Uniformize spam-checker API:

- Some callbacks should return `True` to allow, `False` to deny, while others should return `True` to deny and `False` to allow. With this PR, all callbacks return `ALLOW` to allow or a `Codes` (typically `Codes.FORBIDDEN`) to deny.
- Similarly, some methods returned `True` to allow, `False` to deny, while others returned `True` to deny and `False` to allow. They now all return `ALLOW` to allow or a `Codes` to deny.
- Spam-checker implementations may now return an explicit code, e.g. to differentiate between "User account has been suspended" (which is in practice required by law in some countries, including UK) and "This message looks like spam".

CHANGES.md

@@ -1,8 +1,105 @@
-Synapse 1.59.0
-==============
+Synapse 1.59.0rc1 (2022-05-10)
+==============================
 
-The non-standard `m.login.jwt` login type has been removed from Synapse. It can be replaced with `org.matrix.login.jwt` for identical behaviour. This is only used if `jwt_config.enabled` is set to `true` in the configuration.
+This release makes several changes that server administrators should be aware of:
 
+- Device name lookup over federation is now disabled by default. ([\#12616](https://github.com/matrix-org/synapse/issues/12616))
+- The `synapse.app.appservice` and `synapse.app.user_dir` worker application types are now deprecated. ([\#12452](https://github.com/matrix-org/synapse/issues/12452), [\#12654](https://github.com/matrix-org/synapse/issues/12654))
+
+See [the upgrade notes](https://github.com/matrix-org/synapse/blob/develop/docs/upgrade.md#upgrading-to-v1590) for more details.
+
+Additionally, this release removes the non-standard `m.login.jwt` login type from Synapse. It can be replaced with `org.matrix.login.jwt` for identical behaviour. This is only used if `jwt_config.enabled` is set to `true` in the configuration. ([\#12597](https://github.com/matrix-org/synapse/issues/12597))
+
+Features
+--------
+
+- Support [MSC3266](https://github.com/matrix-org/matrix-doc/pull/3266) room summaries over federation. ([\#11507](https://github.com/matrix-org/synapse/issues/11507))
+- Implement [changes](https://github.com/matrix-org/matrix-spec-proposals/pull/2285/commits/4a77139249c2e830aec3c7d6bd5501a514d1cc27) to [MSC2285 (hidden read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). Contributed by @SimonBrandner. ([\#12168](https://github.com/matrix-org/synapse/issues/12168), [\#12635](https://github.com/matrix-org/synapse/issues/12635), [\#12636](https://github.com/matrix-org/synapse/issues/12636), [\#12670](https://github.com/matrix-org/synapse/issues/12670))
+- Extend the [module API](https://github.com/matrix-org/synapse/blob/release-v1.59/synapse/module_api/__init__.py) to allow modules to change actions for existing push rules of local users. ([\#12406](https://github.com/matrix-org/synapse/issues/12406))
+- Add the `notify_appservices_from_worker` configuration option (superseding `notify_appservices`) to allow a generic worker to be designated as the worker to send traffic to Application Services. ([\#12452](https://github.com/matrix-org/synapse/issues/12452))
+- Add the `update_user_directory_from_worker` configuration option (superseding `update_user_directory`) to allow a generic worker to be designated as the worker to update the user directory. ([\#12654](https://github.com/matrix-org/synapse/issues/12654))
+- Add new `enable_registration_token_3pid_bypass` configuration option to allow registrations via token as an alternative to verifying a 3pid. ([\#12526](https://github.com/matrix-org/synapse/issues/12526))
+- Implement [MSC3786](https://github.com/matrix-org/matrix-spec-proposals/pull/3786): Add a default push rule to ignore `m.room.server_acl` events. ([\#12601](https://github.com/matrix-org/synapse/issues/12601))
+- Add new `mau_appservice_trial_days` configuration option to specify a different trial period for users registered via an appservice. ([\#12619](https://github.com/matrix-org/synapse/issues/12619))
+
+
+Bugfixes
+--------
+
+- Fix a bug introduced in Synapse 1.48.0 where the latest thread reply provided failed to include the proper bundled aggregations. ([\#12273](https://github.com/matrix-org/synapse/issues/12273))
+- Fix a bug introduced in Synapse 1.22.0 where attempting to send a large amount of read receipts to an application service all at once would result in duplicate content and abnormally high memory usage. Contributed by Brad & Nick @ Beeper. ([\#12544](https://github.com/matrix-org/synapse/issues/12544))
+- Fix a bug introduced in Synapse 1.57.0 which could cause `Failed to calculate hosts in room` errors to be logged for outbound federation. ([\#12570](https://github.com/matrix-org/synapse/issues/12570))
+- Fix a long-standing bug where status codes would almost always get logged as `200!`, irrespective of the actual status code, when clients disconnect before a request has finished processing. ([\#12580](https://github.com/matrix-org/synapse/issues/12580))
+- Fix race when persisting an event and deleting a room that could lead to outbound federation breaking. ([\#12594](https://github.com/matrix-org/synapse/issues/12594))
+- Fix a bug introduced in Synapse 1.53.0 where bundled aggregations for annotations/edits were incorrectly calculated. ([\#12633](https://github.com/matrix-org/synapse/issues/12633))
+- Fix a long-standing bug where rooms containing power levels with string values could not be upgraded. ([\#12657](https://github.com/matrix-org/synapse/issues/12657))
+- Prevent memory leak from reoccurring when presence is disabled. ([\#12656](https://github.com/matrix-org/synapse/issues/12656))
+
+
+Updates to the Docker image
+---------------------------
+
+- Explicitly opt-in to using [BuildKit-specific features](https://github.com/moby/buildkit/blob/master/frontend/dockerfile/docs/syntax.md) in the Dockerfile. This fixes issues with building images in some GitLab CI environments. ([\#12541](https://github.com/matrix-org/synapse/issues/12541))
+- Update the "Build docker images" GitHub Actions workflow to use `docker/metadata-action` to generate docker image tags, instead of a custom shell script. Contributed by @henryclw. ([\#12573](https://github.com/matrix-org/synapse/issues/12573))
+
+
+Improved Documentation
+----------------------
+
+- Update SQL statements and replace use of old table `user_stats_historical` in docs for Synapse Admins. ([\#12536](https://github.com/matrix-org/synapse/issues/12536))
+- Add missing linebreak to `pipx` install instructions. ([\#12579](https://github.com/matrix-org/synapse/issues/12579))
+- Add information about the TCP replication module to docs. ([\#12621](https://github.com/matrix-org/synapse/issues/12621))
+- Fixes to the formatting of `README.rst`. ([\#12627](https://github.com/matrix-org/synapse/issues/12627))
+- Fix docs on how to run specific Complement tests using the `complement.sh` test runner. ([\#12664](https://github.com/matrix-org/synapse/issues/12664))
+
+
+Deprecations and Removals
+-------------------------
+
+- Remove unstable identifiers from [MSC3069](https://github.com/matrix-org/matrix-doc/pull/3069). ([\#12596](https://github.com/matrix-org/synapse/issues/12596))
+- Remove the unspecified `m.login.jwt` login type and the unstable `uk.half-shot.msc2778.login.application_service` from 
+  [MSC2778](https://github.com/matrix-org/matrix-doc/pull/2778). ([\#12597](https://github.com/matrix-org/synapse/issues/12597))
+- Synapse now requires at least Python 3.7.1 (up from 3.7.0), for compatibility with the latest Twisted trunk. ([\#12613](https://github.com/matrix-org/synapse/issues/12613))
+
+
+Internal Changes
+----------------
+
+- Use supervisord to supervise Postgres and Caddy in the Complement image to reduce restart time. ([\#12480](https://github.com/matrix-org/synapse/issues/12480))
+- Immediately retry any requests that have backed off when a server comes back online. ([\#12500](https://github.com/matrix-org/synapse/issues/12500))
+- Use `make_awaitable` instead of `defer.succeed` for return values of mocks in tests. ([\#12505](https://github.com/matrix-org/synapse/issues/12505))
+- Consistently check if an object is a `frozendict`. ([\#12564](https://github.com/matrix-org/synapse/issues/12564))
+- Protect module callbacks with read semantics against cancellation. ([\#12568](https://github.com/matrix-org/synapse/issues/12568))
+- Improve comments and error messages around access tokens. ([\#12577](https://github.com/matrix-org/synapse/issues/12577))
+- Improve docstrings for the receipts store. ([\#12581](https://github.com/matrix-org/synapse/issues/12581))
+- Use constants for read-receipts in tests. ([\#12582](https://github.com/matrix-org/synapse/issues/12582))
+- Log status code of cancelled requests as 499 and avoid logging stack traces for them. ([\#12587](https://github.com/matrix-org/synapse/issues/12587), [\#12663](https://github.com/matrix-org/synapse/issues/12663))
+- Remove special-case for `twisted` logger from default log config. ([\#12589](https://github.com/matrix-org/synapse/issues/12589))
+- Use `getClientAddress` instead of the deprecated `getClientIP`. ([\#12599](https://github.com/matrix-org/synapse/issues/12599))
+- Add link to documentation in Grafana Dashboard. ([\#12602](https://github.com/matrix-org/synapse/issues/12602))
+- Reduce log spam when running multiple event persisters. ([\#12610](https://github.com/matrix-org/synapse/issues/12610))
+- Add extra debug logging to federation sender. ([\#12614](https://github.com/matrix-org/synapse/issues/12614))
+- Prevent remote homeservers from requesting local user device names by default. ([\#12616](https://github.com/matrix-org/synapse/issues/12616))
+- Add a consistency check on events which we read from the database. ([\#12620](https://github.com/matrix-org/synapse/issues/12620))
+- Remove use of the `constantly` library and switch to enums for `EventRedactBehaviour`. Contributed by @andrewdoh. ([\#12624](https://github.com/matrix-org/synapse/issues/12624))
+- Remove unused code related to receipts. ([\#12632](https://github.com/matrix-org/synapse/issues/12632))
+- Minor improvements to the scripts for running Synapse in worker mode under Complement. ([\#12637](https://github.com/matrix-org/synapse/issues/12637))
+- Move `pympler` back in to the `all` extras. ([\#12652](https://github.com/matrix-org/synapse/issues/12652))
+- Fix spelling of `M_UNRECOGNIZED` in comments. ([\#12665](https://github.com/matrix-org/synapse/issues/12665))
+- Release script: confirm the commit to be tagged before tagging. ([\#12556](https://github.com/matrix-org/synapse/issues/12556))
+- Fix a typo in the announcement text generated by the Synapse release development script. ([\#12612](https://github.com/matrix-org/synapse/issues/12612))
+
+### Typechecking
+
+- Fix scripts-dev to pass typechecking. ([\#12356](https://github.com/matrix-org/synapse/issues/12356))
+- Add some type hints to datastore. ([\#12485](https://github.com/matrix-org/synapse/issues/12485))
+- Remove unused `# type: ignore`s. ([\#12531](https://github.com/matrix-org/synapse/issues/12531))
+- Allow unused `# type: ignore` comments in bleeding edge CI jobs. ([\#12576](https://github.com/matrix-org/synapse/issues/12576))
+- Remove redundant lines of config from `mypy.ini`. ([\#12608](https://github.com/matrix-org/synapse/issues/12608))
+- Update to mypy 0.950. ([\#12650](https://github.com/matrix-org/synapse/issues/12650))
+- Use `Concatenate` to better annotate `_do_execute`. ([\#12666](https://github.com/matrix-org/synapse/issues/12666))
+- Use `ParamSpec` to refine type hints. ([\#12667](https://github.com/matrix-org/synapse/issues/12667))
+- Fix mypy against latest pillow stubs. ([\#12671](https://github.com/matrix-org/synapse/issues/12671))
 
 Synapse 1.58.1 (2022-05-05)
 ===========================

changelog.d/11507.feature

@@ -1 +0,0 @@
-Support [MSC3266](https://github.com/matrix-org/matrix-doc/pull/3266) room summaries over federation.

changelog.d/12168.feature

@@ -1 +0,0 @@
-Implement [changes](https://github.com/matrix-org/matrix-spec-proposals/pull/2285/commits/4a77139249c2e830aec3c7d6bd5501a514d1cc27) to [MSC2285 (hidden read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). Contributed by @SimonBrandner.

changelog.d/12273.bugfix

@@ -1 +0,0 @@
-Fix a bug introduced in Synapse v1.48.0 where latest thread reply provided failed to include the proper bundled aggregations.

changelog.d/12356.misc

@@ -1 +0,0 @@
-Fix scripts-dev to pass typechecking.

changelog.d/12406.feature

@@ -1 +0,0 @@
-Add a module API to allow modules to change actions for existing push rules of local users.

changelog.d/12452.feature

@@ -1 +0,0 @@
-Add the `notify_appservices_from_worker` configuration option (superseding `notify_appservices`) to allow a generic worker to be designated as the worker to send traffic to Application Services.

changelog.d/12480.misc

@@ -1 +0,0 @@
-Use supervisord to supervise Postgres and Caddy in the Complement image to reduce restart time.

changelog.d/12505.misc

@@ -1 +0,0 @@
-Use `make_awaitable` instead of `defer.succeed` for return values of mocks in tests.

changelog.d/12526.feature

@@ -1 +0,0 @@
-Add new `enable_registration_token_3pid_bypass` configuration option to allow registrations via token as an alternative to verifying a 3pid.

changelog.d/12531.misc

@@ -1 +0,0 @@
-Remove unused `# type: ignore`s.

changelog.d/12541.docker

@@ -1 +0,0 @@
-Explicitly opt-in to using [BuildKit-specific features](https://github.com/moby/buildkit/blob/master/frontend/dockerfile/docs/syntax.md) in the Dockerfile. This fixes issues with building images in some GitLab CI environments.

changelog.d/12544.bugfix

@@ -1 +0,0 @@
-Fix a bug where attempting to send a large amount of read receipts to an application service all at once would result in duplicate content and abnormally high memory usage. Contributed by Brad & Nick @ Beeper.

changelog.d/12556.misc

@@ -1 +0,0 @@
-Release script: confirm the commit to be tagged before tagging.

changelog.d/12564.misc

@@ -1 +0,0 @@
-Consistently check if an object is a `frozendict`.

changelog.d/12568.misc

@@ -1 +0,0 @@
-Protect module callbacks with read semantics against cancellation.

changelog.d/12570.bugfix

@@ -1 +0,0 @@
-Fix a bug introduced in Synapse 1.57 which could cause `Failed to calculate hosts in room` errors to be logged for outbound federation.

changelog.d/12573.docker

@@ -1 +0,0 @@
-Update the "Build docker images" GitHub Actions workflow to use `docker/metadata-action` to generate docker image tags, instead of a custom shell script. Contributed by henryclw.

changelog.d/12576.misc

@@ -1 +0,0 @@
-Allow unused `#type: ignore` comments in bleeding edge CI jobs.

changelog.d/12577.misc

@@ -1 +0,0 @@
-Improve comments and error messages around access tokens.

changelog.d/12579.doc

@@ -1 +0,0 @@
-Add missing linebreak to pipx install instructions.

changelog.d/12580.bugfix

@@ -1 +0,0 @@
-Fix a long standing bug where status codes would almost always get logged as 200!, irrespective of the actual status code, when clients disconnect before a request has finished processing.

changelog.d/12581.misc

@@ -1 +0,0 @@
-Improve docstrings for the receipts store.

changelog.d/12582.misc

@@ -1 +0,0 @@
-Use constants for read-receipts in tests.

changelog.d/12586.misc

@@ -0,0 +1 @@
+Add `@cancellable` decorator, for use on endpoint methods that can be cancelled when clients disconnect.

changelog.d/12587.misc

@@ -1 +0,0 @@
-Log status code of cancelled requests as 499 and avoid logging stack traces for them.

changelog.d/12588.misc

@@ -0,0 +1 @@
+Add ability to cancel disconnected requests to `SynapseRequest`.

changelog.d/12589.misc

@@ -1 +0,0 @@
-Remove special-case for `twisted` logger from default log config.

changelog.d/12594.bugfix

@@ -1 +0,0 @@
-Fix race when persisting an event and deleting a room that could lead to outbound federation breaking.

changelog.d/12596.removal

@@ -1 +0,0 @@
-Remove unstable identifiers from [MSC3069](https://github.com/matrix-org/matrix-doc/pull/3069).

changelog.d/12597.removal

@@ -1,2 +0,0 @@
-Remove the unspecified `m.login.jwt` login type and the unstable `uk.half-shot.msc2778.login.application_service` from 
-[MSC2778](https://github.com/matrix-org/matrix-doc/pull/2778).

changelog.d/12599.misc

@@ -1 +0,0 @@
-Use `getClientAddress` instead of the deprecated `getClientIP`.

changelog.d/12602.misc

@@ -1 +0,0 @@
-Add link to documentation in Grafana Dashboard.

changelog.d/12608.misc

@@ -1 +0,0 @@
-Remove redundant lines of config from `mypy.ini`.

changelog.d/12610.misc

@@ -1 +0,0 @@
-Reduce log spam when running multiple event persisters.

changelog.d/12612.bugfix

@@ -1 +0,0 @@
-Fix a typo in the announcement text generated by the Synapse release development script.

changelog.d/12613.removal

@@ -1 +0,0 @@
-Synapse now requires at least Python 3.7.1 (up from 3.7.0), for compatibility with the latest Twisted trunk.

changelog.d/12614.misc

@@ -1 +0,0 @@
-Add extra debug logging to federation sender.

changelog.d/12616.misc

@@ -1 +0,0 @@
-Prevent remote homeservers from requesting local user device names by default.

changelog.d/12619.feature

@@ -1 +0,0 @@
-Add new `mau_appservice_trial_days` configuration option to specify a different trial period for users registered via an appservice.

changelog.d/12620.misc

@@ -1 +0,0 @@
-Add a consistency check on events which we read from the database.

changelog.d/12624.misc

@@ -1 +0,0 @@
-Remove use of constantly library and switch to enums for EventRedactBehaviour. Contributed by @andrewdoh.

changelog.d/12627.doc

@@ -1 +0,0 @@
-Fixes to the formatting of README.rst.

changelog.d/12630.misc

@@ -0,0 +1 @@
+Add a helper class for testing request cancellation.

changelog.d/12632.misc

@@ -1 +0,0 @@
-Remove unused code related to receipts.

changelog.d/12633.bugfix

@@ -1 +0,0 @@
-Fix a bug introduced in Synapse v1.53.0 where bundled aggregations for annotations/edits were incorrectly calculated.

changelog.d/12635.feature

@@ -1 +0,0 @@
-Implement [changes](https://github.com/matrix-org/matrix-spec-proposals/pull/2285/commits/4a77139249c2e830aec3c7d6bd5501a514d1cc27) to [MSC2285 (hidden read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). Contributed by @SimonBrandner.

changelog.d/12636.feature

@@ -1 +0,0 @@
-Implement [changes](https://github.com/matrix-org/matrix-spec-proposals/pull/2285/commits/4a77139249c2e830aec3c7d6bd5501a514d1cc27) to [MSC2285 (hidden read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). Contributed by @SimonBrandner.

changelog.d/12639.bugfix

@@ -1 +0,0 @@
-Add new `enable_registration_token_3pid_bypass` configuration option to allow registrations via token as an alternative to verifying a 3pid.

changelog.d/12650.misc

@@ -1 +0,0 @@
-Update to mypy 0.950.

changelog.d/12652.misc

@@ -1 +0,0 @@
-Move `pympler` back in to the `all` extras.

changelog.d/12656.misc

@@ -1 +0,0 @@
-Prevent memory leak from reoccurring when presence is disabled.

changelog.d/12657.bugfix

@@ -1 +0,0 @@
-Fix a long-standing bug where rooms containing power levels with string values could not be upgraded.

changelog.d/12663.misc

@@ -1 +0,0 @@
-Log status code of cancelled requests as 499 and avoid logging stack traces for them.

changelog.d/12664.doc

@@ -1 +0,0 @@
-Fix docs on how to run specific Complement tests using the `complement.sh` test runner.

changelog.d/12665.misc

@@ -1 +0,0 @@
-Fix spelling of `M_UNRECOGNIZED` in comments.

changelog.d/12666.misc

@@ -1 +0,0 @@
-Use `Concatenate` to better annotate `_do_execute`.

changelog.d/12667.misc

@@ -1 +0,0 @@
-Use `ParamSpec` to refine type hints.

changelog.d/12670.feature

@@ -1 +0,0 @@
-Implement [changes](https://github.com/matrix-org/matrix-spec-proposals/pull/2285/commits/4a77139249c2e830aec3c7d6bd5501a514d1cc27) to [MSC2285 (hidden read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). Contributed by @SimonBrandner.

changelog.d/12671.misc

@@ -1 +0,0 @@
-Fix mypy against latest pillow stubs.

changelog.d/12676.misc

@@ -0,0 +1 @@
+Improve documentation of the `synapse.push` module.

changelog.d/12677.misc

@@ -0,0 +1 @@
+Refactor functions to on `PushRuleEvaluatorForEvent`.

changelog.d/12679.misc

@@ -0,0 +1 @@
+Preparation for database schema simplifications: stop writing to `event_reference_hashes`.

changelog.d/12683.bugfix

@@ -0,0 +1 @@
+Fix a bug introduced in Synapse 1.57.0 where `/messages` would throw a 500 error when querying for a non-existent room.

changelog.d/12689.misc

@@ -0,0 +1 @@
+Refactor `EventContext` class.

changelog.d/12694.misc

@@ -0,0 +1 @@
+Capture the `Deferred` for request cancellation in `_AsyncResource`.

changelog.d/12695.misc

@@ -0,0 +1 @@
+Fixes an incorrect type hint for `Filter._check_event_relations`.

debian/changelog

@@ -1,10 +1,11 @@
-matrix-synapse-py3 (1.58.2) UNRELEASED; urgency=medium
+matrix-synapse-py3 (1.59.0~rc1) stable; urgency=medium
 
   * Adjust how the `exported-requirements.txt` file is generated as part of
     the process of building these packages. This affects the package
     maintainers only; end-users are unaffected.
+  * New Synapse release 1.59.0rc1.
 
- -- Synapse Packaging team <packages@matrix.org>  Fri, 06 May 2022 13:49:29 +0100
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 10 May 2022 10:45:08 +0100
 
 matrix-synapse-py3 (1.58.1) stable; urgency=medium
 

docker/conf-workers/supervisord.conf.j2

@@ -9,7 +9,7 @@ user=root
 files = /etc/supervisor/conf.d/*.conf
 
 [program:nginx]
-command=/usr/sbin/nginx -g "daemon off;"
+command=/usr/local/bin/prefix-log /usr/sbin/nginx -g "daemon off;"
 priority=500
 stdout_logfile=/dev/stdout
 stdout_logfile_maxbytes=0
@@ -19,7 +19,7 @@ username=www-data
 autorestart=true
 
 [program:redis]
-command=/usr/bin/redis-server /etc/redis/redis.conf --daemonize no
+command=/usr/local/bin/prefix-log /usr/bin/redis-server /etc/redis/redis.conf --daemonize no
 priority=1
 stdout_logfile=/dev/stdout
 stdout_logfile_maxbytes=0
@@ -29,7 +29,7 @@ username=redis
 autorestart=true
 
 [program:synapse_main]
-command=/usr/local/bin/python -m synapse.app.homeserver --config-path="{{ main_config_path }}" --config-path=/conf/workers/shared.yaml
+command=/usr/local/bin/prefix-log /usr/local/bin/python -m synapse.app.homeserver --config-path="{{ main_config_path }}" --config-path=/conf/workers/shared.yaml
 priority=10
 # Log startup failures to supervisord's stdout/err
 # Regular synapse logs will still go in the configured data directory

docs/modules/spam_checker_callbacks.md

@@ -18,6 +18,17 @@ async def check_event_for_spam(event: "synapse.events.EventBase") -> Union[bool,
 
 Called when receiving an event from a client or via federation. The callback must return
 either:
+  - on `Decision.ALLOW`, the action is permitted.
+  - on `Decision.DENY`, the action is rejected with a default error message/code.
+  - on `Codes`, the action is rejected with a specific error message/code. In case
+      of doubt, use `Codes.FORBIDDEN`.
+  - (deprecated) on `False`, behave as `Decision.ALLOW`. Deprecated as methods in
+      this API are inconsistent, some expect `True` for `ALLOW` and others `True`
+      for `DENY`.
+  - (deprecated) on `True`, behave as `Decision.DENY`. Deprecated as methods in
+      this API are inconsistent, some expect `True` for `ALLOW` and others `True`
+      for `DENY`.
+
 - an error message string, to indicate the event must be rejected because of spam and 
   give a rejection reason to forward to clients;
 - the boolean `True`, to indicate that the event is spammy, but not provide further details; or

docs/replication.md

@@ -35,3 +35,8 @@ See [the TCP replication documentation](tcp_replication.md).
 There are read-only version of the synapse storage layer in
 `synapse/replication/slave/storage` that use the response of the
 replication API to invalidate their caches.
+
+### The TCP Replication Module
+Information about how the tcp replication module is structured, including how
+the classes interact, can be found in
+`synapse/replication/tcp/__init__.py`

docs/upgrade.md

@@ -101,29 +101,36 @@ To re-enable this functionality, set the
 homeserver config option to `true`.
 
 
-## Deprecation of the `synapse.app.appservice` worker application type
+## Deprecation of the `synapse.app.appservice` and `synapse.app.user_dir` worker application types
 
 The `synapse.app.appservice` worker application type allowed you to configure a
 single worker to use to notify application services of new events, as long
 as this functionality was disabled on the main process with `notify_appservices: False`.
+Further, the `synapse.app.user_dir` worker application type allowed you to configure
+a single worker to be responsible for updating the user directory, as long as this
+was disabled on the main process with `update_user_directory: False`.
 
 To unify Synapse's worker types, the `synapse.app.appservice` worker application
 type and the `notify_appservices` configuration option have been deprecated.
+The `synapse.app.user_dir` worker application type and `update_user_directory`
+configuration option have also been deprecated.
 
-To get the same functionality, it's now recommended that the `synapse.app.generic_worker`
-worker application type is used and that the `notify_appservices_from_worker` option
-is set to the name of a worker.
+To get the same functionality as was provided by the deprecated options, it's now recommended that the `synapse.app.generic_worker`
+worker application type is used and that the `notify_appservices_from_worker` and/or
+`update_user_directory_from_worker` options are set to the name of a worker.
 
-For the time being, `notify_appservices_from_worker` can be used alongside
-`synapse.app.appservice` and `notify_appservices` to make it easier to transition
-between the two configurations, however please note that:
+For the time being, the old options can be used alongside the new options to make
+it easier to transition between the two configurations, however please note that:
 
 - the options must not contradict each other (otherwise Synapse won't start); and
-- the `notify_appservices` option will be removed in a future release of Synapse.
+- the `notify_appservices` and `update_user_directory` options will be removed in a future release of Synapse.
 
-Please see [the relevant section of the worker documentation][v1_59_notify_ases_from] for more information.
+Please see the [*Notifying Application Services*][v1_59_notify_ases_from] and
+[*Updating the User Directory*][v1_59_update_user_dir] sections of the worker
+documentation for more information.
 
 [v1_59_notify_ases_from]: workers.md#notifying-application-services
+[v1_59_update_user_dir]: workers.md#updating-the-user-directory
 
 
 # Upgrading to v1.58.0

docs/usage/administration/useful_sql_for_admins.md

@@ -1,7 +1,10 @@
 ## Some useful SQL queries for Synapse Admins
 
 ## Size of full matrix db
-`SELECT pg_size_pretty( pg_database_size( 'matrix' ) );`
+```sql
+SELECT pg_size_pretty( pg_database_size( 'matrix' ) );
+```
+
 ### Result example:
 ``` 
 pg_size_pretty 
@@ -9,39 +12,19 @@ pg_size_pretty
  6420 MB
 (1 row)
 ```
-## Show top 20 larger rooms by state events count
-```sql
-SELECT r.name, s.room_id, s.current_state_events
-  FROM room_stats_current s
-  LEFT JOIN room_stats_state r USING (room_id)
-  ORDER BY current_state_events DESC
-  LIMIT 20;
-```
-
-and by state_group_events count:
-```sql
-SELECT rss.name, s.room_id, count(s.room_id) FROM state_groups_state s
-LEFT JOIN room_stats_state rss USING (room_id)
-GROUP BY s.room_id, rss.name        
-ORDER BY count(s.room_id) DESC
-LIMIT 20;
-```
-plus same, but with join removed for performance reasons:
-```sql
-SELECT s.room_id, count(s.room_id) FROM state_groups_state s
-GROUP BY s.room_id        
-ORDER BY count(s.room_id) DESC
-LIMIT 20;
-```
 
 ## Show top 20 larger tables by row count
 ```sql
-SELECT relname, n_live_tup as rows
-  FROM pg_stat_user_tables 
+SELECT relname, n_live_tup AS "rows"
+  FROM pg_stat_user_tables
   ORDER BY n_live_tup DESC
   LIMIT 20;
 ```
-This query is quick, but may be very approximate, for exact number of rows use `SELECT COUNT(*) FROM <table_name>`.
+This query is quick, but may be very approximate, for exact number of rows use:
+```sql
+SELECT COUNT(*) FROM <table_name>;
+```
+
 ### Result example:
 ```
 state_groups_state - 161687170
@@ -66,46 +49,19 @@ device_lists_stream - 326903
 user_directory_search - 316433
 ```
 
-## Show top 20 rooms by new events count in last 1 day:
-```sql
-SELECT e.room_id, r.name, COUNT(e.event_id) cnt FROM events e
-LEFT JOIN room_stats_state r USING (room_id)
-WHERE e.origin_server_ts >= DATE_PART('epoch', NOW() - INTERVAL '1 day') * 1000 GROUP BY e.room_id, r.name ORDER BY cnt DESC LIMIT 20;
-```
-
-## Show top 20 users on homeserver by sent events (messages) at last month:
-```sql
-SELECT user_id, SUM(total_events) 
-   FROM user_stats_historical
-   WHERE TO_TIMESTAMP(end_ts/1000) AT TIME ZONE 'UTC' > date_trunc('day', now() - interval '1 month')
-   GROUP BY user_id
-   ORDER BY SUM(total_events) DESC 
-   LIMIT 20;
-```
-
-## Show last 100 messages from needed user, with room names:
-```sql
-SELECT e.room_id, r.name, e.event_id, e.type, e.content, j.json FROM events e
-  LEFT JOIN event_json j USING (room_id)
-  LEFT JOIN room_stats_state r USING (room_id)
-  WHERE sender = '@LOGIN:example.com'
-  AND e.type = 'm.room.message'
-  ORDER BY stream_ordering DESC
-  LIMIT 100;
-```
-
 ## Show top 20 larger tables by storage size
 ```sql
 SELECT nspname || '.' || relname AS "relation",
-    pg_size_pretty(pg_total_relation_size(C.oid)) AS "total_size"
-  FROM pg_class C
-  LEFT JOIN pg_namespace N ON (N.oid = C.relnamespace)
+    pg_size_pretty(pg_total_relation_size(c.oid)) AS "total_size"
+  FROM pg_class c
+  LEFT JOIN pg_namespace n ON (n.oid = c.relnamespace)
   WHERE nspname NOT IN ('pg_catalog', 'information_schema')
-    AND C.relkind <> 'i'
+    AND c.relkind <> 'i'
     AND nspname !~ '^pg_toast'
-  ORDER BY pg_total_relation_size(C.oid) DESC
+  ORDER BY pg_total_relation_size(c.oid) DESC
   LIMIT 20;
 ```
+
 ### Result example:
 ```
 public.state_groups_state - 27 GB
@@ -130,8 +86,93 @@ public.device_lists_remote_cache - 124 MB
 public.state_group_edges - 122 MB
 ```
 
+## Show top 20 larger rooms by state events count
+You get the same information when you use the
+[admin API](../../admin_api/rooms.md#list-room-api)
+and set parameter `order_by=state_events`.
+
+```sql
+SELECT r.name, s.room_id, s.current_state_events
+  FROM room_stats_current s
+  LEFT JOIN room_stats_state r USING (room_id)
+  ORDER BY current_state_events DESC
+  LIMIT 20;
+```
+
+and by state_group_events count:
+```sql
+SELECT rss.name, s.room_id, COUNT(s.room_id)
+  FROM state_groups_state s
+  LEFT JOIN room_stats_state rss USING (room_id)
+  GROUP BY s.room_id, rss.name
+  ORDER BY COUNT(s.room_id) DESC
+  LIMIT 20;
+```
+
+plus same, but with join removed for performance reasons:
+```sql
+SELECT s.room_id, COUNT(s.room_id)
+  FROM state_groups_state s
+  GROUP BY s.room_id 
+  ORDER BY COUNT(s.room_id) DESC
+  LIMIT 20;
+```
+
+## Show top 20 rooms by new events count in last 1 day:
+```sql
+SELECT e.room_id, r.name, COUNT(e.event_id) cnt
+  FROM events e
+  LEFT JOIN room_stats_state r USING (room_id)
+  WHERE e.origin_server_ts >= DATE_PART('epoch', NOW() - INTERVAL '1 day') * 1000
+  GROUP BY e.room_id, r.name 
+  ORDER BY cnt DESC
+  LIMIT 20;
+```
+
+## Show top 20 users on homeserver by sent events (messages) at last month:
+Caution. This query does not use any indexes, can be slow and create load on the database.
+```sql
+SELECT COUNT(*), sender
+  FROM events
+  WHERE (type = 'm.room.encrypted' OR type = 'm.room.message')
+    AND origin_server_ts >= DATE_PART('epoch', NOW() - INTERVAL '1 month') * 1000
+  GROUP BY sender
+  ORDER BY COUNT(*) DESC
+  LIMIT 20;
+```
+
+## Show last 100 messages from needed user, with room names:
+```sql
+SELECT e.room_id, r.name, e.event_id, e.type, e.content, j.json
+  FROM events e
+  LEFT JOIN event_json j USING (room_id)
+  LEFT JOIN room_stats_state r USING (room_id)
+  WHERE sender = '@LOGIN:example.com'
+    AND e.type = 'm.room.message'
+  ORDER BY stream_ordering DESC
+  LIMIT 100;
+```
+
 ## Show rooms with names, sorted by events in this rooms
-`echo "select event_json.room_id,room_stats_state.name from event_json,room_stats_state where room_stats_state.room_id=event_json.room_id" | psql synapse | sort | uniq -c | sort -n`
+
+**Sort and order with bash**
+```bash
+echo "SELECT event_json.room_id, room_stats_state.name FROM event_json, room_stats_state \
+WHERE room_stats_state.room_id = event_json.room_id" | psql -d synapse -h localhost -U synapse_user -t \
+| sort | uniq -c | sort -n
+```
+Documentation for `psql` command line parameters: https://www.postgresql.org/docs/current/app-psql.html
+
+**Sort and order with SQL**
+```sql
+SELECT COUNT(*), event_json.room_id, room_stats_state.name
+  FROM event_json, room_stats_state
+  WHERE room_stats_state.room_id = event_json.room_id
+  GROUP BY event_json.room_id, room_stats_state.name
+  ORDER BY COUNT(*) DESC
+  LIMIT 50;
+```
+
 ### Result example:
 ```
    9459  !FPUfgzXYWTKgIrwKxW:matrix.org              | This Week in Matrix
@@ -145,12 +186,22 @@ public.state_group_edges - 122 MB
 ```
 
 ## Lookup room state info by list of room_id
+You get the same information when you use the
+[admin API](../../admin_api/rooms.md#room-details-api).
 ```sql
-SELECT rss.room_id, rss.name, rss.canonical_alias, rss.topic, rss.encryption, rsc.joined_members, rsc.local_users_in_room, rss.join_rules
-FROM room_stats_state rss
-LEFT JOIN room_stats_current rsc USING (room_id)
-WHERE room_id IN (WHERE room_id IN (
- '!OGEhHVWSdvArJzumhm:matrix.org',
- '!YTvKGNlinIzlkMTVRl:matrix.org'
-)
-```
+SELECT rss.room_id, rss.name, rss.canonical_alias, rss.topic, rss.encryption,
+    rsc.joined_members, rsc.local_users_in_room, rss.join_rules
+  FROM room_stats_state rss
+  LEFT JOIN room_stats_current rsc USING (room_id)
+  WHERE room_id IN ( WHERE room_id IN (
+    '!OGEhHVWSdvArJzumhm:matrix.org',
+    '!YTvKGNlinIzlkMTVRl:matrix.org' 
+  );
+```
+
+## Show users and devices that have not been online for a while
+```sql
+SELECT user_id, device_id, user_agent, TO_TIMESTAMP(last_seen / 1000) AS "last_seen"
+  FROM devices
+  WHERE last_seen < DATE_PART('epoch', NOW() - INTERVAL '3 month') * 1000;
+```

docs/workers.md

@@ -426,7 +426,7 @@ the shared configuration would include:
 run_background_tasks_on: background_worker
 ```
 
-You might also wish to investigate the `update_user_directory` and
+You might also wish to investigate the `update_user_directory_from_worker` and
 `media_instance_running_background_jobs` settings.
 
 An example for a dedicated background worker instance:
@@ -435,9 +435,26 @@ An example for a dedicated background worker instance:
 {{#include systemd-with-workers/workers/background_worker.yaml}}
 ```
 
+#### Updating the User Directory
+
+You can designate one generic worker to update the user directory.
+
+Specify its name in the shared configuration as follows:
+
+```yaml
+update_user_directory_from_worker: worker_name
+```
+
+This work cannot be load-balanced; please ensure the main process is restarted
+after setting this option in the shared configuration!
+
+This style of configuration supersedes the legacy `synapse.app.user_dir`
+worker application type.
+
+
 #### Notifying Application Services
 
-You can designate one worker to send output traffic to Application Services.
+You can designate one generic worker to send output traffic to Application Services.
 
 Specify its name in the shared configuration as follows:
 
@@ -470,7 +487,7 @@ pusher_instances:
 
 ### `synapse.app.appservice`
 
-**Deprecated as of Synapse v1.58.** [Use `synapse.app.generic_worker` with the
+**Deprecated as of Synapse v1.59.** [Use `synapse.app.generic_worker` with the
 `notify_appservices_from_worker` option instead.](#notifying-application-services)
 
 Handles sending output traffic to Application Services. Doesn't handle any
@@ -540,6 +557,9 @@ Note that if a reverse proxy is used , then `/_matrix/media/` must be routed for
 
 ### `synapse.app.user_dir`
 
+**Deprecated as of Synapse v1.59.** [Use `synapse.app.generic_worker` with the
+`update_user_directory_from_worker` option instead.](#updating-the-user-directory)
+
 Handles searches in the user directory. It can handle REST endpoints matching
 the following regular expressions:
 

pyproject.toml

@@ -54,7 +54,7 @@ skip_gitignore = true
 
 [tool.poetry]
 name = "matrix-synapse"
-version = "1.58.1"
+version = "1.59.0rc1"
 description = "Homeserver for the Matrix decentralised comms protocol"
 authors = ["Matrix.org Team and Contributors <packages@matrix.org>"]
 license = "Apache-2.0"

scripts-dev/complement.sh

@@ -43,6 +43,8 @@ fi
 # Build the base Synapse image from the local checkout
 docker build -t matrixdotorg/synapse -f "docker/Dockerfile" .
 
+extra_test_args=()
+
 # If we're using workers, modify the docker files slightly.
 if [[ -n "$WORKERS" ]]; then
   # Build the workers docker image (from the base Synapse image).
@@ -52,7 +54,14 @@ if [[ -n "$WORKERS" ]]; then
   COMPLEMENT_DOCKERFILE=SynapseWorkers.Dockerfile
 
   # And provide some more configuration to complement.
-  export COMPLEMENT_SPAWN_HS_TIMEOUT_SECS=60
+
+  # It can take quite a while to spin up a worker-mode Synapse for the first
+  # time (the main problem is that we start 14 python processes for each test,
+  # and complement likes to do two of them in parallel).
+  export COMPLEMENT_SPAWN_HS_TIMEOUT_SECS=120
+
+  # ... and it takes longer than 10m to run the whole suite.
+  extra_test_args+=("-timeout=60m")
 else
   export COMPLEMENT_BASE_IMAGE=complement-synapse
   COMPLEMENT_DOCKERFILE=Dockerfile
@@ -64,4 +73,4 @@ docker build -t $COMPLEMENT_BASE_IMAGE -f "docker/complement/$COMPLEMENT_DOCKERF
 # Run the tests!
 echo "Images built; running complement"
 cd "$COMPLEMENT_DIR"
-go test -v -tags synapse_blacklist,msc2716,msc3030,faster_joins -count=1 "$@" ./tests/...
+go test -v -tags synapse_blacklist,msc2716,msc3030,faster_joins -count=1 "${extra_test_args[@]}" "$@" ./tests/...

synapse/api/filtering.py

@@ -19,6 +19,7 @@ from typing import (
     TYPE_CHECKING,
     Awaitable,
     Callable,
+    Collection,
     Dict,
     Iterable,
     List,
@@ -444,9 +445,9 @@ class Filter:
         return room_ids
 
     async def _check_event_relations(
-        self, events: Iterable[FilterEvent]
+        self, events: Collection[FilterEvent]
     ) -> List[FilterEvent]:
-        # The event IDs to check, mypy doesn't understand the ifinstance check.
+        # The event IDs to check, mypy doesn't understand the isinstance check.
         event_ids = [event.event_id for event in events if isinstance(event, EventBase)]  # type: ignore[attr-defined]
         event_ids_to_keep = set(
             await self._store.events_have_relations(

synapse/app/admin_cmd.py

@@ -210,7 +210,7 @@ def start(config_options: List[str]) -> None:
         config.logging.no_redirect_stdio = True
 
     # Explicitly disable background processes
-    config.server.update_user_directory = False
+    config.worker.should_update_user_directory = False
     config.worker.run_background_tasks = False
     config.worker.start_pushers = False
     config.worker.pusher_shard_config.instances = []

synapse/app/generic_worker.py

@@ -441,22 +441,6 @@ def start(config_options: List[str]) -> None:
         "synapse.app.user_dir",
     )
 
-    if config.worker.worker_app == "synapse.app.user_dir":
-        if config.server.update_user_directory:
-            sys.stderr.write(
-                "\nThe update_user_directory must be disabled in the main synapse process"
-                "\nbefore they can be run in a separate worker."
-                "\nPlease add ``update_user_directory: false`` to the main config"
-                "\n"
-            )
-            sys.exit(1)
-
-        # Force the pushers to start since they will be disabled in the main config
-        config.server.update_user_directory = True
-    else:
-        # For other worker types we force this to off.
-        config.server.update_user_directory = False
-
     synapse.events.USE_FROZEN_DICTS = config.server.use_frozen_dicts
     synapse.util.caches.TRACK_MEMORY_USAGE = config.caches.track_memory_usage
 

synapse/config/experimental.py

@@ -81,3 +81,6 @@ class ExperimentalConfig(Config):
 
         # MSC2815 (allow room moderators to view redacted event content)
         self.msc2815_enabled: bool = experimental.get("msc2815_enabled", False)
+
+        # MSC3786 (Add a default push rule to ignore m.room.server_acl events)
+        self.msc3786_enabled: bool = experimental.get("msc3786_enabled", False)

synapse/config/server.py

@@ -319,10 +319,6 @@ class ServerConfig(Config):
                 self.presence_router_config,
             ) = load_module(presence_router_config, ("presence", "presence_router"))
 
-        # Whether to update the user directory or not. This should be set to
-        # false only if we are updating the user directory in a worker
-        self.update_user_directory = config.get("update_user_directory", True)
-
         # whether to enable the media repository endpoints. This should be set
         # to false if the media repository is running as a separate endpoint;
         # doing so ensures that we will not run cache cleanup jobs on the

synapse/config/workers.py

@@ -311,6 +311,13 @@ class WorkerConfig(Config):
             new_option_name="notify_appservices_from_worker",
         )
 
+        self.should_update_user_directory = self._should_this_worker_perform_duty(
+            config,
+            legacy_master_option_name="update_user_directory",
+            legacy_worker_app_name="synapse.app.user_dir",
+            new_option_name="update_user_directory_from_worker",
+        )
+
     def _should_this_worker_perform_duty(
         self,
         config: Dict[str, Any],

synapse/events/snapshot.py

@@ -15,12 +15,10 @@ from typing import TYPE_CHECKING, List, Optional, Tuple, Union
 
 import attr
 from frozendict import frozendict
-
-from twisted.internet.defer import Deferred
+from typing_extensions import Literal
 
 from synapse.appservice import ApplicationService
 from synapse.events import EventBase
-from synapse.logging.context import make_deferred_yieldable, run_in_background
 from synapse.types import JsonDict, StateMap
 
 if TYPE_CHECKING:
@@ -60,6 +58,9 @@ class EventContext:
             If ``state_group`` is None (ie, the event is an outlier),
             ``state_group_before_event`` will always also be ``None``.
 
+        state_delta_due_to_event: If `state_group` and `state_group_before_event` are not None
+            then this is the delta of the state between the two groups.
+
         prev_group: If it is known, ``state_group``'s prev_group. Note that this being
             None does not necessarily mean that ``state_group`` does not have
             a prev_group!
@@ -78,73 +79,47 @@ class EventContext:
         app_service: If this event is being sent by a (local) application service, that
             app service.
 
-        _current_state_ids: The room state map, including this event - ie, the state
-            in ``state_group``.
-
-            (type, state_key) -> event_id
-
-            For an outlier, this is {}
-
-            Note that this is a private attribute: it should be accessed via
-            ``get_current_state_ids``. _AsyncEventContext impl calculates this
-            on-demand: it will be None until that happens.
-
-        _prev_state_ids: The room state map, excluding this event - ie, the state
-            in ``state_group_before_event``. For a non-state
-            event, this will be the same as _current_state_events.
-
-            Note that it is a completely different thing to prev_group!
-
-            (type, state_key) -> event_id
-
-            For an outlier, this is {}
-
-            As with _current_state_ids, this is a private attribute. It should be
-            accessed via get_prev_state_ids.
-
         partial_state: if True, we may be storing this event with a temporary,
             incomplete state.
     """
 
-    rejected: Union[bool, str] = False
+    _storage: "Storage"
+    rejected: Union[Literal[False], str] = False
     _state_group: Optional[int] = None
     state_group_before_event: Optional[int] = None
+    _state_delta_due_to_event: Optional[StateMap[str]] = None
     prev_group: Optional[int] = None
     delta_ids: Optional[StateMap[str]] = None
     app_service: Optional[ApplicationService] = None
 
-    _current_state_ids: Optional[StateMap[str]] = None
-    _prev_state_ids: Optional[StateMap[str]] = None
-
     partial_state: bool = False
 
     @staticmethod
     def with_state(
+        storage: "Storage",
         state_group: Optional[int],
         state_group_before_event: Optional[int],
-        current_state_ids: Optional[StateMap[str]],
-        prev_state_ids: Optional[StateMap[str]],
+        state_delta_due_to_event: Optional[StateMap[str]],
         partial_state: bool,
         prev_group: Optional[int] = None,
         delta_ids: Optional[StateMap[str]] = None,
     ) -> "EventContext":
         return EventContext(
-            current_state_ids=current_state_ids,
-            prev_state_ids=prev_state_ids,
+            storage=storage,
             state_group=state_group,
             state_group_before_event=state_group_before_event,
+            state_delta_due_to_event=state_delta_due_to_event,
             prev_group=prev_group,
             delta_ids=delta_ids,
             partial_state=partial_state,
         )
 
     @staticmethod
-    def for_outlier() -> "EventContext":
+    def for_outlier(
+        storage: "Storage",
+    ) -> "EventContext":
         """Return an EventContext instance suitable for persisting an outlier event"""
-        return EventContext(
-            current_state_ids={},
-            prev_state_ids={},
-        )
+        return EventContext(storage=storage)
 
     async def serialize(self, event: EventBase, store: "DataStore") -> JsonDict:
         """Converts self to a type that can be serialized as JSON, and then
@@ -157,24 +132,14 @@ class EventContext:
             The serialized event.
         """
 
-        # We don't serialize the full state dicts, instead they get pulled out
-        # of the DB on the other side. However, the other side can't figure out
-        # the prev_state_ids, so if we're a state event we include the event
-        # id that we replaced in the state.
-        if event.is_state():
-            prev_state_ids = await self.get_prev_state_ids()
-            prev_state_id = prev_state_ids.get((event.type, event.state_key))
-        else:
-            prev_state_id = None
-
         return {
-            "prev_state_id": prev_state_id,
-            "event_type": event.type,
-            "event_state_key": event.get_state_key(),
             "state_group": self._state_group,
             "state_group_before_event": self.state_group_before_event,
             "rejected": self.rejected,
             "prev_group": self.prev_group,
+            "state_delta_due_to_event": _encode_state_dict(
+                self._state_delta_due_to_event
+            ),
             "delta_ids": _encode_state_dict(self.delta_ids),
             "app_service_id": self.app_service.id if self.app_service else None,
             "partial_state": self.partial_state,
@@ -192,16 +157,16 @@ class EventContext:
         Returns:
             The event context.
         """
-        context = _AsyncEventContextImpl(
+        context = EventContext(
             # We use the state_group and prev_state_id stuff to pull the
             # current_state_ids out of the DB and construct prev_state_ids.
             storage=storage,
-            prev_state_id=input["prev_state_id"],
-            event_type=input["event_type"],
-            event_state_key=input["event_state_key"],
             state_group=input["state_group"],
             state_group_before_event=input["state_group_before_event"],
             prev_group=input["prev_group"],
+            state_delta_due_to_event=_decode_state_dict(
+                input["state_delta_due_to_event"]
+            ),
             delta_ids=_decode_state_dict(input["delta_ids"]),
             rejected=input["rejected"],
             partial_state=input.get("partial_state", False),
@@ -249,8 +214,15 @@ class EventContext:
         if self.rejected:
             raise RuntimeError("Attempt to access state_ids of rejected event")
 
-        await self._ensure_fetched()
-        return self._current_state_ids
+        assert self._state_delta_due_to_event is not None
+
+        prev_state_ids = await self.get_prev_state_ids()
+
+        if self._state_delta_due_to_event:
+            prev_state_ids = dict(prev_state_ids)
+            prev_state_ids.update(self._state_delta_due_to_event)
+
+        return prev_state_ids
 
     async def get_prev_state_ids(self) -> StateMap[str]:
         """
@@ -265,94 +237,10 @@ class EventContext:
             Maps a (type, state_key) to the event ID of the state event matching
             this tuple.
         """
-        await self._ensure_fetched()
-        # There *should* be previous state IDs now.
-        assert self._prev_state_ids is not None
-        return self._prev_state_ids
-
-    def get_cached_current_state_ids(self) -> Optional[StateMap[str]]:
-        """Gets the current state IDs if we have them already cached.
-
-        It is an error to access this for a rejected event, since rejected state should
-        not make it into the room state. This method will raise an exception if
-        ``rejected`` is set.
-
-        Returns:
-            Returns None if we haven't cached the state or if state_group is None
-            (which happens when the associated event is an outlier).
-
-            Otherwise, returns the the current state IDs.
-        """
-        if self.rejected:
-            raise RuntimeError("Attempt to access state_ids of rejected event")
-
-        return self._current_state_ids
-
-    async def _ensure_fetched(self) -> None:
-        return None
-
-
-@attr.s(slots=True)
-class _AsyncEventContextImpl(EventContext):
-    """
-    An implementation of EventContext which fetches _current_state_ids and
-    _prev_state_ids from the database on demand.
-
-    Attributes:
-
-        _storage
-
-        _fetching_state_deferred: Resolves when *_state_ids have been calculated.
-            None if we haven't started calculating yet
-
-        _event_type: The type of the event the context is associated with.
-
-        _event_state_key: The state_key of the event the context is associated with.
-
-        _prev_state_id: If the event associated with the context is a state event,
-            then `_prev_state_id` is the event_id of the state that was replaced.
-    """
-
-    # This needs to have a default as we're inheriting
-    _storage: "Storage" = attr.ib(default=None)
-    _prev_state_id: Optional[str] = attr.ib(default=None)
-    _event_type: str = attr.ib(default=None)
-    _event_state_key: Optional[str] = attr.ib(default=None)
-    _fetching_state_deferred: Optional["Deferred[None]"] = attr.ib(default=None)
-
-    async def _ensure_fetched(self) -> None:
-        if not self._fetching_state_deferred:
-            self._fetching_state_deferred = run_in_background(self._fill_out_state)
-
-        await make_deferred_yieldable(self._fetching_state_deferred)
-
-    async def _fill_out_state(self) -> None:
-        """Called to populate the _current_state_ids and _prev_state_ids
-        attributes by loading from the database.
-        """
-        if self.state_group is None:
-            # No state group means the event is an outlier. Usually the state_ids dicts are also
-            # pre-set to empty dicts, but they get reset when the context is serialized, so set
-            # them to empty dicts again here.
-            self._current_state_ids = {}
-            self._prev_state_ids = {}
-            return
-
-        current_state_ids = await self._storage.state.get_state_ids_for_group(
-            self.state_group
+        assert self.state_group_before_event is not None
+        return await self._storage.state.get_state_ids_for_group(
+            self.state_group_before_event
         )
-        # Set this separately so mypy knows current_state_ids is not None.
-        self._current_state_ids = current_state_ids
-        if self._event_state_key is not None:
-            self._prev_state_ids = dict(current_state_ids)
-
-            key = (self._event_type, self._event_state_key)
-            if self._prev_state_id:
-                self._prev_state_ids[key] = self._prev_state_id
-            else:
-                self._prev_state_ids.pop(key, None)
-        else:
-            self._prev_state_ids = current_state_ids
 
 
 def _encode_state_dict(

synapse/events/spamcheck.py

@@ -27,9 +27,10 @@ from typing import (
     Union,
 )
 
+from synapse.api.errors import Codes
 from synapse.rest.media.v1._base import FileInfo
 from synapse.rest.media.v1.media_storage import ReadableFileWrapper
-from synapse.spam_checker_api import RegistrationBehaviour
+from synapse.spam_checker_api import ALLOW, Decision, RegistrationBehaviour
 from synapse.types import RoomAlias, UserProfile
 from synapse.util.async_helpers import delay_cancellation, maybe_awaitable
 
@@ -39,17 +40,34 @@ if TYPE_CHECKING:
 
 logger = logging.getLogger(__name__)
 
+
+DEPRECATED_BOOL = bool
+
 CHECK_EVENT_FOR_SPAM_CALLBACK = Callable[
     ["synapse.events.EventBase"],
-    Awaitable[Union[bool, str]],
+    Awaitable[Union[ALLOW, Codes, str, DEPRECATED_BOOL]],
+]
+USER_MAY_JOIN_ROOM_CALLBACK = Callable[
+    [str, str, bool], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+USER_MAY_INVITE_CALLBACK = Callable[
+    [str, str, str], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+USER_MAY_SEND_3PID_INVITE_CALLBACK = Callable[
+    [str, str, str, str], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+USER_MAY_CREATE_ROOM_CALLBACK = Callable[
+    [str], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+USER_MAY_CREATE_ROOM_ALIAS_CALLBACK = Callable[
+    [str, RoomAlias], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+USER_MAY_PUBLISH_ROOM_CALLBACK = Callable[
+    [str, str], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
+]
+CHECK_USERNAME_FOR_SPAM_CALLBACK = Callable[
+    [UserProfile], Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]]
 ]
-USER_MAY_JOIN_ROOM_CALLBACK = Callable[[str, str, bool], Awaitable[bool]]
-USER_MAY_INVITE_CALLBACK = Callable[[str, str, str], Awaitable[bool]]
-USER_MAY_SEND_3PID_INVITE_CALLBACK = Callable[[str, str, str, str], Awaitable[bool]]
-USER_MAY_CREATE_ROOM_CALLBACK = Callable[[str], Awaitable[bool]]
-USER_MAY_CREATE_ROOM_ALIAS_CALLBACK = Callable[[str, RoomAlias], Awaitable[bool]]
-USER_MAY_PUBLISH_ROOM_CALLBACK = Callable[[str, str], Awaitable[bool]]
-CHECK_USERNAME_FOR_SPAM_CALLBACK = Callable[[UserProfile], Awaitable[bool]]
 LEGACY_CHECK_REGISTRATION_FOR_SPAM_CALLBACK = Callable[
     [
         Optional[dict],
@@ -65,11 +83,11 @@ CHECK_REGISTRATION_FOR_SPAM_CALLBACK = Callable[
         Collection[Tuple[str, str]],
         Optional[str],
     ],
-    Awaitable[RegistrationBehaviour],
+    Awaitable[Union[RegistrationBehaviour, Codes]],
 ]
 CHECK_MEDIA_FILE_FOR_SPAM_CALLBACK = Callable[
     [ReadableFileWrapper, FileInfo],
-    Awaitable[bool],
+    Awaitable[Union[ALLOW, Codes, DEPRECATED_BOOL]],
 ]
 
 
@@ -240,7 +258,7 @@ class SpamChecker:
 
     async def check_event_for_spam(
         self, event: "synapse.events.EventBase"
-    ) -> Union[bool, str]:
+    ) -> Union[ALLOW, Codes, str]:
         """Checks if a given event is considered "spammy" by this server.
 
         If the server considers an event spammy, then it will be rejected if
@@ -251,19 +269,29 @@ class SpamChecker:
             event: the event to be checked
 
         Returns:
-            True or a string if the event is spammy. If a string is returned it
-            will be used as the error message returned to the user.
+            - on `ALLOW`, the event is considered good (non-spammy) and should
+                be let through. Other spamcheck filters may still reject it.
+            - on `Codes`, the event is considered spammy and is rejected with a specific
+                error message/code.
+            - on `str`, the event is considered spammy and the string is used as error
+                message.
         """
         for callback in self._check_event_for_spam_callbacks:
-            res: Union[bool, str] = await delay_cancellation(callback(event))
-            if res:
+            res: Union[ALLOW, Codes, str, DEPRECATED_BOOL] = await delay_cancellation(
+                callback(event)
+            )
+            if res is False or res is ALLOW:
+                continue
+            elif res is True:
+                return Codes.FORBIDDEN
+            else:
                 return res
 
-        return False
+        return ALLOW
 
     async def user_may_join_room(
         self, user_id: str, room_id: str, is_invited: bool
-    ) -> bool:
+    ) -> Decision:
         """Checks if a given users is allowed to join a room.
         Not called when a user creates a room.
 
@@ -273,48 +301,54 @@ class SpamChecker:
             is_invited: Whether the user is invited into the room
 
         Returns:
-            Whether the user may join the room
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_join_room_callbacks:
             may_join_room = await delay_cancellation(
                 callback(user_id, room_id, is_invited)
             )
-            if may_join_room is False:
-                return False
+            if may_join_room is True or may_join_room is ALLOW:
+                continue
+            elif may_join_room is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_join_room
 
-        return True
+        return ALLOW
 
     async def user_may_invite(
         self, inviter_userid: str, invitee_userid: str, room_id: str
-    ) -> bool:
+    ) -> Decision:
         """Checks if a given user may send an invite
 
-        If this method returns false, the invite will be rejected.
-
         Args:
             inviter_userid: The user ID of the sender of the invitation
             invitee_userid: The user ID targeted in the invitation
             room_id: The room ID
 
         Returns:
-            True if the user may send an invite, otherwise False
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_invite_callbacks:
             may_invite = await delay_cancellation(
                 callback(inviter_userid, invitee_userid, room_id)
             )
-            if may_invite is False:
-                return False
+            if may_invite is True or may_invite is ALLOW:
+                continue
+            elif may_invite is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_invite
 
-        return True
+        return ALLOW
 
     async def user_may_send_3pid_invite(
         self, inviter_userid: str, medium: str, address: str, room_id: str
-    ) -> bool:
+    ) -> Decision:
         """Checks if a given user may invite a given threepid into the room
 
-        If this method returns false, the threepid invite will be rejected.
-
         Note that if the threepid is already associated with a Matrix user ID, Synapse
         will call user_may_invite with said user ID instead.
 
@@ -325,78 +359,94 @@ class SpamChecker:
             room_id: The room ID
 
         Returns:
-            True if the user may send the invite, otherwise False
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_send_3pid_invite_callbacks:
             may_send_3pid_invite = await delay_cancellation(
                 callback(inviter_userid, medium, address, room_id)
             )
-            if may_send_3pid_invite is False:
-                return False
+            if may_send_3pid_invite is True or may_send_3pid_invite is ALLOW:
+                continue
+            elif may_send_3pid_invite is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_send_3pid_invite
 
-        return True
+        return ALLOW
 
-    async def user_may_create_room(self, userid: str) -> bool:
+    async def user_may_create_room(self, userid: str) -> Decision:
         """Checks if a given user may create a room
 
-        If this method returns false, the creation request will be rejected.
-
         Args:
             userid: The ID of the user attempting to create a room
 
         Returns:
-            True if the user may create a room, otherwise False
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_create_room_callbacks:
             may_create_room = await delay_cancellation(callback(userid))
-            if may_create_room is False:
-                return False
+            if may_create_room is True or may_create_room is ALLOW:
+                continue
+            elif may_create_room is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_create_room
 
-        return True
+        return ALLOW
 
     async def user_may_create_room_alias(
         self, userid: str, room_alias: RoomAlias
-    ) -> bool:
+    ) -> Decision:
         """Checks if a given user may create a room alias
 
-        If this method returns false, the association request will be rejected.
-
         Args:
             userid: The ID of the user attempting to create a room alias
             room_alias: The alias to be created
 
         Returns:
-            True if the user may create a room alias, otherwise False
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_create_room_alias_callbacks:
             may_create_room_alias = await delay_cancellation(
                 callback(userid, room_alias)
             )
-            if may_create_room_alias is False:
-                return False
+            if may_create_room_alias is True or may_create_room_alias is ALLOW:
+                continue
+            elif may_create_room_alias is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_create_room_alias
 
-        return True
+        return ALLOW
 
-    async def user_may_publish_room(self, userid: str, room_id: str) -> bool:
+    async def user_may_publish_room(
+        self, userid: str, room_id: str
+    ) -> Union[ALLOW, Codes, DEPRECATED_BOOL]:
         """Checks if a given user may publish a room to the directory
 
-        If this method returns false, the publish request will be rejected.
-
         Args:
             userid: The user ID attempting to publish the room
             room_id: The ID of the room that would be published
 
         Returns:
-            True if the user may publish the room, otherwise False
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._user_may_publish_room_callbacks:
             may_publish_room = await delay_cancellation(callback(userid, room_id))
-            if may_publish_room is False:
-                return False
+            if may_publish_room is True or may_publish_room is ALLOW:
+                continue
+            elif may_publish_room is False:
+                return Codes.FORBIDDEN
+            else:
+                return may_publish_room
 
-        return True
+        return ALLOW
 
-    async def check_username_for_spam(self, user_profile: UserProfile) -> bool:
+    async def check_username_for_spam(self, user_profile: UserProfile) -> Decision:
         """Checks if a user ID or display name are considered "spammy" by this server.
 
         If the server considers a username spammy, then it will not be included in
@@ -409,15 +459,21 @@ class SpamChecker:
                 * avatar_url
 
         Returns:
-            True if the user is spammy.
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
         for callback in self._check_username_for_spam_callbacks:
             # Make a copy of the user profile object to ensure the spam checker cannot
             # modify it.
-            if await delay_cancellation(callback(user_profile.copy())):
-                return True
+            is_spam = await delay_cancellation(callback(user_profile.copy()))
+            if is_spam is False or is_spam is ALLOW:
+                continue
+            elif is_spam is True:
+                return Codes.FORBIDDEN
+            else:
+                return is_spam
 
-        return False
+        return ALLOW
 
     async def check_registration_for_spam(
         self,
@@ -445,6 +501,8 @@ class SpamChecker:
             behaviour = await delay_cancellation(
                 callback(email_threepid, username, request_info, auth_provider_id)
             )
+            if isinstance(behaviour, Codes):
+                return behaviour
             assert isinstance(behaviour, RegistrationBehaviour)
             if behaviour != RegistrationBehaviour.ALLOW:
                 return behaviour
@@ -453,7 +511,7 @@ class SpamChecker:
 
     async def check_media_file_for_spam(
         self, file_wrapper: ReadableFileWrapper, file_info: FileInfo
-    ) -> bool:
+    ) -> Decision:
         """Checks if a piece of newly uploaded media should be blocked.
 
         This will be called for local uploads, downloads of remote media, each
@@ -475,19 +533,22 @@ class SpamChecker:
 
                 return False
 
-
         Args:
             file: An object that allows reading the contents of the media.
             file_info: Metadata about the file.
 
         Returns:
-            True if the media should be blocked or False if it should be
-            allowed.
+            - on `ALLOW`, the action is permitted.
+            - on `Codes`, the action is rejected with a specific error message/code.
         """
 
         for callback in self._check_media_file_for_spam_callbacks:
-            spam = await delay_cancellation(callback(file_wrapper, file_info))
-            if spam:
-                return True
+            is_spam = await delay_cancellation(callback(file_wrapper, file_info))
+            if is_spam is False or is_spam is ALLOW:
+                continue
+            elif is_spam is True:
+                return Codes.FORBIDDEN
+            else:
+                return is_spam
 
-        return False
+        return ALLOW

synapse/federation/federation_base.py

@@ -15,6 +15,7 @@
 import logging
 from typing import TYPE_CHECKING
 
+import synapse
 from synapse.api.constants import MAX_DEPTH, EventContentFields, EventTypes, Membership
 from synapse.api.errors import Codes, SynapseError
 from synapse.api.room_versions import EventFormatVersions, RoomVersion
@@ -98,9 +99,9 @@ class FederationBase:
                 )
             return redacted_event
 
-        result = await self.spam_checker.check_event_for_spam(pdu)
+        spam_check = await self.spam_checker.check_event_for_spam(pdu)
 
-        if result:
+        if spam_check is not synapse.spam_checker_api.ALLOW:
             logger.warning("Event contains spam, soft-failing %s", pdu.event_id)
             # we redact (to save disk space) as well as soft-failing (to stop
             # using the event in prev_events).

synapse/handlers/directory.py

@@ -16,6 +16,7 @@ import logging
 import string
 from typing import TYPE_CHECKING, Iterable, List, Optional
 
+import synapse
 from synapse.api.constants import MAX_ALIAS_LENGTH, EventTypes
 from synapse.api.errors import (
     AuthError,
@@ -137,10 +138,13 @@ class DirectoryHandler:
                         403, "You must be in the room to create an alias for it"
                     )
 
-            if not await self.spam_checker.user_may_create_room_alias(
+            spam_check = await self.spam_checker.user_may_create_room_alias(
                 user_id, room_alias
-            ):
-                raise AuthError(403, "This user is not permitted to create this alias")
+            )
+            if spam_check is not synapse.spam_checker_api.ALLOW:
+                raise AuthError(
+                    403, "This alias creation request has been rejected", spam_check
+                )
 
             if not self.config.roomdirectory.is_alias_creation_allowed(
                 user_id, room_id, room_alias_str
@@ -426,9 +430,12 @@ class DirectoryHandler:
         """
         user_id = requester.user.to_string()
 
-        if not await self.spam_checker.user_may_publish_room(user_id, room_id):
+        spam_check = await self.spam_checker.user_may_publish_room(user_id, room_id)
+        if spam_check is not synapse.spam_checker_api.ALLOW:
             raise AuthError(
-                403, "This user is not permitted to publish rooms to the room list"
+                403,
+                "This request to publish a room to the room list has been rejected",
+                spam_check,
             )
 
         if requester.is_guest:

synapse/handlers/federation.py

@@ -27,6 +27,7 @@ from signedjson.key import decode_verify_key_bytes
 from signedjson.sign import verify_signed_json
 from unpaddedbase64 import decode_base64
 
+import synapse
 from synapse import event_auth
 from synapse.api.constants import EventContentFields, EventTypes, Membership
 from synapse.api.errors import (
@@ -659,7 +660,7 @@ class FederationHandler:
         # in the invitee's sync stream. It is stripped out for all other local users.
         event.unsigned["knock_room_state"] = stripped_room_state["knock_state_events"]
 
-        context = EventContext.for_outlier()
+        context = EventContext.for_outlier(self.storage)
         stream_id = await self._federation_event_handler.persist_events_and_notify(
             event.room_id, [(event, context)]
         )
@@ -799,11 +800,14 @@ class FederationHandler:
         if self.hs.config.server.block_non_admin_invites:
             raise SynapseError(403, "This server does not accept room invites")
 
-        if not await self.spam_checker.user_may_invite(
+        spam_check = await self.spam_checker.user_may_invite(
             event.sender, event.state_key, event.room_id
-        ):
+        )
+        if spam_check is not synapse.spam_checker_api.ALLOW:
             raise SynapseError(
-                403, "This user is not permitted to send invites to this server/user"
+                403,
+                "This user is not permitted to send invites to this server/user",
+                spam_check,
             )
 
         membership = event.content.get("membership")
@@ -848,7 +852,7 @@ class FederationHandler:
             )
         )
 
-        context = EventContext.for_outlier()
+        context = EventContext.for_outlier(self.storage)
         await self._federation_event_handler.persist_events_and_notify(
             event.room_id, [(event, context)]
         )
@@ -877,7 +881,7 @@ class FederationHandler:
 
         await self.federation_client.send_leave(host_list, event)
 
-        context = EventContext.for_outlier()
+        context = EventContext.for_outlier(self.storage)
         stream_id = await self._federation_event_handler.persist_events_and_notify(
             event.room_id, [(event, context)]
         )

synapse/handlers/federation_event.py

@@ -1423,7 +1423,7 @@ class FederationEventHandler:
                 # we're not bothering about room state, so flag the event as an outlier.
                 event.internal_metadata.outlier = True
 
-                context = EventContext.for_outlier()
+                context = EventContext.for_outlier(self._storage)
                 try:
                     validate_event_for_room_version(room_version_obj, event)
                     check_auth_rules_for_event(room_version_obj, event, auth)
@@ -1874,10 +1874,10 @@ class FederationEventHandler:
         )
 
         return EventContext.with_state(
+            storage=self._storage,
             state_group=state_group,
             state_group_before_event=context.state_group_before_event,
-            current_state_ids=current_state_ids,
-            prev_state_ids=prev_state_ids,
+            state_delta_due_to_event=state_updates,
             prev_group=prev_group,
             delta_ids=state_updates,
             partial_state=context.partial_state,

synapse/handlers/message.py

@@ -23,6 +23,7 @@ from canonicaljson import encode_canonical_json
 
 from twisted.internet.interfaces import IDelayedCall
 
+import synapse
 from synapse import event_auth
 from synapse.api.constants import (
     EventContentFields,
@@ -757,6 +758,10 @@ class EventCreationHandler:
             The previous version of the event is returned, if it is found in the
             event context. Otherwise, None is returned.
         """
+        if event.internal_metadata.is_outlier():
+            # This can happen due to out of band memberships
+            return None
+
         prev_state_ids = await context.get_prev_state_ids()
         prev_event_id = prev_state_ids.get((event.type, event.state_key))
         if not prev_event_id:
@@ -877,11 +882,11 @@ class EventCreationHandler:
                 event.sender,
             )
 
-            spam_error = await self.spam_checker.check_event_for_spam(event)
-            if spam_error:
-                if not isinstance(spam_error, str):
-                    spam_error = "Spam is not permitted here"
-                raise SynapseError(403, spam_error, Codes.FORBIDDEN)
+            spam_check = await self.spam_checker.check_event_for_spam(event)
+            if spam_check is not synapse.spam_checker_api.ALLOW:
+                raise SynapseError(
+                    403, "This message had been rejected as probable spam", spam_check
+                )
 
             ev = await self.handle_new_client_event(
                 requester=requester,
@@ -1001,7 +1006,7 @@ class EventCreationHandler:
         # after it is created
         if builder.internal_metadata.outlier:
             event.internal_metadata.outlier = True
-            context = EventContext.for_outlier()
+            context = EventContext.for_outlier(self.storage)
         elif (
             event.type == EventTypes.MSC2716_INSERTION
             and state_event_ids

synapse/handlers/pagination.py

@@ -448,7 +448,7 @@ class PaginationHandler:
             )
             # We expect `/messages` to use historic pagination tokens by default but
             # `/messages` should still works with live tokens when manually provided.
-            assert from_token.room_key.topological
+            assert from_token.room_key.topological is not None
 
         if pagin_config.limit is None:
             # This shouldn't happen as we've set a default limit before this

synapse/handlers/room.py

@@ -33,6 +33,7 @@ from typing import (
 import attr
 from typing_extensions import TypedDict
 
+import synapse
 from synapse.api.constants import (
     EventContentFields,
     EventTypes,
@@ -407,9 +408,10 @@ class RoomCreationHandler:
         """
         user_id = requester.user.to_string()
 
-        if not await self.spam_checker.user_may_create_room(user_id):
+        spam_check = await self.spam_checker.user_may_create_room(user_id)
+        if spam_check is not synapse.spam_checker_api.ALLOW:
             raise SynapseError(
-                403, "You are not permitted to create rooms", Codes.FORBIDDEN
+                403, "This room creation request has been rejected", spam_check
             )
 
         creation_content: JsonDict = {

synapse/handlers/room_member.py

@@ -18,6 +18,7 @@ import random
 from http import HTTPStatus
 from typing import TYPE_CHECKING, Iterable, List, Optional, Set, Tuple
 
+import synapse
 from synapse import types
 from synapse.api.constants import (
     AccountDataTypes,
@@ -679,8 +680,6 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
             if target_id == self._server_notices_mxid:
                 raise SynapseError(HTTPStatus.FORBIDDEN, "Cannot invite this user")
 
-            block_invite = False
-
             if (
                 self._server_notices_mxid is not None
                 and requester.user.to_string() == self._server_notices_mxid
@@ -697,16 +696,18 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
                         "Blocking invite: user is not admin and non-admin "
                         "invites disabled"
                     )
-                    block_invite = True
+                    raise SynapseError(403, "Invites have been disabled on this server")
 
-                if not await self.spam_checker.user_may_invite(
+                spam_check = await self.spam_checker.user_may_invite(
                     requester.user.to_string(), target_id, room_id
-                ):
+                )
+                if spam_check is not synapse.spam_checker_api.ALLOW:
                     logger.info("Blocking invite due to spam checker")
-                    block_invite = True
-
-            if block_invite:
-                raise SynapseError(403, "Invites have been disabled on this server")
+                    raise SynapseError(
+                        403,
+                        "This invite has been rejected as probable spam",
+                        spam_check,
+                    )
 
         # An empty prev_events list is allowed as long as the auth_event_ids are present
         if prev_event_ids is not None:
@@ -814,11 +815,14 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
                 # We assume that if the spam checker allowed the user to create
                 # a room then they're allowed to join it.
                 and not new_room
-                and not await self.spam_checker.user_may_join_room(
+            ):
+                spam_check = await self.spam_checker.user_may_join_room(
                     target.to_string(), room_id, is_invited=inviter is not None
                 )
-            ):
-                raise SynapseError(403, "Not allowed to join this room")
+                if spam_check is not synapse.spam_checker_api.ALLOW:
+                    raise SynapseError(
+                        403, "This request to join room has been rejected", spam_check
+                    )
 
             # Check if a remote join should be performed.
             remote_join, remote_room_hosts = await self._should_perform_remote_join(
@@ -1372,13 +1376,14 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
             )
         else:
             # Check if the spamchecker(s) allow this invite to go through.
-            if not await self.spam_checker.user_may_send_3pid_invite(
+            spam_check = await self.spam_checker.user_may_send_3pid_invite(
                 inviter_userid=requester.user.to_string(),
                 medium=medium,
                 address=address,
                 room_id=room_id,
-            ):
-                raise SynapseError(403, "Cannot send threepid invite")
+            )
+            if spam_check is not synapse.spam_checker_api.ALLOW:
+                raise SynapseError(403, "Cannot send threepid invite", spam_check)
 
             stream_id = await self._make_and_store_3pid_invite(
                 requester,

synapse/handlers/user_directory.py

@@ -60,7 +60,7 @@ class UserDirectoryHandler(StateDeltasHandler):
         self.clock = hs.get_clock()
         self.notifier = hs.get_notifier()
         self.is_mine_id = hs.is_mine_id
-        self.update_user_directory = hs.config.server.update_user_directory
+        self.update_user_directory = hs.config.worker.should_update_user_directory
         self.search_all_users = hs.config.userdirectory.user_directory_search_all_users
         self.spam_checker = hs.get_spam_checker()
         # The current position in the current_state_delta stream
@@ -100,7 +100,8 @@ class UserDirectoryHandler(StateDeltasHandler):
         # Remove any spammy users from the results.
         non_spammy_users = []
         for user in results["results"]:
-            if not await self.spam_checker.check_username_for_spam(user):
+            spam_check = await self.spam_checker.check_username_for_spam(user)
+            if spam_check is synapse.spam_checker_api.ALLOW:
                 non_spammy_users.append(user)
         results["results"] = non_spammy_users
 

synapse/http/matrixfederationclient.py

@@ -73,7 +73,7 @@ from synapse.logging.context import make_deferred_yieldable, run_in_background
 from synapse.logging.opentracing import set_tag, start_active_span, tags
 from synapse.types import JsonDict
 from synapse.util import json_decoder
-from synapse.util.async_helpers import timeout_deferred
+from synapse.util.async_helpers import AwakenableSleeper, timeout_deferred
 from synapse.util.metrics import Measure
 
 if TYPE_CHECKING:
@@ -353,6 +353,13 @@ class MatrixFederationHttpClient:
 
         self._cooperator = Cooperator(scheduler=schedule)
 
+        self._sleeper = AwakenableSleeper(self.reactor)
+
+    def wake_destination(self, destination: str) -> None:
+        """Called when the remote server may have come back online."""
+
+        self._sleeper.wake(destination)
+
     async def _send_request_with_optional_trailing_slash(
         self,
         request: MatrixFederationRequest,
@@ -474,6 +481,8 @@ class MatrixFederationHttpClient:
             self._store,
             backoff_on_404=backoff_on_404,
             ignore_backoff=ignore_backoff,
+            notifier=self.hs.get_notifier(),
+            replication_client=self.hs.get_replication_command_handler(),
         )
 
         method_bytes = request.method.encode("ascii")
@@ -664,7 +673,9 @@ class MatrixFederationHttpClient:
                             delay,
                         )
 
-                        await self.clock.sleep(delay)
+                        # Sleep for the calculated delay, or wake up immediately
+                        # if we get notified that the server is back up.
+                        await self._sleeper.sleep(request.destination, delay * 1000)
                         retries_left -= 1
                     else:
                         raise

synapse/http/server.py

@@ -33,6 +33,7 @@ from typing import (
     Optional,
     Pattern,
     Tuple,
+    TypeVar,
     Union,
 )
 
@@ -92,6 +93,66 @@ HTML_ERROR_TEMPLATE = """<!DOCTYPE html>
 HTTP_STATUS_REQUEST_CANCELLED = 499
 
 
+F = TypeVar("F", bound=Callable[..., Any])
+
+
+_cancellable_method_names = frozenset(
+    {
+        # `RestServlet`, `BaseFederationServlet` and `BaseFederationServerServlet`
+        # methods
+        "on_GET",
+        "on_PUT",
+        "on_POST",
+        "on_DELETE",
+        # `_AsyncResource`, `DirectServeHtmlResource` and `DirectServeJsonResource`
+        # methods
+        "_async_render_GET",
+        "_async_render_PUT",
+        "_async_render_POST",
+        "_async_render_DELETE",
+        "_async_render_OPTIONS",
+        # `ReplicationEndpoint` methods
+        "_handle_request",
+    }
+)
+
+
+def cancellable(method: F) -> F:
+    """Marks a servlet method as cancellable.
+
+    Methods with this decorator will be cancelled if the client disconnects before we
+    finish processing the request.
+
+    During cancellation, `Deferred.cancel()` will be invoked on the `Deferred` wrapping
+    the method. The `cancel()` call will propagate down to the `Deferred` that is
+    currently being waited on. That `Deferred` will raise a `CancelledError`, which will
+    propagate up, as per normal exception handling.
+
+    Before applying this decorator to a new endpoint, you MUST recursively check
+    that all `await`s in the function are on `async` functions or `Deferred`s that
+    handle cancellation cleanly, otherwise a variety of bugs may occur, ranging from
+    premature logging context closure, to stuck requests, to database corruption.
+
+    Usage:
+        class SomeServlet(RestServlet):
+            @cancellable
+            async def on_GET(self, request: SynapseRequest) -> ...:
+                ...
+    """
+    if method.__name__ not in _cancellable_method_names:
+        raise ValueError(
+            "@cancellable decorator can only be applied to servlet methods."
+        )
+
+    method.cancellable = True  # type: ignore[attr-defined]
+    return method
+
+
+def is_method_cancellable(method: Callable[..., Any]) -> bool:
+    """Checks whether a servlet method has the `@cancellable` flag."""
+    return getattr(method, "cancellable", False)
+
+
 def return_json_error(f: failure.Failure, request: SynapseRequest) -> None:
     """Sends a JSON error response to clients."""
 
@@ -283,7 +344,9 @@ class _AsyncResource(resource.Resource, metaclass=abc.ABCMeta):
 
     def render(self, request: SynapseRequest) -> int:
         """This gets called by twisted every time someone sends us a request."""
-        defer.ensureDeferred(self._async_render_wrapper(request))
+        request.render_deferred = defer.ensureDeferred(
+            self._async_render_wrapper(request)
+        )
         return NOT_DONE_YET
 
     @wrap_async_request_handler

synapse/http/site.py

@@ -19,6 +19,7 @@ from typing import TYPE_CHECKING, Any, Generator, Optional, Tuple, Union
 import attr
 from zope.interface import implementer
 
+from twisted.internet.defer import Deferred
 from twisted.internet.interfaces import IAddress, IReactorTime
 from twisted.python.failure import Failure
 from twisted.web.http import HTTPChannel
@@ -91,6 +92,14 @@ class SynapseRequest(Request):
         # we can't yet create the logcontext, as we don't know the method.
         self.logcontext: Optional[LoggingContext] = None
 
+        # The `Deferred` to cancel if the client disconnects early and
+        # `is_render_cancellable` is set. Expected to be set by `Resource.render`.
+        self.render_deferred: Optional["Deferred[None]"] = None
+        # A boolean indicating whether `render_deferred` should be cancelled if the
+        # client disconnects early. Expected to be set by the coroutine started by
+        # `Resource.render`, if rendering is asynchronous.
+        self.is_render_cancellable = False
+
         global _next_request_seq
         self.request_seq = _next_request_seq
         _next_request_seq += 1
@@ -357,7 +366,21 @@ class SynapseRequest(Request):
                     {"event": "client connection lost", "reason": str(reason.value)}
                 )
 
-            if not self._is_processing:
+            if self._is_processing:
+                if self.is_render_cancellable:
+                    if self.render_deferred is not None:
+                        # Throw a cancellation into the request processing, in the hope
+                        # that it will finish up sooner than it normally would.
+                        # The `self.processing()` context manager will call
+                        # `_finished_processing()` when done.
+                        with PreserveLoggingContext():
+                            self.render_deferred.cancel()
+                    else:
+                        logger.error(
+                            "Connection from client lost, but have no Deferred to "
+                            "cancel even though the request is marked as cancellable."
+                        )
+            else:
                 self._finished_processing()
 
     def _started_processing(self, servlet_name: str) -> None:

synapse/notifier.py

@@ -228,9 +228,7 @@ class Notifier:
         # Called when there are new things to stream over replication
         self.replication_callbacks: List[Callable[[], None]] = []
 
-        # Called when remote servers have come back online after having been
-        # down.
-        self.remote_server_up_callbacks: List[Callable[[str], None]] = []
+        self._federation_client = hs.get_federation_http_client()
 
         self._third_party_rules = hs.get_third_party_event_rules()
 
@@ -731,3 +729,7 @@ class Notifier:
         # circular dependencies.
         if self.federation_sender:
             self.federation_sender.wake_destination(server)
+
+        # Tell the federation client about the fact the server is back up, so
+        # that any in flight requests can be immediately retried.
+        self._federation_client.wake_destination(server)

synapse/push/__init__.py

@@ -12,6 +12,85 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+"""
+This module implements the push rules & notifications portion of the Matrix
+specification.
+
+There's a few related features:
+
+* Push notifications (i.e. email or outgoing requests to a Push Gateway).
+* Calculation of unread notifications (for /sync and /notifications).
+
+When Synapse receives a new event (locally, via the Client-Server API, or via
+federation), the following occurs:
+
+1. The push rules get evaluated to generate a set of per-user actions.
+2. The event is persisted into the database.
+3. (In the background) The notifier is notified about the new event.
+
+The per-user actions are initially stored in the event_push_actions_staging table,
+before getting moved into the event_push_actions table when the event is persisted.
+The event_push_actions table is periodically summarised into the event_push_summary
+and event_push_summary_stream_ordering tables.
+
+Since push actions block an event from being persisted the generation of push
+actions is performance sensitive.
+
+The general interaction of the classes are:
+
+        +---------------------------------------------+
+        | FederationEventHandler/EventCreationHandler |
+        +---------------------------------------------+
+                |
+                v
+        +-----------------+
+        | ActionGenerator |
+        +-----------------+
+                |
+                v
+        +-----------------------+     +---------------------------+
+        | BulkPushRuleEvaluator |---->| PushRuleEvaluatorForEvent |
+        +-----------------------+     +---------------------------+
+                |
+                v
+        +-----------------------------+
+        | EventPushActionsWorkerStore |
+        +-----------------------------+
+
+The notifier notifies the pusher pool of the new event, which checks for affected
+users. Each user-configured pusher of the affected users then performs the
+previously calculated action.
+
+The general interaction of the classes are:
+
+        +----------+
+        | Notifier |
+        +----------+
+                |
+                v
+        +------------+     +--------------+
+        | PusherPool |---->| PusherConfig |
+        +------------+     +--------------+
+                |
+                |     +---------------+
+                +<--->| PusherFactory |
+                |     +---------------+
+                v
+        +------------------------+     +-----------------------------------------------+
+        | EmailPusher/HttpPusher |---->| EventPushActionsWorkerStore/PusherWorkerStore |
+        +------------------------+     +-----------------------------------------------+
+                |
+                v
+        +-------------------------+
+        | Mailer/SimpleHttpClient |
+        +-------------------------+
+
+The Pusher instance also calls out to various utilities for generating payloads
+(or email templates), but those interactions are not detailed in this diagram
+(and are specific to the type of pusher).
+
+"""
+
 import abc
 from typing import TYPE_CHECKING, Any, Dict, Optional
 

synapse/push/action_generator.py

@@ -40,5 +40,9 @@ class ActionGenerator:
     async def handle_push_actions_for_event(
         self, event: EventBase, context: EventContext
     ) -> None:
+        if event.internal_metadata.is_outlier():
+            # This can happen due to out of band memberships
+            return
+
         with Measure(self.clock, "action_for_event_by_user"):
             await self.bulk_evaluator.action_for_event_by_user(event, context)