navigation should be used instead.
+-->
+<body id='manpage'>
+  <div class='mp' id='man'>
+
+  <div class='man-navigation' style='display:none'>
+    <a href="#NAME">NAME</a>
+    <a href="#SYNOPSIS">SYNOPSIS</a>
+    <a href="#DESCRIPTION">DESCRIPTION</a>
+    <a href="#FILES">FILES</a>
+    <a href="#OPTIONS">OPTIONS</a>
+    <a href="#EXAMPLES">EXAMPLES</a>
+    <a href="#COPYRIGHT">COPYRIGHT</a>
+    <a href="#SEE-ALSO">SEE ALSO</a>
+  </div>
+
+  <ol class='man-decor man-head man head'>
+    <li class='tl'>hash_password(1)</li>
+    <li class='tc'></li>
+    <li class='tr'>hash_password(1)</li>
+  </ol>
+
+  
+
+<h2 id="NAME">NAME</h2>
+<p class="man-name">
+  <code>hash_password</code> - <span class="man-whatis">Calculate the hash of a new password, so that passwords can be reset</span>
+</p>
+<h2 id="SYNOPSIS">SYNOPSIS</h2>
+
+<table>
+  <tbody>
+    <tr>
+      <td>
+<code>hash_password</code> [<code>-p</code>
+</td>
+      <td>
+<code>--password</code> [password]] [<code>-c</code>
+</td>
+      <td>
+<code>--config</code> <var>file</var>]</td>
+    </tr>
+  </tbody>
+</table>
+
+<h2 id="DESCRIPTION">DESCRIPTION</h2>
+
+<p><strong>hash_password</strong> calculates the hash of a supplied password using bcrypt.</p>
+
+<p><code>hash_password</code> takes a password as an parameter either on the command line
+or the <code>STDIN</code> if not supplied.</p>
+
+<p>It accepts an YAML file which can be used to specify parameters like the
+number of rounds for bcrypt and password_config section having the pepper
+value used for the hashing. By default <code>bcrypt_rounds</code> is set to <strong>12</strong>.</p>
+
+<p>The hashed password is written on the <code>STDOUT</code>.</p>
+
+<h2 id="FILES">FILES</h2>
+
+<p>A sample YAML file accepted by <code>hash_password</code> is described below:</p>
+
+<p>bcrypt_rounds: 17
+  password_config:
+    pepper: "random hashing pepper"</p>
+
+<h2 id="OPTIONS">OPTIONS</h2>
+
+<dl>
+<dt>
+<code>-p</code>, <code>--password</code>
+</dt>
+<dd>Read the password form the command line if [password] is supplied, or from <code>STDIN</code>.
+If not, prompt the user and read the password from the tty prompt.
+It is not recommended to type the password on the command line
+directly. Use the STDIN instead.</dd>
+<dt>
+<code>-c</code>, <code>--config</code>
+</dt>
+<dd>Read the supplied YAML <var>file</var> containing the options <code>bcrypt_rounds</code>
+and the <code>password_config</code> section containing the <code>pepper</code> value.</dd>
+</dl>
+
+<h2 id="EXAMPLES">EXAMPLES</h2>
+
+<p>Hash from the command line:</p>
+
+<pre><code>$ hash_password -p "p@ssw0rd"
+$2b$12$VJNqWQYfsWTEwcELfoSi4Oa8eA17movHqqi8.X8fWFpum7SxZ9MFe
+</code></pre>
+
+<p>Hash from the stdin:</p>
+
+<pre><code>$ cat password_file | hash_password
+Password:
+Confirm password:
+$2b$12$AszlvfmJl2esnyhmn8m/kuR2tdXgROWtWxnX.rcuAbM8ErLoUhybG
+</code></pre>
+
+<p>Hash from the prompt:</p>
+
+<pre><code>$ hash_password
+Password:
+Confirm password:
+$2b$12$AszlvfmJl2esnyhmn8m/kuR2tdXgROWtWxnX.rcuAbM8ErLoUhybG
+</code></pre>
+
+<p>Using a config file:</p>
+
+<pre><code>$ hash_password -c config.yml
+Password:
+Confirm password:
+$2b$12$CwI.wBNr.w3kmiUlV3T5s.GT2wH7uebDCovDrCOh18dFedlANK99O
+</code></pre>
+
+<h2 id="COPYRIGHT">COPYRIGHT</h2>
+
+<p>This man page was written by Rahul De «rahulde@swecha.net»
+for Debian GNU/Linux distribution.</p>
+
+<h2 id="SEE-ALSO">SEE ALSO</h2>
+
+<p><span class="man-ref">synctl<span class="s">(1)</span></span>, <span class="man-ref">synapse_port_db<span class="s">(1)</span></span>, <span class="man-ref">register_new_matrix_user<span class="s">(1)</span></span>, <span class="man-ref">synapse_review_recent_signups<span class="s">(1)</span></span></p>
+
+  <ol class='man-decor man-foot man foot'>
+    <li class='tl'></li>
+    <li class='tc'>August 2024</li>
+    <li class='tr'>hash_password(1)</li>
+  </ol>
+
+  </div>
+</body>
+</html>

debian/hash_password.ronn

@@ -29,8 +29,8 @@ A sample YAML file accepted by `hash_password` is described below:
 ## OPTIONS
 
   * `-p`, `--password`:
-    Read the password form the command line if [password] is supplied.
-    If not, prompt the user and read the password form the `STDIN`.
+    Read the password form the command line if [password] is supplied, or from `STDIN`.
+    If not, prompt the user and read the password from the tty prompt.
     It is not recommended to type the password on the command line
     directly. Use the STDIN instead.
 
@@ -45,7 +45,14 @@ Hash from the command line:
     $ hash_password -p "p@ssw0rd"
     $2b$12$VJNqWQYfsWTEwcELfoSi4Oa8eA17movHqqi8.X8fWFpum7SxZ9MFe
 
-Hash from the STDIN:
+Hash from the stdin:
+
+    $ cat password_file | hash_password
+    Password:
+    Confirm password:
+    $2b$12$AszlvfmJl2esnyhmn8m/kuR2tdXgROWtWxnX.rcuAbM8ErLoUhybG
+
+Hash from the prompt:
 
     $ hash_password
     Password:

docker/complement/conf/workers-shared-extra.yaml.j2

@@ -111,6 +111,9 @@ server_notices:
   system_mxid_avatar_url: ""
   room_name: "Server Alert"
 
+# Enable delayed events (msc4140)
+max_event_delay_duration: 24h
+
 
 # Disable sync cache so that initial `/sync` requests are up-to-date.
 caches:

docs/admin_api/user_admin_api.md

@@ -1361,3 +1361,83 @@ Returns a `404` HTTP status code if no user was found, with a response body like
 ```
 
 _Added in Synapse 1.72.0._
+
+
+## Redact all the events of a user
+
+The API is 
+```
+POST /_synapse/admin/v1/user/$user_id/redact
+
+{
+  "rooms": ["!roomid1", "!roomid2"]
+}
+```
+If an empty list is provided as the key for `rooms`, all events in all the rooms the user is member of will be redacted, 
+otherwise all the events in the rooms provided in the request will be redacted. 
+
+The API starts redaction process running, and returns immediately with a JSON body with
+a redact id which can be used to query the status of the redaction process:
+
+```json
+{
+    "redact_id": "<opaque id>"
+}
+```
+
+**Parameters**
+
+The following parameters should be set in the URL:
+
+- `user_id` - The fully qualified MXID of the user: for example, `@user:server.com`.
+
+The following JSON body parameter must be provided:
+
+-  `rooms` - A list of rooms to redact the user's events in. If an empty list is provided all events in all rooms
+  the user is a member of will be redacted
+
+_Added in Synapse 1.116.0._
+
+The following JSON body parameters are optional:
+
+- `reason` - Reason the redaction is being requested, ie "spam", "abuse", etc. This will be included in each redaction event, and be visible to users.
+- `limit` - a limit on the number of the user's events to search for ones that can be redacted (events are redacted newest to oldest) in each room, defaults to 1000 if not provided
+
+
+## Check the status of a redaction process
+
+It is possible to query the status of the background task for redacting a user's events.
+The status can be queried up to 24 hours after completion of the task,
+or until Synapse is restarted (whichever happens first).
+
+The API is:
+
+```
+GET /_synapse/admin/v1/user/redact_status/$redact_id
+```
+
+A response body like the following is returned:
+
+```
+{
+  "status": "active",
+  "failed_redactions": [],
+}
+```
+
+**Parameters**
+
+The following parameters should be set in the URL:
+
+* `redact_id` - string - The ID for this redaction process, provided when the redaction was requested.
+
+
+**Response**
+
+The following fields are returned in the JSON response body:
+
+- `status` - string - one of scheduled/active/completed/failed, indicating the status of the redaction job
+- `failed_redactions` - dictionary - the keys of the dict are event ids the process was unable to redact, if any, and the values are 
+  the corresponding error that caused the redaction to fail
+
+_Added in Synapse 1.116.0._

docs/code_style.md

@@ -8,9 +8,7 @@ errors in code.
 
 The necessary tools are:
 
-- [black](https://black.readthedocs.io/en/stable/), a source code formatter;
-- [isort](https://pycqa.github.io/isort/), which organises each file's imports;
-- [ruff](https://github.com/charliermarsh/ruff), which can spot common errors; and
+- [ruff](https://github.com/charliermarsh/ruff), which can spot common errors and enforce a consistent style; and
 - [mypy](https://mypy.readthedocs.io/en/stable/), a type checker.
 
 See [the contributing guide](development/contributing_guide.md#run-the-linters) for instructions

docs/setup/installation.md

@@ -52,8 +52,6 @@ architecture via <https://packages.matrix.org/debian/>.
 
 To install the latest release:
 
-TODO UPDATE ALL THIS
-
 ```sh
 sudo apt install -y lsb-release wget apt-transport-https
 sudo wget -O /usr/share/keyrings/matrix-org-archive-keyring.gpg https://packages.matrix.org/debian/matrix-org-archive-keyring.gpg
@@ -316,7 +314,7 @@ sudo dnf group install "Development Tools"
 
 *Note: The term "RHEL" below refers to both Red Hat Enterprise Linux and Rocky Linux. The distributions are 1:1 binary compatible.*
 
-It's recommended to use the latest Python versions. 
+It's recommended to use the latest Python versions.
 
 RHEL 8 in particular ships with Python 3.6 by default which is EOL and therefore no longer supported by Synapse. RHEL 9 ship with Python 3.9 which is still supported by the Python core team as of this writing. However, newer Python versions provide significant performance improvements and they're available in official distributions' repositories. Therefore it's recommended to use them.
 
@@ -346,7 +344,7 @@ dnf install python3.12 python3.12-devel
 ```
 Finally, install common prerequisites
 ```bash
-dnf install libicu libicu-devel libpq5 libpq5-devel lz4 pkgconf 
+dnf install libicu libicu-devel libpq5 libpq5-devel lz4 pkgconf
 dnf group install "Development Tools"
 ```
 ###### Using venv module instead of virtualenv command
@@ -355,7 +353,7 @@ It's recommended to use Python venv module directly rather than the virtualenv c
 * On RHEL 9, virtualenv is only available on [EPEL](https://docs.fedoraproject.org/en-US/epel/).
 * On RHEL 8, virtualenv is based on Python 3.6. It does not support creating 3.11/3.12 virtual environments.
 
-Here's an example of creating Python 3.12 virtual environment and installing Synapse from PyPI. 
+Here's an example of creating Python 3.12 virtual environment and installing Synapse from PyPI.
 
 ```bash
 mkdir -p ~/synapse

docs/usage/configuration/config_documentation.md

@@ -509,7 +509,8 @@ Unix socket support (_Added in Synapse 1.89.0_):
 
 Valid resource names are:
 
-* `client`: the client-server API (/_matrix/client), and the synapse admin API (/_synapse/admin). Also implies `media` and `static`.
+* `client`: the client-server API (/_matrix/client). Also implies `media` and `static`.
+  If configuring the main process, the Synapse Admin API (/_synapse/admin) is also implied.
 
 * `consent`: user consent forms (/_matrix/consent). See [here](../../consent_tracking.md) for more.
 
@@ -760,6 +761,19 @@ email:
     password_reset: "[%(server_name)s] Password reset"
     email_validation: "[%(server_name)s] Validate your email"
 ```
+---
+### `max_event_delay_duration`
+
+The maximum allowed duration by which sent events can be delayed, as per
+[MSC4140](https://github.com/matrix-org/matrix-spec-proposals/pull/4140).
+Must be a positive value if set.
+
+Defaults to no duration (`null`), which disallows sending delayed events.
+
+Example configuration:
+```yaml
+max_event_delay_duration: 24h
+```
 
 ## Homeserver blocking
 Useful options for Synapse admins.
@@ -1765,7 +1779,7 @@ rc_3pid_validation:
 
 This option sets ratelimiting how often invites can be sent in a room or to a
 specific user. `per_room` defaults to `per_second: 0.3`, `burst_count: 10`,
-`per_user` defaults to `per_second: 0.003`, `burst_count: 5`, and `per_issuer` 
+`per_user` defaults to `per_second: 0.003`, `burst_count: 5`, and `per_issuer`
 defaults to `per_second: 0.3`, `burst_count: 10`.
 
 Client requests that invite user(s) when [creating a
@@ -1966,7 +1980,7 @@ max_image_pixels: 35M
 ---
 ### `remote_media_download_burst_count`
 
-Remote media downloads are ratelimited using a [leaky bucket algorithm](https://en.wikipedia.org/wiki/Leaky_bucket), where a given "bucket" is keyed to the IP address of the requester when requesting remote media downloads. This configuration option sets the size of the bucket against which the size in bytes of downloads are penalized - if the bucket is full, ie a given number of bytes have already been downloaded, further downloads will be denied until the bucket drains.  Defaults to 500MiB. See also `remote_media_download_per_second` which determines the rate at which the "bucket" is emptied and thus has available space to authorize new requests.  
+Remote media downloads are ratelimited using a [leaky bucket algorithm](https://en.wikipedia.org/wiki/Leaky_bucket), where a given "bucket" is keyed to the IP address of the requester when requesting remote media downloads. This configuration option sets the size of the bucket against which the size in bytes of downloads are penalized - if the bucket is full, ie a given number of bytes have already been downloaded, further downloads will be denied until the bucket drains.  Defaults to 500MiB. See also `remote_media_download_per_second` which determines the rate at which the "bucket" is emptied and thus has available space to authorize new requests.
 
 Example configuration:
 ```yaml
@@ -2314,6 +2328,22 @@ Example configuration:
 ```yaml
 turn_shared_secret: "YOUR_SHARED_SECRET"
 ```
+---
+### `turn_shared_secret_path`
+
+An alternative to [`turn_shared_secret`](#turn_shared_secret):
+allows the shared secret to be specified in an external file.
+
+The file should be a plain text file, containing only the shared secret.
+Synapse reads the shared secret from the given file once at startup.
+
+Example configuration:
+```yaml
+turn_shared_secret_path: /path/to/secrets/file
+```
+
+_Added in Synapse 1.116.0._
+
 ---
 ### `turn_username` and `turn_password`
 
@@ -3302,8 +3332,8 @@ saml2_config:
     contact_person:
       - given_name: Bob
         sur_name: "the Sysadmin"
-        email_address": ["admin@example.com"]
-        contact_type": technical
+        email_address: ["admin@example.com"]
+        contact_type: technical
 
   saml_session_lifetime: 5m
 

docs/workers.md

@@ -290,6 +290,7 @@ information.
 Additionally, the following REST endpoints can be handled for GET requests:
 
     ^/_matrix/client/(api/v1|r0|v3|unstable)/pushrules/
+    ^/_matrix/client/unstable/org.matrix.msc4140/delayed_events
 
 Pagination requests can also be handled, but all requests for a given
 room must be routed to the same instance. Additionally, care must be taken to

pylint.cfg

@@ -1,280 +0,0 @@
-[MASTER]
-
-# Specify a configuration file.
-#rcfile=
-
-# Python code to execute, usually for sys.path manipulation such as
-# pygtk.require().
-#init-hook=
-
-# Profiled execution.
-profile=no
-
-# Add files or directories to the blacklist. They should be base names, not
-# paths.
-ignore=CVS
-
-# Pickle collected data for later comparisons.
-persistent=yes
-
-# List of plugins (as comma separated values of python modules names) to load,
-# usually to register additional checkers.
-load-plugins=
-
-
-[MESSAGES CONTROL]
-
-# Enable the message, report, category or checker with the given id(s). You can
-# either give multiple identifier separated by comma (,) or put this option
-# multiple time. See also the "--disable" option for examples.
-#enable=
-
-# Disable the message, report, category or checker with the given id(s). You
-# can either give multiple identifiers separated by comma (,) or put this
-# option multiple times (only on the command line, not in the configuration
-# file where it should appear only once).You can also use "--disable=all" to
-# disable everything first and then reenable specific checks. For example, if
-# you want to run only the similarities checker, you can use "--disable=all
-# --enable=similarities". If you want to run only the classes checker, but have
-# no Warning level messages displayed, use"--disable=all --enable=classes
-# --disable=W"
-disable=missing-docstring
-
-
-[REPORTS]
-
-# Set the output format. Available formats are text, parseable, colorized, msvs
-# (visual studio) and html. You can also give a reporter class, eg
-# mypackage.mymodule.MyReporterClass.
-output-format=text
-
-# Put messages in a separate file for each module / package specified on the
-# command line instead of printing them on stdout. Reports (if any) will be
-# written in a file name "pylint_global.[txt|html]".
-files-output=no
-
-# Tells whether to display a full report or only the messages
-reports=yes
-
-# Python expression which should return a note less than 10 (10 is the highest
-# note). You have access to the variables errors warning, statement which
-# respectively contain the number of errors / warnings messages and the total
-# number of statements analyzed. This is used by the global evaluation report
-# (RP0004).
-evaluation=10.0 - ((float(5 * error + warning + refactor + convention) / statement) * 10)
-
-# Add a comment according to your evaluation note. This is used by the global
-# evaluation report (RP0004).
-comment=no
-
-# Template used to display messages. This is a python new-style format string
-# used to format the message information. See doc for all details
-#msg-template=
-
-
-[TYPECHECK]
-
-# Tells whether missing members accessed in mixin class should be ignored. A
-# mixin class is detected if its name ends with "mixin" (case insensitive).
-ignore-mixin-members=yes
-
-# List of classes names for which member attributes should not be checked
-# (useful for classes with attributes dynamically set).
-ignored-classes=SQLObject
-
-# When zope mode is activated, add a predefined set of Zope acquired attributes
-# to generated-members.
-zope=no
-
-# List of members which are set dynamically and missed by pylint inference
-# system, and so shouldn't trigger E0201 when accessed. Python regular
-# expressions are accepted.
-generated-members=REQUEST,acl_users,aq_parent
-
-
-[MISCELLANEOUS]
-
-# List of note tags to take in consideration, separated by a comma.
-notes=FIXME,XXX,TODO
-
-
-[SIMILARITIES]
-
-# Minimum lines number of a similarity.
-min-similarity-lines=4
-
-# Ignore comments when computing similarities.
-ignore-comments=yes
-
-# Ignore docstrings when computing similarities.
-ignore-docstrings=yes
-
-# Ignore imports when computing similarities.
-ignore-imports=no
-
-
-[VARIABLES]
-
-# Tells whether we should check for unused import in __init__ files.
-init-import=no
-
-# A regular expression matching the beginning of the name of dummy variables
-# (i.e. not used).
-dummy-variables-rgx=_$|dummy
-
-# List of additional names supposed to be defined in builtins. Remember that
-# you should avoid to define new builtins when possible.
-additional-builtins=
-
-
-[BASIC]
-
-# Required attributes for module, separated by a comma
-required-attributes=
-
-# List of builtins function names that should not be used, separated by a comma
-bad-functions=map,filter,apply,input
-
-# Regular expression which should only match correct module names
-module-rgx=(([a-z_][a-z0-9_]*)|([A-Z][a-zA-Z0-9]+))$
-
-# Regular expression which should only match correct module level names
-const-rgx=(([A-Z_][A-Z0-9_]*)|(__.*__))$
-
-# Regular expression which should only match correct class names
-class-rgx=[A-Z_][a-zA-Z0-9]+$
-
-# Regular expression which should only match correct function names
-function-rgx=[a-z_][a-z0-9_]{2,30}$
-
-# Regular expression which should only match correct method names
-method-rgx=[a-z_][a-z0-9_]{2,30}$
-
-# Regular expression which should only match correct instance attribute names
-attr-rgx=[a-z_][a-z0-9_]{2,30}$
-
-# Regular expression which should only match correct argument names
-argument-rgx=[a-z_][a-z0-9_]{2,30}$
-
-# Regular expression which should only match correct variable names
-variable-rgx=[a-z_][a-z0-9_]{2,30}$
-
-# Regular expression which should only match correct attribute names in class
-# bodies
-class-attribute-rgx=([A-Za-z_][A-Za-z0-9_]{2,30}|(__.*__))$
-
-# Regular expression which should only match correct list comprehension /
-# generator expression variable names
-inlinevar-rgx=[A-Za-z_][A-Za-z0-9_]*$
-
-# Good variable names which should always be accepted, separated by a comma
-good-names=i,j,k,ex,Run,_
-
-# Bad variable names which should always be refused, separated by a comma
-bad-names=foo,bar,baz,toto,tutu,tata
-
-# Regular expression which should only match function or class names that do
-# not require a docstring.
-no-docstring-rgx=__.*__
-
-# Minimum line length for functions/classes that require docstrings, shorter
-# ones are exempt.
-docstring-min-length=-1
-
-
-[FORMAT]
-
-# Maximum number of characters on a single line.
-max-line-length=80
-
-# Regexp for a line that is allowed to be longer than the limit.
-ignore-long-lines=^\s*(# )?<?https?://\S+>?$
-
-# Allow the body of an if to be on the same line as the test if there is no
-# else.
-single-line-if-stmt=no
-
-# List of optional constructs for which whitespace checking is disabled
-no-space-check=trailing-comma,dict-separator
-
-# Maximum number of lines in a module
-max-module-lines=1000
-
-# String used as indentation unit. This is usually " " (4 spaces) or "\t" (1
-# tab).
-indent-string='    '
-
-
-[DESIGN]
-
-# Maximum number of arguments for function / method
-max-args=5
-
-# Argument names that match this expression will be ignored. Default to name
-# with leading underscore
-ignored-argument-names=_.*
-
-# Maximum number of locals for function / method body
-max-locals=15
-
-# Maximum number of return / yield for function / method body
-max-returns=6
-
-# Maximum number of branch for function / method body
-max-branches=12
-
-# Maximum number of statements in function / method body
-max-statements=50
-
-# Maximum number of parents for a class (see R0901).
-max-parents=7
-
-# Maximum number of attributes for a class (see R0902).
-max-attributes=7
-
-# Minimum number of public methods for a class (see R0903).
-min-public-methods=2
-
-# Maximum number of public methods for a class (see R0904).
-max-public-methods=20
-
-
-[IMPORTS]
-
-# Deprecated modules which should not be used, separated by a comma
-deprecated-modules=regsub,TERMIOS,Bastion,rexec
-
-# Create a graph of every (i.e. internal and external) dependencies in the
-# given file (report RP0402 must not be disabled)
-import-graph=
-
-# Create a graph of external dependencies in the given file (report RP0402 must
-# not be disabled)
-ext-import-graph=
-
-# Create a graph of internal dependencies in the given file (report RP0402 must
-# not be disabled)
-int-import-graph=
-
-
-[CLASSES]
-
-# List of interface methods to ignore, separated by a comma. This is used for
-# instance to not check methods defines in Zope's Interface base class.
-ignore-iface-methods=isImplementedBy,deferred,extends,names,namesAndDescriptions,queryDescriptionFor,getBases,getDescriptionFor,getDoc,getName,getTaggedValue,getTaggedValueTags,isEqualOrExtendedBy,setTaggedValue,isImplementedByInstancesOf,adaptWith,is_implemented_by
-
-# List of method names used to declare (i.e. assign) instance attributes.
-defining-attr-methods=__init__,__new__,setUp
-
-# List of valid names for the first argument in a class method.
-valid-classmethod-first-arg=cls
-
-# List of valid names for the first argument in a metaclass class method.
-valid-metaclass-classmethod-first-arg=mcs
-
-
-[EXCEPTIONS]
-
-# Exceptions that will emit a warning when being caught. Defaults to
-# "Exception"
-overgeneral-exceptions=Exception

pyproject.toml

@@ -34,14 +34,9 @@
         name = "Internal Changes"
         showcontent = true
 
-[tool.black]
-target-version = ['py38', 'py39', 'py310', 'py311']
-# black ignores everything in .gitignore by default, see
-# https://black.readthedocs.io/en/stable/usage_and_configuration/file_collection_and_discovery.html#gitignore
-# Use `extend-exclude` if you want to exclude something in addition to this.
-
 [tool.ruff]
 line-length = 88
+target-version = "py38"
 
 [tool.ruff.lint]
 # See https://beta.ruff.rs/docs/rules/#error-e
@@ -63,6 +58,8 @@ select = [
     "W",
     # pyflakes
     "F",
+    # isort
+    "I001",
     # flake8-bugbear
     "B0",
     # flake8-comprehensions
@@ -79,17 +76,20 @@ select = [
     "EXE",
 ]
 
-[tool.isort]
-line_length = 88
-sections = ["FUTURE", "STDLIB", "THIRDPARTY", "TWISTED", "FIRSTPARTY", "TESTS", "LOCALFOLDER"]
-default_section = "THIRDPARTY"
-known_first_party = ["synapse"]
-known_tests = ["tests"]
-known_twisted = ["twisted", "OpenSSL"]
-multi_line_output = 3
-include_trailing_comma = true
-combine_as_imports = true
-skip_gitignore = true
+[tool.ruff.lint.isort]
+combine-as-imports = true
+section-order = ["future", "standard-library", "third-party", "twisted", "first-party", "testing", "local-folder"]
+known-first-party = ["synapse"]
+
+[tool.ruff.lint.isort.sections]
+twisted = ["twisted", "OpenSSL"]
+testing = ["tests"]
+
+[tool.ruff.format]
+quote-style = "double"
+indent-style = "space"
+skip-magic-trailing-comma = false
+line-ending = "auto"
 
 [tool.maturin]
 manifest-path = "rust/Cargo.toml"
@@ -97,7 +97,7 @@ module-name = "synapse.synapse_rust"
 
 [tool.poetry]
 name = "matrix-synapse"
-version = "1.114.0rc1"
+version = "1.116.0"
 description = "Homeserver for the Matrix decentralised comms protocol"
 authors = ["Matrix.org Team and Contributors <packages@matrix.org>"]
 license = "AGPL-3.0-or-later"
@@ -320,9 +320,7 @@ all = [
 # failing on new releases. Keeping lower bounds loose here means that dependabot
 # can bump versions without having to update the content-hash in the lockfile.
 # This helps prevents merge conflicts when running a batch of dependabot updates.
-isort = ">=5.10.1"
-black = ">=22.7.0"
-ruff = "0.5.5"
+ruff = "0.6.8"
 # Type checking only works with the pydantic.v1 compat module from pydantic v2
 pydantic = "^2"
 

scripts-dev/check_pydantic_models.py

@@ -31,6 +31,7 @@ Pydantic does not yet offer a strict mode, but it is planned for pydantic v2. Se
 until then, this script is a best effort to stop us from introducing type coersion bugs
 (like the infamous stringy power levels fixed in room version 10).
 """
+
 import argparse
 import contextlib
 import functools
@@ -44,7 +45,6 @@ import traceback
 import unittest.mock
 from contextlib import contextmanager
 from typing import (
-    TYPE_CHECKING,
     Any,
     Callable,
     Dict,
@@ -56,30 +56,17 @@ from typing import (
 )
 
 from parameterized import parameterized
-
-from synapse._pydantic_compat import HAS_PYDANTIC_V2
-
-if TYPE_CHECKING or HAS_PYDANTIC_V2:
-    from pydantic.v1 import (
-        BaseModel as PydanticBaseModel,
-        conbytes,
-        confloat,
-        conint,
-        constr,
-    )
-    from pydantic.v1.typing import get_args
-else:
-    from pydantic import (
-        BaseModel as PydanticBaseModel,
-        conbytes,
-        confloat,
-        conint,
-        constr,
-    )
-    from pydantic.typing import get_args
-
 from typing_extensions import ParamSpec
 
+from synapse._pydantic_compat import (
+    BaseModel as PydanticBaseModel,
+    conbytes,
+    confloat,
+    conint,
+    constr,
+    get_args,
+)
+
 logger = logging.getLogger(__name__)
 
 CONSTRAINED_TYPE_FACTORIES_WITH_STRICT_FLAG: List[Callable] = [
@@ -182,22 +169,16 @@ def monkeypatch_pydantic() -> Generator[None, None, None]:
         # Most Synapse code ought to import the patched objects directly from
         # `pydantic`. But we also patch their containing modules `pydantic.main` and
         # `pydantic.types` for completeness.
-        patch_basemodel1 = unittest.mock.patch(
-            "pydantic.BaseModel", new=PatchedBaseModel
+        patch_basemodel = unittest.mock.patch(
+            "synapse._pydantic_compat.BaseModel", new=PatchedBaseModel
         )
-        patch_basemodel2 = unittest.mock.patch(
-            "pydantic.main.BaseModel", new=PatchedBaseModel
-        )
-        patches.enter_context(patch_basemodel1)
-        patches.enter_context(patch_basemodel2)
+        patches.enter_context(patch_basemodel)
         for factory in CONSTRAINED_TYPE_FACTORIES_WITH_STRICT_FLAG:
             wrapper: Callable = make_wrapper(factory)
-            patch1 = unittest.mock.patch(f"pydantic.{factory.__name__}", new=wrapper)
-            patch2 = unittest.mock.patch(
-                f"pydantic.types.{factory.__name__}", new=wrapper
+            patch = unittest.mock.patch(
+                f"synapse._pydantic_compat.{factory.__name__}", new=wrapper
             )
-            patches.enter_context(patch1)
-            patches.enter_context(patch2)
+            patches.enter_context(patch)
         yield
 
 

scripts-dev/complement.sh

@@ -220,9 +220,11 @@ test_packages=(
     ./tests/msc3874
     ./tests/msc3890
     ./tests/msc3391
+    ./tests/msc3757
     ./tests/msc3930
     ./tests/msc3902
     ./tests/msc3967
+    ./tests/msc4140
 )
 
 # Enable dirty runs, so tests will reuse the same container where possible.

scripts-dev/lint.sh

@@ -1,8 +1,9 @@
 #!/usr/bin/env bash
 #
 # Runs linting scripts over the local Synapse checkout
-# black - opinionated code formatter
 # ruff - lints and finds mistakes
+# mypy - typechecks python code
+# cargo clippy - lints rust code
 
 set -e
 
@@ -101,12 +102,6 @@ echo
 # Print out the commands being run
 set -x
 
-# Ensure the sort order of imports.
-isort "${files[@]}"
-
-# Ensure Python code conforms to an opinionated style.
-python3 -m black "${files[@]}"
-
 # Ensure the sample configuration file conforms to style checks.
 ./scripts-dev/config-lint.sh
 
@@ -114,6 +109,9 @@ python3 -m black "${files[@]}"
 # --quiet suppresses the update check.
 ruff check --quiet --fix "${files[@]}"
 
+# Reformat Python code.
+ruff format --quiet "${files[@]}"
+
 # Catch any common programming mistakes in Rust code.
 #
 # --bins, --examples, --lib, --tests combined explicitly disable checking

scripts-dev/release.py

@@ -20,8 +20,7 @@
 #
 #
 
-"""An interactive script for doing a release. See `cli()` below.
-"""
+"""An interactive script for doing a release. See `cli()` below."""
 
 import glob
 import json

stubs/txredisapi.pyi

@@ -13,8 +13,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-"""Contains *incomplete* type hints for txredisapi.
-"""
+"""Contains *incomplete* type hints for txredisapi."""
+
 from typing import Any, List, Optional, Type, Union
 
 from twisted.internet import protocol

synapse/__init__.py

@@ -20,8 +20,7 @@
 #
 #
 
-""" This is an implementation of a Matrix homeserver.
-"""
+"""This is an implementation of a Matrix homeserver."""
 
 import os
 import sys

synapse/_pydantic_compat.py

@@ -19,6 +19,8 @@
 #
 #
 
+from typing import TYPE_CHECKING
+
 from packaging.version import Version
 
 try:
@@ -30,4 +32,64 @@ except ImportError:
 
 HAS_PYDANTIC_V2: bool = Version(pydantic_version).major == 2
 
-__all__ = ("HAS_PYDANTIC_V2",)
+if TYPE_CHECKING or HAS_PYDANTIC_V2:
+    from pydantic.v1 import (
+        BaseModel,
+        Extra,
+        Field,
+        MissingError,
+        PydanticValueError,
+        StrictBool,
+        StrictInt,
+        StrictStr,
+        ValidationError,
+        conbytes,
+        confloat,
+        conint,
+        constr,
+        parse_obj_as,
+        validator,
+    )
+    from pydantic.v1.error_wrappers import ErrorWrapper
+    from pydantic.v1.typing import get_args
+else:
+    from pydantic import (
+        BaseModel,
+        Extra,
+        Field,
+        MissingError,
+        PydanticValueError,
+        StrictBool,
+        StrictInt,
+        StrictStr,
+        ValidationError,
+        conbytes,
+        confloat,
+        conint,
+        constr,
+        parse_obj_as,
+        validator,
+    )
+    from pydantic.error_wrappers import ErrorWrapper
+    from pydantic.typing import get_args
+
+__all__ = (
+    "HAS_PYDANTIC_V2",
+    "BaseModel",
+    "constr",
+    "conbytes",
+    "conint",
+    "confloat",
+    "ErrorWrapper",
+    "Extra",
+    "Field",
+    "get_args",
+    "MissingError",
+    "parse_obj_as",
+    "PydanticValueError",
+    "StrictBool",
+    "StrictInt",
+    "StrictStr",
+    "ValidationError",
+    "validator",
+)

synapse/_scripts/generate_workers_map.py

@@ -171,7 +171,7 @@ def elide_http_methods_if_unconflicting(
     """
 
     def paths_to_methods_dict(
-        methods_and_paths: Iterable[Tuple[str, str]]
+        methods_and_paths: Iterable[Tuple[str, str]],
     ) -> Dict[str, Set[str]]:
         """
         Given (method, path) pairs, produces a dict from path to set of methods
@@ -201,7 +201,7 @@ def elide_http_methods_if_unconflicting(
 
 
 def simplify_path_regexes(
-    registrations: Dict[Tuple[str, str], EndpointDescription]
+    registrations: Dict[Tuple[str, str], EndpointDescription],
 ) -> Dict[Tuple[str, str], EndpointDescription]:
     """
     Simplify all the path regexes for the dict of endpoint descriptions,

synapse/_scripts/hash_password.py

@@ -56,7 +56,9 @@ def main() -> None:
     password_pepper = password_config.get("pepper", password_pepper)
     password = args.password
 
-    if not password:
+    if not password and not sys.stdin.isatty():
+        password = sys.stdin.readline().strip()
+    elif not password:
         password = prompt_for_pass()
 
     # On Python 2, make sure we decode it to Unicode before we normalise it

synapse/_scripts/review_recent_signups.py

@@ -40,6 +40,7 @@ from synapse.storage.engines import create_engine
 
 class ReviewConfig(RootConfig):
     "A config class that just pulls out the database config"
+
     config_classes = [DatabaseConfig]
 
 
@@ -160,7 +161,11 @@ def main() -> None:
 
     with make_conn(database_config, engine, "review_recent_signups") as db_conn:
         # This generates a type of Cursor, not LoggingTransaction.
-        user_infos = get_recent_users(db_conn.cursor(), since_ms, exclude_users_with_appservice)  # type: ignore[arg-type]
+        user_infos = get_recent_users(
+            db_conn.cursor(),
+            since_ms,  # type: ignore[arg-type]
+            exclude_users_with_appservice,
+        )
 
     for user_info in user_infos:
         if exclude_users_with_email and user_info.emails:

synapse/_scripts/synapse_port_db.py

@@ -717,9 +717,7 @@ class Porter:
                 return
 
             # Check if all background updates are done, abort if not.
-            updates_complete = (
-                await self.sqlite_store.db_pool.updates.has_completed_background_updates()
-            )
+            updates_complete = await self.sqlite_store.db_pool.updates.has_completed_background_updates()
             if not updates_complete:
                 end_error = (
                     "Pending background updates exist in the SQLite3 database."
@@ -1095,10 +1093,10 @@ class Porter:
         return done, remaining + done
 
     async def _setup_state_group_id_seq(self) -> None:
-        curr_id: Optional[int] = (
-            await self.sqlite_store.db_pool.simple_select_one_onecol(
-                table="state_groups", keyvalues={}, retcol="MAX(id)", allow_none=True
-            )
+        curr_id: Optional[
+            int
+        ] = await self.sqlite_store.db_pool.simple_select_one_onecol(
+            table="state_groups", keyvalues={}, retcol="MAX(id)", allow_none=True
         )
 
         if not curr_id:
@@ -1186,13 +1184,13 @@ class Porter:
         )
 
     async def _setup_auth_chain_sequence(self) -> None:
-        curr_chain_id: Optional[int] = (
-            await self.sqlite_store.db_pool.simple_select_one_onecol(
-                table="event_auth_chains",
-                keyvalues={},
-                retcol="MAX(chain_id)",
-                allow_none=True,
-            )
+        curr_chain_id: Optional[
+            int
+        ] = await self.sqlite_store.db_pool.simple_select_one_onecol(
+            table="event_auth_chains",
+            keyvalues={},
+            retcol="MAX(chain_id)",
+            allow_none=True,
         )
 
         def r(txn: LoggingTransaction) -> None:

synapse/api/auth/msc3861_delegated.py

@@ -121,7 +121,9 @@ class MSC3861DelegatedAuth(BaseAuth):
         self._hostname = hs.hostname
         self._admin_token = self._config.admin_token
 
-        self._issuer_metadata = RetryOnExceptionCachedCall(self._load_metadata)
+        self._issuer_metadata = RetryOnExceptionCachedCall[OpenIDProviderMetadata](
+            self._load_metadata
+        )
 
         if isinstance(auth_method, PrivateKeyJWTWithKid):
             # Use the JWK as the client secret when using the private_key_jwt method
@@ -145,6 +147,33 @@ class MSC3861DelegatedAuth(BaseAuth):
         # metadata.validate_introspection_endpoint()
         return metadata
 
+    async def issuer(self) -> str:
+        """
+        Get the configured issuer
+
+        This will use the issuer value set in the metadata,
+        falling back to the one set in the config if not set in the metadata
+        """
+        metadata = await self._issuer_metadata.get()
+        return metadata.issuer or self._config.issuer
+
+    async def account_management_url(self) -> Optional[str]:
+        """
+        Get the configured account management URL
+
+        This will discover the account management URL from the issuer if it's not set in the config
+        """
+        if self._config.account_management_url is not None:
+            return self._config.account_management_url
+
+        try:
+            metadata = await self._issuer_metadata.get()
+            return metadata.get("account_management_uri", None)
+        # We don't want to raise here if we can't load the metadata
+        except Exception:
+            logger.warning("Failed to load metadata:", exc_info=True)
+            return None
+
     async def _introspection_endpoint(self) -> str:
         """
         Returns the introspection endpoint of the issuer
@@ -154,7 +183,7 @@ class MSC3861DelegatedAuth(BaseAuth):
         if self._config.introspection_endpoint is not None:
             return self._config.introspection_endpoint
 
-        metadata = await self._load_metadata()
+        metadata = await self._issuer_metadata.get()
         return metadata.get("introspection_endpoint")
 
     async def _introspect_token(self, token: str) -> IntrospectionToken:
@@ -309,7 +338,7 @@ class MSC3861DelegatedAuth(BaseAuth):
             logger.exception("Failed to introspect token")
             raise SynapseError(503, "Unable to introspect the access token")
 
-        logger.info(f"Introspection result: {introspection_result!r}")
+        logger.debug("Introspection result: %r", introspection_result)
 
         # TODO: introspection verification should be more extensive, especially:
         #   - verify the audience

synapse/api/constants.py

@@ -230,6 +230,8 @@ class EventContentFields:
 
     ROOM_NAME: Final = "name"
 
+    MEMBERSHIP: Final = "membership"
+
     # Used in m.room.guest_access events.
     GUEST_ACCESS: Final = "guest_access"
 

synapse/api/room_versions.py

@@ -107,6 +107,8 @@ class RoomVersion:
     # support the flag. Unknown flags are ignored by the evaluator, making conditions
     # fail if used.
     msc3931_push_features: Tuple[str, ...]  # values from PushRuleRoomFlag
+    # MSC3757: Restricting who can overwrite a state event
+    msc3757_enabled: bool
 
 
 class RoomVersions:
@@ -128,6 +130,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V2 = RoomVersion(
         "2",
@@ -147,6 +150,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V3 = RoomVersion(
         "3",
@@ -166,6 +170,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V4 = RoomVersion(
         "4",
@@ -185,6 +190,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V5 = RoomVersion(
         "5",
@@ -204,6 +210,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V6 = RoomVersion(
         "6",
@@ -223,6 +230,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V7 = RoomVersion(
         "7",
@@ -242,6 +250,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V8 = RoomVersion(
         "8",
@@ -261,6 +270,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V9 = RoomVersion(
         "9",
@@ -280,6 +290,7 @@ class RoomVersions:
         knock_restricted_join_rule=False,
         enforce_int_power_levels=False,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     V10 = RoomVersion(
         "10",
@@ -299,6 +310,7 @@ class RoomVersions:
         knock_restricted_join_rule=True,
         enforce_int_power_levels=True,
         msc3931_push_features=(),
+        msc3757_enabled=False,
     )
     MSC1767v10 = RoomVersion(
         # MSC1767 (Extensible Events) based on room version "10"
@@ -319,6 +331,28 @@ class RoomVersions:
         knock_restricted_join_rule=True,
         enforce_int_power_levels=True,
         msc3931_push_features=(PushRuleRoomFlag.EXTENSIBLE_EVENTS,),
+        msc3757_enabled=False,
+    )
+    MSC3757v10 = RoomVersion(
+        # MSC3757 (Restricting who can overwrite a state event) based on room version "10"
+        "org.matrix.msc3757.10",
+        RoomDisposition.UNSTABLE,
+        EventFormatVersions.ROOM_V4_PLUS,
+        StateResolutionVersions.V2,
+        enforce_key_validity=True,
+        special_case_aliases_auth=False,
+        strict_canonicaljson=True,
+        limit_notifications_power_levels=True,
+        implicit_room_creator=False,
+        updated_redaction_rules=False,
+        restricted_join_rule=True,
+        restricted_join_rule_fix=True,
+        knock_join_rule=True,
+        msc3389_relation_redactions=False,
+        knock_restricted_join_rule=True,
+        enforce_int_power_levels=True,
+        msc3931_push_features=(),
+        msc3757_enabled=True,
     )
     V11 = RoomVersion(
         "11",
@@ -338,6 +372,28 @@ class RoomVersions:
         knock_restricted_join_rule=True,
         enforce_int_power_levels=True,
         msc3931_push_features=(),
+        msc3757_enabled=False,
+    )
+    MSC3757v11 = RoomVersion(
+        # MSC3757 (Restricting who can overwrite a state event) based on room version "11"
+        "org.matrix.msc3757.11",
+        RoomDisposition.UNSTABLE,
+        EventFormatVersions.ROOM_V4_PLUS,
+        StateResolutionVersions.V2,
+        enforce_key_validity=True,
+        special_case_aliases_auth=False,
+        strict_canonicaljson=True,
+        limit_notifications_power_levels=True,
+        implicit_room_creator=True,  # Used by MSC3820
+        updated_redaction_rules=True,  # Used by MSC3820
+        restricted_join_rule=True,
+        restricted_join_rule_fix=True,
+        knock_join_rule=True,
+        msc3389_relation_redactions=False,
+        knock_restricted_join_rule=True,
+        enforce_int_power_levels=True,
+        msc3931_push_features=(),
+        msc3757_enabled=True,
     )
 
 
@@ -355,6 +411,8 @@ KNOWN_ROOM_VERSIONS: Dict[str, RoomVersion] = {
         RoomVersions.V9,
         RoomVersions.V10,
         RoomVersions.V11,
+        RoomVersions.MSC3757v10,
+        RoomVersions.MSC3757v11,
     )
 }
 

synapse/api/urls.py

@@ -19,7 +19,8 @@
 #
 #
 
-"""Contains the URL paths to prefix various aspects of the server with. """
+"""Contains the URL paths to prefix various aspects of the server with."""
+
 import hmac
 from hashlib import sha256
 from urllib.parse import urlencode

synapse/app/generic_worker.py

@@ -65,6 +65,7 @@ from synapse.storage.databases.main.appservice import (
 )
 from synapse.storage.databases.main.censor_events import CensorEventsStore
 from synapse.storage.databases.main.client_ips import ClientIpWorkerStore
+from synapse.storage.databases.main.delayed_events import DelayedEventsStore
 from synapse.storage.databases.main.deviceinbox import DeviceInboxWorkerStore
 from synapse.storage.databases.main.devices import DeviceWorkerStore
 from synapse.storage.databases.main.directory import DirectoryWorkerStore
@@ -161,6 +162,7 @@ class GenericWorkerStore(
     TaskSchedulerWorkerStore,
     ExperimentalFeaturesStore,
     SlidingSyncStore,
+    DelayedEventsStore,
 ):
     # Properties that multiple storage classes define. Tell mypy what the
     # expected type is.

synapse/appservice/scheduler.py

@@ -54,6 +54,7 @@ UP & quit           +---------- YES                       SUCCESS
 This is all tied together by the AppServiceScheduler which DIs the required
 components.
 """
+
 import logging
 from typing import (
     TYPE_CHECKING,

synapse/config/_util.py

@@ -18,17 +18,11 @@
 # [This file includes modifications made by New Vector Limited]
 #
 #
-from typing import TYPE_CHECKING, Any, Dict, Type, TypeVar
+from typing import Any, Dict, Type, TypeVar
 
 import jsonschema
 
-from synapse._pydantic_compat import HAS_PYDANTIC_V2
-
-if TYPE_CHECKING or HAS_PYDANTIC_V2:
-    from pydantic.v1 import BaseModel, ValidationError, parse_obj_as
-else:
-    from pydantic import BaseModel, ValidationError, parse_obj_as
-
+from synapse._pydantic_compat import BaseModel, ValidationError, parse_obj_as
 from synapse.config._base import ConfigError
 from synapse.types import JsonDict, StrSequence
 

synapse/config/experimental.py

@@ -338,8 +338,10 @@ class ExperimentalConfig(Config):
         # MSC3391: Removing account data.
         self.msc3391_enabled = experimental.get("msc3391_enabled", False)
 
-        # MSC3575 (Sliding Sync API endpoints)
-        self.msc3575_enabled: bool = experimental.get("msc3575_enabled", False)
+        # MSC3575 (Sliding Sync) alternate endpoints, c.f. MSC4186.
+        #
+        # This is enabled by default as a replacement for the sliding sync proxy.
+        self.msc3575_enabled: bool = experimental.get("msc3575_enabled", True)
 
         # MSC3773: Thread notifications
         self.msc3773_enabled: bool = experimental.get("msc3773_enabled", False)
@@ -445,6 +447,3 @@ class ExperimentalConfig(Config):
 
         # MSC4151: Report room API (Client-Server API)
         self.msc4151_enabled: bool = experimental.get("msc4151_enabled", False)
-
-        # MSC4156: Migrate server_name to via
-        self.msc4156_enabled: bool = experimental.get("msc4156_enabled", False)

synapse/config/key.py

@@ -200,16 +200,13 @@ class KeyConfig(Config):
             )
             form_secret = 'form_secret: "%s"' % random_string_with_symbols(50)
 
-        return (
-            """\
+        return """\
         %(macaroon_secret_key)s
         %(form_secret)s
         signing_key_path: "%(base_key_name)s.signing.key"
         trusted_key_servers:
           - server_name: "matrix.org"
-        """
-            % locals()
-        )
+        """ % locals()
 
     def read_signing_keys(self, signing_key_path: str, name: str) -> List[SigningKey]:
         """Read the signing keys in the given path.
@@ -249,7 +246,9 @@ class KeyConfig(Config):
             if is_signing_algorithm_supported(key_id):
                 key_base64 = key_data["key"]
                 key_bytes = decode_base64(key_base64)
-                verify_key: "VerifyKeyWithExpiry" = decode_verify_key_bytes(key_id, key_bytes)  # type: ignore[assignment]
+                verify_key: "VerifyKeyWithExpiry" = decode_verify_key_bytes(
+                    key_id, key_bytes
+                )  # type: ignore[assignment]
                 verify_key.expired = key_data["expired_ts"]
                 keys[key_id] = verify_key
             else:

synapse/config/logger.py

@@ -157,12 +157,9 @@ class LoggingConfig(Config):
         self, config_dir_path: str, server_name: str, **kwargs: Any
     ) -> str:
         log_config = os.path.join(config_dir_path, server_name + ".log.config")
-        return (
-            """\
+        return """\
         log_config: "%(log_config)s"
-        """
-            % locals()
-        )
+        """ % locals()
 
     def read_arguments(self, args: argparse.Namespace) -> None:
         if args.no_redirect_stdio is not None:

synapse/config/server.py

@@ -2,7 +2,7 @@
 # This file is licensed under the Affero General Public License (AGPL) version 3.
 #
 # Copyright 2014-2021 The Matrix.org Foundation C.I.C.
-# Copyright (C) 2023 New Vector, Ltd
+# Copyright (C) 2023-2024 New Vector, Ltd
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License as
@@ -780,6 +780,17 @@ class ServerConfig(Config):
         else:
             self.delete_stale_devices_after = None
 
+        # The maximum allowed delay duration for delayed events (MSC4140).
+        max_event_delay_duration = config.get("max_event_delay_duration")
+        if max_event_delay_duration is not None:
+            self.max_event_delay_ms: Optional[int] = self.parse_duration(
+                max_event_delay_duration
+            )
+            if self.max_event_delay_ms <= 0:
+                raise ConfigError("max_event_delay_duration must be a positive value")
+        else:
+            self.max_event_delay_ms = None
+
     def has_tls_listener(self) -> bool:
         return any(listener.is_tls() for listener in self.listeners)
 
@@ -828,13 +839,10 @@ class ServerConfig(Config):
             ).lstrip()
 
         if not unsecure_listeners:
-            unsecure_http_bindings = (
-                """- port: %(unsecure_port)s
+            unsecure_http_bindings = """- port: %(unsecure_port)s
             tls: false
             type: http
-            x_forwarded: true"""
-                % locals()
-            )
+            x_forwarded: true""" % locals()
 
             if not open_private_ports:
                 unsecure_http_bindings += (
@@ -853,16 +861,13 @@ class ServerConfig(Config):
         if not secure_listeners:
             secure_http_bindings = ""
 
-        return (
-            """\
+        return """\
         server_name: "%(server_name)s"
         pid_file: %(pid_file)s
         listeners:
           %(secure_http_bindings)s
           %(unsecure_http_bindings)s
-        """
-            % locals()
-        )
+        """ % locals()
 
     def read_arguments(self, args: argparse.Namespace) -> None:
         if args.manhole is not None:

synapse/config/voip.py

@@ -23,7 +23,12 @@ from typing import Any
 
 from synapse.types import JsonDict
 
-from ._base import Config
+from ._base import Config, ConfigError, read_file
+
+CONFLICTING_SHARED_SECRET_OPTS_ERROR = """\
+You have configured both `turn_shared_secret` and `turn_shared_secret_path`.
+These are mutually incompatible.
+"""
 
 
 class VoipConfig(Config):
@@ -32,6 +37,13 @@ class VoipConfig(Config):
     def read_config(self, config: JsonDict, **kwargs: Any) -> None:
         self.turn_uris = config.get("turn_uris", [])
         self.turn_shared_secret = config.get("turn_shared_secret")
+        turn_shared_secret_path = config.get("turn_shared_secret_path")
+        if turn_shared_secret_path:
+            if self.turn_shared_secret:
+                raise ConfigError(CONFLICTING_SHARED_SECRET_OPTS_ERROR)
+            self.turn_shared_secret = read_file(
+                turn_shared_secret_path, ("turn_shared_secret_path",)
+            ).strip()
         self.turn_username = config.get("turn_username")
         self.turn_password = config.get("turn_password")
         self.turn_user_lifetime = self.parse_duration(

synapse/config/workers.py

@@ -22,17 +22,17 @@
 
 import argparse
 import logging
-from typing import TYPE_CHECKING, Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Union
 
 import attr
 
-from synapse._pydantic_compat import HAS_PYDANTIC_V2
-
-if TYPE_CHECKING or HAS_PYDANTIC_V2:
-    from pydantic.v1 import BaseModel, Extra, StrictBool, StrictInt, StrictStr
-else:
-    from pydantic import BaseModel, Extra, StrictBool, StrictInt, StrictStr
-
+from synapse._pydantic_compat import (
+    BaseModel,
+    Extra,
+    StrictBool,
+    StrictInt,
+    StrictStr,
+)
 from synapse.config._base import (
     Config,
     ConfigError,
@@ -328,10 +328,11 @@ class WorkerConfig(Config):
                 )
 
         # type-ignore: the expression `Union[A, B]` is not a Type[Union[A, B]] currently
-        self.instance_map: Dict[
-            str, InstanceLocationConfig
-        ] = parse_and_validate_mapping(
-            instance_map, InstanceLocationConfig  # type: ignore[arg-type]
+        self.instance_map: Dict[str, InstanceLocationConfig] = (
+            parse_and_validate_mapping(
+                instance_map,
+                InstanceLocationConfig,  # type: ignore[arg-type]
+            )
         )
 
         # Map from type of streams to source, c.f. WriterLocations.

synapse/event_auth.py

@@ -388,6 +388,7 @@ LENIENT_EVENT_BYTE_LIMITS_ROOM_VERSIONS = {
     RoomVersions.V9,
     RoomVersions.V10,
     RoomVersions.MSC1767v10,
+    RoomVersions.MSC3757v10,
 }
 
 
@@ -790,9 +791,10 @@ def get_send_level(
 
 
 def _can_send_event(event: "EventBase", auth_events: StateMap["EventBase"]) -> bool:
+    state_key = event.get_state_key()
     power_levels_event = get_power_level_event(auth_events)
 
-    send_level = get_send_level(event.type, event.get("state_key"), power_levels_event)
+    send_level = get_send_level(event.type, state_key, power_levels_event)
     user_level = get_user_power_level(event.user_id, auth_events)
 
     if user_level < send_level:
@@ -803,11 +805,34 @@ def _can_send_event(event: "EventBase", auth_events: StateMap["EventBase"]) -> b
             errcode=Codes.INSUFFICIENT_POWER,
         )
 
-    # Check state_key
-    if hasattr(event, "state_key"):
-        if event.state_key.startswith("@"):
-            if event.state_key != event.user_id:
-                raise AuthError(403, "You are not allowed to set others state")
+    if (
+        state_key is not None
+        and state_key.startswith("@")
+        and state_key != event.user_id
+    ):
+        if event.room_version.msc3757_enabled:
+            try:
+                colon_idx = state_key.index(":", 1)
+                suffix_idx = state_key.find("_", colon_idx + 1)
+                state_key_user_id = (
+                    state_key[:suffix_idx] if suffix_idx != -1 else state_key
+                )
+                if not UserID.is_valid(state_key_user_id):
+                    raise ValueError
+            except ValueError:
+                raise SynapseError(
+                    400,
+                    "State key neither equals a valid user ID, nor starts with one plus an underscore",
+                    errcode=Codes.BAD_JSON,
+                )
+            if (
+                # sender is owner of the state key
+                state_key_user_id == event.user_id
+                # sender has higher PL than the owner of the state key
+                or user_level > get_user_power_level(state_key_user_id, auth_events)
+            ):
+                return True
+        raise AuthError(403, "You are not allowed to set others state")
 
     return True
 
@@ -887,7 +912,8 @@ def _check_power_levels(
                     raise SynapseError(400, f"{v!r} must be an integer.")
             if k in {"events", "notifications", "users"}:
                 if not isinstance(v, collections.abc.Mapping) or not all(
-                    type(v) is int for v in v.values()  # noqa: E721
+                    type(v) is int
+                    for v in v.values()  # noqa: E721
                 ):
                     raise SynapseError(
                         400,

synapse/events/presence_router.py

@@ -80,7 +80,7 @@ def load_legacy_presence_router(hs: "HomeServer") -> None:
     # All methods that the module provides should be async, but this wasn't enforced
     # in the old module system, so we wrap them if needed
     def async_wrapper(
-        f: Optional[Callable[P, R]]
+        f: Optional[Callable[P, R]],
     ) -> Optional[Callable[P, Awaitable[R]]]:
         # f might be None if the callback isn't implemented by the module. In this
         # case we don't want to register a callback at all so we return None.

synapse/events/snapshot.py

@@ -504,7 +504,7 @@ class UnpersistedEventContext(UnpersistedEventContextBase):
 
 
 def _encode_state_group_delta(
-    state_group_delta: Dict[Tuple[int, int], StateMap[str]]
+    state_group_delta: Dict[Tuple[int, int], StateMap[str]],
 ) -> List[Tuple[int, int, Optional[List[Tuple[str, str, str]]]]]:
     if not state_group_delta:
         return []
@@ -517,7 +517,7 @@ def _encode_state_group_delta(
 
 
 def _decode_state_group_delta(
-    input: List[Tuple[int, int, List[Tuple[str, str, str]]]]
+    input: List[Tuple[int, int, List[Tuple[str, str, str]]]],
 ) -> Dict[Tuple[int, int], StateMap[str]]:
     if not input:
         return {}
@@ -544,7 +544,7 @@ def _encode_state_dict(
 
 
 def _decode_state_dict(
-    input: Optional[List[Tuple[str, str, str]]]
+    input: Optional[List[Tuple[str, str, str]]],
 ) -> Optional[StateMap[str]]:
     """Decodes a state dict encoded using `_encode_state_dict` above"""
     if input is None:

synapse/events/validator.py

@@ -19,17 +19,11 @@
 #
 #
 import collections.abc
-from typing import TYPE_CHECKING, List, Type, Union, cast
+from typing import List, Type, Union, cast
 
 import jsonschema
 
-from synapse._pydantic_compat import HAS_PYDANTIC_V2
-
-if TYPE_CHECKING or HAS_PYDANTIC_V2:
-    from pydantic.v1 import Field, StrictBool, StrictStr
-else:
-    from pydantic import Field, StrictBool, StrictStr
-
+from synapse._pydantic_compat import Field, StrictBool, StrictStr
 from synapse.api.constants import (
     MAX_ALIAS_LENGTH,
     EventContentFields,

synapse/federation/__init__.py

@@ -19,5 +19,4 @@
 #
 #
 
-""" This package includes all the federation specific logic.
-"""
+"""This package includes all the federation specific logic."""

synapse/federation/persistence.py

@@ -20,7 +20,7 @@
 #
 #
 
-""" This module contains all the persistence actions done by the federation
+"""This module contains all the persistence actions done by the federation
 package.
 
 These actions are mostly only used by the :py:mod:`.replication` module.

synapse/federation/transport/server/federation.py

@@ -859,7 +859,6 @@ class FederationMediaThumbnailServlet(BaseFederationServerServlet):
         request: SynapseRequest,
         media_id: str,
     ) -> None:
-
         width = parse_integer(request, "width", required=True)
         height = parse_integer(request, "height", required=True)
         method = parse_string(request, "method", "scale")

synapse/federation/units.py

@@ -19,7 +19,7 @@
 #
 #
 
-""" Defines the JSON structure of the protocol units used by the server to
+"""Defines the JSON structure of the protocol units used by the server to
 server protocol.
 """
 

synapse/handlers/account.py

@@ -118,10 +118,10 @@ class AccountHandler:
             }
 
             if self._use_account_validity_in_account_status:
-                status["org.matrix.expired"] = (
-                    await self._account_validity_handler.is_user_expired(
-                        user_id.to_string()
-                    )
+                status[
+                    "org.matrix.expired"
+                ] = await self._account_validity_handler.is_user_expired(
+                    user_id.to_string()
                 )
 
         return status

synapse/handlers/account_data.py

@@ -33,7 +33,7 @@ from synapse.replication.http.account_data import (
     ReplicationRemoveUserAccountDataRestServlet,
 )
 from synapse.streams import EventSource
-from synapse.types import JsonDict, StrCollection, StreamKeyType, UserID
+from synapse.types import JsonDict, JsonMapping, StrCollection, StreamKeyType, UserID
 
 if TYPE_CHECKING:
     from synapse.server import HomeServer
@@ -253,7 +253,7 @@ class AccountDataHandler:
             return response["max_stream_id"]
 
     async def add_tag_to_room(
-        self, user_id: str, room_id: str, tag: str, content: JsonDict
+        self, user_id: str, room_id: str, tag: str, content: JsonMapping
     ) -> int:
         """Add a tag to a room for a user.
 

synapse/handlers/admin.py

@@ -21,13 +21,34 @@
 
 import abc
 import logging
-from typing import TYPE_CHECKING, Any, Dict, List, Mapping, Optional, Sequence, Set
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    List,
+    Mapping,
+    Optional,
+    Sequence,
+    Set,
+    Tuple,
+)
 
 import attr
 
-from synapse.api.constants import Direction, Membership
+from synapse.api.constants import Direction, EventTypes, Membership
+from synapse.api.errors import SynapseError
 from synapse.events import EventBase
-from synapse.types import JsonMapping, RoomStreamToken, StateMap, UserID, UserInfo
+from synapse.types import (
+    JsonMapping,
+    Requester,
+    RoomStreamToken,
+    ScheduledTask,
+    StateMap,
+    TaskStatus,
+    UserID,
+    UserInfo,
+    create_requester,
+)
 from synapse.visibility import filter_events_for_client
 
 if TYPE_CHECKING:
@@ -35,6 +56,8 @@ if TYPE_CHECKING:
 
 logger = logging.getLogger(__name__)
 
+REDACT_ALL_EVENTS_ACTION_NAME = "redact_all_events"
+
 
 class AdminHandler:
     def __init__(self, hs: "HomeServer"):
@@ -43,6 +66,20 @@ class AdminHandler:
         self._storage_controllers = hs.get_storage_controllers()
         self._state_storage_controller = self._storage_controllers.state
         self._msc3866_enabled = hs.config.experimental.msc3866.enabled
+        self.event_creation_handler = hs.get_event_creation_handler()
+        self._task_scheduler = hs.get_task_scheduler()
+
+        self._task_scheduler.register_action(
+            self._redact_all_events, REDACT_ALL_EVENTS_ACTION_NAME
+        )
+
+    async def get_redact_task(self, redact_id: str) -> Optional[ScheduledTask]:
+        """Get the current status of an active redaction process
+
+        Args:
+            redact_id: redact_id returned by start_redact_events.
+        """
+        return await self._task_scheduler.get_task(redact_id)
 
     async def get_whois(self, user: UserID) -> JsonMapping:
         connections = []
@@ -197,14 +234,16 @@ class AdminHandler:
             # events that we have and then filtering, this isn't the most
             # efficient method perhaps but it does guarantee we get everything.
             while True:
-                events, _ = (
-                    await self._store.paginate_room_events_by_topological_ordering(
-                        room_id=room_id,
-                        from_key=from_key,
-                        to_key=to_key,
-                        limit=100,
-                        direction=Direction.FORWARDS,
-                    )
+                (
+                    events,
+                    _,
+                    _,
+                ) = await self._store.paginate_room_events_by_topological_ordering(
+                    room_id=room_id,
+                    from_key=from_key,
+                    to_key=to_key,
+                    limit=100,
+                    direction=Direction.FORWARDS,
                 )
                 if not events:
                     break
@@ -311,6 +350,153 @@ class AdminHandler:
 
         return writer.finished()
 
+    async def start_redact_events(
+        self,
+        user_id: str,
+        rooms: list,
+        requester: JsonMapping,
+        reason: Optional[str],
+        limit: Optional[int],
+    ) -> str:
+        """
+        Start a task redacting the events of the given user in the given rooms
+
+        Args:
+            user_id: the user ID of the user whose events should be redacted
+            rooms: the rooms in which to redact the user's events
+            requester: the user requesting the events
+            reason: reason for requesting the redaction, ie spam, etc
+            limit: limit on the number of events in each room to redact
+
+        Returns:
+            a unique ID which can be used to query the status of the task
+        """
+        active_tasks = await self._task_scheduler.get_tasks(
+            actions=[REDACT_ALL_EVENTS_ACTION_NAME],
+            resource_id=user_id,
+            statuses=[TaskStatus.ACTIVE],
+        )
+
+        if len(active_tasks) > 0:
+            raise SynapseError(
+                400, "Redact already in progress for user %s" % (user_id,)
+            )
+
+        if not limit:
+            limit = 1000
+
+        redact_id = await self._task_scheduler.schedule_task(
+            REDACT_ALL_EVENTS_ACTION_NAME,
+            resource_id=user_id,
+            params={
+                "rooms": rooms,
+                "requester": requester,
+                "user_id": user_id,
+                "reason": reason,
+                "limit": limit,
+            },
+        )
+
+        logger.info(
+            "starting redact events with redact_id %s",
+            redact_id,
+        )
+
+        return redact_id
+
+    async def _redact_all_events(
+        self, task: ScheduledTask
+    ) -> Tuple[TaskStatus, Optional[Mapping[str, Any]], Optional[str]]:
+        """
+        Task to redact all of a users events in the given rooms, tracking which, if any, events
+        whose redaction failed
+        """
+
+        assert task.params is not None
+        rooms = task.params.get("rooms")
+        assert rooms is not None
+
+        r = task.params.get("requester")
+        assert r is not None
+        admin = Requester.deserialize(self._store, r)
+
+        user_id = task.params.get("user_id")
+        assert user_id is not None
+
+        requester = create_requester(
+            user_id, authenticated_entity=admin.user.to_string()
+        )
+
+        reason = task.params.get("reason")
+        limit = task.params.get("limit")
+        assert limit is not None
+
+        result: Mapping[str, Any] = (
+            task.result if task.result else {"failed_redactions": {}}
+        )
+        for room in rooms:
+            room_version = await self._store.get_room_version(room)
+            event_ids = await self._store.get_events_sent_by_user_in_room(
+                user_id,
+                room,
+                limit,
+                ["m.room.member", "m.room.message"],
+            )
+            if not event_ids:
+                # there's nothing to redact
+                return TaskStatus.COMPLETE, result, None
+
+            events = await self._store.get_events_as_list(event_ids)
+            for event in events:
+                # we care about join events but not other membership events
+                if event.type == "m.room.member":
+                    content = event.content
+                    if content:
+                        if content.get("membership") == Membership.JOIN:
+                            pass
+                        else:
+                            continue
+                relations = await self._store.get_relations_for_event(
+                    room, event.event_id, event, event_type=EventTypes.Redaction
+                )
+
+                # if we've already successfully redacted this event then skip processing it
+                if relations[0]:
+                    continue
+
+                event_dict = {
+                    "type": EventTypes.Redaction,
+                    "content": {"reason": reason} if reason else {},
+                    "room_id": room,
+                    "sender": user_id,
+                }
+                if room_version.updated_redaction_rules:
+                    event_dict["content"]["redacts"] = event.event_id
+                else:
+                    event_dict["redacts"] = event.event_id
+
+                try:
+                    # set the prev event to the offending message to allow for redactions
+                    # to be processed in the case where the user has been kicked/banned before
+                    # redactions are requested
+                    (
+                        redaction,
+                        _,
+                    ) = await self.event_creation_handler.create_and_send_nonmember_event(
+                        requester,
+                        event_dict,
+                        prev_event_ids=[event.event_id],
+                        ratelimit=False,
+                    )
+                except Exception as ex:
+                    logger.info(
+                        f"Redaction of event {event.event_id} failed due to: {ex}"
+                    )
+                    result["failed_redactions"][event.event_id] = str(ex)
+                    await self._task_scheduler.update_task(task.id, result=result)
+
+        return TaskStatus.COMPLETE, result, None
+
 
 class ExfiltrationWriter(metaclass=abc.ABCMeta):
     """Interface used to specify how to write exported data."""

synapse/handlers/auth.py

@@ -166,8 +166,7 @@ def login_id_phone_to_thirdparty(identifier: JsonDict) -> Dict[str, str]:
     if "country" not in identifier or (
         # The specification requires a "phone" field, while Synapse used to require a "number"
         # field. Accept both for backwards compatibility.
-        "phone" not in identifier
-        and "number" not in identifier
+        "phone" not in identifier and "number" not in identifier
     ):
         raise SynapseError(
             400, "Invalid phone-type identifier", errcode=Codes.INVALID_PARAM

synapse/handlers/delayed_events.py

@@ -0,0 +1,484 @@
+import logging
+from typing import TYPE_CHECKING, List, Optional, Set, Tuple
+
+from twisted.internet.interfaces import IDelayedCall
+
+from synapse.api.constants import EventTypes
+from synapse.api.errors import ShadowBanError
+from synapse.config.workers import MAIN_PROCESS_INSTANCE_NAME
+from synapse.logging.opentracing import set_tag
+from synapse.metrics import event_processing_positions
+from synapse.metrics.background_process_metrics import run_as_background_process
+from synapse.replication.http.delayed_events import (
+    ReplicationAddedDelayedEventRestServlet,
+)
+from synapse.storage.databases.main.delayed_events import (
+    DelayedEventDetails,
+    DelayID,
+    EventType,
+    StateKey,
+    Timestamp,
+    UserLocalpart,
+)
+from synapse.storage.databases.main.state_deltas import StateDelta
+from synapse.types import (
+    JsonDict,
+    Requester,
+    RoomID,
+    UserID,
+    create_requester,
+)
+from synapse.util.events import generate_fake_event_id
+from synapse.util.metrics import Measure
+
+if TYPE_CHECKING:
+    from synapse.server import HomeServer
+
+logger = logging.getLogger(__name__)
+
+
+class DelayedEventsHandler:
+    def __init__(self, hs: "HomeServer"):
+        self._store = hs.get_datastores().main
+        self._storage_controllers = hs.get_storage_controllers()
+        self._config = hs.config
+        self._clock = hs.get_clock()
+        self._request_ratelimiter = hs.get_request_ratelimiter()
+        self._event_creation_handler = hs.get_event_creation_handler()
+        self._room_member_handler = hs.get_room_member_handler()
+
+        self._next_delayed_event_call: Optional[IDelayedCall] = None
+
+        # The current position in the current_state_delta stream
+        self._event_pos: Optional[int] = None
+
+        # Guard to ensure we only process event deltas one at a time
+        self._event_processing = False
+
+        if hs.config.worker.worker_app is None:
+            self._repl_client = None
+
+            async def _schedule_db_events() -> None:
+                # We kick this off to pick up outstanding work from before the last restart.
+                # Block until we're up to date.
+                await self._unsafe_process_new_event()
+                hs.get_notifier().add_replication_callback(self.notify_new_event)
+                # Kick off again (without blocking) to catch any missed notifications
+                # that may have fired before the callback was added.
+                self._clock.call_later(0, self.notify_new_event)
+
+                # Delayed events that are already marked as processed on startup might not have been
+                # sent properly on the last run of the server, so unmark them to send them again.
+                # Caveat: this will double-send delayed events that successfully persisted, but failed
+                # to be removed from the DB table of delayed events.
+                # TODO: To avoid double-sending, scan the timeline to find which of these events were
+                # already sent. To do so, must store delay_ids in sent events to retrieve them later.
+                await self._store.unprocess_delayed_events()
+
+                events, next_send_ts = await self._store.process_timeout_delayed_events(
+                    self._get_current_ts()
+                )
+
+                if next_send_ts:
+                    self._schedule_next_at(next_send_ts)
+
+                # Can send the events in background after having awaited on marking them as processed
+                run_as_background_process(
+                    "_send_events",
+                    self._send_events,
+                    events,
+                )
+
+            self._initialized_from_db = run_as_background_process(
+                "_schedule_db_events", _schedule_db_events
+            )
+        else:
+            self._repl_client = ReplicationAddedDelayedEventRestServlet.make_client(hs)
+
+    @property
+    def _is_master(self) -> bool:
+        return self._repl_client is None
+
+    def notify_new_event(self) -> None:
+        """
+        Called when there may be more state event deltas to process,
+        which should cancel pending delayed events for the same state.
+        """
+        if self._event_processing:
+            return
+
+        self._event_processing = True
+
+        async def process() -> None:
+            try:
+                await self._unsafe_process_new_event()
+            finally:
+                self._event_processing = False
+
+        run_as_background_process("delayed_events.notify_new_event", process)
+
+    async def _unsafe_process_new_event(self) -> None:
+        # If self._event_pos is None then means we haven't fetched it from the DB yet
+        if self._event_pos is None:
+            self._event_pos = await self._store.get_delayed_events_stream_pos()
+            room_max_stream_ordering = self._store.get_room_max_stream_ordering()
+            if self._event_pos > room_max_stream_ordering:
+                # apparently, we've processed more events than exist in the database!
+                # this can happen if events are removed with history purge or similar.
+                logger.warning(
+                    "Event stream ordering appears to have gone backwards (%i -> %i): "
+                    "rewinding delayed events processor",
+                    self._event_pos,
+                    room_max_stream_ordering,
+                )
+                self._event_pos = room_max_stream_ordering
+
+        # Loop round handling deltas until we're up to date
+        while True:
+            with Measure(self._clock, "delayed_events_delta"):
+                room_max_stream_ordering = self._store.get_room_max_stream_ordering()
+                if self._event_pos == room_max_stream_ordering:
+                    return
+
+                logger.debug(
+                    "Processing delayed events %s->%s",
+                    self._event_pos,
+                    room_max_stream_ordering,
+                )
+                (
+                    max_pos,
+                    deltas,
+                ) = await self._storage_controllers.state.get_current_state_deltas(
+                    self._event_pos, room_max_stream_ordering
+                )
+
+                logger.debug(
+                    "Handling %d state deltas for delayed events processing",
+                    len(deltas),
+                )
+                await self._handle_state_deltas(deltas)
+
+                self._event_pos = max_pos
+
+                # Expose current event processing position to prometheus
+                event_processing_positions.labels("delayed_events").set(max_pos)
+
+                await self._store.update_delayed_events_stream_pos(max_pos)
+
+    async def _handle_state_deltas(self, deltas: List[StateDelta]) -> None:
+        """
+        Process current state deltas to cancel pending delayed events
+        that target the same state.
+        """
+        for delta in deltas:
+            logger.debug(
+                "Handling: %r %r, %s", delta.event_type, delta.state_key, delta.event_id
+            )
+
+            next_send_ts = await self._store.cancel_delayed_state_events(
+                room_id=delta.room_id,
+                event_type=delta.event_type,
+                state_key=delta.state_key,
+            )
+
+            if self._next_send_ts_changed(next_send_ts):
+                self._schedule_next_at_or_none(next_send_ts)
+
+    async def add(
+        self,
+        requester: Requester,
+        *,
+        room_id: str,
+        event_type: str,
+        state_key: Optional[str],
+        origin_server_ts: Optional[int],
+        content: JsonDict,
+        delay: int,
+    ) -> str:
+        """
+        Creates a new delayed event and schedules its delivery.
+
+        Args:
+            requester: The requester of the delayed event, who will be its owner.
+            room_id: The ID of the room where the event should be sent to.
+            event_type: The type of event to be sent.
+            state_key: The state key of the event to be sent, or None if it is not a state event.
+            origin_server_ts: The custom timestamp to send the event with.
+                If None, the timestamp will be the actual time when the event is sent.
+            content: The content of the event to be sent.
+            delay: How long (in milliseconds) to wait before automatically sending the event.
+
+        Returns: The ID of the added delayed event.
+
+        Raises:
+            SynapseError: if the delayed event fails validation checks.
+        """
+        await self._request_ratelimiter.ratelimit(requester)
+
+        self._event_creation_handler.validator.validate_builder(
+            self._event_creation_handler.event_builder_factory.for_room_version(
+                await self._store.get_room_version(room_id),
+                {
+                    "type": event_type,
+                    "content": content,
+                    "room_id": room_id,
+                    "sender": str(requester.user),
+                    **({"state_key": state_key} if state_key is not None else {}),
+                },
+            )
+        )
+
+        creation_ts = self._get_current_ts()
+
+        delay_id, next_send_ts = await self._store.add_delayed_event(
+            user_localpart=requester.user.localpart,
+            device_id=requester.device_id,
+            creation_ts=creation_ts,
+            room_id=room_id,
+            event_type=event_type,
+            state_key=state_key,
+            origin_server_ts=origin_server_ts,
+            content=content,
+            delay=delay,
+        )
+
+        if self._repl_client is not None:
+            # NOTE: If this throws, the delayed event will remain in the DB and
+            # will be picked up once the main worker gets another delayed event.
+            await self._repl_client(
+                instance_name=MAIN_PROCESS_INSTANCE_NAME,
+                next_send_ts=next_send_ts,
+            )
+        elif self._next_send_ts_changed(next_send_ts):
+            self._schedule_next_at(next_send_ts)
+
+        return delay_id
+
+    def on_added(self, next_send_ts: int) -> None:
+        next_send_ts = Timestamp(next_send_ts)
+        if self._next_send_ts_changed(next_send_ts):
+            self._schedule_next_at(next_send_ts)
+
+    async def cancel(self, requester: Requester, delay_id: str) -> None:
+        """
+        Cancels the scheduled delivery of the matching delayed event.
+
+        Args:
+            requester: The owner of the delayed event to act on.
+            delay_id: The ID of the delayed event to act on.
+
+        Raises:
+            NotFoundError: if no matching delayed event could be found.
+        """
+        assert self._is_master
+        await self._request_ratelimiter.ratelimit(requester)
+        await self._initialized_from_db
+
+        next_send_ts = await self._store.cancel_delayed_event(
+            delay_id=delay_id,
+            user_localpart=requester.user.localpart,
+        )
+
+        if self._next_send_ts_changed(next_send_ts):
+            self._schedule_next_at_or_none(next_send_ts)
+
+    async def restart(self, requester: Requester, delay_id: str) -> None:
+        """
+        Restarts the scheduled delivery of the matching delayed event.
+
+        Args:
+            requester: The owner of the delayed event to act on.
+            delay_id: The ID of the delayed event to act on.
+
+        Raises:
+            NotFoundError: if no matching delayed event could be found.
+        """
+        assert self._is_master
+        await self._request_ratelimiter.ratelimit(requester)
+        await self._initialized_from_db
+
+        next_send_ts = await self._store.restart_delayed_event(
+            delay_id=delay_id,
+            user_localpart=requester.user.localpart,
+            current_ts=self._get_current_ts(),
+        )
+
+        if self._next_send_ts_changed(next_send_ts):
+            self._schedule_next_at(next_send_ts)
+
+    async def send(self, requester: Requester, delay_id: str) -> None:
+        """
+        Immediately sends the matching delayed event, instead of waiting for its scheduled delivery.
+
+        Args:
+            requester: The owner of the delayed event to act on.
+            delay_id: The ID of the delayed event to act on.
+
+        Raises:
+            NotFoundError: if no matching delayed event could be found.
+        """
+        assert self._is_master
+        await self._request_ratelimiter.ratelimit(requester)
+        await self._initialized_from_db
+
+        event, next_send_ts = await self._store.process_target_delayed_event(
+            delay_id=delay_id,
+            user_localpart=requester.user.localpart,
+        )
+
+        if self._next_send_ts_changed(next_send_ts):
+            self._schedule_next_at_or_none(next_send_ts)
+
+        await self._send_event(
+            DelayedEventDetails(
+                delay_id=DelayID(delay_id),
+                user_localpart=UserLocalpart(requester.user.localpart),
+                room_id=event.room_id,
+                type=event.type,
+                state_key=event.state_key,
+                origin_server_ts=event.origin_server_ts,
+                content=event.content,
+                device_id=event.device_id,
+            )
+        )
+
+    async def _send_on_timeout(self) -> None:
+        self._next_delayed_event_call = None
+
+        events, next_send_ts = await self._store.process_timeout_delayed_events(
+            self._get_current_ts()
+        )
+
+        if next_send_ts:
+            self._schedule_next_at(next_send_ts)
+
+        await self._send_events(events)
+
+    async def _send_events(self, events: List[DelayedEventDetails]) -> None:
+        sent_state: Set[Tuple[RoomID, EventType, StateKey]] = set()
+        for event in events:
+            if event.state_key is not None:
+                state_info = (event.room_id, event.type, event.state_key)
+                if state_info in sent_state:
+                    continue
+            else:
+                state_info = None
+            try:
+                # TODO: send in background if message event or non-conflicting state event
+                await self._send_event(event)
+                if state_info is not None:
+                    sent_state.add(state_info)
+            except Exception:
+                logger.exception("Failed to send delayed event")
+
+            for room_id, event_type, state_key in sent_state:
+                await self._store.delete_processed_delayed_state_events(
+                    room_id=str(room_id),
+                    event_type=event_type,
+                    state_key=state_key,
+                )
+
+    def _schedule_next_at_or_none(self, next_send_ts: Optional[Timestamp]) -> None:
+        if next_send_ts is not None:
+            self._schedule_next_at(next_send_ts)
+        elif self._next_delayed_event_call is not None:
+            self._next_delayed_event_call.cancel()
+            self._next_delayed_event_call = None
+
+    def _schedule_next_at(self, next_send_ts: Timestamp) -> None:
+        delay = next_send_ts - self._get_current_ts()
+        delay_sec = delay / 1000 if delay > 0 else 0
+
+        if self._next_delayed_event_call is None:
+            self._next_delayed_event_call = self._clock.call_later(
+                delay_sec,
+                run_as_background_process,
+                "_send_on_timeout",
+                self._send_on_timeout,
+            )
+        else:
+            self._next_delayed_event_call.reset(delay_sec)
+
+    async def get_all_for_user(self, requester: Requester) -> List[JsonDict]:
+        """Return all pending delayed events requested by the given user."""
+        await self._request_ratelimiter.ratelimit(requester)
+        return await self._store.get_all_delayed_events_for_user(
+            requester.user.localpart
+        )
+
+    async def _send_event(
+        self,
+        event: DelayedEventDetails,
+        txn_id: Optional[str] = None,
+    ) -> None:
+        user_id = UserID(event.user_localpart, self._config.server.server_name)
+        user_id_str = user_id.to_string()
+        # Create a new requester from what data is currently available
+        requester = create_requester(
+            user_id,
+            is_guest=await self._store.is_guest(user_id_str),
+            device_id=event.device_id,
+        )
+
+        try:
+            if event.state_key is not None and event.type == EventTypes.Member:
+                membership = event.content.get("membership")
+                assert membership is not None
+                event_id, _ = await self._room_member_handler.update_membership(
+                    requester,
+                    target=UserID.from_string(event.state_key),
+                    room_id=event.room_id.to_string(),
+                    action=membership,
+                    content=event.content,
+                    origin_server_ts=event.origin_server_ts,
+                )
+            else:
+                event_dict: JsonDict = {
+                    "type": event.type,
+                    "content": event.content,
+                    "room_id": event.room_id.to_string(),
+                    "sender": user_id_str,
+                }
+
+                if event.origin_server_ts is not None:
+                    event_dict["origin_server_ts"] = event.origin_server_ts
+
+                if event.state_key is not None:
+                    event_dict["state_key"] = event.state_key
+
+                (
+                    sent_event,
+                    _,
+                ) = await self._event_creation_handler.create_and_send_nonmember_event(
+                    requester,
+                    event_dict,
+                    txn_id=txn_id,
+                )
+                event_id = sent_event.event_id
+        except ShadowBanError:
+            event_id = generate_fake_event_id()
+        finally:
+            # TODO: If this is a temporary error, retry. Otherwise, consider notifying clients of the failure
+            try:
+                await self._store.delete_processed_delayed_event(
+                    event.delay_id, event.user_localpart
+                )
+            except Exception:
+                logger.exception("Failed to delete processed delayed event")
+
+        set_tag("event_id", event_id)
+
+    def _get_current_ts(self) -> Timestamp:
+        return Timestamp(self._clock.time_msec())
+
+    def _next_send_ts_changed(self, next_send_ts: Optional[Timestamp]) -> bool:
+        # The DB alone knows if the next send time changed after adding/modifying
+        # a delayed event, but if we were to ever miss updating our delayed call's
+        # firing time, we may miss other updates. So, keep track of changes to the
+        # the next send time here instead of in the DB.
+        cached_next_send_ts = (
+            int(self._next_delayed_event_call.getTime() * 1000)
+            if self._next_delayed_event_call is not None
+            else None
+        )
+        return next_send_ts != cached_next_send_ts

synapse/handlers/device.py

@@ -48,6 +48,7 @@ from synapse.metrics.background_process_metrics import (
     wrap_as_background_process,
 )
 from synapse.storage.databases.main.client_ips import DeviceLastConnectionInfo
+from synapse.storage.databases.main.roommember import EventIdMembership
 from synapse.storage.databases.main.state_deltas import StateDelta
 from synapse.types import (
     DeviceListUpdates,
@@ -222,7 +223,6 @@ class DeviceWorkerHandler:
         return changed
 
     @trace
-    @measure_func("device.get_user_ids_changed")
     @cancellable
     async def get_user_ids_changed(
         self, user_id: str, from_token: StreamToken
@@ -290,9 +290,11 @@ class DeviceWorkerHandler:
                 memberships_to_fetch.add(delta.prev_event_id)
 
         # Fetch all the memberships for the membership events
-        event_id_to_memberships = await self.store.get_membership_from_event_ids(
-            memberships_to_fetch
-        )
+        event_id_to_memberships: Mapping[str, Optional[EventIdMembership]] = {}
+        if memberships_to_fetch:
+            event_id_to_memberships = await self.store.get_membership_from_event_ids(
+                memberships_to_fetch
+            )
 
         joined_invited_knocked = (
             Membership.JOIN,
@@ -349,7 +351,6 @@ class DeviceWorkerHandler:
 
         return device_list_updates
 
-    @measure_func("_generate_sync_entry_for_device_list")
     async def generate_sync_entry_for_device_list(
         self,
         user_id: str,

synapse/handlers/directory.py

@@ -265,9 +265,9 @@ class DirectoryHandler:
     async def get_association(self, room_alias: RoomAlias) -> JsonDict:
         room_id = None
         if self.hs.is_mine(room_alias):
-            result: Optional[RoomAliasMapping] = (
-                await self.get_association_from_room_alias(room_alias)
-            )
+            result: Optional[
+                RoomAliasMapping
+            ] = await self.get_association_from_room_alias(room_alias)
 
             if result:
                 room_id = result.room_id
@@ -512,11 +512,9 @@ class DirectoryHandler:
                 raise SynapseError(403, "Not allowed to publish room")
 
             # Check if publishing is blocked by a third party module
-            allowed_by_third_party_rules = (
-                await (
-                    self._third_party_event_rules.check_visibility_can_be_modified(
-                        room_id, visibility
-                    )
+            allowed_by_third_party_rules = await (
+                self._third_party_event_rules.check_visibility_can_be_modified(
+                    room_id, visibility
                 )
             )
             if not allowed_by_third_party_rules:

synapse/handlers/federation.py

@@ -1001,11 +1001,11 @@ class FederationHandler:
                     )
 
                 if include_auth_user_id:
-                    event_content[EventContentFields.AUTHORISING_USER] = (
-                        await self._event_auth_handler.get_user_which_could_invite(
-                            room_id,
-                            state_ids,
-                        )
+                    event_content[
+                        EventContentFields.AUTHORISING_USER
+                    ] = await self._event_auth_handler.get_user_which_could_invite(
+                        room_id,
+                        state_ids,
                     )
 
         builder = self.event_builder_factory.for_room_version(

synapse/handlers/identity.py

@@ -21,6 +21,7 @@
 #
 
 """Utilities for interacting with Identity Servers"""
+
 import logging
 import urllib.parse
 from typing import TYPE_CHECKING, Awaitable, Callable, Dict, List, Optional, Tuple

synapse/handlers/message.py

@@ -1225,10 +1225,9 @@ class EventCreationHandler:
             )
 
         if prev_event_ids is not None:
-            assert (
-                len(prev_event_ids) <= 10
-            ), "Attempting to create an event with %i prev_events" % (
-                len(prev_event_ids),
+            assert len(prev_event_ids) <= 10, (
+                "Attempting to create an event with %i prev_events"
+                % (len(prev_event_ids),)
             )
         else:
             prev_event_ids = await self.store.get_prev_events_for_room(builder.room_id)

synapse/handlers/pagination.py

@@ -507,15 +507,17 @@ class PaginationHandler:
 
         # Initially fetch the events from the database. With any luck, we can return
         # these without blocking on backfill (handled below).
-        events, next_key = (
-            await self.store.paginate_room_events_by_topological_ordering(
-                room_id=room_id,
-                from_key=from_token.room_key,
-                to_key=to_room_key,
-                direction=pagin_config.direction,
-                limit=pagin_config.limit,
-                event_filter=event_filter,
-            )
+        (
+            events,
+            next_key,
+            _,
+        ) = await self.store.paginate_room_events_by_topological_ordering(
+            room_id=room_id,
+            from_key=from_token.room_key,
+            to_key=to_room_key,
+            direction=pagin_config.direction,
+            limit=pagin_config.limit,
+            event_filter=event_filter,
         )
 
         if pagin_config.direction == Direction.BACKWARDS:
@@ -584,15 +586,17 @@ class PaginationHandler:
                 # If we did backfill something, refetch the events from the database to
                 # catch anything new that might have been added since we last fetched.
                 if did_backfill:
-                    events, next_key = (
-                        await self.store.paginate_room_events_by_topological_ordering(
-                            room_id=room_id,
-                            from_key=from_token.room_key,
-                            to_key=to_room_key,
-                            direction=pagin_config.direction,
-                            limit=pagin_config.limit,
-                            event_filter=event_filter,
-                        )
+                    (
+                        events,
+                        next_key,
+                        _,
+                    ) = await self.store.paginate_room_events_by_topological_ordering(
+                        room_id=room_id,
+                        from_key=from_token.room_key,
+                        to_key=to_room_key,
+                        direction=pagin_config.direction,
+                        limit=pagin_config.limit,
+                        event_filter=event_filter,
                     )
             else:
                 # Otherwise, we can backfill in the background for eventual

synapse/handlers/presence.py

@@ -71,6 +71,7 @@ user state; this device follows the normal timeout logic (see above) and will
 automatically be replaced with any information from currently available devices.
 
 """
+
 import abc
 import contextlib
 import itertools
@@ -493,9 +494,9 @@ class WorkerPresenceHandler(BasePresenceHandler):
 
         # The number of ongoing syncs on this process, by (user ID, device ID).
         # Empty if _presence_enabled is false.
-        self._user_device_to_num_current_syncs: Dict[Tuple[str, Optional[str]], int] = (
-            {}
-        )
+        self._user_device_to_num_current_syncs: Dict[
+            Tuple[str, Optional[str]], int
+        ] = {}
 
         self.notifier = hs.get_notifier()
         self.instance_id = hs.get_instance_id()
@@ -818,9 +819,9 @@ class PresenceHandler(BasePresenceHandler):
 
         # Keeps track of the number of *ongoing* syncs on this process. While
         # this is non zero a user will never go offline.
-        self._user_device_to_num_current_syncs: Dict[Tuple[str, Optional[str]], int] = (
-            {}
-        )
+        self._user_device_to_num_current_syncs: Dict[
+            Tuple[str, Optional[str]], int
+        ] = {}
 
         # Keeps track of the number of *ongoing* syncs on other processes.
         #

synapse/handlers/profile.py

@@ -351,9 +351,9 @@ class ProfileHandler:
             server_name = host
 
         if self._is_mine_server_name(server_name):
-            media_info: Optional[Union[LocalMedia, RemoteMedia]] = (
-                await self.store.get_local_media(media_id)
-            )
+            media_info: Optional[
+                Union[LocalMedia, RemoteMedia]
+            ] = await self.store.get_local_media(media_id)
         else:
             media_info = await self.store.get_cached_remote_media(server_name, media_id)
 

synapse/handlers/relations.py

@@ -188,13 +188,13 @@ class RelationsHandler:
         if include_original_event:
             # Do not bundle aggregations when retrieving the original event because
             # we want the content before relations are applied to it.
-            return_value["original_event"] = (
-                await self._event_serializer.serialize_event(
-                    event,
-                    now,
-                    bundle_aggregations=None,
-                    config=serialize_options,
-                )
+            return_value[
+                "original_event"
+            ] = await self._event_serializer.serialize_event(
+                event,
+                now,
+                bundle_aggregations=None,
+                config=serialize_options,
             )
 
         if next_token:

synapse/handlers/room.py

@@ -20,6 +20,7 @@
 #
 
 """Contains functions for performing actions on rooms."""
+
 import itertools
 import logging
 import math
@@ -900,11 +901,9 @@ class RoomCreationHandler:
         )
 
         # Check whether this visibility value is blocked by a third party module
-        allowed_by_third_party_rules = (
-            await (
-                self._third_party_event_rules.check_visibility_can_be_modified(
-                    room_id, visibility
-                )
+        allowed_by_third_party_rules = await (
+            self._third_party_event_rules.check_visibility_can_be_modified(
+                room_id, visibility
             )
         )
         if not allowed_by_third_party_rules:
@@ -1754,7 +1753,7 @@ class RoomEventSource(EventSource[RoomStreamToken, EventBase]):
             )
 
             events = list(room_events)
-            events.extend(e for evs, _ in room_to_events.values() for e in evs)
+            events.extend(e for evs, _, _ in room_to_events.values() for e in evs)
 
             # We know stream_ordering must be not None here, as its been
             # persisted, but mypy doesn't know that

synapse/handlers/room_member.py

@@ -1302,11 +1302,11 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
         # If this is going to be a local join, additional information must
         # be included in the event content in order to efficiently validate
         # the event.
-        content[EventContentFields.AUTHORISING_USER] = (
-            await self.event_auth_handler.get_user_which_could_invite(
-                room_id,
-                state_before_join,
-            )
+        content[
+            EventContentFields.AUTHORISING_USER
+        ] = await self.event_auth_handler.get_user_which_could_invite(
+            room_id,
+            state_before_join,
         )
 
         return False, []
@@ -1415,9 +1415,9 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
 
         if requester is not None:
             sender = UserID.from_string(event.sender)
-            assert (
-                sender == requester.user
-            ), "Sender (%s) must be same as requester (%s)" % (sender, requester.user)
+            assert sender == requester.user, (
+                "Sender (%s) must be same as requester (%s)" % (sender, requester.user)
+            )
             assert self.hs.is_mine(sender), "Sender must be our own: %s" % (sender,)
         else:
             requester = types.create_requester(target_user)

synapse/handlers/room_summary.py

@@ -183,8 +183,13 @@ class RoomSummaryHandler:
     ) -> JsonDict:
         """See docstring for SpaceSummaryHandler.get_room_hierarchy."""
 
-        # First of all, check that the room is accessible.
-        if not await self._is_local_room_accessible(requested_room_id, requester):
+        # If the room is available locally, quickly check that the user can access it.
+        local_room = await self._store.is_host_joined(
+            requested_room_id, self._server_name
+        )
+        if local_room and not await self._is_local_room_accessible(
+            requested_room_id, requester
+        ):
             raise UnstableSpecAuthError(
                 403,
                 "User %s not in room %s, and room previews are disabled"
@@ -192,6 +197,22 @@ class RoomSummaryHandler:
                 errcode=Codes.NOT_JOINED,
             )
 
+        if not local_room:
+            room_hierarchy = await self._summarize_remote_room_hierarchy(
+                _RoomQueueEntry(requested_room_id, ()),
+                False,
+            )
+            root_room_entry = room_hierarchy[0]
+            if not root_room_entry or not await self._is_remote_room_accessible(
+                requester, requested_room_id, root_room_entry.room
+            ):
+                raise UnstableSpecAuthError(
+                    403,
+                    "User %s not in room %s, and room previews are disabled"
+                    % (requester, requested_room_id),
+                    errcode=Codes.NOT_JOINED,
+                )
+
         # If this is continuing a previous session, pull the persisted data.
         if from_token:
             try:

synapse/handlers/search.py

@@ -423,9 +423,9 @@ class SearchHandler:
             }
 
         if search_result.room_groups and "room_id" in group_keys:
-            rooms_cat_res.setdefault("groups", {})[
-                "room_id"
-            ] = search_result.room_groups
+            rooms_cat_res.setdefault("groups", {})["room_id"] = (
+                search_result.room_groups
+            )
 
         if sender_group and "sender" in group_keys:
             rooms_cat_res.setdefault("groups", {})["sender"] = sender_group

synapse/handlers/sliding_sync/extensions.py

@@ -12,19 +12,33 @@
 # <https://www.gnu.org/licenses/agpl-3.0.html>.
 #
 
+import itertools
 import logging
-from typing import TYPE_CHECKING, Dict, List, Mapping, Optional, Sequence, Set
+from typing import (
+    TYPE_CHECKING,
+    AbstractSet,
+    ChainMap,
+    Dict,
+    Mapping,
+    MutableMapping,
+    Optional,
+    Sequence,
+    Set,
+    cast,
+)
 
 from typing_extensions import assert_never
 
-from synapse.api.constants import AccountDataTypes
+from synapse.api.constants import AccountDataTypes, EduTypes
 from synapse.handlers.receipts import ReceiptEventSource
 from synapse.logging.opentracing import trace
+from synapse.storage.databases.main.receipts import ReceiptInRoom
 from synapse.types import (
     DeviceListUpdates,
     JsonMapping,
     MultiWriterStreamToken,
     SlidingSyncStreamToken,
+    StrCollection,
     StreamToken,
 )
 from synapse.types.handlers.sliding_sync import (
@@ -35,6 +49,7 @@ from synapse.types.handlers.sliding_sync import (
     SlidingSyncConfig,
     SlidingSyncResult,
 )
+from synapse.util.async_helpers import concurrently_execute
 
 if TYPE_CHECKING:
     from synapse.server import HomeServer
@@ -57,9 +72,9 @@ class SlidingSyncExtensionHandler:
         sync_config: SlidingSyncConfig,
         previous_connection_state: "PerConnectionState",
         new_connection_state: "MutablePerConnectionState",
-        actual_lists: Dict[str, SlidingSyncResult.SlidingWindowList],
+        actual_lists: Mapping[str, SlidingSyncResult.SlidingWindowList],
         actual_room_ids: Set[str],
-        actual_room_response_map: Dict[str, SlidingSyncResult.RoomResult],
+        actual_room_response_map: Mapping[str, SlidingSyncResult.RoomResult],
         to_token: StreamToken,
         from_token: Optional[SlidingSyncStreamToken],
     ) -> SlidingSyncResult.Extensions:
@@ -103,6 +118,8 @@ class SlidingSyncExtensionHandler:
         if sync_config.extensions.account_data is not None:
             account_data_response = await self.get_account_data_extension_response(
                 sync_config=sync_config,
+                previous_connection_state=previous_connection_state,
+                new_connection_state=new_connection_state,
                 actual_lists=actual_lists,
                 actual_room_ids=actual_room_ids,
                 account_data_request=sync_config.extensions.account_data,
@@ -146,10 +163,10 @@ class SlidingSyncExtensionHandler:
 
     def find_relevant_room_ids_for_extension(
         self,
-        requested_lists: Optional[List[str]],
-        requested_room_ids: Optional[List[str]],
-        actual_lists: Dict[str, SlidingSyncResult.SlidingWindowList],
-        actual_room_ids: Set[str],
+        requested_lists: Optional[StrCollection],
+        requested_room_ids: Optional[StrCollection],
+        actual_lists: Mapping[str, SlidingSyncResult.SlidingWindowList],
+        actual_room_ids: AbstractSet[str],
     ) -> Set[str]:
         """
         Handle the reserved `lists`/`rooms` keys for extensions. Extensions should only
@@ -345,7 +362,9 @@ class SlidingSyncExtensionHandler:
     async def get_account_data_extension_response(
         self,
         sync_config: SlidingSyncConfig,
-        actual_lists: Dict[str, SlidingSyncResult.SlidingWindowList],
+        previous_connection_state: "PerConnectionState",
+        new_connection_state: "MutablePerConnectionState",
+        actual_lists: Mapping[str, SlidingSyncResult.SlidingWindowList],
         actual_room_ids: Set[str],
         account_data_request: SlidingSyncConfig.Extensions.AccountDataExtension,
         to_token: StreamToken,
@@ -377,29 +396,39 @@ class SlidingSyncExtensionHandler:
                 )
             )
 
+            # TODO: This should take into account the `from_token` and `to_token`
             have_push_rules_changed = await self.store.have_push_rules_changed_for_user(
                 user_id, from_token.stream_token.push_rules_key
             )
             if have_push_rules_changed:
-                global_account_data_map = dict(global_account_data_map)
                 # TODO: This should take into account the `from_token` and `to_token`
-                global_account_data_map[AccountDataTypes.PUSH_RULES] = (
-                    await self.push_rules_handler.push_rules_for_user(sync_config.user)
-                )
+                global_account_data_map[
+                    AccountDataTypes.PUSH_RULES
+                ] = await self.push_rules_handler.push_rules_for_user(sync_config.user)
         else:
             # TODO: This should take into account the `to_token`
-            all_global_account_data = await self.store.get_global_account_data_for_user(
-                user_id
+            immutable_global_account_data_map = (
+                await self.store.get_global_account_data_for_user(user_id)
             )
 
-            global_account_data_map = dict(all_global_account_data)
-            # TODO: This should take into account the  `to_token`
-            global_account_data_map[AccountDataTypes.PUSH_RULES] = (
-                await self.push_rules_handler.push_rules_for_user(sync_config.user)
+            # Use a `ChainMap` to avoid copying the immutable data from the cache
+            global_account_data_map = ChainMap(
+                {
+                    # TODO: This should take into account the `to_token`
+                    AccountDataTypes.PUSH_RULES: await self.push_rules_handler.push_rules_for_user(
+                        sync_config.user
+                    )
+                },
+                # Cast is safe because `ChainMap` only mutates the top-most map,
+                # see https://github.com/python/typeshed/issues/8430
+                cast(
+                    MutableMapping[str, JsonMapping], immutable_global_account_data_map
+                ),
             )
 
         # Fetch room account data
-        account_data_by_room_map: Mapping[str, Mapping[str, JsonMapping]] = {}
+        #
+        account_data_by_room_map: MutableMapping[str, Mapping[str, JsonMapping]] = {}
         relevant_room_ids = self.find_relevant_room_ids_for_extension(
             requested_lists=account_data_request.lists,
             requested_room_ids=account_data_request.rooms,
@@ -407,25 +436,154 @@ class SlidingSyncExtensionHandler:
             actual_room_ids=actual_room_ids,
         )
         if len(relevant_room_ids) > 0:
+            # We need to handle the different cases depending on if we have sent
+            # down account data previously or not, so we split the relevant
+            # rooms up into different collections based on status.
+            live_rooms = set()
+            previously_rooms: Dict[str, int] = {}
+            initial_rooms = set()
+
+            for room_id in relevant_room_ids:
+                if not from_token:
+                    initial_rooms.add(room_id)
+                    continue
+
+                room_status = previous_connection_state.account_data.have_sent_room(
+                    room_id
+                )
+                if room_status.status == HaveSentRoomFlag.LIVE:
+                    live_rooms.add(room_id)
+                elif room_status.status == HaveSentRoomFlag.PREVIOUSLY:
+                    assert room_status.last_token is not None
+                    previously_rooms[room_id] = room_status.last_token
+                elif room_status.status == HaveSentRoomFlag.NEVER:
+                    initial_rooms.add(room_id)
+                else:
+                    assert_never(room_status.status)
+
+            # We fetch all room account data since the from_token. This is so
+            # that we can record which rooms have updates that haven't been sent
+            # down.
+            #
+            # Mapping from room_id to mapping of `type` to `content` of room account
+            # data events.
+            all_updates_since_the_from_token: Mapping[
+                str, Mapping[str, JsonMapping]
+            ] = {}
             if from_token is not None:
                 # TODO: This should take into account the `from_token` and `to_token`
-                account_data_by_room_map = (
+                all_updates_since_the_from_token = (
                     await self.store.get_updated_room_account_data_for_user(
                         user_id, from_token.stream_token.account_data_key
                     )
                 )
-            else:
-                # TODO: This should take into account the `to_token`
-                account_data_by_room_map = (
-                    await self.store.get_room_account_data_for_user(user_id)
+
+                # Add room tags
+                #
+                # TODO: This should take into account the `from_token` and `to_token`
+                tags_by_room = await self.store.get_updated_tags(
+                    user_id, from_token.stream_token.account_data_key
+                )
+                for room_id, tags in tags_by_room.items():
+                    all_updates_since_the_from_token.setdefault(room_id, {})[
+                        AccountDataTypes.TAG
+                    ] = {"tags": tags}
+
+            # For live rooms we just get the updates from `all_updates_since_the_from_token`
+            if live_rooms:
+                for room_id in all_updates_since_the_from_token.keys() & live_rooms:
+                    account_data_by_room_map[room_id] = (
+                        all_updates_since_the_from_token[room_id]
+                    )
+
+            # For previously and initial rooms we query each room individually.
+            if previously_rooms or initial_rooms:
+
+                async def handle_previously(room_id: str) -> None:
+                    # Either get updates or all account data in the room
+                    # depending on if the room state is PREVIOUSLY or NEVER.
+                    previous_token = previously_rooms.get(room_id)
+                    if previous_token is not None:
+                        room_account_data = await (
+                            self.store.get_updated_room_account_data_for_user_for_room(
+                                user_id=user_id,
+                                room_id=room_id,
+                                from_stream_id=previous_token,
+                                to_stream_id=to_token.account_data_key,
+                            )
+                        )
+
+                        # Add room tags
+                        changed = await self.store.has_tags_changed_for_room(
+                            user_id=user_id,
+                            room_id=room_id,
+                            from_stream_id=previous_token,
+                            to_stream_id=to_token.account_data_key,
+                        )
+                        if changed:
+                            # XXX: Ideally, this should take into account the `to_token`
+                            # and return the set of tags at that time but we don't track
+                            # changes to tags so we just have to return all tags for the
+                            # room.
+                            immutable_tag_map = await self.store.get_tags_for_room(
+                                user_id, room_id
+                            )
+                            room_account_data[AccountDataTypes.TAG] = {
+                                "tags": immutable_tag_map
+                            }
+
+                        # Only add an entry if there were any updates.
+                        if room_account_data:
+                            account_data_by_room_map[room_id] = room_account_data
+                    else:
+                        # TODO: This should take into account the `to_token`
+                        immutable_room_account_data = (
+                            await self.store.get_account_data_for_room(user_id, room_id)
+                        )
+
+                        # Add room tags
+                        #
+                        # XXX: Ideally, this should take into account the `to_token`
+                        # and return the set of tags at that time but we don't track
+                        # changes to tags so we just have to return all tags for the
+                        # room.
+                        immutable_tag_map = await self.store.get_tags_for_room(
+                            user_id, room_id
+                        )
+
+                        account_data_by_room_map[room_id] = ChainMap(
+                            {AccountDataTypes.TAG: {"tags": immutable_tag_map}}
+                            if immutable_tag_map
+                            else {},
+                            # Cast is safe because `ChainMap` only mutates the top-most map,
+                            # see https://github.com/python/typeshed/issues/8430
+                            cast(
+                                MutableMapping[str, JsonMapping],
+                                immutable_room_account_data,
+                            ),
+                        )
+
+                # We handle these rooms concurrently to speed it up.
+                await concurrently_execute(
+                    handle_previously,
+                    previously_rooms.keys() | initial_rooms,
+                    limit=20,
                 )
 
-        # Filter down to the relevant rooms
-        account_data_by_room_map = {
-            room_id: account_data_map
-            for room_id, account_data_map in account_data_by_room_map.items()
-            if room_id in relevant_room_ids
-        }
+            # Now record which rooms are now up to data, and which rooms have
+            # pending updates to send.
+            new_connection_state.account_data.record_sent_rooms(previously_rooms.keys())
+            new_connection_state.account_data.record_sent_rooms(initial_rooms)
+            missing_updates = (
+                all_updates_since_the_from_token.keys() - relevant_room_ids
+            )
+            if missing_updates:
+                # If we have missing updates then we must have had a from_token.
+                assert from_token is not None
+
+                new_connection_state.account_data.record_unsent_rooms(
+                    missing_updates, from_token.stream_token.account_data_key
+                )
 
         return SlidingSyncResult.Extensions.AccountDataExtension(
             global_account_data_map=global_account_data_map,
@@ -438,9 +596,9 @@ class SlidingSyncExtensionHandler:
         sync_config: SlidingSyncConfig,
         previous_connection_state: "PerConnectionState",
         new_connection_state: "MutablePerConnectionState",
-        actual_lists: Dict[str, SlidingSyncResult.SlidingWindowList],
+        actual_lists: Mapping[str, SlidingSyncResult.SlidingWindowList],
         actual_room_ids: Set[str],
-        actual_room_response_map: Dict[str, SlidingSyncResult.RoomResult],
+        actual_room_response_map: Mapping[str, SlidingSyncResult.RoomResult],
         receipts_request: SlidingSyncConfig.Extensions.ReceiptsExtension,
         to_token: StreamToken,
         from_token: Optional[SlidingSyncStreamToken],
@@ -486,15 +644,21 @@ class SlidingSyncExtensionHandler:
                     initial_rooms.add(room_id)
                     continue
 
-                # If we're sending down the room from scratch again for some reason, we
-                # should always resend the receipts as well (regardless of if
-                # we've sent them down before). This is to mimic the behaviour
-                # of what happens on initial sync, where you get a chunk of
-                # timeline with all of the corresponding receipts for the events in the timeline.
+                # If we're sending down the room from scratch again for some
+                # reason, we should always resend the receipts as well
+                # (regardless of if we've sent them down before). This is to
+                # mimic the behaviour of what happens on initial sync, where you
+                # get a chunk of timeline with all of the corresponding receipts
+                # for the events in the timeline.
+                #
+                # We also resend down receipts when we "expand" the timeline,
+                # (see the "XXX: Odd behavior" in
+                # `synapse.handlers.sliding_sync`).
                 room_result = actual_room_response_map.get(room_id)
-                if room_result is not None and room_result.initial:
-                    initial_rooms.add(room_id)
-                    continue
+                if room_result is not None:
+                    if room_result.initial or room_result.unstable_expanded_timeline:
+                        initial_rooms.add(room_id)
+                        continue
 
                 room_status = previous_connection_state.receipts.have_sent_room(room_id)
                 if room_status.status == HaveSentRoomFlag.LIVE:
@@ -525,7 +689,10 @@ class SlidingSyncExtensionHandler:
             # For rooms we've previously sent down, but aren't up to date, we
             # need to use the from token from the room status.
             if previously_rooms:
-                for room_id, receipt_token in previously_rooms.items():
+                # Fetch any missing rooms concurrently.
+
+                async def handle_previously_room(room_id: str) -> None:
+                    receipt_token = previously_rooms[room_id]
                     # TODO: Limit the number of receipts we're about to send down
                     # for the room, if its too many we should TODO
                     previously_receipts = (
@@ -537,21 +704,53 @@ class SlidingSyncExtensionHandler:
                     )
                     fetched_receipts.extend(previously_receipts)
 
-            # For rooms we haven't previously sent down, we could send all receipts
-            # from that room but we only want to include receipts for events
-            # in the timeline to avoid bloating and blowing up the sync response
-            # as the number of users in the room increases. (this behavior is part of the spec)
-            initial_rooms_and_event_ids = [
-                (room_id, event.event_id)
-                for room_id in initial_rooms
-                if room_id in actual_room_response_map
-                for event in actual_room_response_map[room_id].timeline_events
-            ]
-            if initial_rooms_and_event_ids:
+                await concurrently_execute(
+                    handle_previously_room, previously_rooms.keys(), 20
+                )
+
+            if initial_rooms:
+                # We also always send down receipts for the current user.
+                user_receipts = (
+                    await self.store.get_linearized_receipts_for_user_in_rooms(
+                        user_id=sync_config.user.to_string(),
+                        room_ids=initial_rooms,
+                        to_key=to_token.receipt_key,
+                    )
+                )
+
+                # For rooms we haven't previously sent down, we could send all receipts
+                # from that room but we only want to include receipts for events
+                # in the timeline to avoid bloating and blowing up the sync response
+                # as the number of users in the room increases. (this behavior is part of the spec)
+                initial_rooms_and_event_ids = [
+                    (room_id, event.event_id)
+                    for room_id in initial_rooms
+                    if room_id in actual_room_response_map
+                    for event in actual_room_response_map[room_id].timeline_events
+                ]
                 initial_receipts = await self.store.get_linearized_receipts_for_events(
                     room_and_event_ids=initial_rooms_and_event_ids,
                 )
-                fetched_receipts.extend(initial_receipts)
+
+                # Combine the receipts for a room and add them to
+                # `fetched_receipts`
+                for room_id in initial_receipts.keys() | user_receipts.keys():
+                    receipt_content = ReceiptInRoom.merge_to_content(
+                        list(
+                            itertools.chain(
+                                initial_receipts.get(room_id, []),
+                                user_receipts.get(room_id, []),
+                            )
+                        )
+                    )
+
+                    fetched_receipts.append(
+                        {
+                            "room_id": room_id,
+                            "type": EduTypes.RECEIPT,
+                            "content": receipt_content,
+                        }
+                    )
 
             fetched_receipts = ReceiptEventSource.filter_out_private_receipts(
                 fetched_receipts, sync_config.user.to_string()
@@ -565,9 +764,10 @@ class SlidingSyncExtensionHandler:
 
                 room_id_to_receipt_map[room_id] = {"type": type, "content": content}
 
-        # Now we update the per-connection state to track which receipts we have
-        # and haven't sent down.
-        new_connection_state.receipts.record_sent_rooms(relevant_room_ids)
+            # Update the per-connection state to track which rooms we have sent
+            # all the receipts for.
+            new_connection_state.receipts.record_sent_rooms(previously_rooms.keys())
+            new_connection_state.receipts.record_sent_rooms(initial_rooms)
 
         if from_token:
             # Now find the set of rooms that may have receipts that we're not sending
@@ -600,9 +800,9 @@ class SlidingSyncExtensionHandler:
     async def get_typing_extension_response(
         self,
         sync_config: SlidingSyncConfig,
-        actual_lists: Dict[str, SlidingSyncResult.SlidingWindowList],
+        actual_lists: Mapping[str, SlidingSyncResult.SlidingWindowList],
         actual_room_ids: Set[str],
-        actual_room_response_map: Dict[str, SlidingSyncResult.RoomResult],
+        actual_room_response_map: Mapping[str, SlidingSyncResult.RoomResult],
         typing_request: SlidingSyncConfig.Extensions.TypingExtension,
         to_token: StreamToken,
         from_token: Optional[SlidingSyncStreamToken],

synapse/handlers/sliding_sync/store.py

@@ -63,7 +63,7 @@ class SlidingSyncConnectionStore:
 
     store: "DataStore"
 
-    async def get_per_connection_state(
+    async def get_and_clear_connection_positions(
         self,
         sync_config: SlidingSyncConfig,
         from_token: Optional[SlidingSyncStreamToken],
@@ -73,6 +73,7 @@ class SlidingSyncConnectionStore:
         Raises:
             SlidingSyncUnknownPosition if the connection_token is unknown
         """
+        # If this is our first request, there is no previous connection state to fetch out of the database
         if from_token is None or from_token.connection_position == 0:
             return PerConnectionState()
 
@@ -81,7 +82,7 @@ class SlidingSyncConnectionStore:
         device_id = sync_config.requester.device_id
         assert device_id is not None
 
-        return await self.store.get_per_connection_state(
+        return await self.store.get_and_clear_connection_positions(
             sync_config.user.to_string(),
             device_id,
             conn_id,
@@ -106,8 +107,12 @@ class SlidingSyncConnectionStore:
             else:
                 return 0
 
-        if from_token is not None and from_token.connection_position == 0:
-            from_token = None
+        # A from token with a zero connection position means there was no
+        # previously stored connection state, so we treat a zero the same as
+        # there being no previous position.
+        previous_connection_position = None
+        if from_token is not None and from_token.connection_position != 0:
+            previous_connection_position = from_token.connection_position
 
         conn_id = sync_config.conn_id or ""
 
@@ -118,6 +123,6 @@ class SlidingSyncConnectionStore:
             sync_config.user.to_string(),
             device_id,
             conn_id,
-            from_token.connection_position if from_token else None,
+            previous_connection_position,
             new_connection_state,
         )

synapse/handlers/sync.py

@@ -183,10 +183,7 @@ class JoinedSyncResult:
         to tell if room needs to be part of the sync result.
         """
         return bool(
-            self.timeline
-            or self.state
-            or self.ephemeral
-            or self.account_data
+            self.timeline or self.state or self.ephemeral or self.account_data
             # nb the notification count does not, er, count: if there's nothing
             # else in the result, we don't need to send it.
         )
@@ -575,10 +572,10 @@ class SyncHandler:
         if timeout == 0 or since_token is None or full_state:
             # we are going to return immediately, so don't bother calling
             # notifier.wait_for_events.
-            result: Union[SyncResult, E2eeSyncResult] = (
-                await self.current_sync_for_user(
-                    sync_config, sync_version, since_token, full_state=full_state
-                )
+            result: Union[
+                SyncResult, E2eeSyncResult
+            ] = await self.current_sync_for_user(
+                sync_config, sync_version, since_token, full_state=full_state
             )
         else:
             # Otherwise, we wait for something to happen and report it to the user.
@@ -673,10 +670,10 @@ class SyncHandler:
 
             # Go through the `/sync` v2 path
             if sync_version == SyncVersion.SYNC_V2:
-                sync_result: Union[SyncResult, E2eeSyncResult] = (
-                    await self.generate_sync_result(
-                        sync_config, since_token, full_state
-                    )
+                sync_result: Union[
+                    SyncResult, E2eeSyncResult
+                ] = await self.generate_sync_result(
+                    sync_config, since_token, full_state
                 )
             # Go through the MSC3575 Sliding Sync `/sync/e2ee` path
             elif sync_version == SyncVersion.E2EE_SYNC:
@@ -909,7 +906,7 @@ class SyncHandler:
                     # Use `stream_ordering` for updates
                     else paginate_room_events_by_stream_ordering
                 )
-                events, end_key = await pagination_method(
+                events, end_key, limited = await pagination_method(
                     room_id=room_id,
                     # The bounds are reversed so we can paginate backwards
                     # (from newer to older events) starting at to_bound.
@@ -917,9 +914,7 @@ class SyncHandler:
                     from_key=end_key,
                     to_key=since_key,
                     direction=Direction.BACKWARDS,
-                    # We add one so we can determine if there are enough events to saturate
-                    # the limit or not (see `limited`)
-                    limit=load_limit + 1,
+                    limit=load_limit,
                 )
                 # We want to return the events in ascending order (the last event is the
                 # most recent).
@@ -974,9 +969,6 @@ class SyncHandler:
                 loaded_recents.extend(recents)
                 recents = loaded_recents
 
-                if len(events) <= load_limit:
-                    limited = False
-                    break
                 max_repeat -= 1
 
             if len(recents) > timeline_limit:
@@ -1488,13 +1480,16 @@ class SyncHandler:
                     # timeline here. The caller will then dedupe any redundant
                     # ones.
 
-                    state_ids = await self._state_storage_controller.get_state_ids_for_event(
-                        batch.events[0].event_id,
-                        # we only want members!
-                        state_filter=StateFilter.from_types(
-                            (EventTypes.Member, member) for member in members_to_fetch
-                        ),
-                        await_full_state=False,
+                    state_ids = (
+                        await self._state_storage_controller.get_state_ids_for_event(
+                            batch.events[0].event_id,
+                            # we only want members!
+                            state_filter=StateFilter.from_types(
+                                (EventTypes.Member, member)
+                                for member in members_to_fetch
+                            ),
+                            await_full_state=False,
+                        )
                     )
             return state_ids
 
@@ -2166,18 +2161,18 @@ class SyncHandler:
 
             if push_rules_changed:
                 global_account_data = dict(global_account_data)
-                global_account_data[AccountDataTypes.PUSH_RULES] = (
-                    await self._push_rules_handler.push_rules_for_user(sync_config.user)
-                )
+                global_account_data[
+                    AccountDataTypes.PUSH_RULES
+                ] = await self._push_rules_handler.push_rules_for_user(sync_config.user)
         else:
             all_global_account_data = await self.store.get_global_account_data_for_user(
                 user_id
             )
 
             global_account_data = dict(all_global_account_data)
-            global_account_data[AccountDataTypes.PUSH_RULES] = (
-                await self._push_rules_handler.push_rules_for_user(sync_config.user)
-            )
+            global_account_data[
+                AccountDataTypes.PUSH_RULES
+            ] = await self._push_rules_handler.push_rules_for_user(sync_config.user)
 
         account_data_for_user = (
             await sync_config.filter_collection.filter_global_account_data(
@@ -2608,7 +2603,7 @@ class SyncHandler:
 
             newly_joined = room_id in newly_joined_rooms
             if room_entry:
-                events, start_key = room_entry
+                events, start_key, _ = room_entry
                 # We want to return the events in ascending order (the last event is the
                 # most recent).
                 events.reverse()

synapse/handlers/worker_lock.py

@@ -183,7 +183,7 @@ class WorkerLocksHandler:
             return
 
         def _wake_all_locks(
-            locks: Collection[Union[WaitingLock, WaitingMultiLock]]
+            locks: Collection[Union[WaitingLock, WaitingMultiLock]],
         ) -> None:
             for lock in locks:
                 deferred = lock.deferred

synapse/http/client.py

@@ -1313,6 +1313,5 @@ def is_unknown_endpoint(
         )
     ) or (
         # Older Synapses returned a 400 error.
-        e.code == 400
-        and synapse_error.errcode == Codes.UNRECOGNIZED
+        e.code == 400 and synapse_error.errcode == Codes.UNRECOGNIZED
     )

synapse/http/matrixfederationclient.py

@@ -19,7 +19,6 @@
 #
 #
 import abc
-import cgi
 import codecs
 import logging
 import random
@@ -792,7 +791,7 @@ class MatrixFederationHttpClient:
                                 url_str,
                                 _flatten_response_never_received(e),
                             )
-                            body = None
+                            body = b""
 
                         exc = HttpResponseException(
                             response.code, response_phrase, body
@@ -1756,8 +1755,10 @@ class MatrixFederationHttpClient:
                 request.destination,
                 str_url,
             )
+            # We don't know how large the response will be upfront, so limit it to
+            # the `max_size` config value.
             length, headers, _, _ = await self._simple_http_client.get_file(
-                str_url, output_stream, expected_size
+                str_url, output_stream, max_size
             )
 
         logger.info(
@@ -1811,8 +1812,9 @@ def check_content_type_is(headers: Headers, expected_content_type: str) -> None:
         )
 
     c_type = content_type_headers[0].decode("ascii")  # only the first header
-    val, options = cgi.parse_header(c_type)
-    if val != expected_content_type:
+    # Extract the 'essence' of the mimetype, removing any parameter
+    c_type_parsed = c_type.split(";", 1)[0].strip()
+    if c_type_parsed != expected_content_type:
         raise RequestSendFailed(
             RuntimeError(
                 f"Remote server sent Content-Type header of '{c_type}', not '{expected_content_type}'",

synapse/http/server.py

@@ -233,7 +233,7 @@ def return_html_error(
 
 
 def wrap_async_request_handler(
-    h: Callable[["_AsyncResource", "SynapseRequest"], Awaitable[None]]
+    h: Callable[["_AsyncResource", "SynapseRequest"], Awaitable[None]],
 ) -> Callable[["_AsyncResource", "SynapseRequest"], "defer.Deferred[None]"]:
     """Wraps an async request handler so that it calls request.processing.
 

synapse/http/servlet.py

@@ -37,19 +37,17 @@ from typing import (
     overload,
 )
 
-from synapse._pydantic_compat import HAS_PYDANTIC_V2
-
-if TYPE_CHECKING or HAS_PYDANTIC_V2:
-    from pydantic.v1 import BaseModel, MissingError, PydanticValueError, ValidationError
-    from pydantic.v1.error_wrappers import ErrorWrapper
-else:
-    from pydantic import BaseModel, MissingError, PydanticValueError, ValidationError
-    from pydantic.error_wrappers import ErrorWrapper
-
 from typing_extensions import Literal
 
 from twisted.web.server import Request
 
+from synapse._pydantic_compat import (
+    BaseModel,
+    ErrorWrapper,
+    MissingError,
+    PydanticValueError,
+    ValidationError,
+)
 from synapse.api.errors import Codes, SynapseError
 from synapse.http import redact_uri
 from synapse.http.server import HttpServer

synapse/logging/_terse_json.py

@@ -22,6 +22,7 @@
 """
 Log formatters that output terse JSON.
 """
+
 import json
 import logging
 

synapse/logging/context.py

@@ -20,7 +20,7 @@
 #
 #
 
-""" Thread-local-alike tracking of log contexts within synapse
+"""Thread-local-alike tracking of log contexts within synapse
 
 This module provides objects and utilities for tracking contexts through
 synapse code, so that log lines can include a request identifier, and so that
@@ -29,6 +29,7 @@ them.
 
 See doc/log_contexts.rst for details on how this works.
 """
+
 import logging
 import threading
 import typing
@@ -751,7 +752,7 @@ def preserve_fn(
     f: Union[
         Callable[P, R],
         Callable[P, Awaitable[R]],
-    ]
+    ],
 ) -> Callable[P, "defer.Deferred[R]"]:
     """Function decorator which wraps the function with run_in_background"""
 

synapse/logging/opentracing.py

@@ -169,6 +169,7 @@ Gotchas
   than one caller? Will all of those calling functions have be in a context
   with an active span?
 """
+
 import contextlib
 import enum
 import inspect
@@ -414,7 +415,7 @@ def ensure_active_span(
     """
 
     def ensure_active_span_inner_1(
-        func: Callable[P, R]
+        func: Callable[P, R],
     ) -> Callable[P, Union[Optional[T], R]]:
         @wraps(func)
         def ensure_active_span_inner_2(
@@ -700,7 +701,7 @@ def set_operation_name(operation_name: str) -> None:
 
 @only_if_tracing
 def force_tracing(
-    span: Union["opentracing.Span", _Sentinel] = _Sentinel.sentinel
+    span: Union["opentracing.Span", _Sentinel] = _Sentinel.sentinel,
 ) -> None:
     """Force sampling for the active/given span and its children.
 
@@ -1032,13 +1033,13 @@ def tag_args(func: Callable[P, R]) -> Callable[P, R]:
     def _wrapping_logic(
         _func: Callable[P, R], *args: P.args, **kwargs: P.kwargs
     ) -> Generator[None, None, None]:
-        # We use `[1:]` to skip the `self` object reference and `start=1` to
-        # make the index line up with `argspec.args`.
-        #
-        # FIXME: We could update this to handle any type of function by ignoring the
-        #   first argument only if it's named `self` or `cls`. This isn't fool-proof
-        #   but handles the idiomatic cases.
-        for i, arg in enumerate(args[1:], start=1):
+        for i, arg in enumerate(args, start=0):
+            if argspec.args[i] in ("self", "cls"):
+                # Ignore `self` and `cls` values. Ideally we'd properly detect
+                # if we were wrapping a method, but that is really non-trivial
+                # and this is good enough.
+                continue
+
             set_tag(SynapseTags.FUNC_ARG_PREFIX + argspec.args[i], str(arg))
         set_tag(SynapseTags.FUNC_ARGS, str(args[len(argspec.args) :]))
         set_tag(SynapseTags.FUNC_KWARGS, str(kwargs))
@@ -1093,9 +1094,10 @@ def trace_servlet(
 
             # Mypy seems to think that start_context.tag below can be Optional[str], but
             # that doesn't appear to be correct and works in practice.
-            request_tags[
-                SynapseTags.REQUEST_TAG
-            ] = request.request_metrics.start_context.tag  # type: ignore[assignment]
+
+            request_tags[SynapseTags.REQUEST_TAG] = (
+                request.request_metrics.start_context.tag  # type: ignore[assignment]
+            )
 
             # set the tags *after* the servlet completes, in case it decided to
             # prioritise the span (tags will get dropped on unprioritised spans)

synapse/media/_base.py

@@ -60,8 +60,6 @@ from synapse.util.stringutils import is_ascii
 
 if TYPE_CHECKING:
     from synapse.server import HomeServer
-    from synapse.storage.databases.main.media_repository import LocalMedia
-
 
 logger = logging.getLogger(__name__)
 
@@ -290,7 +288,9 @@ async def respond_with_multipart_responder(
     clock: Clock,
     request: SynapseRequest,
     responder: "Optional[Responder]",
-    media_info: "LocalMedia",
+    media_type: str,
+    media_length: Optional[int],
+    upload_name: Optional[str],
 ) -> None:
     """
     Responds to requests originating from the federation media `/download` endpoint by
@@ -314,7 +314,7 @@ async def respond_with_multipart_responder(
             )
             return
 
-        if media_info.media_type.lower().split(";", 1)[0] in INLINE_CONTENT_TYPES:
+        if media_type.lower().split(";", 1)[0] in INLINE_CONTENT_TYPES:
             disposition = "inline"
         else:
             disposition = "attachment"
@@ -322,16 +322,16 @@ async def respond_with_multipart_responder(
         def _quote(x: str) -> str:
             return urllib.parse.quote(x.encode("utf-8"))
 
-        if media_info.upload_name:
-            if _can_encode_filename_as_token(media_info.upload_name):
+        if upload_name:
+            if _can_encode_filename_as_token(upload_name):
                 disposition = "%s; filename=%s" % (
                     disposition,
-                    media_info.upload_name,
+                    upload_name,
                 )
             else:
                 disposition = "%s; filename*=utf-8''%s" % (
                     disposition,
-                    _quote(media_info.upload_name),
+                    _quote(upload_name),
                 )
 
         from synapse.media.media_storage import MultipartFileConsumer
@@ -341,14 +341,14 @@ async def respond_with_multipart_responder(
         multipart_consumer = MultipartFileConsumer(
             clock,
             request,
-            media_info.media_type,
+            media_type,
             {},
             disposition,
-            media_info.media_length,
+            media_length,
         )
 
         logger.debug("Responding to media request with responder %s", responder)
-        if media_info.media_length is not None:
+        if media_length is not None:
             content_length = multipart_consumer.content_length()
             assert content_length is not None
             request.setHeader(b"Content-Length", b"%d" % (content_length,))

synapse/media/media_repository.py

@@ -471,7 +471,7 @@ class MediaRepository:
         responder = await self.media_storage.fetch_media(file_info)
         if federation:
             await respond_with_multipart_responder(
-                self.clock, request, responder, media_info
+                self.clock, request, responder, media_type, media_length, upload_name
             )
         else:
             await respond_with_responder(
@@ -1008,7 +1008,7 @@ class MediaRepository:
         t_method: str,
         t_type: str,
         url_cache: bool,
-    ) -> Optional[str]:
+    ) -> Optional[Tuple[str, FileInfo]]:
         input_path = await self.media_storage.ensure_media_is_in_local_cache(
             FileInfo(None, media_id, url_cache=url_cache)
         )
@@ -1070,7 +1070,7 @@ class MediaRepository:
                 t_len,
             )
 
-            return output_path
+            return output_path, file_info
 
         # Could not generate thumbnail.
         return None

synapse/media/thumbnailer.py

@@ -348,7 +348,12 @@ class ThumbnailProvider:
                 if responder:
                     if for_federation:
                         await respond_with_multipart_responder(
-                            self.hs.get_clock(), request, responder, media_info
+                            self.hs.get_clock(),
+                            request,
+                            responder,
+                            info.type,
+                            info.length,
+                            None,
                         )
                         return
                     else:
@@ -360,7 +365,7 @@ class ThumbnailProvider:
         logger.debug("We don't have a thumbnail of that size. Generating")
 
         # Okay, so we generate one.
-        file_path = await self.media_repo.generate_local_exact_thumbnail(
+        thumbnail_result = await self.media_repo.generate_local_exact_thumbnail(
             media_id,
             desired_width,
             desired_height,
@@ -369,13 +374,18 @@ class ThumbnailProvider:
             url_cache=bool(media_info.url_cache),
         )
 
-        if file_path:
+        if thumbnail_result:
+            file_path, file_info = thumbnail_result
+            assert file_info.thumbnail is not None
+
             if for_federation:
                 await respond_with_multipart_responder(
                     self.hs.get_clock(),
                     request,
                     FileResponder(self.hs, open(file_path, "rb")),
-                    media_info,
+                    file_info.thumbnail.type,
+                    file_info.thumbnail.length,
+                    None,
                 )
             else:
                 await respond_with_file(self.hs, request, desired_type, file_path)
@@ -580,7 +590,12 @@ class ThumbnailProvider:
                 if for_federation:
                     assert media_info is not None
                     await respond_with_multipart_responder(
-                        self.hs.get_clock(), request, responder, media_info
+                        self.hs.get_clock(),
+                        request,
+                        responder,
+                        file_info.thumbnail.type,
+                        file_info.thumbnail.length,
+                        None,
                     )
                     return
                 else:
@@ -634,7 +649,12 @@ class ThumbnailProvider:
             if for_federation:
                 assert media_info is not None
                 await respond_with_multipart_responder(
-                    self.hs.get_clock(), request, responder, media_info
+                    self.hs.get_clock(),
+                    request,
+                    responder,
+                    file_info.thumbnail.type,
+                    file_info.thumbnail.length,
+                    None,
                 )
             else:
                 await respond_with_responder(

synapse/metrics/background_process_metrics.py

@@ -293,7 +293,7 @@ def wrap_as_background_process(
     """
 
     def wrap_as_background_process_inner(
-        func: Callable[P, Awaitable[Optional[R]]]
+        func: Callable[P, Awaitable[Optional[R]]],
     ) -> Callable[P, "defer.Deferred[Optional[R]]"]:
         @wraps(func)
         def wrap_as_background_process_inner_2(

synapse/notifier.py

@@ -41,6 +41,7 @@ import attr
 from prometheus_client import Counter
 
 from twisted.internet import defer
+from twisted.internet.defer import Deferred
 
 from synapse.api.constants import EduTypes, EventTypes, HistoryVisibility, Membership
 from synapse.api.errors import AuthError
@@ -52,6 +53,7 @@ from synapse.logging.opentracing import log_kv, start_active_span
 from synapse.metrics import LaterGauge
 from synapse.streams.config import PaginationConfig
 from synapse.types import (
+    ISynapseReactor,
     JsonDict,
     MultiWriterStreamToken,
     PersistedEventPosition,
@@ -61,8 +63,11 @@ from synapse.types import (
     StreamToken,
     UserID,
 )
-from synapse.util.async_helpers import ObservableDeferred, timeout_deferred
+from synapse.util.async_helpers import (
+    timeout_deferred,
+)
 from synapse.util.metrics import Measure
+from synapse.util.stringutils import shortstr
 from synapse.visibility import filter_events_for_client
 
 if TYPE_CHECKING:
@@ -89,18 +94,6 @@ def count(func: Callable[[T], bool], it: Iterable[T]) -> int:
     return n
 
 
-class _NotificationListener:
-    """This represents a single client connection to the events stream.
-    The events stream handler will have yielded to the deferred, so to
-    notify the handler it is sufficient to resolve the deferred.
-    """
-
-    __slots__ = ["deferred"]
-
-    def __init__(self, deferred: "defer.Deferred"):
-        self.deferred = deferred
-
-
 class _NotifierUserStream:
     """This represents a user connected to the event stream.
     It tracks the most recent stream token for that user.
@@ -113,59 +106,49 @@ class _NotifierUserStream:
 
     def __init__(
         self,
+        reactor: ISynapseReactor,
         user_id: str,
         rooms: StrCollection,
         current_token: StreamToken,
         time_now_ms: int,
     ):
+        self.reactor = reactor
         self.user_id = user_id
         self.rooms = set(rooms)
-        self.current_token = current_token
 
         # The last token for which we should wake up any streams that have a
         # token that comes before it. This gets updated every time we get poked.
         # We start it at the current token since if we get any streams
         # that have a token from before we have no idea whether they should be
         # woken up or not, so lets just wake them up.
-        self.last_notified_token = current_token
+        self.current_token = current_token
         self.last_notified_ms = time_now_ms
 
-        self.notify_deferred: ObservableDeferred[StreamToken] = ObservableDeferred(
-            defer.Deferred()
-        )
+        # Set of listeners that we need to wake up when there has been a change.
+        self.listeners: Set[Deferred[StreamToken]] = set()
 
-    def notify(
+    def update_and_fetch_deferreds(
         self,
-        stream_key: StreamKeyType,
-        stream_id: Union[int, RoomStreamToken, MultiWriterStreamToken],
+        current_token: StreamToken,
         time_now_ms: int,
-    ) -> None:
-        """Notify any listeners for this user of a new event from an
-        event source.
+    ) -> Collection["Deferred[StreamToken]"]:
+        """Update the stream for this user because of a new event from an
+        event source, and return the set of deferreds to wake up.
+
         Args:
-            stream_key: The stream the event came from.
-            stream_id: The new id for the stream the event came from.
+            current_token: The new current token.
             time_now_ms: The current time in milliseconds.
+
+        Returns:
+            The set of deferreds that need to be called.
         """
-        self.current_token = self.current_token.copy_and_advance(stream_key, stream_id)
-        self.last_notified_token = self.current_token
+        self.current_token = current_token
         self.last_notified_ms = time_now_ms
-        notify_deferred = self.notify_deferred
 
-        log_kv(
-            {
-                "notify": self.user_id,
-                "stream": stream_key,
-                "stream_id": stream_id,
-                "listeners": self.count_listeners(),
-            }
-        )
+        listeners = self.listeners
+        self.listeners = set()
 
-        users_woken_by_stream_counter.labels(stream_key).inc()
-
-        with PreserveLoggingContext():
-            self.notify_deferred = ObservableDeferred(defer.Deferred())
-            notify_deferred.callback(self.current_token)
+        return listeners
 
     def remove(self, notifier: "Notifier") -> None:
         """Remove this listener from all the indexes in the Notifier
@@ -179,9 +162,9 @@ class _NotifierUserStream:
         notifier.user_to_user_stream.pop(self.user_id)
 
     def count_listeners(self) -> int:
-        return len(self.notify_deferred.observers())
+        return len(self.listeners)
 
-    def new_listener(self, token: StreamToken) -> _NotificationListener:
+    def new_listener(self, token: StreamToken) -> "Deferred[StreamToken]":
         """Returns a deferred that is resolved when there is a new token
         greater than the given token.
 
@@ -191,10 +174,17 @@ class _NotifierUserStream:
         """
         # Immediately wake up stream if something has already since happened
         # since their last token.
-        if self.last_notified_token != token:
-            return _NotificationListener(defer.succeed(self.current_token))
-        else:
-            return _NotificationListener(self.notify_deferred.observe())
+        if token != self.current_token:
+            return defer.succeed(self.current_token)
+
+        # Create a new deferred and add it to the set of listeners. We add a
+        # cancel handler to remove it from the set again, to handle timeouts.
+        deferred: "Deferred[StreamToken]" = Deferred(
+            canceller=lambda d: self.listeners.discard(d)
+        )
+        self.listeners.add(deferred)
+
+        return deferred
 
 
 @attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -247,6 +237,7 @@ class Notifier:
         # List of callbacks to be notified when a lock is released
         self._lock_released_callback: List[Callable[[str, str, str], None]] = []
 
+        self.reactor = hs.get_reactor()
         self.clock = hs.get_clock()
         self.appservice_handler = hs.get_application_service_handler()
         self._pusher_pool = hs.get_pusherpool()
@@ -342,14 +333,25 @@ class Notifier:
         # Wake up all related user stream notifiers
         user_streams = self.room_to_user_streams.get(room_id, set())
         time_now_ms = self.clock.time_msec()
+        current_token = self.event_sources.get_current_token()
+
+        listeners: List["Deferred[StreamToken]"] = []
         for user_stream in user_streams:
             try:
-                user_stream.notify(
-                    StreamKeyType.UN_PARTIAL_STATED_ROOMS, new_token, time_now_ms
+                listeners.extend(
+                    user_stream.update_and_fetch_deferreds(current_token, time_now_ms)
                 )
             except Exception:
                 logger.exception("Failed to notify listener")
 
+        with PreserveLoggingContext():
+            for listener in listeners:
+                listener.callback(current_token)
+
+        users_woken_by_stream_counter.labels(StreamKeyType.UN_PARTIAL_STATED_ROOMS).inc(
+            len(user_streams)
+        )
+
         # Poke the replication so that other workers also see the write to
         # the un-partial-stated rooms stream.
         self.notify_replication()
@@ -519,12 +521,16 @@ class Notifier:
         rooms = rooms or []
 
         with Measure(self.clock, "on_new_event"):
-            user_streams = set()
+            user_streams: Set[_NotifierUserStream] = set()
 
             log_kv(
                 {
                     "waking_up_explicit_users": len(users),
                     "waking_up_explicit_rooms": len(rooms),
+                    "users": shortstr(users),
+                    "rooms": shortstr(rooms),
+                    "stream": stream_key,
+                    "stream_id": new_token,
                 }
             )
 
@@ -544,12 +550,27 @@ class Notifier:
                 )
 
             time_now_ms = self.clock.time_msec()
+            current_token = self.event_sources.get_current_token()
+            listeners: List["Deferred[StreamToken]"] = []
             for user_stream in user_streams:
                 try:
-                    user_stream.notify(stream_key, new_token, time_now_ms)
+                    listeners.extend(
+                        user_stream.update_and_fetch_deferreds(
+                            current_token, time_now_ms
+                        )
+                    )
                 except Exception:
                     logger.exception("Failed to notify listener")
 
+            # We resolve all these deferreds in one go so that we only need to
+            # call `PreserveLoggingContext` once, as it has a bunch of overhead
+            # (to calculate performance stats)
+            with PreserveLoggingContext():
+                for listener in listeners:
+                    listener.callback(current_token)
+
+            users_woken_by_stream_counter.labels(stream_key).inc(len(user_streams))
+
             self.notify_replication()
 
             # Notify appservices.
@@ -586,6 +607,7 @@ class Notifier:
             if room_ids is None:
                 room_ids = await self.store.get_rooms_for_user(user_id)
             user_stream = _NotifierUserStream(
+                reactor=self.reactor,
                 user_id=user_id,
                 rooms=room_ids,
                 current_token=current_token,
@@ -608,8 +630,8 @@ class Notifier:
                         # Now we wait for the _NotifierUserStream to be told there
                         # is a new token.
                         listener = user_stream.new_listener(prev_token)
-                        listener.deferred = timeout_deferred(
-                            listener.deferred,
+                        listener = timeout_deferred(
+                            listener,
                             (end_time - now) / 1000.0,
                             self.hs.get_reactor(),
                         )
@@ -622,7 +644,7 @@ class Notifier:
                         )
 
                         with PreserveLoggingContext():
-                            await listener.deferred
+                            await listener
 
                         log_kv(
                             {