1.65.0

In state res v2, we apply two passes of iterative auth checks. The first
pass replays power events and events in their auth chains, but only
those belonging to the full conflicted set. The source code as written
suggests that we want only those belonging to the auth difference (which
is a smaller set of events).

At runtime we were doing the correct thing anyway, because the only
callsite of `_reverse_topological_power_sort` passes in the
`full_conflicted_set`. So this really is just a rename.

.ci/scripts/test_old_deps.sh

@@ -69,7 +69,7 @@ with open('pyproject.toml', 'w') as f:
 "
 python3 -c "$REMOVE_DEV_DEPENDENCIES"
 
-pipx install poetry==1.1.12
+pipx install poetry==1.1.14
 ~/.local/bin/poetry lock
 
 echo "::group::Patched pyproject.toml"

.git-blame-ignore-revs

@@ -1,3 +1,16 @@
+# Commits in this file will be removed from GitHub blame results.
+#
+# To use this file locally, use:
+#   git blame --ignore-revs-file="path/to/.git-blame-ignore-revs" <files>
+#
+# or configure the `blame.ignoreRevsFile` option in your git config.
+#
+# If ignoring a pull request that was not squash merged, only the merge
+# commit needs to be put here. Child commits will be resolved from it.
+
+# Run black (#3679).
+8b3d9b6b199abb87246f982d5db356f1966db925
+
 # Black reformatting (#5482).
 32e7c9e7f20b57dd081023ac42d6931a8da9b3a3
 

.github/workflows/latest_deps.yml

@@ -135,11 +135,42 @@ jobs:
             /logs/**/*.log*
 
 
-  # TODO: run complement (as with twisted trunk, see #12473).
+  complement:
+    if: "${{ !failure() && !cancelled() }}"
+    runs-on: ubuntu-latest
 
-  # open an issue if the build fails, so we know about it.
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - arrangement: monolith
+            database: SQLite
+
+          - arrangement: monolith
+            database: Postgres
+
+          - arrangement: workers
+            database: Postgres
+
+    steps:
+      - name: Run actions/checkout@v2 for synapse
+        uses: actions/checkout@v2
+        with:
+          path: synapse
+
+      - name: Prepare Complement's Prerequisites
+        run: synapse/.ci/scripts/setup_complement_prerequisites.sh
+
+      - run: |
+          set -o pipefail
+          TEST_ONLY_IGNORE_POETRY_LOCKFILE=1 POSTGRES=${{ (matrix.database == 'Postgres') && 1 || '' }} WORKERS=${{ (matrix.arrangement == 'workers') && 1 || '' }} COMPLEMENT_DIR=`pwd`/complement synapse/scripts-dev/complement.sh -json 2>&1 | gotestfmt
+        shell: bash
+        name: Run Complement Tests
+
+  # Open an issue if the build fails, so we know about it.
+  # Only do this if we're not experimenting with this action in a PR.
   open-issue:
-    if: failure()
+    if: "failure() && github.event_name != 'push' && github.event_name != 'pull_request'"
     needs:
       # TODO: should mypy be included here? It feels more brittle than the other two.
       - mypy

.github/workflows/tests.yml

@@ -328,6 +328,9 @@ jobs:
           - arrangement: monolith
             database: Postgres
 
+          - arrangement: workers
+            database: Postgres
+
     steps:
       - name: Run actions/checkout@v2 for synapse
         uses: actions/checkout@v2
@@ -343,30 +346,6 @@ jobs:
         shell: bash
         name: Run Complement Tests
 
-  # XXX When complement with workers is stable, move this back into the standard
-  #     "complement" matrix above.
-  #
-  # See https://github.com/matrix-org/synapse/issues/13161
-  complement-workers:
-    if: "${{ !failure() && !cancelled() }}"
-    needs: linting-done
-    runs-on: ubuntu-latest
-
-    steps:
-      - name: Run actions/checkout@v2 for synapse
-        uses: actions/checkout@v2
-        with:
-          path: synapse
-
-      - name: Prepare Complement's Prerequisites
-        run: synapse/.ci/scripts/setup_complement_prerequisites.sh
-
-      - run: |
-          set -o pipefail
-          POSTGRES=1 WORKERS=1 COMPLEMENT_DIR=`pwd`/complement synapse/scripts-dev/complement.sh -json 2>&1 | gotestfmt
-        shell: bash
-        name: Run Complement Tests
-
   # a job which marks all the other jobs as complete, thus allowing PRs to be merged.
   tests-done:
     if: ${{ always() }}

.github/workflows/twisted_trunk.yml

@@ -127,12 +127,12 @@ jobs:
         run: |
           set -x
           DEBIAN_FRONTEND=noninteractive sudo apt-get install -yqq python3 pipx
-          pipx install poetry==1.1.12
+          pipx install poetry==1.1.14
 
           poetry remove -n twisted
           poetry add -n --extras tls git+https://github.com/twisted/twisted.git#trunk
           poetry lock --no-update
-          # NOT IN 1.1.12 poetry lock --check
+          # NOT IN 1.1.14 poetry lock --check
         working-directory: synapse
 
       - run: |

CHANGES.md

@@ -1,7 +1,222 @@
-Synapse vNext
-=============
+Synapse 1.65.0 (2022-08-16)
+===========================
+
+No significant changes since 1.65.0rc2.
+
+
+Synapse 1.65.0rc2 (2022-08-11)
+==============================
+
+Internal Changes
+----------------
+
+- Revert 'Remove the unspecced `room_id` field in the `/hierarchy` response. ([\#13365](https://github.com/matrix-org/synapse/issues/13365))' to give more time for clients to update. ([\#13501](https://github.com/matrix-org/synapse/issues/13501))
+
+
+Synapse 1.65.0rc1 (2022-08-09)
+==============================
+
+Features
+--------
+
+- Add support for stable prefixes for [MSC2285 (private read receipts)](https://github.com/matrix-org/matrix-spec-proposals/pull/2285). ([\#13273](https://github.com/matrix-org/synapse/issues/13273))
+- Add new unstable error codes `ORG.MATRIX.MSC3848.ALREADY_JOINED`, `ORG.MATRIX.MSC3848.NOT_JOINED`, and `ORG.MATRIX.MSC3848.INSUFFICIENT_POWER` described in [MSC3848](https://github.com/matrix-org/matrix-spec-proposals/pull/3848). ([\#13343](https://github.com/matrix-org/synapse/issues/13343))
+- Use stable prefixes for [MSC3827](https://github.com/matrix-org/matrix-spec-proposals/pull/3827). ([\#13370](https://github.com/matrix-org/synapse/issues/13370))
+- Add a new module API method to translate a room alias into a room ID. ([\#13428](https://github.com/matrix-org/synapse/issues/13428))
+- Add a new module API method to create a room. ([\#13429](https://github.com/matrix-org/synapse/issues/13429))
+- Add remote join capability to the module API's `update_room_membership` method (in a backwards compatible manner). ([\#13441](https://github.com/matrix-org/synapse/issues/13441))
+
+
+Bugfixes
+--------
+
+- Update the version of the LDAP3 auth provider module included in the `matrixdotorg/synapse` DockerHub images and the Debian packages hosted on packages.matrix.org to 0.2.2. This version fixes a regression in the module. ([\#13470](https://github.com/matrix-org/synapse/issues/13470))
+- Fix a bug introduced in Synapse v1.41.0 where the `/hierarchy` API returned non-standard information (a `room_id` field under each entry in `children_state`). ([\#13365](https://github.com/matrix-org/synapse/issues/13365))
+- Fix a bug introduced in Synapse 0.24.0 that would respond with the wrong error status code to `/joined_members` requests when the requester is not a current member of the room. Contributed by @andrewdoh. ([\#13374](https://github.com/matrix-org/synapse/issues/13374))
+- Fix bug in handling of typing events for appservices. Contributed by Nick @ Beeper (@fizzadar). ([\#13392](https://github.com/matrix-org/synapse/issues/13392))
+- Fix a bug introduced in Synapse 1.57.0 where rooms listed in `exclude_rooms_from_sync` in the configuration file would not be properly excluded from incremental syncs. ([\#13408](https://github.com/matrix-org/synapse/issues/13408))
+- Fix a bug in the experimental faster-room-joins support which could cause it to get stuck in an infinite loop. ([\#13353](https://github.com/matrix-org/synapse/issues/13353))
+- Faster room joins: fix a bug which caused rejected events to become un-rejected during state syncing. ([\#13413](https://github.com/matrix-org/synapse/issues/13413))
+- Faster room joins: fix error when running out of servers to sync partial state with, so that Synapse raises the intended error instead. ([\#13432](https://github.com/matrix-org/synapse/issues/13432))
+
+
+Updates to the Docker image
+---------------------------
+
+- Make Docker images build on armv7 by installing cryptography dependencies in the 'requirements' stage. Contributed by Jasper Spaans. ([\#13372](https://github.com/matrix-org/synapse/issues/13372))
+
+
+Improved Documentation
+----------------------
+
+- Update the 'registration tokens' page to acknowledge that the relevant MSC was merged into version 1.2 of the Matrix specification. Contributed by @moan0s. ([\#11897](https://github.com/matrix-org/synapse/issues/11897))
+- Document which HTTP resources support gzip compression. ([\#13221](https://github.com/matrix-org/synapse/issues/13221))
+- Add steps describing how to elevate an existing user to administrator by manipulating the database. ([\#13230](https://github.com/matrix-org/synapse/issues/13230))
+- Fix wrong headline for `url_preview_accept_language` in documentation. ([\#13437](https://github.com/matrix-org/synapse/issues/13437))
+- Remove redundant 'Contents' section from the Configuration Manual. Contributed by @dklimpel. ([\#13438](https://github.com/matrix-org/synapse/issues/13438))
+- Update documentation for config setting `macaroon_secret_key`. ([\#13443](https://github.com/matrix-org/synapse/issues/13443))
+- Update outdated information on `sso_mapping_providers` documentation. ([\#13449](https://github.com/matrix-org/synapse/issues/13449))
+- Fix example code in module documentation of `password_auth_provider_callbacks`. ([\#13450](https://github.com/matrix-org/synapse/issues/13450))
+- Make the configuration for the cache clearer. ([\#13481](https://github.com/matrix-org/synapse/issues/13481))
+
+
+Internal Changes
+----------------
+
+- Extend the release script to automatically push a new SyTest branch, rather than having that be a manual process. ([\#12978](https://github.com/matrix-org/synapse/issues/12978))
+- Make minor clarifications to the error messages given when we fail to join a room via any server. ([\#13160](https://github.com/matrix-org/synapse/issues/13160))
+- Enable Complement CI tests in the 'latest deps' test run. ([\#13213](https://github.com/matrix-org/synapse/issues/13213))
+- Fix long-standing bugged logic which was never hit in `get_pdu` asking every remote destination even after it finds an event. ([\#13346](https://github.com/matrix-org/synapse/issues/13346))
+- Faster room joins: avoid blocking when pulling events with partially missing prev events. ([\#13355](https://github.com/matrix-org/synapse/issues/13355))
+- Instrument `/messages` for understandable traces in Jaeger. ([\#13368](https://github.com/matrix-org/synapse/issues/13368))
+- Remove an unused argument to `get_relations_for_event`. ([\#13383](https://github.com/matrix-org/synapse/issues/13383))
+- Add a `merge-back` command to the release script, which automates merging the correct branches after a release. ([\#13393](https://github.com/matrix-org/synapse/issues/13393))
+- Adding missing type hints to tests. ([\#13397](https://github.com/matrix-org/synapse/issues/13397))
+- Faster Room Joins: don't leave a stuck room partial state flag if the join fails. ([\#13403](https://github.com/matrix-org/synapse/issues/13403))
+- Refactor `_resolve_state_at_missing_prevs` to compute an `EventContext` instead. ([\#13404](https://github.com/matrix-org/synapse/issues/13404), [\#13431](https://github.com/matrix-org/synapse/issues/13431))
+- Faster Room Joins: prevent Synapse from answering federated join requests for a room which it has not fully joined yet. ([\#13416](https://github.com/matrix-org/synapse/issues/13416))
+- Re-enable running Complement tests against Synapse with workers. ([\#13420](https://github.com/matrix-org/synapse/issues/13420))
+- Prevent unnecessary lookups to any external `get_event` cache. Contributed by Nick @ Beeper (@fizzadar). ([\#13435](https://github.com/matrix-org/synapse/issues/13435))
+- Add some tracing to give more insight into local room joins. ([\#13439](https://github.com/matrix-org/synapse/issues/13439))
+- Rename class `RateLimitConfig` to `RatelimitSettings` and `FederationRateLimitConfig` to `FederationRatelimitSettings`. ([\#13442](https://github.com/matrix-org/synapse/issues/13442))
+- Add some comments about how event push actions are stored. ([\#13445](https://github.com/matrix-org/synapse/issues/13445), [\#13455](https://github.com/matrix-org/synapse/issues/13455))
+- Improve rebuild speed for the "synapse-workers" docker image. ([\#13447](https://github.com/matrix-org/synapse/issues/13447))
+- Fix `@tag_args` being off-by-one with the arguments when tagging a span (tracing). ([\#13452](https://github.com/matrix-org/synapse/issues/13452))
+- Update type of `EventContext.rejected`. ([\#13460](https://github.com/matrix-org/synapse/issues/13460))
+- Use literals in place of `HTTPStatus` constants in tests. ([\#13463](https://github.com/matrix-org/synapse/issues/13463), [\#13469](https://github.com/matrix-org/synapse/issues/13469))
+- Correct a misnamed argument in state res v2 internals. ([\#13467](https://github.com/matrix-org/synapse/issues/13467))
+
+
+Synapse 1.64.0 (2022-08-02)
+===========================
+
+No significant changes since 1.64.0rc2.
+
+
+Deprecation Warning
+-------------------
+
+Synapse v1.66.0 will remove the ability to delegate the tasks of verifying email address ownership, and password reset confirmation, to an identity server.
+
+If you require your homeserver to verify e-mail addresses or to support password resets via e-mail, please configure your homeserver with SMTP access so that it can send e-mails on its own behalf.
+[Consult the configuration documentation for more information.](https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html#email)
+
+
+Synapse 1.64.0rc2 (2022-07-29)
+==============================
+
+This RC reintroduces support for `account_threepid_delegates.email`, which was removed in 1.64.0rc1. It remains deprecated and will be removed altogether in Synapse v1.66.0. ([\#13406](https://github.com/matrix-org/synapse/issues/13406))
+
+
+Synapse 1.64.0rc1 (2022-07-26)
+==============================
+
+This RC removed the ability to delegate the tasks of verifying email address ownership, and password reset confirmation, to an identity server.
+
+We have also stopped building `.deb` packages for Ubuntu 21.10 as it is no longer an active version of Ubuntu.
+
+
+Features
+--------
+
+- Improve error messages when media thumbnails cannot be served. ([\#13038](https://github.com/matrix-org/synapse/issues/13038))
+- Allow pagination from remote event after discovering it from [MSC3030](https://github.com/matrix-org/matrix-spec-proposals/pull/3030) `/timestamp_to_event`. ([\#13205](https://github.com/matrix-org/synapse/issues/13205))
+- Add a `room_type` field in the responses for the list room and room details admin APIs. Contributed by @andrewdoh. ([\#13208](https://github.com/matrix-org/synapse/issues/13208))
+- Add support for room version 10. ([\#13220](https://github.com/matrix-org/synapse/issues/13220))
+- Add per-room rate limiting for room joins. For each room, Synapse now monitors the rate of join events in that room, and throttles additional joins if that rate grows too large. ([\#13253](https://github.com/matrix-org/synapse/issues/13253), [\#13254](https://github.com/matrix-org/synapse/issues/13254), [\#13255](https://github.com/matrix-org/synapse/issues/13255), [\#13276](https://github.com/matrix-org/synapse/issues/13276))
+- Support Implicit TLS (TLS without using a STARTTLS upgrade, typically on port 465) for sending emails, enabled by the new option `force_tls`. Contributed by Jan Schär. ([\#13317](https://github.com/matrix-org/synapse/issues/13317))
+
+
+Bugfixes
+--------
+
+- Fix a bug introduced in Synapse 1.15.0 where adding a user through the Synapse Admin API with a phone number would fail if the `enable_email_notifs` and `email_notifs_for_new_users` options were enabled. Contributed by @thomasweston12. ([\#13263](https://github.com/matrix-org/synapse/issues/13263))
+- Fix a bug introduced in Synapse 1.40.0 where a user invited to a restricted room would be briefly unable to join. ([\#13270](https://github.com/matrix-org/synapse/issues/13270))
+- Fix a long-standing bug where, in rare instances, Synapse could store the incorrect state for a room after a state resolution. ([\#13278](https://github.com/matrix-org/synapse/issues/13278))
+- Fix a bug introduced in v1.18.0 where the `synapse_pushers` metric would overcount pushers when they are replaced. ([\#13296](https://github.com/matrix-org/synapse/issues/13296))
+- Disable autocorrection and autocapitalisation on the username text field shown during registration when using SSO. ([\#13350](https://github.com/matrix-org/synapse/issues/13350))
+- Update locked version of `frozendict` to 2.3.3, which has fixes for memory leaks affecting `/sync`. ([\#13284](https://github.com/matrix-org/synapse/issues/13284), [\#13352](https://github.com/matrix-org/synapse/issues/13352))
+
+
+Improved Documentation
+----------------------
+
+- Provide an example of using the Admin API. Contributed by @jejo86. ([\#13231](https://github.com/matrix-org/synapse/issues/13231))
+- Move the documentation for how URL previews work to the URL preview module. ([\#13233](https://github.com/matrix-org/synapse/issues/13233), [\#13261](https://github.com/matrix-org/synapse/issues/13261))
+- Add another `contrib` script to help set up worker processes. Contributed by @villepeh. ([\#13271](https://github.com/matrix-org/synapse/issues/13271))
+- Document that certain config options were added or changed in Synapse 1.62. Contributed by @behrmann. ([\#13314](https://github.com/matrix-org/synapse/issues/13314))
+- Document the new `rc_invites.per_issuer` throttling option added in Synapse 1.63. ([\#13333](https://github.com/matrix-org/synapse/issues/13333))
+- Mention that BuildKit is needed when building Docker images for tests. ([\#13338](https://github.com/matrix-org/synapse/issues/13338))
+- Improve Caddy reverse proxy documentation. ([\#13344](https://github.com/matrix-org/synapse/issues/13344))
+
+
+Deprecations and Removals
+-------------------------
+
+- Drop tables that were formerly used for groups/communities. ([\#12967](https://github.com/matrix-org/synapse/issues/12967))
+- Drop support for delegating email verification to an external server. ([\#13192](https://github.com/matrix-org/synapse/issues/13192))
+- Drop support for calling `/_matrix/client/v3/account/3pid/bind` without an `id_access_token`, which was not permitted by the spec. Contributed by @Vetchu. ([\#13239](https://github.com/matrix-org/synapse/issues/13239))
+- Stop building `.deb` packages for Ubuntu 21.10 (Impish Indri), which has reached end of life. ([\#13326](https://github.com/matrix-org/synapse/issues/13326))
+
+
+Internal Changes
+----------------
+
+- Use lower transaction isolation level when purging rooms to avoid serialization errors. Contributed by Nick @ Beeper. ([\#12942](https://github.com/matrix-org/synapse/issues/12942))
+- Remove code which incorrectly attempted to reconcile state with remote servers when processing incoming events. ([\#12943](https://github.com/matrix-org/synapse/issues/12943))
+- Make the AS login method call `Auth.get_user_by_req` for checking the AS token. ([\#13094](https://github.com/matrix-org/synapse/issues/13094))
+- Always use a version of canonicaljson that supports the C implementation of frozendict. ([\#13172](https://github.com/matrix-org/synapse/issues/13172))
+- Add prometheus counters for ephemeral events and to device messages pushed to app services. Contributed by Brad @ Beeper. ([\#13175](https://github.com/matrix-org/synapse/issues/13175))
+- Refactor receipts servlet logic to avoid duplicated code. ([\#13198](https://github.com/matrix-org/synapse/issues/13198))
+- Preparation for database schema simplifications: populate `state_key` and `rejection_reason` for existing rows in the `events` table. ([\#13215](https://github.com/matrix-org/synapse/issues/13215))
+- Remove unused database table `event_reference_hashes`. ([\#13218](https://github.com/matrix-org/synapse/issues/13218))
+- Further reduce queries used sending events when creating new rooms. Contributed by Nick @ Beeper (@fizzadar). ([\#13224](https://github.com/matrix-org/synapse/issues/13224))
+- Call the v2 identity service `/3pid/unbind` endpoint, rather than v1. Contributed by @Vetchu. ([\#13240](https://github.com/matrix-org/synapse/issues/13240))
+- Use an asynchronous cache wrapper for the get event cache. Contributed by Nick @ Beeper (@fizzadar). ([\#13242](https://github.com/matrix-org/synapse/issues/13242), [\#13308](https://github.com/matrix-org/synapse/issues/13308))
+- Optimise federation sender and appservice pusher event stream processing queries. Contributed by Nick @ Beeper (@fizzadar). ([\#13251](https://github.com/matrix-org/synapse/issues/13251))
+- Log the stack when waiting for an entire room to be un-partial stated. ([\#13257](https://github.com/matrix-org/synapse/issues/13257))
+- Fix spurious warning when fetching state after a missing prev event. ([\#13258](https://github.com/matrix-org/synapse/issues/13258))
+- Clean-up tests for notifications. ([\#13260](https://github.com/matrix-org/synapse/issues/13260))
+- Do not fail build if complement with workers fails. ([\#13266](https://github.com/matrix-org/synapse/issues/13266))
+- Don't pull out state in `compute_event_context` for unconflicted state. ([\#13267](https://github.com/matrix-org/synapse/issues/13267), [\#13274](https://github.com/matrix-org/synapse/issues/13274))
+- Reduce the rebuild time for the complement-synapse docker image. ([\#13279](https://github.com/matrix-org/synapse/issues/13279))
+- Don't pull out the full state when creating an event. ([\#13281](https://github.com/matrix-org/synapse/issues/13281), [\#13307](https://github.com/matrix-org/synapse/issues/13307))
+- Upgrade from Poetry 1.1.12 to 1.1.14, to fix bugs when locking packages. ([\#13285](https://github.com/matrix-org/synapse/issues/13285))
+- Make `DictionaryCache` expire full entries if they haven't been queried in a while, even if specific keys have been queried recently. ([\#13292](https://github.com/matrix-org/synapse/issues/13292))
+- Use `HTTPStatus` constants in place of literals in tests. ([\#13297](https://github.com/matrix-org/synapse/issues/13297))
+- Improve performance of query  `_get_subset_users_in_room_with_profiles`. ([\#13299](https://github.com/matrix-org/synapse/issues/13299))
+- Up batch size of `bulk_get_push_rules` and `_get_joined_profiles_from_event_ids`. ([\#13300](https://github.com/matrix-org/synapse/issues/13300))
+- Remove unnecessary `json.dumps` from tests. ([\#13303](https://github.com/matrix-org/synapse/issues/13303))
+- Reduce memory usage of sending dummy events. ([\#13310](https://github.com/matrix-org/synapse/issues/13310))
+- Prevent formatting changes of [#3679](https://github.com/matrix-org/synapse/pull/3679) from appearing in `git blame`. ([\#13311](https://github.com/matrix-org/synapse/issues/13311))
+- Change `get_users_in_room` and `get_rooms_for_user` caches to enable pruning of old entries. ([\#13313](https://github.com/matrix-org/synapse/issues/13313))
+- Validate federation destinations and log an error if a destination is invalid. ([\#13318](https://github.com/matrix-org/synapse/issues/13318))
+- Fix `FederationClient.get_pdu()` returning events from the cache as `outliers` instead of original events we saw over federation. ([\#13320](https://github.com/matrix-org/synapse/issues/13320))
+- Reduce memory usage of state caches. ([\#13323](https://github.com/matrix-org/synapse/issues/13323))
+- Reduce the amount of state we store in the `state_cache`. ([\#13324](https://github.com/matrix-org/synapse/issues/13324))
+- Add missing type hints to open tracing module. ([\#13328](https://github.com/matrix-org/synapse/issues/13328), [\#13345](https://github.com/matrix-org/synapse/issues/13345), [\#13362](https://github.com/matrix-org/synapse/issues/13362))
+- Remove old base slaved store and de-duplicate cache ID generators. Contributed by Nick @ Beeper (@fizzadar). ([\#13329](https://github.com/matrix-org/synapse/issues/13329), [\#13349](https://github.com/matrix-org/synapse/issues/13349))
+- When reporting metrics is enabled, use ~8x less data to describe DB transaction metrics. ([\#13342](https://github.com/matrix-org/synapse/issues/13342))
+- Faster room joins: skip soft fail checks while Synapse only has partial room state, since the current membership of event senders may not be accurately known. ([\#13354](https://github.com/matrix-org/synapse/issues/13354))
+
+
+Synapse 1.63.1 (2022-07-20)
+===========================
+
+Bugfixes
+--------
+
+- Fix a bug introduced in Synapse 1.63.0 where push actions were incorrectly calculated for appservice users. This caused performance issues on servers with large numbers of appservices. ([\#13332](https://github.com/matrix-org/synapse/issues/13332))
+
+
+Synapse 1.63.0 (2022-07-19)
+===========================
+
+Improved Documentation
+----------------------
+
+- Clarify that homeserver server names are included in the reported data when the `report_stats` config option is enabled. ([\#13321](https://github.com/matrix-org/synapse/issues/13321))
 
-As of this release, Synapse no longer allows the tasks of verifying email address ownership, and password reset confirmation, to be delegated to an identity server. For more information, see the [upgrade notes](https://matrix-org.github.io/synapse/v1.64/upgrade.html#upgrading-to-v1640).
 
 Synapse 1.63.0rc1 (2022-07-12)
 ==============================
@@ -11,7 +226,7 @@ Features
 
 - Add a rate limit for local users sending invites. ([\#13125](https://github.com/matrix-org/synapse/issues/13125))
 - Implement [MSC3827](https://github.com/matrix-org/matrix-spec-proposals/pull/3827): Filtering of `/publicRooms` by room type. ([\#13031](https://github.com/matrix-org/synapse/issues/13031))
-- Improve validation logic in Synapse's REST endpoints. ([\#13148](https://github.com/matrix-org/synapse/issues/13148))
+- Improve validation logic in the account data REST endpoints. ([\#13148](https://github.com/matrix-org/synapse/issues/13148))
 
 
 Bugfixes
@@ -39,7 +254,7 @@ Improved Documentation
 - Add an explanation of the `--report-stats` argument to the docs. ([\#13029](https://github.com/matrix-org/synapse/issues/13029))
 - Add a helpful example bash script to the contrib directory for creating multiple worker configuration files of the same type. Contributed by @villepeh. ([\#13032](https://github.com/matrix-org/synapse/issues/13032))
 - Add missing links to config options. ([\#13166](https://github.com/matrix-org/synapse/issues/13166))
-- Add documentation for anonymised homeserver statistics collection. ([\#13086](https://github.com/matrix-org/synapse/issues/13086))
+- Add documentation for homeserver usage statistics collection. ([\#13086](https://github.com/matrix-org/synapse/issues/13086))
 - Add documentation for the existing `databases` option in the homeserver configuration manual. ([\#13212](https://github.com/matrix-org/synapse/issues/13212))
 - Clean up references to sample configuration and redirect users to the configuration manual instead. ([\#13077](https://github.com/matrix-org/synapse/issues/13077), [\#13139](https://github.com/matrix-org/synapse/issues/13139))
 - Document how the Synapse team does reviews. ([\#13132](https://github.com/matrix-org/synapse/issues/13132))
@@ -124,7 +339,7 @@ Bugfixes
 - Update [MSC3786](https://github.com/matrix-org/matrix-spec-proposals/pull/3786) implementation to check `state_key`. ([\#12939](https://github.com/matrix-org/synapse/issues/12939))
 - Fix a bug introduced in Synapse 1.58 where Synapse would not report full version information when installed from a git checkout. This is a best-effort affair and not guaranteed to be stable. ([\#12973](https://github.com/matrix-org/synapse/issues/12973))
 - Fix a bug introduced in Synapse 1.60 where Synapse would fail to start if the `sqlite3` module was not available. ([\#12979](https://github.com/matrix-org/synapse/issues/12979))
-- Fix a bug where non-standard information was required when requesting the `/hierarchy` API over federation. Introduced 
+- Fix a bug where non-standard information was required when requesting the `/hierarchy` API over federation. Introduced
   in Synapse v1.41.0. ([\#12991](https://github.com/matrix-org/synapse/issues/12991))
 - Fix a long-standing bug which meant that rate limiting was not restrictive enough in some cases. ([\#13018](https://github.com/matrix-org/synapse/issues/13018))
 - Fix a bug introduced in Synapse 1.58 where profile requests for a malformed user ID would ccause an internal error. Synapse now returns 400 Bad Request in this situation. ([\#13041](https://github.com/matrix-org/synapse/issues/13041))

changelog.d/12943.misc

@@ -1 +0,0 @@
-Remove code which incorrectly attempted to reconcile state with remote servers when processing incoming events.

changelog.d/12967.removal

@@ -1 +0,0 @@
-Drop tables used for groups/communities.

changelog.d/13094.misc

@@ -1 +0,0 @@
-Make the AS login method call `Auth.get_user_by_req` for checking the AS token.

changelog.d/13175.misc

@@ -1 +0,0 @@
-Add prometheus counters for ephemeral events and to device messages pushed to app services. Contributed by Brad @ Beeper.

changelog.d/13192.removal

@@ -1 +0,0 @@
-Drop support for delegating email verification to an external server.

changelog.d/13198.misc

@@ -1 +0,0 @@
-Refactor receipts servlet logic to avoid duplicated code.

changelog.d/13208.feature

@@ -1 +0,0 @@
-Add a `room_type` field in the responses for the list room and room details admin API. Contributed by @andrewdoh.

changelog.d/13215.misc

@@ -1 +0,0 @@
-Preparation for database schema simplifications: populate `state_key` and `rejection_reason` for existing rows in the `events` table.

changelog.d/13218.misc

@@ -1 +0,0 @@
-Remove unused database table `event_reference_hashes`.

changelog.d/13220.feature

@@ -1 +0,0 @@
-Add support for room version 10.

changelog.d/13224.misc

@@ -1 +0,0 @@
-Further reduce queries used sending events when creating new rooms. Contributed by Nick @ Beeper (@fizzadar).

changelog.d/13231.doc

@@ -1 +0,0 @@
-Provide an example of using the Admin API. Contributed by @jejo86.

changelog.d/13233.doc

@@ -1 +0,0 @@
-Move the documentation for how URL previews work to the URL preview module.

changelog.d/13239.removal

@@ -1 +0,0 @@
-Drop support for calling `/_matrix/client/v3/account/3pid/bind` without an `id_access_token`, which was not permitted by the spec. Contributed by @Vetchu.

changelog.d/13240.misc

@@ -1 +0,0 @@
-Call the v2 identity service `/3pid/unbind` endpoint, rather than v1.

changelog.d/13242.misc

@@ -1 +0,0 @@
-Use an asynchronous cache wrapper for the get event cache. Contributed by Nick @ Beeper (@fizzadar).

changelog.d/13251.misc

@@ -1 +0,0 @@
-Optimise federation sender and appservice pusher event stream processing queries. Contributed by Nick @ Beeper (@fizzadar).

changelog.d/13253.misc

@@ -1 +0,0 @@
-Preparatory work for a per-room rate limiter on joins.

changelog.d/13254.misc

@@ -1 +0,0 @@
-Preparatory work for a per-room rate limiter on joins.

changelog.d/13255.misc

@@ -1 +0,0 @@
-Preparatory work for a per-room rate limiter on joins.

changelog.d/13257.misc

@@ -1 +0,0 @@
-Log the stack when waiting for an entire room to be un-partial stated.

changelog.d/13260.misc

@@ -1 +0,0 @@
-Clean-up tests for notifications.

changelog.d/13261.doc

@@ -1 +0,0 @@
-Move the documentation for how URL previews work to the URL preview module.

changelog.d/13263.bugfix

@@ -1 +0,0 @@
-Fix a bug introduced in Synapse 1.15.0 where adding a user through the Synapse Admin API with a phone number would fail if the "enable_email_notifs" and "email_notifs_for_new_users" options were enabled. Contributed by @thomasweston12.

changelog.d/13266.misc

@@ -1 +0,0 @@
-Do not fail build if complement with workers fails.

changelog.d/13267.misc

@@ -1 +0,0 @@
-Don't pull out state in `compute_event_context` for unconflicted state.

changelog.d/13270.bugfix

@@ -1 +0,0 @@
-Fix a bug introduced in Synapse 1.40 where a user invited to a restricted room would be briefly unable to join.

changelog.d/13274.misc

@@ -1 +0,0 @@
-Don't pull out state in `compute_event_context` for unconflicted state.

changelog.d/13278.bugfix

@@ -1 +0,0 @@
-Fix long-standing bug where in rare instances Synapse could store the incorrect state for a room after a state resolution.

changelog.d/13279.misc

@@ -1 +0,0 @@
-Reduce the rebuild time for the complement-synapse docker image.

changelog.d/13284.misc

@@ -1 +0,0 @@
-Update locked version of `frozendict` to 2.3.2, which has a fix for a memory leak.

contrib/workers-bash-scripts/create-multiple-generic-workers.md

@@ -1,4 +1,4 @@
-# Creating multiple workers with a bash script
+# Creating multiple generic workers with a bash script
 
 Setting up multiple worker configuration files manually can be time-consuming.
 You can alternatively create multiple worker configuration files with a simple `bash` script. For example:

contrib/workers-bash-scripts/create-multiple-stream-writers.md

@@ -0,0 +1,145 @@
+# Creating multiple stream writers with a bash script
+
+This script creates multiple [stream writer](https://github.com/matrix-org/synapse/blob/develop/docs/workers.md#stream-writers) workers.
+
+Stream writers require both replication and HTTP listeners.
+
+It also prints out the example lines for Synapse main configuration file.
+
+Remember to route necessary endpoints directly to a worker associated with it.
+
+If you run the script as-is, it will create workers with the replication listener starting from port 8034 and another, regular http listener starting from 8044. If you don't need all of the stream writers listed in the script, just remove them from the ```STREAM_WRITERS``` array.
+
+```sh
+#!/bin/bash
+
+# Start with these replication and http ports.
+# The script loop starts with the exact port and then increments it by one.
+REP_START_PORT=8034
+HTTP_START_PORT=8044
+
+# Stream writer workers to generate. Feel free to add or remove them as you wish.
+# Event persister ("events") isn't included here as it does not require its
+# own HTTP listener.
+
+STREAM_WRITERS+=( "presence" "typing" "receipts" "to_device" "account_data" )
+
+NUM_WRITERS=$(expr ${#STREAM_WRITERS[@]})
+
+i=0
+
+while [ $i -lt "$NUM_WRITERS" ]
+do
+cat << EOF > ${STREAM_WRITERS[$i]}_stream_writer.yaml
+worker_app: synapse.app.generic_worker
+worker_name: ${STREAM_WRITERS[$i]}_stream_writer
+
+# The replication listener on the main synapse process.
+worker_replication_host: 127.0.0.1
+worker_replication_http_port: 9093
+
+worker_listeners:
+  - type: http
+    port: $(expr $REP_START_PORT + $i)
+    resources:
+      - names: [replication]
+
+  - type: http
+    port: $(expr $HTTP_START_PORT + $i)
+    resources:
+      - names: [client]
+
+worker_log_config: /etc/matrix-synapse/stream-writer-log.yaml
+EOF
+HOMESERVER_YAML_INSTANCE_MAP+=$"  ${STREAM_WRITERS[$i]}_stream_writer:
+    host: 127.0.0.1
+    port: $(expr $REP_START_PORT + $i)
+"
+
+HOMESERVER_YAML_STREAM_WRITERS+=$"  ${STREAM_WRITERS[$i]}: ${STREAM_WRITERS[$i]}_stream_writer
+"
+
+((i++))
+done
+
+cat << EXAMPLECONFIG
+# Add these lines to your homeserver.yaml.
+# Don't forget to configure your reverse proxy and
+# necessary endpoints to their respective worker.
+
+# See https://github.com/matrix-org/synapse/blob/develop/docs/workers.md
+# for more information.
+
+# Remember: Under NO circumstances should the replication
+# listener be exposed to the public internet;
+# it has no authentication and is unencrypted.
+
+instance_map:
+$HOMESERVER_YAML_INSTANCE_MAP
+stream_writers:
+$HOMESERVER_YAML_STREAM_WRITERS
+EXAMPLECONFIG
+```
+
+Copy the code above save it to a file ```create_stream_writers.sh``` (for example).
+
+Make the script executable by running ```chmod +x create_stream_writers.sh```.
+
+## Run the script to create workers and print out a sample configuration
+
+Simply run the script to create YAML files in the current folder and print out the required configuration for ```homeserver.yaml```.
+
+```console
+$ ./create_stream_writers.sh
+
+# Add these lines to your homeserver.yaml.
+# Don't forget to configure your reverse proxy and
+# necessary endpoints to their respective worker.
+
+# See https://github.com/matrix-org/synapse/blob/develop/docs/workers.md
+# for more information
+
+# Remember: Under NO circumstances should the replication
+# listener be exposed to the public internet;
+# it has no authentication and is unencrypted.
+
+instance_map:
+  presence_stream_writer:
+    host: 127.0.0.1
+    port: 8034
+  typing_stream_writer:
+    host: 127.0.0.1
+    port: 8035
+  receipts_stream_writer:
+    host: 127.0.0.1
+    port: 8036
+  to_device_stream_writer:
+    host: 127.0.0.1
+    port: 8037
+  account_data_stream_writer:
+    host: 127.0.0.1
+    port: 8038
+
+stream_writers:
+  presence: presence_stream_writer
+  typing: typing_stream_writer
+  receipts: receipts_stream_writer
+  to_device: to_device_stream_writer
+  account_data: account_data_stream_writer
+```
+
+Simply copy-and-paste the output to an appropriate place in your Synapse main configuration file.
+
+## Write directly to Synapse configuration file
+
+You could also write the output directly to homeserver main configuration file. **This, however, is not recommended** as even a small typo (such as replacing >> with >) can erase the entire ```homeserver.yaml```. 
+
+If you do this, back up your original configuration file first:
+
+```console
+# Back up homeserver.yaml first
+cp /etc/matrix-synapse/homeserver.yaml /etc/matrix-synapse/homeserver.yaml.bak 
+
+# Create workers and write output to your homeserver.yaml
+./create_stream_writers.sh >> /etc/matrix-synapse/homeserver.yaml 
+```

debian/changelog

@@ -1,3 +1,53 @@
+matrix-synapse-py3 (1.65.0) stable; urgency=medium
+
+  * New Synapse release 1.65.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 16 Aug 2022 16:51:26 +0100
+
+matrix-synapse-py3 (1.65.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.65.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Thu, 11 Aug 2022 11:38:18 +0100
+
+matrix-synapse-py3 (1.65.0~rc1) stable; urgency=medium
+
+  * New Synapse release 1.65.0rc1.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 09 Aug 2022 11:39:29 +0100
+
+matrix-synapse-py3 (1.64.0) stable; urgency=medium
+
+  * New Synapse release 1.64.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 02 Aug 2022 10:32:30 +0100
+
+matrix-synapse-py3 (1.64.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.64.0rc2.
+
+ -- Synapse Packaging team <packages@matrix.org>  Fri, 29 Jul 2022 12:22:53 +0100
+
+matrix-synapse-py3 (1.64.0~rc1) stable; urgency=medium
+
+  * New Synapse release 1.64.0rc1.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 26 Jul 2022 12:11:49 +0100
+
+matrix-synapse-py3 (1.63.1) stable; urgency=medium
+
+  * New Synapse release 1.63.1.
+
+ -- Synapse Packaging team <packages@matrix.org>  Wed, 20 Jul 2022 13:36:52 +0100
+
+matrix-synapse-py3 (1.63.0) stable; urgency=medium
+
+  * Clarify that homeserver server names are included in the data reported
+    by opt-in server stats reporting (`report_stats` homeserver config option).
+  * New Synapse release 1.63.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 19 Jul 2022 14:42:24 +0200
+
 matrix-synapse-py3 (1.63.0~rc1) stable; urgency=medium
 
   * New Synapse release 1.63.0rc1.

debian/matrix-synapse-py3.postinst

@@ -31,7 +31,7 @@ EOF
 # This file is autogenerated, and will be recreated on upgrade if it is deleted.
 # Any changes you make will be preserved.
 
-# Whether to report anonymized homeserver usage statistics.
+# Whether to report homeserver usage statistics.
 report_stats: false
 EOF
     fi

debian/po/templates.pot

@@ -37,7 +37,7 @@ msgstr ""
 #. Type: boolean
 #. Description
 #: ../templates:2001
-msgid "Report anonymous statistics?"
+msgid "Report homeserver usage statistics?"
 msgstr ""
 
 #. Type: boolean
@@ -45,11 +45,11 @@ msgstr ""
 #: ../templates:2001
 msgid ""
 "Developers of Matrix and Synapse really appreciate helping the project out "
-"by reporting anonymized usage statistics from this homeserver. Only very "
-"basic aggregate data (e.g. number of users) will be reported, but it helps "
-"track the growth of the Matrix community, and helps in making Matrix a "
-"success, as well as to convince other networks that they should peer with "
-"Matrix."
+"by reporting homeserver usage statistics from this homeserver. Your "
+"homeserver's server name, along with very basic aggregate data (e.g. "
+"number of users) will be reported. But it helps track the growth of the "
+"Matrix community, and helps in making Matrix a success, as well as to "
+"convince other networks that they should peer with Matrix."
 msgstr ""
 
 #. Type: boolean

debian/templates

@@ -10,12 +10,13 @@ _Description: Name of the server:
 Template: matrix-synapse/report-stats
 Type: boolean
 Default: false
-_Description: Report anonymous statistics?
+_Description: Report homeserver usage statistics?
  Developers of Matrix and Synapse really appreciate helping the
- project out by reporting anonymized usage statistics from this
- homeserver. Only very basic aggregate data (e.g. number of users)
- will be reported, but it helps track the growth of the Matrix
- community, and helps in making Matrix a success, as well as to
- convince other networks that they should peer with Matrix.
+ project out by reporting homeserver usage statistics from this
+ homeserver. Your homeserver's server name, along with very basic
+ aggregate data (e.g. number of users) will be reported. But it
+ helps track the growth of the Matrix community, and helps in
+ making Matrix a success, as well as to convince other networks
+ that they should peer with Matrix.
  .
  Thank you.

docker/Dockerfile

@@ -40,12 +40,13 @@ FROM docker.io/python:${PYTHON_VERSION}-slim as requirements
 RUN \
    --mount=type=cache,target=/var/cache/apt,sharing=locked \
    --mount=type=cache,target=/var/lib/apt,sharing=locked \
- apt-get update -qq && apt-get install -yqq git \
+    apt-get update -qq && apt-get install -yqq \
+      build-essential cargo git libffi-dev libssl-dev \
     && rm -rf /var/lib/apt/lists/*
 
 # We install poetry in its own build stage to avoid its dependencies conflicting with
 # synapse's dependencies.
-# We use a specific commit from poetry's master branch instead of our usual 1.1.12,
+# We use a specific commit from poetry's master branch instead of our usual 1.1.14,
 # to incorporate fixes to some bugs in `poetry export`. This commit corresponds to
 #    https://github.com/python-poetry/poetry/pull/5156 and
 #    https://github.com/python-poetry/poetry/issues/5141 ;
@@ -68,7 +69,18 @@ COPY pyproject.toml poetry.lock /synapse/
 # reason, such as when a git repository is used directly as a dependency.
 ARG TEST_ONLY_SKIP_DEP_HASH_VERIFICATION
 
-RUN /root/.local/bin/poetry export --extras all -o /synapse/requirements.txt ${TEST_ONLY_SKIP_DEP_HASH_VERIFICATION:+--without-hashes}
+# If specified, we won't use the Poetry lockfile.
+# Instead, we'll just install what a regular `pip install` would from PyPI.
+ARG TEST_ONLY_IGNORE_POETRY_LOCKFILE
+
+# Export the dependencies, but only if we're actually going to use the Poetry lockfile.
+# Otherwise, just create an empty requirements file so that the Dockerfile can
+# proceed.
+RUN if [ -z "$TEST_ONLY_IGNORE_POETRY_LOCKFILE" ]; then \
+    /root/.local/bin/poetry export --extras all -o /synapse/requirements.txt ${TEST_ONLY_SKIP_DEP_HASH_VERIFICATION:+--without-hashes}; \
+  else \
+    touch /synapse/requirements.txt; \
+  fi
 
 ###
 ### Stage 1: builder
@@ -108,8 +120,17 @@ COPY synapse /synapse/synapse/
 # ... and what we need to `pip install`.
 COPY pyproject.toml README.rst /synapse/
 
+# Repeat of earlier build argument declaration, as this is a new build stage.
+ARG TEST_ONLY_IGNORE_POETRY_LOCKFILE
+
 # Install the synapse package itself.
-RUN pip install --prefix="/install" --no-deps --no-warn-script-location /synapse
+# If we have populated requirements.txt, we don't install any dependencies
+# as we should already have those from the previous `pip install` step.
+RUN if [ -z "$TEST_ONLY_IGNORE_POETRY_LOCKFILE" ]; then \
+    pip install --prefix="/install" --no-deps --no-warn-script-location /synapse[all]; \
+  else \
+    pip install --prefix="/install" --no-warn-script-location /synapse[all]; \
+  fi
 
 ###
 ### Stage 2: runtime

docker/Dockerfile-workers

@@ -1,38 +1,62 @@
-# Inherit from the official Synapse docker image
+# syntax=docker/dockerfile:1
+
 ARG SYNAPSE_VERSION=latest
+
+# first of all, we create a base image with an nginx which we can copy into the
+# target image. For repeated rebuilds, this is much faster than apt installing
+# each time.
+
+FROM debian:bullseye-slim AS deps_base
+    RUN \
+       --mount=type=cache,target=/var/cache/apt,sharing=locked \
+       --mount=type=cache,target=/var/lib/apt,sharing=locked \
+      apt-get update -qq && \
+      DEBIAN_FRONTEND=noninteractive apt-get install -yqq --no-install-recommends \
+          redis-server nginx-light
+
+# Similarly, a base to copy the redis server from.
+#
+# The redis docker image has fewer dynamic libraries than the debian package,
+# which makes it much easier to copy (but we need to make sure we use an image
+# based on the same debian version as the synapse image, to make sure we get
+# the expected version of libc.
+FROM redis:6-bullseye AS redis_base
+
+# now build the final image, based on the the regular Synapse docker image
 FROM matrixdotorg/synapse:$SYNAPSE_VERSION
 
-# Install deps
-RUN \
-   --mount=type=cache,target=/var/cache/apt,sharing=locked \
-   --mount=type=cache,target=/var/lib/apt,sharing=locked \
-  apt-get update -qq && \
-  DEBIAN_FRONTEND=noninteractive apt-get install -yqq --no-install-recommends \
-     redis-server nginx-light
+    # Install supervisord with pip instead of apt, to avoid installing a second
+    # copy of python.
+    RUN --mount=type=cache,target=/root/.cache/pip \
+        pip install supervisor~=4.2
+    RUN mkdir -p /etc/supervisor/conf.d
 
-# Install supervisord with pip instead of apt, to avoid installing a second
-# copy of python.
-RUN --mount=type=cache,target=/root/.cache/pip \
-    pip install supervisor~=4.2
+    # Copy over redis and nginx
+    COPY --from=redis_base /usr/local/bin/redis-server /usr/local/bin
 
-# Disable the default nginx sites
-RUN rm /etc/nginx/sites-enabled/default
+    COPY --from=deps_base /usr/sbin/nginx /usr/sbin
+    COPY --from=deps_base /usr/share/nginx /usr/share/nginx
+    COPY --from=deps_base /usr/lib/nginx /usr/lib/nginx
+    COPY --from=deps_base /etc/nginx /etc/nginx
+    RUN rm /etc/nginx/sites-enabled/default
+    RUN mkdir /var/log/nginx /var/lib/nginx
+    RUN chown www-data /var/log/nginx /var/lib/nginx
 
-# Copy Synapse worker, nginx and supervisord configuration template files
-COPY ./docker/conf-workers/* /conf/
+    # Copy Synapse worker, nginx and supervisord configuration template files
+    COPY ./docker/conf-workers/* /conf/
 
-# Copy a script to prefix log lines with the supervisor program name
-COPY ./docker/prefix-log /usr/local/bin/
+    # Copy a script to prefix log lines with the supervisor program name
+    COPY ./docker/prefix-log /usr/local/bin/
 
-# Expose nginx listener port
-EXPOSE 8080/tcp
+    # Expose nginx listener port
+    EXPOSE 8080/tcp
 
-# A script to read environment variables and create the necessary
-# files to run the desired worker configuration. Will start supervisord.
-COPY ./docker/configure_workers_and_start.py /configure_workers_and_start.py
-ENTRYPOINT ["/configure_workers_and_start.py"]
+    # A script to read environment variables and create the necessary
+    # files to run the desired worker configuration. Will start supervisord.
+    COPY ./docker/configure_workers_and_start.py /configure_workers_and_start.py
+    ENTRYPOINT ["/configure_workers_and_start.py"]
 
-# Replace the healthcheck with one which checks *all* the workers. The script
-# is generated by configure_workers_and_start.py.
-HEALTHCHECK --start-period=5s --interval=15s --timeout=5s \
-    CMD /bin/sh /healthcheck.sh
+    # Replace the healthcheck with one which checks *all* the workers. The script
+    # is generated by configure_workers_and_start.py.
+    HEALTHCHECK --start-period=5s --interval=15s --timeout=5s \
+        CMD /bin/sh /healthcheck.sh

docker/README-testing.md

@@ -22,6 +22,10 @@ Consult the [contributing guide][guideComplementSh] for instructions on how to u
 Under some circumstances, you may wish to build the images manually.
 The instructions below will lead you to doing that.
 
+Note that these images can only be built using [BuildKit](https://docs.docker.com/develop/develop-images/build_enhancements/),
+therefore BuildKit needs to be enabled when calling `docker build`. This can be done by
+setting `DOCKER_BUILDKIT=1` in your environment.
+
 Start by building the base Synapse docker image. If you wish to run tests with the latest
 release of Synapse, instead of your current checkout, you can skip this step. From the
 root of the repository:

docker/complement/Dockerfile

@@ -1,3 +1,4 @@
+# syntax=docker/dockerfile:1
 # This dockerfile builds on top of 'docker/Dockerfile-workers' in matrix-org/synapse
 # by including a built-in postgres instance, as well as setting up the homeserver so
 # that it is ready for testing via Complement.

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

@@ -67,6 +67,10 @@ rc_joins:
     per_second: 9999
     burst_count: 9999
 
+rc_joins_per_room:
+    per_second: 9999
+    burst_count: 9999
+
 rc_3pid_validation:
   per_second: 1000
   burst_count: 1000

docker/conf-workers/supervisord.conf.j2

@@ -19,7 +19,7 @@ username=www-data
 autorestart=true
 
 [program:redis]
-command=/usr/local/bin/prefix-log /usr/bin/redis-server /etc/redis/redis.conf --daemonize no
+command=/usr/local/bin/prefix-log /usr/local/bin/redis-server
 priority=1
 stdout_logfile=/dev/stdout
 stdout_logfile_maxbytes=0

docs/SUMMARY.md

@@ -68,7 +68,7 @@
       - [Federation](usage/administration/admin_api/federation.md)
     - [Manhole](manhole.md)
     - [Monitoring](metrics-howto.md)
-      - [Reporting Anonymised Statistics](usage/administration/monitoring/reporting_anonymised_statistics.md)
+      - [Reporting Homeserver Usage Statistics](usage/administration/monitoring/reporting_homeserver_usage_statistics.md)
     - [Understanding Synapse Through Grafana Graphs](usage/administration/understanding_synapse_through_grafana_graphs.md)
     - [Useful SQL for Admins](usage/administration/useful_sql_for_admins.md)
     - [Database Maintenance Tools](usage/administration/database_maintenance_tools.md)

docs/development/dependencies.md

@@ -237,3 +237,28 @@ poetry run pip install build && poetry run python -m build
 because [`build`](https://github.com/pypa/build) is a standardish tool which
 doesn't require poetry. (It's what we use in CI too). However, you could try
 `poetry build` too.
+
+
+# Troubleshooting
+
+## Check the version of poetry with `poetry --version`.
+
+At the time of writing, the 1.2 series is beta only. We have seen some examples
+where the lockfiles generated by 1.2 prereleasese aren't interpreted correctly
+by poetry 1.1.x. For now, use poetry 1.1.14, which includes a critical
+[change](https://github.com/python-poetry/poetry/pull/5973) needed to remain
+[compatible with PyPI](https://github.com/pypi/warehouse/pull/11775).
+
+It can also be useful to check the version of `poetry-core` in use. If you've
+installed `poetry` with `pipx`, try `pipx runpip poetry list | grep poetry-core`.
+
+## Clear caches: `poetry cache clear --all pypi`.
+
+Poetry caches a bunch of information about packages that isn't readily available
+from PyPI. (This is what makes poetry seem slow when doing the first
+`poetry install`.) Try `poetry cache list` and `poetry cache clear --all
+<name of cache>` to see if that fixes things.
+
+## Try `--verbose` or `--dry-run` arguments.
+
+Sometimes useful to see what poetry's internal logic is.

docs/modules/password_auth_provider_callbacks.md

@@ -263,7 +263,7 @@ class MyAuthProvider:
             return None
 
         if self.credentials.get(username) == login_dict.get("my_field"):
-            return self.api.get_qualified_user_id(username)
+            return (self.api.get_qualified_user_id(username), None)
 
     async def check_pass(
         self,
@@ -280,5 +280,5 @@ class MyAuthProvider:
             return None
 
         if self.credentials.get(username) == login_dict.get("password"):
-            return self.api.get_qualified_user_id(username)
+            return (self.api.get_qualified_user_id(username), None)
 ```

docs/reverse_proxy.md

@@ -79,63 +79,32 @@ server {
 }
 ```
 
-### Caddy v1
-
-```
-matrix.example.com {
-  proxy /_matrix http://localhost:8008 {
-    transparent
-  }
-
-  proxy /_synapse/client http://localhost:8008 {
-    transparent
-  }
-}
-
-example.com:8448 {
-  proxy / http://localhost:8008 {
-    transparent
-  }
-}
-```
-
 ### Caddy v2
 
 ```
 matrix.example.com {
-  reverse_proxy /_matrix/* http://localhost:8008
-  reverse_proxy /_synapse/client/* http://localhost:8008
+  reverse_proxy /_matrix/* localhost:8008
+  reverse_proxy /_synapse/client/* localhost:8008
 }
 
 example.com:8448 {
-  reverse_proxy http://localhost:8008
+  reverse_proxy localhost:8008
 }
 ```
+
 [Delegation](delegate.md) example:
+
 ```
-(matrix-well-known-header) {
-    # Headers
-    header Access-Control-Allow-Origin "*"
-    header Access-Control-Allow-Methods "GET, POST, PUT, DELETE, OPTIONS"
-    header Access-Control-Allow-Headers "Origin, X-Requested-With, Content-Type, Accept, Authorization"
-    header Content-Type "application/json"
-}
-
 example.com {
-    handle /.well-known/matrix/server {
-        import matrix-well-known-header
-        respond `{"m.server":"matrix.example.com:443"}`
-    }
-
-    handle /.well-known/matrix/client {
-        import matrix-well-known-header
-        respond `{"m.homeserver":{"base_url":"https://matrix.example.com"},"m.identity_server":{"base_url":"https://identity.example.com"}}`
-    }
+	header /.well-known/matrix/* Content-Type application/json
+	header /.well-known/matrix/* Access-Control-Allow-Origin *
+	respond /.well-known/matrix/server `{"m.server": "matrix.example.com:443"}`
+	respond /.well-known/matrix/client `{"m.homeserver":{"base_url":"https://matrix.example.com"},"m.identity_server":{"base_url":"https://identity.example.com"}}`
 }
 
 matrix.example.com {
-    reverse_proxy /_matrix/* http://localhost:8008
-    reverse_proxy /_synapse/client/* http://localhost:8008
+    reverse_proxy /_matrix/* localhost:8008
+    reverse_proxy /_synapse/client/* localhost:8008
 }
 ```
 

docs/sso_mapping_providers.md

@@ -22,7 +22,7 @@ choose their own username.
 In the first case - where users are automatically allocated a Matrix ID - it is
 the responsibility of the mapping provider to normalise the SSO attributes and
 map them to a valid Matrix ID. The [specification for Matrix
-IDs](https://matrix.org/docs/spec/appendices#user-identifiers) has some
+IDs](https://spec.matrix.org/latest/appendices/#user-identifiers) has some
 information about what is considered valid.
 
 If the mapping provider does not assign a Matrix ID, then Synapse will
@@ -37,9 +37,10 @@ as Synapse). The Synapse config is then modified to point to the mapping provide
 ## OpenID Mapping Providers
 
 The OpenID mapping provider can be customized by editing the
-`oidc_config.user_mapping_provider.module` config option.
+[`oidc_providers.user_mapping_provider.module`](usage/configuration/config_documentation.md#oidc_providers)
+config option.
 
-`oidc_config.user_mapping_provider.config` allows you to provide custom
+`oidc_providers.user_mapping_provider.config` allows you to provide custom
 configuration options to the module. Check with the module's documentation for
 what options it provides (if any). The options listed by default are for the
 user mapping provider built in to Synapse. If using a custom module, you should
@@ -58,7 +59,7 @@ A custom mapping provider must specify the following methods:
     - This method should have the `@staticmethod` decoration.
     - Arguments:
         - `config` - A `dict` representing the parsed content of the
-          `oidc_config.user_mapping_provider.config` homeserver config option.
+          `oidc_providers.user_mapping_provider.config` homeserver config option.
            Runs on homeserver startup. Providers should extract and validate
            any option values they need here.
     - Whatever is returned will be passed back to the user mapping provider module's
@@ -102,7 +103,7 @@ A custom mapping provider must specify the following methods:
       will be returned as part of the response during a successful login.
 
       Note that care should be taken to not overwrite any of the parameters
-      usually returned as part of the [login response](https://matrix.org/docs/spec/client_server/latest#post-matrix-client-r0-login).
+      usually returned as part of the [login response](https://spec.matrix.org/latest/client-server-api/#post_matrixclientv3login).
 
 ### Default OpenID Mapping Provider
 
@@ -113,7 +114,8 @@ specified in the config. It is located at
 ## SAML Mapping Providers
 
 The SAML mapping provider can be customized by editing the
-`saml2_config.user_mapping_provider.module` config option.
+[`saml2_config.user_mapping_provider.module`](docs/usage/configuration/config_documentation.md#saml2_config)
+config option.
 
 `saml2_config.user_mapping_provider.config` allows you to provide custom
 configuration options to the module. Check with the module's documentation for

docs/upgrade.md

@@ -91,18 +91,34 @@ process, for example:
 
 # Upgrading to v1.64.0
 
-## Delegation of email validation no longer supported
+## Deprecation of the ability to delegate e-mail verification to identity servers
 
-As of this version, Synapse no longer allows the tasks of verifying email address
-ownership, and password reset confirmation, to be delegated to an identity server.
+Synapse v1.66.0 will remove the ability to delegate the tasks of verifying email address ownership, and password reset confirmation, to an identity server.
 
-To continue to allow users to add email addresses to their homeserver accounts,
-and perform password resets, make sure that Synapse is configured with a
-working email server in the `email` configuration section (including, at a
-minimum, a `notif_from` setting.)
+If you require your homeserver to verify e-mail addresses or to support password resets via e-mail, please configure your homeserver with SMTP access so that it can send e-mails on its own behalf.
+[Consult the configuration documentation for more information.](https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html#email)
 
-Specifying an `email` setting under `account_threepid_delegates` will now cause
-an error at startup.
+The option that will be removed is `account_threepid_delegates.email`.
+
+
+## Changes to the event replication streams
+
+Synapse now includes a flag indicating if an event is an outlier when
+replicating it to other workers. This is a forwards- and backwards-incompatible
+change: v1.63 and workers cannot process events replicated by v1.64 workers, and
+vice versa.
+
+Once all workers are upgraded to v1.64 (or downgraded to v1.63), event
+replication will resume as normal.
+
+## frozendict release
+
+[frozendict 2.3.3](https://github.com/Marco-Sulla/python-frozendict/releases/tag/v2.3.3)
+has recently been released, which fixes a memory leak that occurs during `/sync`
+requests. We advise server administrators who installed Synapse via pip to upgrade
+frozendict with `pip install --upgrade frozendict`. The Docker image
+`matrixdotorg/synapse` and the Debian packages from `packages.matrix.org` already
+include the updated library.
 
 # Upgrading to v1.62.0
 

docs/usage/administration/admin_api/README.md

@@ -5,8 +5,9 @@
 Many of the API calls in the admin api will require an `access_token` for a
 server admin. (Note that a server admin is distinct from a room admin.)
 
-A user can be marked as a server admin by updating the database directly, e.g.:
+An existing user can be marked as a server admin by updating the database directly.
 
+Check your [database settings](config_documentation.md#database) in the configuration file, connect to the correct database using either `psql [database name]` (if using PostgreSQL) or `sqlite3 path/to/your/database.db` (if using SQLite) and elevate the user `@foo:bar.com` to administrator.
 ```sql
 UPDATE users SET admin = 1 WHERE name = '@foo:bar.com';
 ```

docs/usage/administration/admin_api/registration_tokens.md

@@ -2,11 +2,11 @@
 
 This API allows you to manage tokens which can be used to authenticate
 registration requests, as proposed in
-[MSC3231](https://github.com/matrix-org/matrix-doc/blob/main/proposals/3231-token-authenticated-registration.md).
+[MSC3231](https://github.com/matrix-org/matrix-doc/blob/main/proposals/3231-token-authenticated-registration.md)
+and stabilised in version 1.2 of the Matrix specification.
 To use it, you will need to enable the `registration_requires_token` config
 option, and authenticate by providing an `access_token` for a server admin:
-see [Admin API](../../usage/administration/admin_api).
-Note that this API is still experimental; not all clients may support it yet.
+see [Admin API](../admin_api).
 
 
 ## Registration token objects

docs/usage/administration/monitoring/reporting_homeserver_usage_statistics.md

@@ -1,11 +1,11 @@
-# Reporting Anonymised Statistics
+# Reporting Homeserver Usage Statistics
 
 When generating your Synapse configuration file, you are asked whether you
-would like to report anonymised statistics to Matrix.org. These statistics
+would like to report usage statistics to Matrix.org. These statistics
 provide the foundation a glimpse into the number of Synapse homeservers
 participating in the network, as well as statistics such as the number of
 rooms being created and messages being sent. This feature is sometimes
-affectionately called "phone-home" stats. Reporting
+affectionately called "phone home" stats. Reporting
 [is optional](../../configuration/config_documentation.md#report_stats)
 and the reporting endpoint
 [can be configured](../../configuration/config_documentation.md#report_stats_endpoint),
@@ -21,9 +21,9 @@ The following statistics are sent to the configured reporting endpoint:
 
 | Statistic Name             | Type   | Description                                                                                                                                                                                                                                                                                     |
 |----------------------------|--------|-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| `homeserver`               | string | The homeserver's server name.                                                                                                                                                                                                                                                                   |
 | `memory_rss`               | int    | The memory usage of the process (in kilobytes on Unix-based systems, bytes on MacOS).                                                                                                                                                                                                           |
 | `cpu_average`              | int    | CPU time in % of a single core (not % of all cores).                                                                                                                                                                                                                                            |              
-| `homeserver`               | string | The homeserver's server name.                                                                                                                                                                                                                                                                   |
 | `server_context`           | string | An arbitrary string used to group statistics from a set of homeservers.                                                                                                                                                                                                                         |
 | `timestamp`                | int    | The current time, represented as the number of seconds since the epoch.                                                                                                                                                                                                                         |                 
 | `uptime_seconds`           | int    | The number of seconds since the homeserver was last started.                                                                                                                                                                                                                                    |

mypy.ini

@@ -84,9 +84,6 @@ disallow_untyped_defs = False
 [mypy-synapse.http.matrixfederationclient]
 disallow_untyped_defs = False
 
-[mypy-synapse.logging.opentracing]
-disallow_untyped_defs = False
-
 [mypy-synapse.metrics._reactor_metrics]
 disallow_untyped_defs = False
 # This module imports select.epoll. That exists on Linux, but doesn't on macOS.

poetry.lock

@@ -177,7 +177,7 @@ optional = false
 python-versions = "*"
 
 [package.extras]
-test = ["flake8 (==3.7.8)", "hypothesis (==3.55.3)"]
+test = ["hypothesis (==3.55.3)", "flake8 (==3.7.8)"]
 
 [[package]]
 name = "constantly"
@@ -290,7 +290,7 @@ importlib-metadata = {version = "*", markers = "python_version < \"3.8\""}
 
 [[package]]
 name = "frozendict"
-version = "2.3.2"
+version = "2.3.3"
 description = "A simple immutable dictionary"
 category = "main"
 optional = false
@@ -435,8 +435,8 @@ optional = false
 python-versions = ">=3.6"
 
 [package.extras]
-test = ["pytest", "pytest-trio", "pytest-asyncio", "testpath", "trio", "async-timeout"]
-trio = ["trio", "async-generator"]
+trio = ["async-generator", "trio"]
+test = ["async-timeout", "trio", "testpath", "pytest-asyncio", "pytest-trio", "pytest"]
 
 [[package]]
 name = "jinja2"
@@ -535,12 +535,12 @@ attrs = "*"
 importlib-metadata = {version = ">=1.4", markers = "python_version < \"3.8\""}
 
 [package.extras]
-dev = ["tox", "twisted", "aiounittest", "mypy (==0.910)", "black (==22.3.0)", "flake8 (==4.0.1)", "isort (==5.9.3)", "build (==0.8.0)", "twine (==4.0.1)"]
-test = ["tox", "twisted", "aiounittest"]
+test = ["aiounittest", "twisted", "tox"]
+dev = ["twine (==4.0.1)", "build (==0.8.0)", "isort (==5.9.3)", "flake8 (==4.0.1)", "black (==22.3.0)", "mypy (==0.910)", "aiounittest", "twisted", "tox"]
 
 [[package]]
 name = "matrix-synapse-ldap3"
-version = "0.2.1"
+version = "0.2.2"
 description = "An LDAP3 auth provider for Synapse"
 category = "main"
 optional = true
@@ -552,7 +552,7 @@ service-identity = "*"
 Twisted = ">=15.1.0"
 
 [package.extras]
-dev = ["matrix-synapse", "tox", "ldaptor", "mypy (==0.910)", "types-setuptools", "black (==22.3.0)", "flake8 (==4.0.1)", "isort (==5.9.3)"]
+dev = ["isort (==5.9.3)", "flake8 (==4.0.1)", "black (==22.3.0)", "types-setuptools", "mypy (==0.910)", "ldaptor", "tox", "matrix-synapse"]
 
 [[package]]
 name = "mccabe"
@@ -820,10 +820,10 @@ optional = false
 python-versions = ">=3.6"
 
 [package.extras]
+tests = ["coverage[toml] (==5.0.4)", "pytest (>=6.0.0,<7.0.0)"]
+docs = ["zope.interface", "sphinx-rtd-theme", "sphinx"]
+dev = ["pre-commit", "mypy", "coverage[toml] (==5.0.4)", "pytest (>=6.0.0,<7.0.0)", "cryptography (>=3.3.1)", "zope.interface", "sphinx-rtd-theme", "sphinx"]
 crypto = ["cryptography (>=3.3.1)"]
-dev = ["sphinx", "sphinx-rtd-theme", "zope.interface", "cryptography (>=3.3.1)", "pytest (>=6.0.0,<7.0.0)", "coverage[toml] (==5.0.4)", "mypy", "pre-commit"]
-docs = ["sphinx", "sphinx-rtd-theme", "zope.interface"]
-tests = ["pytest (>=6.0.0,<7.0.0)", "coverage[toml] (==5.0.4)"]
 
 [[package]]
 name = "pymacaroons"
@@ -1563,7 +1563,7 @@ url_preview = ["lxml"]
 [metadata]
 lock-version = "1.1"
 python-versions = "^3.7.1"
-content-hash = "e96625923122e29b6ea5964379828e321b6cede2b020fc32c6f86c09d86d1ae8"
+content-hash = "c24bbcee7e86dbbe7cdbf49f91a25b310bf21095452641e7440129f59b077f78"
 
 [metadata.files]
 attrs = [
@@ -1753,23 +1753,23 @@ flake8-comprehensions = [
     {file = "flake8_comprehensions-3.8.0-py3-none-any.whl", hash = "sha256:9406314803abe1193c064544ab14fdc43c58424c0882f6ff8a581eb73fc9bb58"},
 ]
 frozendict = [
-    {file = "frozendict-2.3.2-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:4fb171d1e84d17335365877e19d17440373b47ca74a73c06f65ac0b16d01e87f"},
-    {file = "frozendict-2.3.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c0a3640e9d7533d164160b758351aa49d9e85bbe0bd76d219d4021e90ffa6a52"},
-    {file = "frozendict-2.3.2-cp310-cp310-win_amd64.whl", hash = "sha256:87cfd00fafbc147d8cd2590d1109b7db8ac8d7d5bdaa708ba46caee132b55d4d"},
-    {file = "frozendict-2.3.2-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:fb09761e093cfabb2f179dbfdb2521e1ec5701df714d1eb5c51fa7849027be19"},
-    {file = "frozendict-2.3.2-cp36-cp36m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:82176dc7adf01cf8f0193e909401939415a230a1853f4a672ec1629a06ceae18"},
-    {file = "frozendict-2.3.2-cp36-cp36m-win_amd64.whl", hash = "sha256:c1c70826aa4a50fa283fe161834ac4a3ac7c753902c980bb8b595b0998a38ddb"},
-    {file = "frozendict-2.3.2-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:1db5035ddbed995badd1a62c4102b5e207b5aeb24472df2c60aba79639d7996b"},
-    {file = "frozendict-2.3.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4246fc4cb1413645ba4d3513939b90d979a5bae724be605a10b2b26ee12f839c"},
-    {file = "frozendict-2.3.2-cp37-cp37m-win_amd64.whl", hash = "sha256:680cd42fb0a255da1ce45678ccbd7f69da750d5243809524ebe8f45b2eda6e6b"},
-    {file = "frozendict-2.3.2-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:6a7f3a181d6722c92a9fab12d0c5c2b006a18ca5666098531f316d1e1c8984e3"},
-    {file = "frozendict-2.3.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:b1cb866eabb3c1384a7fe88e1e1033e2b6623073589012ab637c552bf03f6364"},
-    {file = "frozendict-2.3.2-cp38-cp38-win_amd64.whl", hash = "sha256:952c5e5e664578c5c2ce8489ee0ab6a1855da02b58ef593ee728fc10d672641a"},
-    {file = "frozendict-2.3.2-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:608b77904cd0117cd816df605a80d0043a5326ee62529327d2136c792165a823"},
-    {file = "frozendict-2.3.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0eed41fd326f0bcc779837d8d9e1374da1bc9857fe3b9f2910195bbd5fff3aeb"},
-    {file = "frozendict-2.3.2-cp39-cp39-win_amd64.whl", hash = "sha256:bde28db6b5868dd3c45b3555f9d1dc5a1cca6d93591502fa5dcecce0dde6a335"},
-    {file = "frozendict-2.3.2-py3-none-any.whl", hash = "sha256:6882a9bbe08ab9b5ff96ce11bdff3fe40b114b9813bc6801261e2a7b45e20012"},
-    {file = "frozendict-2.3.2.tar.gz", hash = "sha256:7fac4542f0a13fbe704db4942f41ba3abffec5af8b100025973e59dff6a09d0d"},
+    {file = "frozendict-2.3.3-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:39942914c1217a5a49c7551495a103b3dbd216e19413687e003b859c6b0ebc12"},
+    {file = "frozendict-2.3.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5589256058b31f2b91419fa30b8dc62dbdefe7710e688a3fd5b43849161eecc9"},
+    {file = "frozendict-2.3.3-cp310-cp310-win_amd64.whl", hash = "sha256:35eb7e59e287c41f4f712d4d3d2333354175b155d217b97c99c201d2d8920790"},
+    {file = "frozendict-2.3.3-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:310aaf81793abf4f471895e6fe65e0e74a28a2aaf7b25c2ba6ccd4e35af06842"},
+    {file = "frozendict-2.3.3-cp36-cp36m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c353c11010a986566a0cb37f9a783c560ffff7d67d5e7fd52221fb03757cdc43"},
+    {file = "frozendict-2.3.3-cp36-cp36m-win_amd64.whl", hash = "sha256:15b5f82aad108125336593cec1b6420c638bf45f449c57e50949fc7654ea5a41"},
+    {file = "frozendict-2.3.3-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:a4737e5257756bd6b877504ff50185b705db577b5330d53040a6cf6417bb3cdb"},
+    {file = "frozendict-2.3.3-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:80a14c11e33e8b0bc09e07bba3732c77a502c39edb8c3959fd9a0e490e031158"},
+    {file = "frozendict-2.3.3-cp37-cp37m-win_amd64.whl", hash = "sha256:027952d1698ac9c766ef43711226b178cdd49d2acbdff396936639ad1d2a5615"},
+    {file = "frozendict-2.3.3-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:ef818d66c85098a37cf42509545a4ba7dd0c4c679d6262123a8dc14cc474bab7"},
+    {file = "frozendict-2.3.3-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:812279f2b270c980112dc4e367b168054f937108f8044eced4199e0ab2945a37"},
+    {file = "frozendict-2.3.3-cp38-cp38-win_amd64.whl", hash = "sha256:c1fb7efbfebc2075f781be3d9774e4ba6ce4fc399148b02097f68d4b3c4bc00a"},
+    {file = "frozendict-2.3.3-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:a0b46d4bf95bce843c0151959d54c3e5b8d0ce29cb44794e820b3ec980d63eee"},
+    {file = "frozendict-2.3.3-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:38c4660f37fcc70a32ff997fe58e40b3fcc60b2017b286e33828efaa16b01308"},
+    {file = "frozendict-2.3.3-cp39-cp39-win_amd64.whl", hash = "sha256:919e3609844fece11ab18bcbf28a3ed20f8108ad4149d7927d413687f281c6c9"},
+    {file = "frozendict-2.3.3-py3-none-any.whl", hash = "sha256:f988b482d08972a196664718167a993a61c9e9f6fe7b0ca2443570b5f20ca44a"},
+    {file = "frozendict-2.3.3.tar.gz", hash = "sha256:398539c52af3c647d103185bbaa1291679f0507ad035fe3bab2a8b0366d52cf1"},
 ]
 gitdb = [
     {file = "gitdb-4.0.9-py3-none-any.whl", hash = "sha256:8033ad4e853066ba6ca92050b9df2f89301b8fc8bf7e9324d412a63f8bf1a8fd"},
@@ -2055,8 +2055,8 @@ matrix-common = [
     {file = "matrix_common-1.2.1.tar.gz", hash = "sha256:a99dcf02a6bd95b24a5a61b354888a2ac92bf2b4b839c727b8dd9da2cdfa3853"},
 ]
 matrix-synapse-ldap3 = [
-    {file = "matrix-synapse-ldap3-0.2.1.tar.gz", hash = "sha256:bfb4390f4a262ffb0d6f057ff3aeb1e46d4e52ff420a064d795fb4f555f00285"},
-    {file = "matrix_synapse_ldap3-0.2.1-py3-none-any.whl", hash = "sha256:1b3310a60f1d06466f35905a269b6df95747fd1305f2b7fe638f373963b2aa2c"},
+    {file = "matrix-synapse-ldap3-0.2.2.tar.gz", hash = "sha256:b388d95693486eef69adaefd0fd9e84463d52fe17b0214a00efcaa669b73cb74"},
+    {file = "matrix_synapse_ldap3-0.2.2-py3-none-any.whl", hash = "sha256:66ee4c85d7952c6c27fd04c09cdfdf4847b8e8b7d6a7ada6ba1100013bda060f"},
 ]
 mccabe = [
     {file = "mccabe-0.6.1-py2.py3-none-any.whl", hash = "sha256:ab8a6258860da4b6677da4bd2fe5dc2c659cff31b3ee4f7f5d64e79735b80d42"},

pyproject.toml

@@ -54,7 +54,7 @@ skip_gitignore = true
 
 [tool.poetry]
 name = "matrix-synapse"
-version = "1.63.0rc1"
+version = "1.65.0"
 description = "Homeserver for the Matrix decentralised comms protocol"
 authors = ["Matrix.org Team and Contributors <packages@matrix.org>"]
 license = "Apache-2.0"
@@ -110,7 +110,9 @@ jsonschema = ">=3.0.0"
 frozendict = ">=1,!=2.1.2"
 # We require 2.1.0 or higher for type hints. Previous guard was >= 1.1.0
 unpaddedbase64 = ">=2.1.0"
-canonicaljson = "^1.4.0"
+# We require 1.5.0 to work around an issue when running against the C implementation of
+# frozendict: https://github.com/matrix-org/python-canonicaljson/issues/36
+canonicaljson = "^1.5.0"
 # we use the type definitions added in signedjson 1.1.
 signedjson = "^1.1.0"
 # validating SSL certs for IP addresses requires service_identity 18.1.

scripts-dev/build_debian_packages.py

@@ -26,7 +26,6 @@ DISTS = (
     "debian:bookworm",
     "debian:sid",
     "ubuntu:focal",  # 20.04 LTS (our EOL forced by Py38 on 2024-10-14)
-    "ubuntu:impish",  # 21.10  (EOL 2022-07)
     "ubuntu:jammy",  # 22.04 LTS (EOL 2027-04)
 )
 

scripts-dev/complement.sh

@@ -101,6 +101,7 @@ if [ -z "$skip_docker_build" ]; then
     echo_if_github "::group::Build Docker image: matrixdotorg/synapse"
     docker build -t matrixdotorg/synapse \
       --build-arg TEST_ONLY_SKIP_DEP_HASH_VERIFICATION \
+      --build-arg TEST_ONLY_IGNORE_POETRY_LOCKFILE \
       -f "docker/Dockerfile" .
     echo_if_github "::endgroup::"
 

scripts-dev/release.py

@@ -32,6 +32,7 @@ import click
 import commonmark
 import git
 from click.exceptions import ClickException
+from git import GitCommandError, Repo
 from github import Github
 from packaging import version
 
@@ -55,9 +56,12 @@ def run_until_successful(
 def cli() -> None:
     """An interactive script to walk through the parts of creating a release.
 
-    Requires the dev dependencies be installed, which can be done via:
+    Requirements:
+      - The dev dependencies be installed, which can be done via:
 
-        pip install -e .[dev]
+            pip install -e .[dev]
+
+      - A checkout of the sytest repository at ../sytest
 
     Then to use:
 
@@ -75,6 +79,8 @@ def cli() -> None:
 
         # Optional: generate some nice links for the announcement
 
+        ./scripts-dev/release.py merge-back
+
         ./scripts-dev/release.py announce
 
     If the env var GH_TOKEN (or GITHUB_TOKEN) is set, or passed into the
@@ -89,10 +95,12 @@ def prepare() -> None:
     """
 
     # Make sure we're in a git repo.
-    repo = get_repo_and_check_clean_checkout()
+    synapse_repo = get_repo_and_check_clean_checkout()
+    sytest_repo = get_repo_and_check_clean_checkout("../sytest", "sytest")
 
-    click.secho("Updating git repo...")
-    repo.remote().fetch()
+    click.secho("Updating Synapse and Sytest git repos...")
+    synapse_repo.remote().fetch()
+    sytest_repo.remote().fetch()
 
     # Get the current version and AST from root Synapse module.
     current_version = get_package_version()
@@ -166,12 +174,12 @@ def prepare() -> None:
     assert not parsed_new_version.is_postrelease
 
     release_branch_name = get_release_branch_name(parsed_new_version)
-    release_branch = find_ref(repo, release_branch_name)
+    release_branch = find_ref(synapse_repo, release_branch_name)
     if release_branch:
         if release_branch.is_remote():
             # If the release branch only exists on the remote we check it out
             # locally.
-            repo.git.checkout(release_branch_name)
+            synapse_repo.git.checkout(release_branch_name)
     else:
         # If a branch doesn't exist we create one. We ask which one branch it
         # should be based off, defaulting to sensible values depending on the
@@ -187,25 +195,34 @@ def prepare() -> None:
             "Which branch should the release be based on?", default=default
         )
 
-        base_branch = find_ref(repo, branch_name)
-        if not base_branch:
-            print(f"Could not find base branch {branch_name}!")
-            click.get_current_context().abort()
+        for repo_name, repo in {"synapse": synapse_repo, "sytest": sytest_repo}.items():
+            base_branch = find_ref(repo, branch_name)
+            if not base_branch:
+                print(f"Could not find base branch {branch_name} for {repo_name}!")
+                click.get_current_context().abort()
 
-        # Check out the base branch and ensure it's up to date
-        repo.head.set_reference(base_branch, "check out the base branch")
-        repo.head.reset(index=True, working_tree=True)
-        if not base_branch.is_remote():
-            update_branch(repo)
+            # Check out the base branch and ensure it's up to date
+            repo.head.set_reference(
+                base_branch, f"check out the base branch for {repo_name}"
+            )
+            repo.head.reset(index=True, working_tree=True)
+            if not base_branch.is_remote():
+                update_branch(repo)
 
-        # Create the new release branch
-        # Type ignore will no longer be needed after GitPython 3.1.28.
-        # See https://github.com/gitpython-developers/GitPython/pull/1419
-        repo.create_head(release_branch_name, commit=base_branch)  # type: ignore[arg-type]
+            # Create the new release branch
+            # Type ignore will no longer be needed after GitPython 3.1.28.
+            # See https://github.com/gitpython-developers/GitPython/pull/1419
+            repo.create_head(release_branch_name, commit=base_branch)  # type: ignore[arg-type]
+
+        # Special-case SyTest: we don't actually prepare any files so we may
+        # as well push it now (and only when we create a release branch;
+        # not on subsequent RCs or full releases).
+        if click.confirm("Push new SyTest branch?", default=True):
+            sytest_repo.git.push("-u", sytest_repo.remote().name, release_branch_name)
 
     # Switch to the release branch and ensure it's up to date.
-    repo.git.checkout(release_branch_name)
-    update_branch(repo)
+    synapse_repo.git.checkout(release_branch_name)
+    update_branch(synapse_repo)
 
     # Update the version specified in pyproject.toml.
     subprocess.check_output(["poetry", "version", new_version])
@@ -230,15 +247,15 @@ def prepare() -> None:
     run_until_successful('dch -M -r -D stable ""', shell=True)
 
     # Show the user the changes and ask if they want to edit the change log.
-    repo.git.add("-u")
+    synapse_repo.git.add("-u")
     subprocess.run("git diff --cached", shell=True)
 
     if click.confirm("Edit changelog?", default=False):
         click.edit(filename="CHANGES.md")
 
     # Commit the changes.
-    repo.git.add("-u")
-    repo.git.commit("-m", new_version)
+    synapse_repo.git.add("-u")
+    synapse_repo.git.commit("-m", new_version)
 
     # We give the option to bail here in case the user wants to make sure things
     # are OK before pushing.
@@ -246,17 +263,21 @@ def prepare() -> None:
         print("")
         print("Run when ready to push:")
         print("")
-        print(f"\tgit push -u {repo.remote().name} {repo.active_branch.name}")
+        print(
+            f"\tgit push -u {synapse_repo.remote().name} {synapse_repo.active_branch.name}"
+        )
         print("")
         sys.exit(0)
 
     # Otherwise, push and open the changelog in the browser.
-    repo.git.push("-u", repo.remote().name, repo.active_branch.name)
+    synapse_repo.git.push(
+        "-u", synapse_repo.remote().name, synapse_repo.active_branch.name
+    )
 
     print("Opening the changelog in your browser...")
     print("Please ask others to give it a check.")
     click.launch(
-        f"https://github.com/matrix-org/synapse/blob/{repo.active_branch.name}/CHANGES.md"
+        f"https://github.com/matrix-org/synapse/blob/{synapse_repo.active_branch.name}/CHANGES.md"
     )
 
 
@@ -423,6 +444,79 @@ def upload() -> None:
     )
 
 
+def _merge_into(repo: Repo, source: str, target: str) -> None:
+    """
+    Merges branch `source` into branch `target`.
+    Pulls both before merging and pushes the result.
+    """
+
+    # Update our branches and switch to the target branch
+    for branch in [source, target]:
+        click.echo(f"Switching to {branch} and pulling...")
+        repo.heads[branch].checkout()
+        # Pull so we're up to date
+        repo.remote().pull()
+
+    assert repo.active_branch.name == target
+
+    try:
+        # TODO This seemed easier than using GitPython directly
+        click.echo(f"Merging {source}...")
+        repo.git.merge(source)
+    except GitCommandError as exc:
+        # If a merge conflict occurs, give some context and try to
+        # make it easy to abort if necessary.
+        click.echo(exc)
+        if not click.confirm(
+            f"Likely merge conflict whilst merging ({source} → {target}). "
+            f"Have you resolved it?"
+        ):
+            repo.git.merge("--abort")
+            return
+
+    # Push result.
+    click.echo("Pushing...")
+    repo.remote().push()
+
+
+@cli.command()
+def merge_back() -> None:
+    """Merge the release branch back into the appropriate branches.
+    All branches will be automatically pulled from the remote and the results
+    will be pushed to the remote."""
+
+    synapse_repo = get_repo_and_check_clean_checkout()
+    branch_name = synapse_repo.active_branch.name
+
+    if not branch_name.startswith("release-v"):
+        raise RuntimeError("Not on a release branch. This does not seem sensible.")
+
+    # Pull so we're up to date
+    synapse_repo.remote().pull()
+
+    current_version = get_package_version()
+
+    if current_version.is_prerelease:
+        # Release candidate
+        if click.confirm(f"Merge {branch_name} → develop?", default=True):
+            _merge_into(synapse_repo, branch_name, "develop")
+    else:
+        # Full release
+        sytest_repo = get_repo_and_check_clean_checkout("../sytest", "sytest")
+
+        if click.confirm(f"Merge {branch_name} → master?", default=True):
+            _merge_into(synapse_repo, branch_name, "master")
+
+        if click.confirm("Merge master → develop?", default=True):
+            _merge_into(synapse_repo, "master", "develop")
+
+        if click.confirm(f"On SyTest, merge {branch_name} → master?", default=True):
+            _merge_into(sytest_repo, branch_name, "master")
+
+        if click.confirm("On SyTest, merge master → develop?", default=True):
+            _merge_into(sytest_repo, "master", "develop")
+
+
 @cli.command()
 def announce() -> None:
     """Generate markdown to announce the release."""
@@ -469,14 +563,18 @@ def get_release_branch_name(version_number: version.Version) -> str:
     return f"release-v{version_number.major}.{version_number.minor}"
 
 
-def get_repo_and_check_clean_checkout() -> git.Repo:
+def get_repo_and_check_clean_checkout(
+    path: str = ".", name: str = "synapse"
+) -> git.Repo:
     """Get the project repo and check it's not got any uncommitted changes."""
     try:
-        repo = git.Repo()
+        repo = git.Repo(path=path)
     except git.InvalidGitRepositoryError:
-        raise click.ClickException("Not in Synapse repo.")
+        raise click.ClickException(
+            f"{path} is not a git repository (expecting a {name} repository)."
+        )
     if repo.is_dirty():
-        raise click.ClickException("Uncommitted changes exist.")
+        raise click.ClickException(f"Uncommitted changes exist in {path}.")
     return repo
 
 

synapse/_scripts/generate_config.py

@@ -33,7 +33,7 @@ def main() -> None:
     parser.add_argument(
         "--report-stats",
         action="store",
-        help="Whether the generated config reports anonymized usage statistics",
+        help="Whether the generated config reports homeserver usage statistics",
         choices=["yes", "no"],
     )
 

synapse/api/auth.py

@@ -26,11 +26,17 @@ from synapse.api.errors import (
     Codes,
     InvalidClientTokenError,
     MissingClientTokenError,
+    UnstableSpecAuthError,
 )
 from synapse.appservice import ApplicationService
 from synapse.http import get_request_user_agent
 from synapse.http.site import SynapseRequest
-from synapse.logging.opentracing import active_span, force_tracing, start_active_span
+from synapse.logging.opentracing import (
+    active_span,
+    force_tracing,
+    start_active_span,
+    trace,
+)
 from synapse.storage.databases.main.registration import TokenLookupResult
 from synapse.types import Requester, UserID, create_requester
 
@@ -106,8 +112,11 @@ class Auth:
                 forgot = await self.store.did_forget(user_id, room_id)
                 if not forgot:
                     return membership, member_event_id
-
-        raise AuthError(403, "User %s not in room %s" % (user_id, room_id))
+        raise UnstableSpecAuthError(
+            403,
+            "User %s not in room %s" % (user_id, room_id),
+            errcode=Codes.NOT_JOINED,
+        )
 
     async def get_user_by_req(
         self,
@@ -563,6 +572,7 @@ class Auth:
 
             return query_params[0].decode("ascii")
 
+    @trace
     async def check_user_in_room_or_world_readable(
         self, room_id: str, user_id: str, allow_departed_users: bool = False
     ) -> Tuple[str, Optional[str]]:
@@ -600,8 +610,9 @@ class Auth:
                 == HistoryVisibility.WORLD_READABLE
             ):
                 return Membership.JOIN, None
-            raise AuthError(
+            raise UnstableSpecAuthError(
                 403,
                 "User %s not in room %s, and room previews are disabled"
                 % (user_id, room_id),
+                errcode=Codes.NOT_JOINED,
             )

synapse/api/constants.py

@@ -257,7 +257,8 @@ class GuestAccess:
 
 class ReceiptTypes:
     READ: Final = "m.read"
-    READ_PRIVATE: Final = "org.matrix.msc2285.read.private"
+    READ_PRIVATE: Final = "m.read.private"
+    UNSTABLE_READ_PRIVATE: Final = "org.matrix.msc2285.read.private"
     FULLY_READ: Final = "m.fully_read"
 
 
@@ -268,4 +269,4 @@ class PublicRoomsFilterFields:
     """
 
     GENERIC_SEARCH_TERM: Final = "generic_search_term"
-    ROOM_TYPES: Final = "org.matrix.msc3827.room_types"
+    ROOM_TYPES: Final = "room_types"

synapse/api/errors.py

@@ -26,6 +26,7 @@ from twisted.web import http
 from synapse.util import json_decoder
 
 if typing.TYPE_CHECKING:
+    from synapse.config.homeserver import HomeServerConfig
     from synapse.types import JsonDict
 
 logger = logging.getLogger(__name__)
@@ -80,6 +81,12 @@ class Codes(str, Enum):
     INVALID_SIGNATURE = "M_INVALID_SIGNATURE"
     USER_DEACTIVATED = "M_USER_DEACTIVATED"
 
+    # Part of MSC3848
+    # https://github.com/matrix-org/matrix-spec-proposals/pull/3848
+    ALREADY_JOINED = "ORG.MATRIX.MSC3848.ALREADY_JOINED"
+    NOT_JOINED = "ORG.MATRIX.MSC3848.NOT_JOINED"
+    INSUFFICIENT_POWER = "ORG.MATRIX.MSC3848.INSUFFICIENT_POWER"
+
     # The account has been suspended on the server.
     # By opposition to `USER_DEACTIVATED`, this is a reversible measure
     # that can possibly be appealed and reverted.
@@ -167,7 +174,7 @@ class SynapseError(CodeMessageException):
         else:
             self._additional_fields = dict(additional_fields)
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, **self._additional_fields)
 
 
@@ -213,7 +220,7 @@ class ConsentNotGivenError(SynapseError):
         )
         self._consent_uri = consent_uri
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, consent_uri=self._consent_uri)
 
 
@@ -307,6 +314,37 @@ class AuthError(SynapseError):
         super().__init__(code, msg, errcode, additional_fields)
 
 
+class UnstableSpecAuthError(AuthError):
+    """An error raised when a new error code is being proposed to replace a previous one.
+    This error will return a "org.matrix.unstable.errcode" property with the new error code,
+    with the previous error code still being defined in the "errcode" property.
+
+    This error will include `org.matrix.msc3848.unstable.errcode` in the C-S error body.
+    """
+
+    def __init__(
+        self,
+        code: int,
+        msg: str,
+        errcode: str,
+        previous_errcode: str = Codes.FORBIDDEN,
+        additional_fields: Optional[dict] = None,
+    ):
+        self.previous_errcode = previous_errcode
+        super().__init__(code, msg, errcode, additional_fields)
+
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
+        fields = {}
+        if config is not None and config.experimental.msc3848_enabled:
+            fields["org.matrix.msc3848.unstable.errcode"] = self.errcode
+        return cs_error(
+            self.msg,
+            self.previous_errcode,
+            **fields,
+            **self._additional_fields,
+        )
+
+
 class InvalidClientCredentialsError(SynapseError):
     """An error raised when there was a problem with the authorisation credentials
     in a client request.
@@ -338,8 +376,8 @@ class InvalidClientTokenError(InvalidClientCredentialsError):
         super().__init__(msg=msg, errcode="M_UNKNOWN_TOKEN")
         self._soft_logout = soft_logout
 
-    def error_dict(self) -> "JsonDict":
-        d = super().error_dict()
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
+        d = super().error_dict(config)
         d["soft_logout"] = self._soft_logout
         return d
 
@@ -362,7 +400,7 @@ class ResourceLimitError(SynapseError):
         self.limit_type = limit_type
         super().__init__(code, msg, errcode=errcode)
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(
             self.msg,
             self.errcode,
@@ -397,7 +435,7 @@ class InvalidCaptchaError(SynapseError):
         super().__init__(code, msg, errcode)
         self.error_url = error_url
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, error_url=self.error_url)
 
 
@@ -414,7 +452,7 @@ class LimitExceededError(SynapseError):
         super().__init__(code, msg, errcode)
         self.retry_after_ms = retry_after_ms
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, retry_after_ms=self.retry_after_ms)
 
 
@@ -429,7 +467,7 @@ class RoomKeysVersionError(SynapseError):
         super().__init__(403, "Wrong room_keys version", Codes.WRONG_ROOM_KEYS_VERSION)
         self.current_version = current_version
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, current_version=self.current_version)
 
 
@@ -469,7 +507,7 @@ class IncompatibleRoomVersionError(SynapseError):
 
         self._room_version = room_version
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         return cs_error(self.msg, self.errcode, room_version=self._room_version)
 
 
@@ -515,7 +553,7 @@ class UnredactedContentDeletedError(SynapseError):
         )
         self.content_keep_ms = content_keep_ms
 
-    def error_dict(self) -> "JsonDict":
+    def error_dict(self, config: Optional["HomeServerConfig"]) -> "JsonDict":
         extra = {}
         if self.content_keep_ms is not None:
             extra = {"fi.mau.msc2815.content_keep_ms": self.content_keep_ms}

synapse/api/ratelimiting.py

@@ -17,7 +17,7 @@ from collections import OrderedDict
 from typing import Hashable, Optional, Tuple
 
 from synapse.api.errors import LimitExceededError
-from synapse.config.ratelimiting import RateLimitConfig
+from synapse.config.ratelimiting import RatelimitSettings
 from synapse.storage.databases.main import DataStore
 from synapse.types import Requester
 from synapse.util import Clock
@@ -314,8 +314,8 @@ class RequestRatelimiter:
         self,
         store: DataStore,
         clock: Clock,
-        rc_message: RateLimitConfig,
-        rc_admin_redaction: Optional[RateLimitConfig],
+        rc_message: RatelimitSettings,
+        rc_admin_redaction: Optional[RatelimitSettings],
     ):
         self.store = store
         self.clock = clock

synapse/app/admin_cmd.py

@@ -28,19 +28,22 @@ from synapse.config.homeserver import HomeServerConfig
 from synapse.config.logger import setup_logging
 from synapse.events import EventBase
 from synapse.handlers.admin import ExfiltrationWriter
-from synapse.replication.slave.storage._base import BaseSlavedStore
-from synapse.replication.slave.storage.account_data import SlavedAccountDataStore
-from synapse.replication.slave.storage.appservice import SlavedApplicationServiceStore
-from synapse.replication.slave.storage.deviceinbox import SlavedDeviceInboxStore
 from synapse.replication.slave.storage.devices import SlavedDeviceStore
 from synapse.replication.slave.storage.events import SlavedEventStore
 from synapse.replication.slave.storage.filtering import SlavedFilteringStore
 from synapse.replication.slave.storage.push_rule import SlavedPushRuleStore
-from synapse.replication.slave.storage.receipts import SlavedReceiptsStore
-from synapse.replication.slave.storage.registration import SlavedRegistrationStore
 from synapse.server import HomeServer
 from synapse.storage.database import DatabasePool, LoggingDatabaseConnection
+from synapse.storage.databases.main.account_data import AccountDataWorkerStore
+from synapse.storage.databases.main.appservice import (
+    ApplicationServiceTransactionWorkerStore,
+    ApplicationServiceWorkerStore,
+)
+from synapse.storage.databases.main.deviceinbox import DeviceInboxWorkerStore
+from synapse.storage.databases.main.receipts import ReceiptsWorkerStore
+from synapse.storage.databases.main.registration import RegistrationWorkerStore
 from synapse.storage.databases.main.room import RoomWorkerStore
+from synapse.storage.databases.main.tags import TagsWorkerStore
 from synapse.types import StateMap
 from synapse.util import SYNAPSE_VERSION
 from synapse.util.logcontext import LoggingContext
@@ -49,16 +52,17 @@ logger = logging.getLogger("synapse.app.admin_cmd")
 
 
 class AdminCmdSlavedStore(
-    SlavedReceiptsStore,
-    SlavedAccountDataStore,
-    SlavedApplicationServiceStore,
-    SlavedRegistrationStore,
     SlavedFilteringStore,
-    SlavedDeviceInboxStore,
     SlavedDeviceStore,
     SlavedPushRuleStore,
     SlavedEventStore,
-    BaseSlavedStore,
+    TagsWorkerStore,
+    DeviceInboxWorkerStore,
+    AccountDataWorkerStore,
+    ApplicationServiceTransactionWorkerStore,
+    ApplicationServiceWorkerStore,
+    RegistrationWorkerStore,
+    ReceiptsWorkerStore,
     RoomWorkerStore,
 ):
     def __init__(

synapse/app/generic_worker.py

@@ -48,20 +48,12 @@ from synapse.http.site import SynapseRequest, SynapseSite
 from synapse.logging.context import LoggingContext
 from synapse.metrics import METRICS_PREFIX, MetricsResource, RegistryProxy
 from synapse.replication.http import REPLICATION_PREFIX, ReplicationRestResource
-from synapse.replication.slave.storage._base import BaseSlavedStore
-from synapse.replication.slave.storage.account_data import SlavedAccountDataStore
-from synapse.replication.slave.storage.appservice import SlavedApplicationServiceStore
-from synapse.replication.slave.storage.deviceinbox import SlavedDeviceInboxStore
 from synapse.replication.slave.storage.devices import SlavedDeviceStore
-from synapse.replication.slave.storage.directory import DirectoryStore
 from synapse.replication.slave.storage.events import SlavedEventStore
 from synapse.replication.slave.storage.filtering import SlavedFilteringStore
 from synapse.replication.slave.storage.keys import SlavedKeyStore
-from synapse.replication.slave.storage.profile import SlavedProfileStore
 from synapse.replication.slave.storage.push_rule import SlavedPushRuleStore
 from synapse.replication.slave.storage.pushers import SlavedPusherStore
-from synapse.replication.slave.storage.receipts import SlavedReceiptsStore
-from synapse.replication.slave.storage.registration import SlavedRegistrationStore
 from synapse.rest.admin import register_servlets_for_media_repo
 from synapse.rest.client import (
     account_data,
@@ -100,8 +92,15 @@ from synapse.rest.key.v2 import KeyApiV2Resource
 from synapse.rest.synapse.client import build_synapse_client_resource_tree
 from synapse.rest.well_known import well_known_resource
 from synapse.server import HomeServer
+from synapse.storage.databases.main.account_data import AccountDataWorkerStore
+from synapse.storage.databases.main.appservice import (
+    ApplicationServiceTransactionWorkerStore,
+    ApplicationServiceWorkerStore,
+)
 from synapse.storage.databases.main.censor_events import CensorEventsStore
 from synapse.storage.databases.main.client_ips import ClientIpWorkerStore
+from synapse.storage.databases.main.deviceinbox import DeviceInboxWorkerStore
+from synapse.storage.databases.main.directory import DirectoryWorkerStore
 from synapse.storage.databases.main.e2e_room_keys import EndToEndRoomKeyStore
 from synapse.storage.databases.main.lock import LockStore
 from synapse.storage.databases.main.media_repository import MediaRepositoryStore
@@ -110,11 +109,15 @@ from synapse.storage.databases.main.monthly_active_users import (
     MonthlyActiveUsersWorkerStore,
 )
 from synapse.storage.databases.main.presence import PresenceStore
+from synapse.storage.databases.main.profile import ProfileWorkerStore
+from synapse.storage.databases.main.receipts import ReceiptsWorkerStore
+from synapse.storage.databases.main.registration import RegistrationWorkerStore
 from synapse.storage.databases.main.room import RoomWorkerStore
 from synapse.storage.databases.main.room_batch import RoomBatchStore
 from synapse.storage.databases.main.search import SearchStore
 from synapse.storage.databases.main.session import SessionStore
 from synapse.storage.databases.main.stats import StatsStore
+from synapse.storage.databases.main.tags import TagsWorkerStore
 from synapse.storage.databases.main.transactions import TransactionWorkerStore
 from synapse.storage.databases.main.ui_auth import UIAuthWorkerStore
 from synapse.storage.databases.main.user_directory import UserDirectoryStore
@@ -227,11 +230,11 @@ class GenericWorkerSlavedStore(
     UIAuthWorkerStore,
     EndToEndRoomKeyStore,
     PresenceStore,
-    SlavedDeviceInboxStore,
+    DeviceInboxWorkerStore,
     SlavedDeviceStore,
-    SlavedReceiptsStore,
     SlavedPushRuleStore,
-    SlavedAccountDataStore,
+    TagsWorkerStore,
+    AccountDataWorkerStore,
     SlavedPusherStore,
     CensorEventsStore,
     ClientIpWorkerStore,
@@ -239,19 +242,20 @@ class GenericWorkerSlavedStore(
     SlavedKeyStore,
     RoomWorkerStore,
     RoomBatchStore,
-    DirectoryStore,
-    SlavedApplicationServiceStore,
-    SlavedRegistrationStore,
-    SlavedProfileStore,
+    DirectoryWorkerStore,
+    ApplicationServiceTransactionWorkerStore,
+    ApplicationServiceWorkerStore,
+    ProfileWorkerStore,
     SlavedFilteringStore,
     MonthlyActiveUsersWorkerStore,
     MediaRepositoryStore,
     ServerMetricsStore,
+    ReceiptsWorkerStore,
+    RegistrationWorkerStore,
     SearchStore,
     TransactionWorkerStore,
     LockStore,
     SessionStore,
-    BaseSlavedStore,
 ):
     # Properties that multiple storage classes define. Tell mypy what the
     # expected type is.

synapse/app/homeserver.py

@@ -44,6 +44,7 @@ from synapse.app._base import (
     register_start,
 )
 from synapse.config._base import ConfigError, format_config_error
+from synapse.config.emailconfig import ThreepidBehaviour
 from synapse.config.homeserver import HomeServerConfig
 from synapse.config.server import ListenerConfig
 from synapse.federation.transport.server import TransportLayerServer
@@ -201,7 +202,7 @@ class SynapseHomeServer(HomeServer):
                 }
             )
 
-            if self.config.email.can_verify_email:
+            if self.config.email.threepid_behaviour_email == ThreepidBehaviour.LOCAL:
                 from synapse.rest.synapse.client.password_reset import (
                     PasswordResetSubmitTokenResource,
                 )

synapse/config/_base.py

@@ -97,16 +97,16 @@ def format_config_error(e: ConfigError) -> Iterator[str]:
 # We split these messages out to allow packages to override with package
 # specific instructions.
 MISSING_REPORT_STATS_CONFIG_INSTRUCTIONS = """\
-Please opt in or out of reporting anonymized homeserver usage statistics, by
-setting the `report_stats` key in your config file to either True or False.
+Please opt in or out of reporting homeserver usage statistics, by setting
+the `report_stats` key in your config file to either True or False.
 """
 
 MISSING_REPORT_STATS_SPIEL = """\
 We would really appreciate it if you could help our project out by reporting
-anonymized usage statistics from your homeserver. Only very basic aggregate
-data (e.g. number of users) will be reported, but it helps us to track the
-growth of the Matrix community, and helps us to make Matrix a success, as well
-as to convince other networks that they should peer with us.
+homeserver usage statistics from your homeserver. Your homeserver's server name,
+along with very basic aggregate data (e.g. number of users) will be reported. But
+it helps us to track the growth of the Matrix community, and helps us to make Matrix
+a success, as well as to convince other networks that they should peer with us.
 
 Thank you.
 """
@@ -621,7 +621,7 @@ class RootConfig:
         generate_group.add_argument(
             "--report-stats",
             action="store",
-            help="Whether the generated config reports anonymized usage statistics.",
+            help="Whether the generated config reports homeserver usage statistics.",
             choices=["yes", "no"],
         )
         generate_group.add_argument(

synapse/config/emailconfig.py

@@ -18,6 +18,7 @@
 import email.utils
 import logging
 import os
+from enum import Enum
 from typing import Any
 
 import attr
@@ -85,14 +86,19 @@ class EmailConfig(Config):
         if email_config is None:
             email_config = {}
 
+        self.force_tls = email_config.get("force_tls", False)
         self.email_smtp_host = email_config.get("smtp_host", "localhost")
-        self.email_smtp_port = email_config.get("smtp_port", 25)
+        self.email_smtp_port = email_config.get(
+            "smtp_port", 465 if self.force_tls else 25
+        )
         self.email_smtp_user = email_config.get("smtp_user", None)
         self.email_smtp_pass = email_config.get("smtp_pass", None)
         self.require_transport_security = email_config.get(
             "require_transport_security", False
         )
         self.enable_smtp_tls = email_config.get("enable_tls", True)
+        if self.force_tls and not self.enable_smtp_tls:
+            raise ConfigError("email.force_tls requires email.enable_tls to be true")
         if self.require_transport_security and not self.enable_smtp_tls:
             raise ConfigError(
                 "email.require_transport_security requires email.enable_tls to be true"
@@ -130,22 +136,40 @@ class EmailConfig(Config):
 
         self.email_enable_notifs = email_config.get("enable_notifs", False)
 
+        self.threepid_behaviour_email = (
+            # Have Synapse handle the email sending if account_threepid_delegates.email
+            # is not defined
+            # msisdn is currently always remote while Synapse does not support any method of
+            # sending SMS messages
+            ThreepidBehaviour.REMOTE
+            if self.root.registration.account_threepid_delegate_email
+            else ThreepidBehaviour.LOCAL
+        )
+
         if config.get("trust_identity_server_for_password_resets"):
             raise ConfigError(
-                'The config option "trust_identity_server_for_password_resets" '
-                "is no longer supported. Please remove it from the config file."
+                'The config option "trust_identity_server_for_password_resets" has been removed.'
+                "Please consult the configuration manual at docs/usage/configuration/config_documentation.md for "
+                "details and update your config file."
             )
 
-        # If we have email config settings, assume that we can verify ownership of
-        # email addresses.
-        self.can_verify_email = email_config != {}
+        self.local_threepid_handling_disabled_due_to_email_config = False
+        if (
+            self.threepid_behaviour_email == ThreepidBehaviour.LOCAL
+            and email_config == {}
+        ):
+            # We cannot warn the user this has happened here
+            # Instead do so when a user attempts to reset their password
+            self.local_threepid_handling_disabled_due_to_email_config = True
+
+            self.threepid_behaviour_email = ThreepidBehaviour.OFF
 
         # Get lifetime of a validation token in milliseconds
         self.email_validation_token_lifetime = self.parse_duration(
             email_config.get("validation_token_lifetime", "1h")
         )
 
-        if self.can_verify_email:
+        if self.threepid_behaviour_email == ThreepidBehaviour.LOCAL:
             missing = []
             if not self.email_notif_from:
                 missing.append("email.notif_from")
@@ -336,3 +360,18 @@ class EmailConfig(Config):
                     "Config option email.invite_client_location must be a http or https URL",
                     path=("email", "invite_client_location"),
                 )
+
+
+class ThreepidBehaviour(Enum):
+    """
+    Enum to define the behaviour of Synapse with regards to when it contacts an identity
+    server for 3pid registration and password resets
+
+    REMOTE = use an external server to send tokens
+    LOCAL = send tokens ourselves
+    OFF = disable registration via 3pid and password resets
+    """
+
+    REMOTE = "remote"
+    LOCAL = "local"
+    OFF = "off"

synapse/config/experimental.py

@@ -32,7 +32,7 @@ class ExperimentalConfig(Config):
         # MSC2716 (importing historical messages)
         self.msc2716_enabled: bool = experimental.get("msc2716_enabled", False)
 
-        # MSC2285 (private read receipts)
+        # MSC2285 (unstable private read receipts)
         self.msc2285_enabled: bool = experimental.get("msc2285_enabled", False)
 
         # MSC3244 (room version capabilities)
@@ -88,5 +88,5 @@ class ExperimentalConfig(Config):
         # MSC3715: dir param on /relations.
         self.msc3715_enabled: bool = experimental.get("msc3715_enabled", False)
 
-        # MSC3827: Filtering of /publicRooms by room type
-        self.msc3827_enabled: bool = experimental.get("msc3827_enabled", False)
+        # MSC3848: Introduce errcodes for specific event sending failures
+        self.msc3848_enabled: bool = experimental.get("msc3848_enabled", False)

synapse/config/ratelimiting.py

@@ -21,7 +21,7 @@ from synapse.types import JsonDict
 from ._base import Config
 
 
-class RateLimitConfig:
+class RatelimitSettings:
     def __init__(
         self,
         config: Dict[str, float],
@@ -34,7 +34,7 @@ class RateLimitConfig:
 
 
 @attr.s(auto_attribs=True)
-class FederationRateLimitConfig:
+class FederationRatelimitSettings:
     window_size: int = 1000
     sleep_limit: int = 10
     sleep_delay: int = 500
@@ -50,11 +50,11 @@ class RatelimitConfig(Config):
         # Load the new-style messages config if it exists. Otherwise fall back
         # to the old method.
         if "rc_message" in config:
-            self.rc_message = RateLimitConfig(
+            self.rc_message = RatelimitSettings(
                 config["rc_message"], defaults={"per_second": 0.2, "burst_count": 10.0}
             )
         else:
-            self.rc_message = RateLimitConfig(
+            self.rc_message = RatelimitSettings(
                 {
                     "per_second": config.get("rc_messages_per_second", 0.2),
                     "burst_count": config.get("rc_message_burst_count", 10.0),
@@ -64,9 +64,9 @@ class RatelimitConfig(Config):
         # Load the new-style federation config, if it exists. Otherwise, fall
         # back to the old method.
         if "rc_federation" in config:
-            self.rc_federation = FederationRateLimitConfig(**config["rc_federation"])
+            self.rc_federation = FederationRatelimitSettings(**config["rc_federation"])
         else:
-            self.rc_federation = FederationRateLimitConfig(
+            self.rc_federation = FederationRatelimitSettings(
                 **{
                     k: v
                     for k, v in {
@@ -80,17 +80,17 @@ class RatelimitConfig(Config):
                 }
             )
 
-        self.rc_registration = RateLimitConfig(config.get("rc_registration", {}))
+        self.rc_registration = RatelimitSettings(config.get("rc_registration", {}))
 
-        self.rc_registration_token_validity = RateLimitConfig(
+        self.rc_registration_token_validity = RatelimitSettings(
             config.get("rc_registration_token_validity", {}),
             defaults={"per_second": 0.1, "burst_count": 5},
         )
 
         rc_login_config = config.get("rc_login", {})
-        self.rc_login_address = RateLimitConfig(rc_login_config.get("address", {}))
-        self.rc_login_account = RateLimitConfig(rc_login_config.get("account", {}))
-        self.rc_login_failed_attempts = RateLimitConfig(
+        self.rc_login_address = RatelimitSettings(rc_login_config.get("address", {}))
+        self.rc_login_account = RatelimitSettings(rc_login_config.get("account", {}))
+        self.rc_login_failed_attempts = RatelimitSettings(
             rc_login_config.get("failed_attempts", {})
         )
 
@@ -101,47 +101,54 @@ class RatelimitConfig(Config):
         rc_admin_redaction = config.get("rc_admin_redaction")
         self.rc_admin_redaction = None
         if rc_admin_redaction:
-            self.rc_admin_redaction = RateLimitConfig(rc_admin_redaction)
+            self.rc_admin_redaction = RatelimitSettings(rc_admin_redaction)
 
-        self.rc_joins_local = RateLimitConfig(
+        self.rc_joins_local = RatelimitSettings(
             config.get("rc_joins", {}).get("local", {}),
             defaults={"per_second": 0.1, "burst_count": 10},
         )
-        self.rc_joins_remote = RateLimitConfig(
+        self.rc_joins_remote = RatelimitSettings(
             config.get("rc_joins", {}).get("remote", {}),
             defaults={"per_second": 0.01, "burst_count": 10},
         )
 
+        # Track the rate of joins to a given room. If there are too many, temporarily
+        # prevent local joins and remote joins via this server.
+        self.rc_joins_per_room = RatelimitSettings(
+            config.get("rc_joins_per_room", {}),
+            defaults={"per_second": 1, "burst_count": 10},
+        )
+
         # Ratelimit cross-user key requests:
         # * For local requests this is keyed by the sending device.
         # * For requests received over federation this is keyed by the origin.
         #
         # Note that this isn't exposed in the configuration as it is obscure.
-        self.rc_key_requests = RateLimitConfig(
+        self.rc_key_requests = RatelimitSettings(
             config.get("rc_key_requests", {}),
             defaults={"per_second": 20, "burst_count": 100},
         )
 
-        self.rc_3pid_validation = RateLimitConfig(
+        self.rc_3pid_validation = RatelimitSettings(
             config.get("rc_3pid_validation") or {},
             defaults={"per_second": 0.003, "burst_count": 5},
         )
 
-        self.rc_invites_per_room = RateLimitConfig(
+        self.rc_invites_per_room = RatelimitSettings(
             config.get("rc_invites", {}).get("per_room", {}),
             defaults={"per_second": 0.3, "burst_count": 10},
         )
-        self.rc_invites_per_user = RateLimitConfig(
+        self.rc_invites_per_user = RatelimitSettings(
             config.get("rc_invites", {}).get("per_user", {}),
             defaults={"per_second": 0.003, "burst_count": 5},
         )
 
-        self.rc_invites_per_issuer = RateLimitConfig(
+        self.rc_invites_per_issuer = RatelimitSettings(
             config.get("rc_invites", {}).get("per_issuer", {}),
             defaults={"per_second": 0.3, "burst_count": 10},
         )
 
-        self.rc_third_party_invite = RateLimitConfig(
+        self.rc_third_party_invite = RatelimitSettings(
             config.get("rc_third_party_invite", {}),
             defaults={
                 "per_second": self.rc_message.per_second,

synapse/config/registration.py

@@ -13,6 +13,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import argparse
+import logging
 from typing import Any, Optional
 
 from synapse.api.constants import RoomCreationPreset
@@ -20,11 +21,15 @@ from synapse.config._base import Config, ConfigError
 from synapse.types import JsonDict, RoomAlias, UserID
 from synapse.util.stringutils import random_string_with_symbols, strtobool
 
-NO_EMAIL_DELEGATE_ERROR = """\
-Delegation of email verification to an identity server is no longer supported. To
+logger = logging.getLogger(__name__)
+
+LEGACY_EMAIL_DELEGATE_WARNING = """\
+Delegation of email verification to an identity server is now deprecated. To
 continue to allow users to add email addresses to their accounts, and use them for
 password resets, configure Synapse with an SMTP server via the `email` setting, and
 remove `account_threepid_delegates.email`.
+
+This will be an error in a future version.
 """
 
 
@@ -59,8 +64,9 @@ class RegistrationConfig(Config):
 
         account_threepid_delegates = config.get("account_threepid_delegates") or {}
         if "email" in account_threepid_delegates:
-            raise ConfigError(NO_EMAIL_DELEGATE_ERROR)
-        # self.account_threepid_delegate_email = account_threepid_delegates.get("email")
+            logger.warning(LEGACY_EMAIL_DELEGATE_WARNING)
+
+        self.account_threepid_delegate_email = account_threepid_delegates.get("email")
         self.account_threepid_delegate_msisdn = account_threepid_delegates.get("msisdn")
         self.default_identity_server = config.get("default_identity_server")
         self.allow_guest_access = config.get("allow_guest_access", False)

synapse/config/repository.py

@@ -42,6 +42,18 @@ THUMBNAIL_SIZE_YAML = """\
         #    method: %(method)s
 """
 
+# A map from the given media type to the type of thumbnail we should generate
+# for it.
+THUMBNAIL_SUPPORTED_MEDIA_FORMAT_MAP = {
+    "image/jpeg": "jpeg",
+    "image/jpg": "jpeg",
+    "image/webp": "jpeg",
+    # Thumbnails can only be jpeg or png. We choose png thumbnails for gif
+    # because it can have transparency.
+    "image/gif": "png",
+    "image/png": "png",
+}
+
 HTTP_PROXY_SET_WARNING = """\
 The Synapse config url_preview_ip_range_blacklist will be ignored as an HTTP(s) proxy is configured."""
 
@@ -79,13 +91,22 @@ def parse_thumbnail_requirements(
         width = size["width"]
         height = size["height"]
         method = size["method"]
-        jpeg_thumbnail = ThumbnailRequirement(width, height, method, "image/jpeg")
-        png_thumbnail = ThumbnailRequirement(width, height, method, "image/png")
-        requirements.setdefault("image/jpeg", []).append(jpeg_thumbnail)
-        requirements.setdefault("image/jpg", []).append(jpeg_thumbnail)
-        requirements.setdefault("image/webp", []).append(jpeg_thumbnail)
-        requirements.setdefault("image/gif", []).append(png_thumbnail)
-        requirements.setdefault("image/png", []).append(png_thumbnail)
+
+        for format, thumbnail_format in THUMBNAIL_SUPPORTED_MEDIA_FORMAT_MAP.items():
+            requirement = requirements.setdefault(format, [])
+            if thumbnail_format == "jpeg":
+                requirement.append(
+                    ThumbnailRequirement(width, height, method, "image/jpeg")
+                )
+            elif thumbnail_format == "png":
+                requirement.append(
+                    ThumbnailRequirement(width, height, method, "image/png")
+                )
+            else:
+                raise Exception(
+                    "Unknown thumbnail mapping from %s to %s. This is a Synapse problem, please report!"
+                    % (format, thumbnail_format)
+                )
     return {
         media_type: tuple(thumbnails) for media_type, thumbnails in requirements.items()
     }

synapse/event_auth.py

@@ -30,7 +30,13 @@ from synapse.api.constants import (
     JoinRules,
     Membership,
 )
-from synapse.api.errors import AuthError, EventSizeError, SynapseError
+from synapse.api.errors import (
+    AuthError,
+    Codes,
+    EventSizeError,
+    SynapseError,
+    UnstableSpecAuthError,
+)
 from synapse.api.room_versions import (
     KNOWN_ROOM_VERSIONS,
     EventFormatVersions,
@@ -291,7 +297,11 @@ def check_state_dependent_auth_rules(
         invite_level = get_named_level(auth_dict, "invite", 0)
 
         if user_level < invite_level:
-            raise AuthError(403, "You don't have permission to invite users")
+            raise UnstableSpecAuthError(
+                403,
+                "You don't have permission to invite users",
+                errcode=Codes.INSUFFICIENT_POWER,
+            )
         else:
             logger.debug("Allowing! %s", event)
             return
@@ -474,7 +484,11 @@ def _is_membership_change_allowed(
             return
 
         if not caller_in_room:  # caller isn't joined
-            raise AuthError(403, "%s not in room %s." % (event.user_id, event.room_id))
+            raise UnstableSpecAuthError(
+                403,
+                "%s not in room %s." % (event.user_id, event.room_id),
+                errcode=Codes.NOT_JOINED,
+            )
 
     if Membership.INVITE == membership:
         # TODO (erikj): We should probably handle this more intelligently
@@ -484,10 +498,18 @@ def _is_membership_change_allowed(
         if target_banned:
             raise AuthError(403, "%s is banned from the room" % (target_user_id,))
         elif target_in_room:  # the target is already in the room.
-            raise AuthError(403, "%s is already in the room." % target_user_id)
+            raise UnstableSpecAuthError(
+                403,
+                "%s is already in the room." % target_user_id,
+                errcode=Codes.ALREADY_JOINED,
+            )
         else:
             if user_level < invite_level:
-                raise AuthError(403, "You don't have permission to invite users")
+                raise UnstableSpecAuthError(
+                    403,
+                    "You don't have permission to invite users",
+                    errcode=Codes.INSUFFICIENT_POWER,
+                )
     elif Membership.JOIN == membership:
         # Joins are valid iff caller == target and:
         # * They are not banned.
@@ -549,15 +571,27 @@ def _is_membership_change_allowed(
     elif Membership.LEAVE == membership:
         # TODO (erikj): Implement kicks.
         if target_banned and user_level < ban_level:
-            raise AuthError(403, "You cannot unban user %s." % (target_user_id,))
+            raise UnstableSpecAuthError(
+                403,
+                "You cannot unban user %s." % (target_user_id,),
+                errcode=Codes.INSUFFICIENT_POWER,
+            )
         elif target_user_id != event.user_id:
             kick_level = get_named_level(auth_events, "kick", 50)
 
             if user_level < kick_level or user_level <= target_level:
-                raise AuthError(403, "You cannot kick user %s." % target_user_id)
+                raise UnstableSpecAuthError(
+                    403,
+                    "You cannot kick user %s." % target_user_id,
+                    errcode=Codes.INSUFFICIENT_POWER,
+                )
     elif Membership.BAN == membership:
         if user_level < ban_level or user_level <= target_level:
-            raise AuthError(403, "You don't have permission to ban")
+            raise UnstableSpecAuthError(
+                403,
+                "You don't have permission to ban",
+                errcode=Codes.INSUFFICIENT_POWER,
+            )
     elif room_version.msc2403_knocking and Membership.KNOCK == membership:
         if join_rule != JoinRules.KNOCK and (
             not room_version.msc3787_knock_restricted_join_rule
@@ -567,7 +601,11 @@ def _is_membership_change_allowed(
         elif target_user_id != event.user_id:
             raise AuthError(403, "You cannot knock for other users")
         elif target_in_room:
-            raise AuthError(403, "You cannot knock on a room you are already in")
+            raise UnstableSpecAuthError(
+                403,
+                "You cannot knock on a room you are already in",
+                errcode=Codes.ALREADY_JOINED,
+            )
         elif caller_invited:
             raise AuthError(403, "You are already invited to this room")
         elif target_banned:
@@ -638,10 +676,11 @@ def _can_send_event(event: "EventBase", auth_events: StateMap["EventBase"]) -> b
     user_level = get_user_power_level(event.user_id, auth_events)
 
     if user_level < send_level:
-        raise AuthError(
+        raise UnstableSpecAuthError(
             403,
             "You don't have permission to post that to the room. "
             + "user_level (%d) < send_level (%d)" % (user_level, send_level),
+            errcode=Codes.INSUFFICIENT_POWER,
         )
 
     # Check state_key
@@ -716,9 +755,10 @@ def check_historical(
     historical_level = get_named_level(auth_events, "historical", 100)
 
     if user_level < historical_level:
-        raise AuthError(
+        raise UnstableSpecAuthError(
             403,
             'You don\'t have permission to send send historical related events ("insertion", "batch", and "marker")',
+            errcode=Codes.INSUFFICIENT_POWER,
         )
 
 

synapse/events/builder.py

@@ -24,9 +24,11 @@ from synapse.api.room_versions import (
     RoomVersion,
 )
 from synapse.crypto.event_signing import add_hashes_and_signatures
+from synapse.event_auth import auth_types_for_event
 from synapse.events import EventBase, _EventInternalMetadata, make_event_from_dict
 from synapse.state import StateHandler
 from synapse.storage.databases.main import DataStore
+from synapse.storage.state import StateFilter
 from synapse.types import EventID, JsonDict
 from synapse.util import Clock
 from synapse.util.stringutils import random_string
@@ -121,7 +123,11 @@ class EventBuilder:
         """
         if auth_event_ids is None:
             state_ids = await self._state.compute_state_after_events(
-                self.room_id, prev_event_ids
+                self.room_id,
+                prev_event_ids,
+                state_filter=StateFilter.from_types(
+                    auth_types_for_event(self.room_version, self)
+                ),
             )
             auth_event_ids = self._event_auth_handler.compute_auth_events(
                 self, state_ids

synapse/events/snapshot.py

@@ -11,11 +11,10 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-from typing import TYPE_CHECKING, List, Optional, Tuple, Union
+from typing import TYPE_CHECKING, List, Optional, Tuple
 
 import attr
 from frozendict import frozendict
-from typing_extensions import Literal
 
 from synapse.appservice import ApplicationService
 from synapse.events import EventBase
@@ -33,7 +32,7 @@ class EventContext:
     Holds information relevant to persisting an event
 
     Attributes:
-        rejected: A rejection reason if the event was rejected, else False
+        rejected: A rejection reason if the event was rejected, else None
 
         _state_group: The ID of the state group for this event. Note that state events
             are persisted with a state group which includes the new event, so this is
@@ -85,7 +84,7 @@ class EventContext:
     """
 
     _storage: "StorageControllers"
-    rejected: Union[Literal[False], str] = False
+    rejected: Optional[str] = None
     _state_group: Optional[int] = None
     state_group_before_event: Optional[int] = None
     _state_delta_due_to_event: Optional[StateMap[str]] = None

synapse/federation/federation_client.py

@@ -53,7 +53,7 @@ from synapse.api.room_versions import (
     RoomVersion,
     RoomVersions,
 )
-from synapse.events import EventBase, builder
+from synapse.events import EventBase, builder, make_event_from_dict
 from synapse.federation.federation_base import (
     FederationBase,
     InvalidEventSignatureError,
@@ -61,6 +61,7 @@ from synapse.federation.federation_base import (
 )
 from synapse.federation.transport.client import SendJoinResponse
 from synapse.http.types import QueryParams
+from synapse.logging.opentracing import trace
 from synapse.types import JsonDict, UserID, get_domain_from_id
 from synapse.util.async_helpers import concurrently_execute
 from synapse.util.caches.expiringcache import ExpiringCache
@@ -217,7 +218,7 @@ class FederationClient(FederationBase):
         )
 
     async def claim_client_keys(
-        self, destination: str, content: JsonDict, timeout: int
+        self, destination: str, content: JsonDict, timeout: Optional[int]
     ) -> JsonDict:
         """Claims one-time keys for a device hosted on a remote server.
 
@@ -233,6 +234,7 @@ class FederationClient(FederationBase):
             destination, content, timeout
         )
 
+    @trace
     async def backfill(
         self, dest: str, room_id: str, limit: int, extremities: Collection[str]
     ) -> Optional[List[EventBase]]:
@@ -299,7 +301,8 @@ class FederationClient(FederationBase):
                 moving to the next destination. None indicates no timeout.
 
         Returns:
-            The requested PDU, or None if we were unable to find it.
+            A copy of the requested PDU that is safe to modify, or None if we
+            were unable to find it.
 
         Raises:
             SynapseError, NotRetryingDestination, FederationDeniedError
@@ -309,7 +312,7 @@ class FederationClient(FederationBase):
         )
 
         logger.debug(
-            "retrieved event id %s from %s: %r",
+            "get_pdu_from_destination_raw: retrieved event id %s from %s: %r",
             event_id,
             destination,
             transaction_data,
@@ -358,54 +361,92 @@ class FederationClient(FederationBase):
             The requested PDU, or None if we were unable to find it.
         """
 
+        logger.debug(
+            "get_pdu: event_id=%s from destinations=%s", event_id, destinations
+        )
+
         # TODO: Rate limit the number of times we try and get the same event.
 
-        ev = self._get_pdu_cache.get(event_id)
-        if ev:
-            return ev
+        # We might need the same event multiple times in quick succession (before
+        # it gets persisted to the database), so we cache the results of the lookup.
+        # Note that this is separate to the regular get_event cache which caches
+        # events once they have been persisted.
+        event = self._get_pdu_cache.get(event_id)
 
-        pdu_attempts = self.pdu_destination_tried.setdefault(event_id, {})
+        # If we don't see the event in the cache, go try to fetch it from the
+        # provided remote federated destinations
+        if not event:
+            pdu_attempts = self.pdu_destination_tried.setdefault(event_id, {})
 
-        signed_pdu = None
-        for destination in destinations:
-            now = self._clock.time_msec()
-            last_attempt = pdu_attempts.get(destination, 0)
-            if last_attempt + PDU_RETRY_TIME_MS > now:
-                continue
+            for destination in destinations:
+                now = self._clock.time_msec()
+                last_attempt = pdu_attempts.get(destination, 0)
+                if last_attempt + PDU_RETRY_TIME_MS > now:
+                    logger.debug(
+                        "get_pdu: skipping destination=%s because we tried it recently last_attempt=%s and we only check every %s (now=%s)",
+                        destination,
+                        last_attempt,
+                        PDU_RETRY_TIME_MS,
+                        now,
+                    )
+                    continue
 
-            try:
-                signed_pdu = await self.get_pdu_from_destination_raw(
-                    destination=destination,
-                    event_id=event_id,
-                    room_version=room_version,
-                    timeout=timeout,
-                )
+                try:
+                    event = await self.get_pdu_from_destination_raw(
+                        destination=destination,
+                        event_id=event_id,
+                        room_version=room_version,
+                        timeout=timeout,
+                    )
 
-                pdu_attempts[destination] = now
+                    pdu_attempts[destination] = now
 
-            except SynapseError as e:
-                logger.info(
-                    "Failed to get PDU %s from %s because %s", event_id, destination, e
-                )
-                continue
-            except NotRetryingDestination as e:
-                logger.info(str(e))
-                continue
-            except FederationDeniedError as e:
-                logger.info(str(e))
-                continue
-            except Exception as e:
-                pdu_attempts[destination] = now
+                    if event:
+                        # Prime the cache
+                        self._get_pdu_cache[event.event_id] = event
 
-                logger.info(
-                    "Failed to get PDU %s from %s because %s", event_id, destination, e
-                )
-                continue
+                        # Now that we have an event, we can break out of this
+                        # loop and stop asking other destinations.
+                        break
 
-        if signed_pdu:
-            self._get_pdu_cache[event_id] = signed_pdu
+                except SynapseError as e:
+                    logger.info(
+                        "Failed to get PDU %s from %s because %s",
+                        event_id,
+                        destination,
+                        e,
+                    )
+                    continue
+                except NotRetryingDestination as e:
+                    logger.info(str(e))
+                    continue
+                except FederationDeniedError as e:
+                    logger.info(str(e))
+                    continue
+                except Exception as e:
+                    pdu_attempts[destination] = now
 
-        return signed_pdu
+                    logger.info(
+                        "Failed to get PDU %s from %s because %s",
+                        event_id,
+                        destination,
+                        e,
+                    )
+                    continue
+
+        if not event:
+            return None
+
+        # `event` now refers to an object stored in `get_pdu_cache`. Our
+        # callers may need to modify the returned object (eg to set
+        # `event.internal_metadata.outlier = true`), so we return a copy
+        # rather than the original object.
+        event_copy = make_event_from_dict(
+            event.get_pdu_json(),
+            event.room_version,
+        )
+
+        return event_copy
 
     async def get_room_state_ids(
         self, destination: str, room_id: str, event_id: str
@@ -686,6 +727,12 @@ class FederationClient(FederationBase):
         if failover_errcodes is None:
             failover_errcodes = ()
 
+        if not destinations:
+            # Give a bit of a clearer message if no servers were specified at all.
+            raise SynapseError(
+                502, f"Failed to {description} via any server: No servers specified."
+            )
+
         for destination in destinations:
             if destination == self.server_name:
                 continue
@@ -735,7 +782,7 @@ class FederationClient(FederationBase):
                     "Failed to %s via %s", description, destination, exc_info=True
                 )
 
-        raise SynapseError(502, "Failed to %s via any server" % (description,))
+        raise SynapseError(502, f"Failed to {description} via any server")
 
     async def make_membership_event(
         self,

synapse/federation/federation_server.py

@@ -118,6 +118,7 @@ class FederationServer(FederationBase):
         self._federation_event_handler = hs.get_federation_event_handler()
         self.state = hs.get_state_handler()
         self._event_auth_handler = hs.get_event_auth_handler()
+        self._room_member_handler = hs.get_room_member_handler()
 
         self._state_storage_controller = hs.get_storage_controllers().state
 
@@ -468,7 +469,7 @@ class FederationServer(FederationBase):
                     )
                     for pdu in pdus_by_room[room_id]:
                         event_id = pdu.event_id
-                        pdu_results[event_id] = e.error_dict()
+                        pdu_results[event_id] = e.error_dict(self.hs.config)
                     return
 
                 for pdu in pdus_by_room[room_id]:
@@ -621,6 +622,15 @@ class FederationServer(FederationBase):
             )
             raise IncompatibleRoomVersionError(room_version=room_version)
 
+        # Refuse the request if that room has seen too many joins recently.
+        # This is in addition to the HS-level rate limiting applied by
+        # BaseFederationServlet.
+        # type-ignore: mypy doesn't seem able to deduce the type of the limiter(!?)
+        await self._room_member_handler._join_rate_per_room_limiter.ratelimit(  # type: ignore[has-type]
+            requester=None,
+            key=room_id,
+            update=False,
+        )
         pdu = await self.handler.on_make_join_request(origin, room_id, user_id)
         return {"event": pdu.get_templated_pdu_json(), "room_version": room_version}
 
@@ -655,6 +665,12 @@ class FederationServer(FederationBase):
         room_id: str,
         caller_supports_partial_state: bool = False,
     ) -> Dict[str, Any]:
+        await self._room_member_handler._join_rate_per_room_limiter.ratelimit(  # type: ignore[has-type]
+            requester=None,
+            key=room_id,
+            update=False,
+        )
+
         event, context = await self._on_send_membership_event(
             origin, content, Membership.JOIN, room_id
         )
@@ -827,8 +843,25 @@ class FederationServer(FederationBase):
                 Codes.BAD_JSON,
             )
 
+        # Note that get_room_version throws if the room does not exist here.
         room_version = await self.store.get_room_version(room_id)
 
+        if await self.store.is_partial_state_room(room_id):
+            # If our server is still only partially joined, we can't give a complete
+            # response to /send_join, /send_knock or /send_leave.
+            # This is because we will not be able to provide the server list (for partial
+            # joins) or the full state (for full joins).
+            # Return a 404 as we would if we weren't in the room at all.
+            logger.info(
+                f"Rejecting /send_{membership_type} to %s because it's a partial state room",
+                room_id,
+            )
+            raise SynapseError(
+                404,
+                f"Unable to handle /send_{membership_type} right now; this server is not fully joined.",
+                errcode=Codes.NOT_FOUND,
+            )
+
         if membership_type == Membership.KNOCK and not room_version.msc2403_knocking:
             raise SynapseError(
                 403,

synapse/federation/transport/client.py

@@ -619,7 +619,7 @@ class TransportLayerClient:
         )
 
     async def claim_client_keys(
-        self, destination: str, query_content: JsonDict, timeout: int
+        self, destination: str, query_content: JsonDict, timeout: Optional[int]
     ) -> JsonDict:
         """Claim one-time keys for a list of devices hosted on a remote server.
 

synapse/federation/transport/server/_base.py

@@ -309,7 +309,7 @@ class BaseFederationServlet:
                 raise
 
             # update the active opentracing span with the authenticated entity
-            set_tag("authenticated_entity", origin)
+            set_tag("authenticated_entity", str(origin))
 
             # if the origin is authenticated and whitelisted, use its span context
             # as the parent.

synapse/handlers/auth.py

@@ -565,7 +565,7 @@ class AuthHandler:
             except LoginError as e:
                 # this step failed. Merge the error dict into the response
                 # so that the client can have another go.
-                errordict = e.error_dict()
+                errordict = e.error_dict(self.hs.config)
 
         creds = await self.store.get_completed_ui_auth_stages(session.session_id)
         for f in flows:

synapse/handlers/device.py

@@ -118,8 +118,8 @@ class DeviceWorkerHandler:
         ips = await self.store.get_last_client_ip_by_device(user_id, device_id)
         _update_device_from_client_ips(device, ips)
 
-        set_tag("device", device)
-        set_tag("ips", ips)
+        set_tag("device", str(device))
+        set_tag("ips", str(ips))
 
         return device
 
@@ -170,7 +170,7 @@ class DeviceWorkerHandler:
         """
 
         set_tag("user_id", user_id)
-        set_tag("from_token", from_token)
+        set_tag("from_token", str(from_token))
         now_room_key = self.store.get_room_max_token()
 
         room_ids = await self.store.get_rooms_for_user(user_id)
@@ -795,7 +795,7 @@ class DeviceListUpdater:
         """
 
         set_tag("origin", origin)
-        set_tag("edu_content", edu_content)
+        set_tag("edu_content", str(edu_content))
         user_id = edu_content.pop("user_id")
         device_id = edu_content.pop("device_id")
         stream_id = str(edu_content.pop("stream_id"))  # They may come as ints

synapse/handlers/e2e_keys.py

@@ -15,7 +15,7 @@
 # limitations under the License.
 
 import logging
-from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Optional, Tuple
+from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Mapping, Optional, Tuple
 
 import attr
 from canonicaljson import encode_canonical_json
@@ -92,7 +92,11 @@ class E2eKeysHandler:
 
     @trace
     async def query_devices(
-        self, query_body: JsonDict, timeout: int, from_user_id: str, from_device_id: str
+        self,
+        query_body: JsonDict,
+        timeout: int,
+        from_user_id: str,
+        from_device_id: Optional[str],
     ) -> JsonDict:
         """Handle a device key query from a client
 
@@ -120,9 +124,7 @@ class E2eKeysHandler:
                 the number of in-flight queries at a time.
         """
         async with self._query_devices_linearizer.queue((from_user_id, from_device_id)):
-            device_keys_query: Dict[str, Iterable[str]] = query_body.get(
-                "device_keys", {}
-            )
+            device_keys_query: Dict[str, List[str]] = query_body.get("device_keys", {})
 
             # separate users by domain.
             # make a map from domain to user_id to device_ids
@@ -136,8 +138,8 @@ class E2eKeysHandler:
                 else:
                     remote_queries[user_id] = device_ids
 
-            set_tag("local_key_query", local_query)
-            set_tag("remote_key_query", remote_queries)
+            set_tag("local_key_query", str(local_query))
+            set_tag("remote_key_query", str(remote_queries))
 
             # First get local devices.
             # A map of destination -> failure response.
@@ -341,7 +343,7 @@ class E2eKeysHandler:
             failure = _exception_to_failure(e)
             failures[destination] = failure
             set_tag("error", True)
-            set_tag("reason", failure)
+            set_tag("reason", str(failure))
 
         return
 
@@ -392,7 +394,7 @@ class E2eKeysHandler:
 
     @trace
     async def query_local_devices(
-        self, query: Dict[str, Optional[List[str]]]
+        self, query: Mapping[str, Optional[List[str]]]
     ) -> Dict[str, Dict[str, dict]]:
         """Get E2E device keys for local users
 
@@ -403,7 +405,7 @@ class E2eKeysHandler:
         Returns:
             A map from user_id -> device_id -> device details
         """
-        set_tag("local_query", query)
+        set_tag("local_query", str(query))
         local_query: List[Tuple[str, Optional[str]]] = []
 
         result_dict: Dict[str, Dict[str, dict]] = {}
@@ -461,7 +463,7 @@ class E2eKeysHandler:
 
     @trace
     async def claim_one_time_keys(
-        self, query: Dict[str, Dict[str, Dict[str, str]]], timeout: int
+        self, query: Dict[str, Dict[str, Dict[str, str]]], timeout: Optional[int]
     ) -> JsonDict:
         local_query: List[Tuple[str, str, str]] = []
         remote_queries: Dict[str, Dict[str, Dict[str, str]]] = {}
@@ -475,8 +477,8 @@ class E2eKeysHandler:
                 domain = get_domain_from_id(user_id)
                 remote_queries.setdefault(domain, {})[user_id] = one_time_keys
 
-        set_tag("local_key_query", local_query)
-        set_tag("remote_key_query", remote_queries)
+        set_tag("local_key_query", str(local_query))
+        set_tag("remote_key_query", str(remote_queries))
 
         results = await self.store.claim_e2e_one_time_keys(local_query)
 
@@ -506,7 +508,7 @@ class E2eKeysHandler:
                 failure = _exception_to_failure(e)
                 failures[destination] = failure
                 set_tag("error", True)
-                set_tag("reason", failure)
+                set_tag("reason", str(failure))
 
         await make_deferred_yieldable(
             defer.gatherResults(
@@ -609,7 +611,7 @@ class E2eKeysHandler:
 
         result = await self.store.count_e2e_one_time_keys(user_id, device_id)
 
-        set_tag("one_time_key_counts", result)
+        set_tag("one_time_key_counts", str(result))
         return {"one_time_key_counts": result}
 
     async def _upload_one_time_keys_for_user(

synapse/handlers/e2e_room_keys.py

@@ -14,7 +14,7 @@
 # limitations under the License.
 
 import logging
-from typing import TYPE_CHECKING, Dict, Optional
+from typing import TYPE_CHECKING, Dict, Optional, cast
 
 from typing_extensions import Literal
 
@@ -97,7 +97,7 @@ class E2eRoomKeysHandler:
                 user_id, version, room_id, session_id
             )
 
-            log_kv(results)
+            log_kv(cast(JsonDict, results))
             return results
 
     @trace

synapse/handlers/federation.py

@@ -59,6 +59,7 @@ from synapse.events.validator import EventValidator
 from synapse.federation.federation_client import InvalidResponseError
 from synapse.http.servlet import assert_params_in_dict
 from synapse.logging.context import nested_logging_context
+from synapse.logging.opentracing import trace
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.module_api import NOT_SPAM
 from synapse.replication.http.federation import (
@@ -180,6 +181,7 @@ class FederationHandler:
                 "resume_sync_partial_state_room", self._resume_sync_partial_state_room
             )
 
+    @trace
     async def maybe_backfill(
         self, room_id: str, current_depth: int, limit: int
     ) -> bool:
@@ -546,9 +548,9 @@ class FederationHandler:
             )
 
             if ret.partial_state:
-                # TODO(faster_joins): roll this back if we don't manage to start the
-                #   background resync (eg process_remote_join fails)
-                #   https://github.com/matrix-org/synapse/issues/12998
+                # Mark the room as having partial state.
+                # The background process is responsible for unmarking this flag,
+                # even if the join fails.
                 await self.store.store_partial_state_room(room_id, ret.servers_in_room)
 
             try:
@@ -574,17 +576,21 @@ class FederationHandler:
                     room_id,
                 )
                 raise LimitExceededError(msg=e.msg, errcode=e.errcode, retry_after_ms=0)
-
-            if ret.partial_state:
-                # Kick off the process of asynchronously fetching the state for this
-                # room.
-                run_as_background_process(
-                    desc="sync_partial_state_room",
-                    func=self._sync_partial_state_room,
-                    initial_destination=origin,
-                    other_destinations=ret.servers_in_room,
-                    room_id=room_id,
-                )
+            finally:
+                # Always kick off the background process that asynchronously fetches
+                # state for the room.
+                # If the join failed, the background process is responsible for
+                # cleaning up — including unmarking the room as a partial state room.
+                if ret.partial_state:
+                    # Kick off the process of asynchronously fetching the state for this
+                    # room.
+                    run_as_background_process(
+                        desc="sync_partial_state_room",
+                        func=self._sync_partial_state_room,
+                        initial_destination=origin,
+                        other_destinations=ret.servers_in_room,
+                        room_id=room_id,
+                    )
 
             # We wait here until this instance has seen the events come down
             # replication (if we're using replication) as the below uses caches.
@@ -748,6 +754,23 @@ class FederationHandler:
         # (and return a 404 otherwise)
         room_version = await self.store.get_room_version(room_id)
 
+        if await self.store.is_partial_state_room(room_id):
+            # If our server is still only partially joined, we can't give a complete
+            # response to /make_join, so return a 404 as we would if we weren't in the
+            # room at all.
+            # The main reason we can't respond properly is that we need to know about
+            # the auth events for the join event that we would return.
+            # We also should not bother entertaining the /make_join since we cannot
+            # handle the /send_join.
+            logger.info(
+                "Rejecting /make_join to %s because it's a partial state room", room_id
+            )
+            raise SynapseError(
+                404,
+                "Unable to handle /make_join right now; this server is not fully joined.",
+                errcode=Codes.NOT_FOUND,
+            )
+
         # now check that we are *still* in the room
         is_in_room = await self._event_auth_handler.check_host_in_room(
             room_id, self.server_name
@@ -1539,15 +1562,16 @@ class FederationHandler:
 
         # Make an infinite iterator of destinations to try. Once we find a working
         # destination, we'll stick with it until it flakes.
+        destinations: Collection[str]
         if initial_destination is not None:
             # Move `initial_destination` to the front of the list.
             destinations = list(other_destinations)
             if initial_destination in destinations:
                 destinations.remove(initial_destination)
             destinations = [initial_destination] + destinations
-            destination_iter = itertools.cycle(destinations)
         else:
-            destination_iter = itertools.cycle(other_destinations)
+            destinations = other_destinations
+        destination_iter = itertools.cycle(destinations)
 
         # `destination` is the current remote homeserver we're pulling from.
         destination = next(destination_iter)

synapse/handlers/federation_event.py

@@ -59,6 +59,7 @@ from synapse.events import EventBase
 from synapse.events.snapshot import EventContext
 from synapse.federation.federation_client import InvalidResponseError
 from synapse.logging.context import nested_logging_context
+from synapse.logging.opentracing import trace
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.replication.http.devices import ReplicationUserDevicesResyncRestServlet
 from synapse.replication.http.federation import (
@@ -278,7 +279,8 @@ class FederationEventHandler:
                 )
 
         try:
-            await self._process_received_pdu(origin, pdu, state_ids=None)
+            context = await self._state_handler.compute_event_context(pdu)
+            await self._process_received_pdu(origin, pdu, context)
         except PartialStateConflictError:
             # The room was un-partial stated while we were processing the PDU.
             # Try once more, with full state this time.
@@ -286,7 +288,8 @@ class FederationEventHandler:
                 "Room %s was un-partial stated while processing the PDU, trying again.",
                 room_id,
             )
-            await self._process_received_pdu(origin, pdu, state_ids=None)
+            context = await self._state_handler.compute_event_context(pdu)
+            await self._process_received_pdu(origin, pdu, context)
 
     async def on_send_membership_event(
         self, origin: str, event: EventBase
@@ -316,6 +319,7 @@ class FederationEventHandler:
             The event and context of the event after inserting it into the room graph.
 
         Raises:
+            RuntimeError if any prev_events are missing
             SynapseError if the event is not accepted into the room
             PartialStateConflictError if the room was un-partial stated in between
                 computing the state at the event and persisting it. The caller should
@@ -376,7 +380,7 @@ class FederationEventHandler:
         # need to.
         await self._event_creation_handler.cache_joined_hosts_for_event(event, context)
 
-        await self._check_for_soft_fail(event, None, origin=origin)
+        await self._check_for_soft_fail(event, context=context, origin=origin)
         await self._run_push_actions_and_persist_event(event, context)
         return event, context
 
@@ -534,32 +538,36 @@ class FederationEventHandler:
             #
             # This is the same operation as we do when we receive a regular event
             # over federation.
-            state_ids = await self._resolve_state_at_missing_prevs(destination, event)
-
-            # build a new state group for it if need be
-            context = await self._state_handler.compute_event_context(
-                event,
-                state_ids_before_event=state_ids,
+            context = await self._compute_event_context_with_maybe_missing_prevs(
+                destination, event
             )
             if context.partial_state:
                 # this can happen if some or all of the event's prev_events still have
-                # partial state - ie, an event has an earlier stream_ordering than one
-                # or more of its prev_events, so we de-partial-state it before its
-                # prev_events.
+                # partial state. We were careful to only pick events from the db without
+                # partial-state prev events, so that implies that a prev event has
+                # been persisted (with partial state) since we did the query.
                 #
-                # TODO(faster_joins): we probably need to be more intelligent, and
-                #    exclude partial-state prev_events from consideration
-                #    https://github.com/matrix-org/synapse/issues/13001
+                # So, let's just ignore `event` for now; when we re-run the db query
+                # we should instead get its partial-state prev event, which we will
+                # de-partial-state, and then come back to event.
                 logger.warning(
-                    "%s still has partial state: can't de-partial-state it yet",
+                    "%s still has prev_events with partial state: can't de-partial-state it yet",
                     event.event_id,
                 )
                 return
+
+            # since the state at this event has changed, we should now re-evaluate
+            # whether it should have been rejected. We must already have all of the
+            # auth events (from last time we went round this path), so there is no
+            # need to pass the origin.
+            await self._check_event_auth(None, event, context)
+
             await self._store.update_state_for_partial_state_event(event, context)
             self._state_storage_controller.notify_event_un_partial_stated(
                 event.event_id
             )
 
+    @trace
     async def backfill(
         self, dest: str, room_id: str, limit: int, extremities: Collection[str]
     ) -> None:
@@ -604,6 +612,7 @@ class FederationEventHandler:
             backfilled=True,
         )
 
+    @trace
     async def _get_missing_events_for_pdu(
         self, origin: str, pdu: EventBase, prevs: Set[str], min_depth: int
     ) -> None:
@@ -704,6 +713,7 @@ class FederationEventHandler:
         logger.info("Got %d prev_events", len(missing_events))
         await self._process_pulled_events(origin, missing_events, backfilled=False)
 
+    @trace
     async def _process_pulled_events(
         self, origin: str, events: Iterable[EventBase], backfilled: bool
     ) -> None:
@@ -742,6 +752,7 @@ class FederationEventHandler:
             with nested_logging_context(ev.event_id):
                 await self._process_pulled_event(origin, ev, backfilled=backfilled)
 
+    @trace
     async def _process_pulled_event(
         self, origin: str, event: EventBase, backfilled: bool
     ) -> None:
@@ -766,10 +777,24 @@ class FederationEventHandler:
         """
         logger.info("Processing pulled event %s", event)
 
-        # these should not be outliers.
-        assert (
-            not event.internal_metadata.is_outlier()
-        ), "pulled event unexpectedly flagged as outlier"
+        # This function should not be used to persist outliers (use something
+        # else) because this does a bunch of operations that aren't necessary
+        # (extra work; in particular, it makes sure we have all the prev_events
+        # and resolves the state across those prev events). If you happen to run
+        # into a situation where the event you're trying to process/backfill is
+        # marked as an `outlier`, then you should update that spot to return an
+        # `EventBase` copy that doesn't have `outlier` flag set.
+        #
+        # `EventBase` is used to represent both an event we have not yet
+        # persisted, and one that we have persisted and now keep in the cache.
+        # In an ideal world this method would only be called with the first type
+        # of event, but it turns out that's not actually the case and for
+        # example, you could get an event from cache that is marked as an
+        # `outlier` (fix up that spot though).
+        assert not event.internal_metadata.is_outlier(), (
+            "Outlier event passed to _process_pulled_event. "
+            "To persist an event as a non-outlier, make sure to pass in a copy without `event.internal_metadata.outlier = true`."
+        )
 
         event_id = event.event_id
 
@@ -779,7 +804,7 @@ class FederationEventHandler:
         if existing:
             if not existing.internal_metadata.is_outlier():
                 logger.info(
-                    "Ignoring received event %s which we have already seen",
+                    "_process_pulled_event: Ignoring received event %s which we have already seen",
                     event_id,
                 )
                 return
@@ -792,29 +817,55 @@ class FederationEventHandler:
             return
 
         try:
-            state_ids = await self._resolve_state_at_missing_prevs(origin, event)
-            # TODO(faster_joins): make sure that _resolve_state_at_missing_prevs does
-            #   not return partial state
-            #   https://github.com/matrix-org/synapse/issues/13002
+            try:
+                context = await self._compute_event_context_with_maybe_missing_prevs(
+                    origin, event
+                )
+                await self._process_received_pdu(
+                    origin,
+                    event,
+                    context,
+                    backfilled=backfilled,
+                )
+            except PartialStateConflictError:
+                # The room was un-partial stated while we were processing the event.
+                # Try once more, with full state this time.
+                context = await self._compute_event_context_with_maybe_missing_prevs(
+                    origin, event
+                )
 
-            await self._process_received_pdu(
-                origin, event, state_ids=state_ids, backfilled=backfilled
-            )
+                # We ought to have full state now, barring some unlikely race where we left and
+                # rejoned the room in the background.
+                if context.partial_state:
+                    raise AssertionError(
+                        f"Event {event.event_id} still has a partial resolved state "
+                        f"after room {event.room_id} was un-partial stated"
+                    )
+
+                await self._process_received_pdu(
+                    origin,
+                    event,
+                    context,
+                    backfilled=backfilled,
+                )
         except FederationError as e:
             if e.code == 403:
                 logger.warning("Pulled event %s failed history check.", event_id)
             else:
                 raise
 
-    async def _resolve_state_at_missing_prevs(
+    async def _compute_event_context_with_maybe_missing_prevs(
         self, dest: str, event: EventBase
-    ) -> Optional[StateMap[str]]:
-        """Calculate the state at an event with missing prev_events.
+    ) -> EventContext:
+        """Build an EventContext structure for a non-outlier event whose prev_events may
+        be missing.
 
-        This is used when we have pulled a batch of events from a remote server, and
-        still don't have all the prev_events.
+        This is used when we have pulled a batch of events from a remote server, and may
+        not have all the prev_events.
 
-        If we already have all the prev_events for `event`, this method does nothing.
+        To build an EventContext, we need to calculate the state before the event. If we
+        already have all the prev_events for `event`, we can simply use the state after
+        the prev_events to calculate the state before `event`.
 
         Otherwise, the missing prevs become new backwards extremities, and we fall back
         to asking the remote server for the state after each missing `prev_event`,
@@ -835,8 +886,7 @@ class FederationEventHandler:
             event: an event to check for missing prevs.
 
         Returns:
-            if we already had all the prev events, `None`. Otherwise, returns
-            the event ids of the state at `event`.
+            The event context.
 
         Raises:
             FederationError if we fail to get the state from the remote server after any
@@ -850,7 +900,7 @@ class FederationEventHandler:
         missing_prevs = prevs - seen
 
         if not missing_prevs:
-            return None
+            return await self._state_handler.compute_event_context(event)
 
         logger.info(
             "Event %s is missing prev_events %s: calculating state for a "
@@ -862,9 +912,15 @@ class FederationEventHandler:
         # resolve them to find the correct state at the current event.
 
         try:
+            # Determine whether we may be about to retrieve partial state
+            # Events may be un-partial stated right after we compute the partial state
+            # flag, but that's okay, as long as the flag errs on the conservative side.
+            partial_state_flags = await self._store.get_partial_state_events(seen)
+            partial_state = any(partial_state_flags.values())
+
             # Get the state of the events we know about
             ours = await self._state_storage_controller.get_state_groups_ids(
-                room_id, seen
+                room_id, seen, await_full_state=False
             )
 
             # state_maps is a list of mappings from (type, state_key) to event_id
@@ -910,7 +966,9 @@ class FederationEventHandler:
                 "We can't get valid state history.",
                 affected=event_id,
             )
-        return state_map
+        return await self._state_handler.compute_event_context(
+            event, state_ids_before_event=state_map, partial_state=partial_state
+        )
 
     async def _get_state_ids_after_missing_prev_event(
         self,
@@ -1037,6 +1095,9 @@ class FederationEventHandler:
         # XXX: this doesn't sound right? it means that we'll end up with incomplete
         #   state.
         failed_to_fetch = desired_events - event_metadata.keys()
+        # `event_id` could be missing from `event_metadata` because it's not necessarily
+        # a state event. We've already checked that we've fetched it above.
+        failed_to_fetch.discard(event_id)
         if failed_to_fetch:
             logger.warning(
                 "Failed to fetch missing state events for %s %s",
@@ -1076,7 +1137,7 @@ class FederationEventHandler:
         self,
         origin: str,
         event: EventBase,
-        state_ids: Optional[StateMap[str]],
+        context: EventContext,
         backfilled: bool = False,
     ) -> None:
         """Called when we have a new non-outlier event.
@@ -1098,24 +1159,18 @@ class FederationEventHandler:
 
             event: event to be persisted
 
-            state_ids: Normally None, but if we are handling a gap in the graph
-                (ie, we are missing one or more prev_events), the resolved state at the
-                event. Must not be partial state.
+            context: The `EventContext` to persist the event with.
 
             backfilled: True if this is part of a historical batch of events (inhibits
                 notification to clients, and validation of device keys.)
 
         PartialStateConflictError: if the room was un-partial stated in between
-            computing the state at the event and persisting it. The caller should retry
-            exactly once in this case. Will never be raised if `state_ids` is provided.
+            computing the state at the event and persisting it. The caller should
+            recompute `context` and retry exactly once when this happens.
         """
         logger.debug("Processing event: %s", event)
         assert not event.internal_metadata.outlier
 
-        context = await self._state_handler.compute_event_context(
-            event,
-            state_ids_before_event=state_ids,
-        )
         try:
             await self._check_event_auth(origin, event, context)
         except AuthError as e:
@@ -1127,7 +1182,7 @@ class FederationEventHandler:
             # For new (non-backfilled and non-outlier) events we check if the event
             # passes auth based on the current state. If it doesn't then we
             # "soft-fail" the event.
-            await self._check_for_soft_fail(event, state_ids, origin=origin)
+            await self._check_for_soft_fail(event, context=context, origin=origin)
 
         await self._run_push_actions_and_persist_event(event, context, backfilled)
 
@@ -1312,6 +1367,53 @@ class FederationEventHandler:
             marker_event,
         )
 
+    async def backfill_event_id(
+        self, destination: str, room_id: str, event_id: str
+    ) -> EventBase:
+        """Backfill a single event and persist it as a non-outlier which means
+        we also pull in all of the state and auth events necessary for it.
+
+        Args:
+            destination: The homeserver to pull the given event_id from.
+            room_id: The room where the event is from.
+            event_id: The event ID to backfill.
+
+        Raises:
+            FederationError if we are unable to find the event from the destination
+        """
+        logger.info(
+            "backfill_event_id: event_id=%s from destination=%s", event_id, destination
+        )
+
+        room_version = await self._store.get_room_version(room_id)
+
+        event_from_response = await self._federation_client.get_pdu(
+            [destination],
+            event_id,
+            room_version,
+        )
+
+        if not event_from_response:
+            raise FederationError(
+                "ERROR",
+                404,
+                "Unable to find event_id=%s from destination=%s to backfill."
+                % (event_id, destination),
+                affected=event_id,
+            )
+
+        # Persist the event we just fetched, including pulling all of the state
+        # and auth events to de-outlier it. This also sets up the necessary
+        # `state_groups` for the event.
+        await self._process_pulled_events(
+            destination,
+            [event_from_response],
+            # Prevent notifications going to clients
+            backfilled=True,
+        )
+
+        return event_from_response
+
     async def _get_events_and_persist(
         self, destination: str, room_id: str, event_ids: Collection[str]
     ) -> None:
@@ -1492,13 +1594,15 @@ class FederationEventHandler:
         )
 
     async def _check_event_auth(
-        self, origin: str, event: EventBase, context: EventContext
+        self, origin: Optional[str], event: EventBase, context: EventContext
     ) -> None:
         """
         Checks whether an event should be rejected (for failing auth checks).
 
         Args:
-            origin: The host the event originates from.
+            origin: The host the event originates from. This is used to fetch
+               any missing auth events. It can be set to None, but only if we are
+               sure that we already have all the auth events.
             event: The event itself.
             context:
                 The event context.
@@ -1641,17 +1745,27 @@ class FederationEventHandler:
     async def _check_for_soft_fail(
         self,
         event: EventBase,
-        state_ids: Optional[StateMap[str]],
+        context: EventContext,
         origin: str,
     ) -> None:
         """Checks if we should soft fail the event; if so, marks the event as
         such.
 
+        Does nothing for events in rooms with partial state, since we may not have an
+        accurate membership event for the sender in the current state.
+
         Args:
             event
-            state_ids: The state at the event if we don't have all the event's prev events
+            context: The `EventContext` which we are about to persist the event with.
             origin: The host the event originates from.
         """
+        if await self._store.is_partial_state_room(event.room_id):
+            # We might not know the sender's membership in the current state, so don't
+            # soft fail anything. Even if we do have a membership for the sender in the
+            # current state, it may have been derived from state resolution between
+            # partial and full state and may not be accurate.
+            return
+
         extrem_ids_list = await self._store.get_latest_event_ids_in_room(event.room_id)
         extrem_ids = set(extrem_ids_list)
         prev_event_ids = set(event.prev_event_ids())
@@ -1668,11 +1782,15 @@ class FederationEventHandler:
         auth_types = auth_types_for_event(room_version_obj, event)
 
         # Calculate the "current state".
-        if state_ids is not None:
-            # If we're explicitly given the state then we won't have all the
-            # prev events, and so we have a gap in the graph. In this case
-            # we want to be a little careful as we might have been down for
-            # a while and have an incorrect view of the current state,
+        seen_event_ids = await self._store.have_events_in_timeline(prev_event_ids)
+        has_missing_prevs = bool(prev_event_ids - seen_event_ids)
+        if has_missing_prevs:
+            # We don't have all the prev_events of this event, which means we have a
+            # gap in the graph, and the new event is going to become a new backwards
+            # extremity.
+            #
+            # In this case we want to be a little careful as we might have been
+            # down for a while and have an incorrect view of the current state,
             # however we still want to do checks as gaps are easy to
             # maliciously manufacture.
             #
@@ -1685,6 +1803,7 @@ class FederationEventHandler:
                 event.room_id, extrem_ids
             )
             state_sets: List[StateMap[str]] = list(state_sets_d.values())
+            state_ids = await context.get_prev_state_ids()
             state_sets.append(state_ids)
             current_state_ids = (
                 await self._state_resolution_handler.resolve_events_with_store(
@@ -1734,7 +1853,7 @@ class FederationEventHandler:
             event.internal_metadata.soft_failed = True
 
     async def _load_or_fetch_auth_events_for_event(
-        self, destination: str, event: EventBase
+        self, destination: Optional[str], event: EventBase
     ) -> Collection[EventBase]:
         """Fetch this event's auth_events, from database or remote
 
@@ -1750,12 +1869,19 @@ class FederationEventHandler:
         Args:
             destination: where to send the /event_auth request. Typically the server
                that sent us `event` in the first place.
+
+               If this is None, no attempt is made to load any missing auth events:
+               rather, an AssertionError is raised if there are any missing events.
+
             event: the event whose auth_events we want
 
         Returns:
             all of the events listed in `event.auth_events_ids`, after deduplication
 
         Raises:
+            AssertionError if some auth events were missing and no `destination` was
+            supplied.
+
             AuthError if we were unable to fetch the auth_events for any reason.
         """
         event_auth_event_ids = set(event.auth_event_ids())
@@ -1767,6 +1893,13 @@ class FederationEventHandler:
         )
         if not missing_auth_event_ids:
             return event_auth_events.values()
+        if destination is None:
+            # this shouldn't happen: destination must be set unless we know we have already
+            # persisted the auth events.
+            raise AssertionError(
+                "_load_or_fetch_auth_events_for_event() called with no destination for "
+                "an event with missing auth_events"
+            )
 
         logger.info(
             "Event %s refers to unknown auth events %s: fetching auth chain",
@@ -1980,6 +2113,10 @@ class FederationEventHandler:
             event, event_pos, max_stream_token, extra_users=extra_users
         )
 
+        if event.type == EventTypes.Member and event.membership == Membership.JOIN:
+            # TODO retrieve the previous state, and exclude join -> join transitions
+            self._notifier.notify_user_joined_room(event.event_id, event.room_id)
+
     def _sanity_check_event(self, ev: EventBase) -> None:
         """
         Do some early sanity checks of a received event

synapse/handlers/identity.py

@@ -26,6 +26,7 @@ from synapse.api.errors import (
     SynapseError,
 )
 from synapse.api.ratelimiting import Ratelimiter
+from synapse.config.emailconfig import ThreepidBehaviour
 from synapse.http import RequestTimedOutError
 from synapse.http.client import SimpleHttpClient
 from synapse.http.site import SynapseRequest
@@ -415,6 +416,48 @@ class IdentityHandler:
 
         return session_id
 
+    async def request_email_token(
+        self,
+        id_server: str,
+        email: str,
+        client_secret: str,
+        send_attempt: int,
+        next_link: Optional[str] = None,
+    ) -> JsonDict:
+        """
+        Request an external server send an email on our behalf for the purposes of threepid
+        validation.
+
+        Args:
+            id_server: The identity server to proxy to
+            email: The email to send the message to
+            client_secret: The unique client_secret sends by the user
+            send_attempt: Which attempt this is
+            next_link: A link to redirect the user to once they submit the token
+
+        Returns:
+            The json response body from the server
+        """
+        params = {
+            "email": email,
+            "client_secret": client_secret,
+            "send_attempt": send_attempt,
+        }
+        if next_link:
+            params["next_link"] = next_link
+
+        try:
+            data = await self.http_client.post_json_get_json(
+                id_server + "/_matrix/identity/api/v1/validate/email/requestToken",
+                params,
+            )
+            return data
+        except HttpResponseException as e:
+            logger.info("Proxied requestToken failed: %r", e)
+            raise e.to_synapse_error()
+        except RequestTimedOutError:
+            raise SynapseError(500, "Timed out contacting identity server")
+
     async def requestMsisdnToken(
         self,
         id_server: str,
@@ -488,7 +531,18 @@ class IdentityHandler:
         validation_session = None
 
         # Try to validate as email
-        if self.hs.config.email.can_verify_email:
+        if self.hs.config.email.threepid_behaviour_email == ThreepidBehaviour.REMOTE:
+            # Remote emails will only be used if a valid identity server is provided.
+            assert (
+                self.hs.config.registration.account_threepid_delegate_email is not None
+            )
+
+            # Ask our delegated email identity server
+            validation_session = await self.threepid_from_creds(
+                self.hs.config.registration.account_threepid_delegate_email,
+                threepid_creds,
+            )
+        elif self.hs.config.email.threepid_behaviour_email == ThreepidBehaviour.LOCAL:
             # Get a validated session matching these details
             validation_session = await self.store.get_threepid_validation_session(
                 "email", client_secret, sid=sid, validated=True

synapse/handlers/initial_sync.py

@@ -143,8 +143,8 @@ class InitialSyncHandler:
             joined_rooms,
             to_key=int(now_token.receipt_key),
         )
-        if self.hs.config.experimental.msc2285_enabled:
-            receipt = ReceiptEventSource.filter_out_private_receipts(receipt, user_id)
+
+        receipt = ReceiptEventSource.filter_out_private_receipts(receipt, user_id)
 
         tags_by_room = await self.store.get_tags_for_user(user_id)
 
@@ -456,11 +456,8 @@ class InitialSyncHandler:
             )
             if not receipts:
                 return []
-            if self.hs.config.experimental.msc2285_enabled:
-                receipts = ReceiptEventSource.filter_out_private_receipts(
-                    receipts, user_id
-                )
-            return receipts
+
+            return ReceiptEventSource.filter_out_private_receipts(receipts, user_id)
 
         presence, receipts, (messages, token) = await make_deferred_yieldable(
             gather_results(

synapse/handlers/message.py

@@ -41,6 +41,7 @@ from synapse.api.errors import (
     NotFoundError,
     ShadowBanError,
     SynapseError,
+    UnstableSpecAuthError,
     UnsupportedRoomVersionError,
 )
 from synapse.api.room_versions import KNOWN_ROOM_VERSIONS
@@ -51,6 +52,7 @@ from synapse.events.builder import EventBuilder
 from synapse.events.snapshot import EventContext
 from synapse.events.validator import EventValidator
 from synapse.handlers.directory import DirectoryHandler
+from synapse.logging import opentracing
 from synapse.logging.context import make_deferred_yieldable, run_in_background
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.replication.http.send_event import ReplicationSendEventRestServlet
@@ -149,7 +151,11 @@ class MessageHandler:
                 "Attempted to retrieve data from a room for a user that has never been in it. "
                 "This should not have happened."
             )
-            raise SynapseError(403, "User not in room", errcode=Codes.FORBIDDEN)
+            raise UnstableSpecAuthError(
+                403,
+                "User not in room",
+                errcode=Codes.NOT_JOINED,
+            )
 
         return data
 
@@ -319,8 +325,10 @@ class MessageHandler:
                 room_id, user_id, allow_departed_users=True
             )
             if membership != Membership.JOIN:
-                raise NotImplementedError(
-                    "Getting joined members after leaving is not implemented"
+                raise SynapseError(
+                    code=403,
+                    errcode=Codes.FORBIDDEN,
+                    msg="Getting joined members while not being a current member of the room is forbidden.",
                 )
 
         users_with_profile = await self.store.get_users_in_room_with_profiles(room_id)
@@ -334,7 +342,11 @@ class MessageHandler:
                     break
             else:
                 # Loop fell through, AS has no interested users in room
-                raise AuthError(403, "Appservice not in room")
+                raise UnstableSpecAuthError(
+                    403,
+                    "Appservice not in room",
+                    errcode=Codes.NOT_JOINED,
+                )
 
         return {
             user_id: {
@@ -463,6 +475,7 @@ class EventCreationHandler:
         )
         self._events_shard_config = self.config.worker.events_shard_config
         self._instance_name = hs.get_instance_name()
+        self._notifier = hs.get_notifier()
 
         self.room_prejoin_state_types = self.hs.config.api.room_prejoin_state
 
@@ -1134,6 +1147,10 @@ class EventCreationHandler:
             context = await self.state.compute_event_context(
                 event,
                 state_ids_before_event=state_map_for_event,
+                # TODO(faster_joins): check how MSC2716 works and whether we can have
+                #   partial state here
+                #   https://github.com/matrix-org/synapse/issues/13003
+                partial_state=False,
             )
         else:
             context = await self.state.compute_event_context(event)
@@ -1358,9 +1375,10 @@ class EventCreationHandler:
         # and `state_groups` because they have `prev_events` that aren't persisted yet
         # (historical messages persisted in reverse-chronological order).
         if not event.internal_metadata.is_historical():
-            await self._bulk_push_rule_evaluator.action_for_event_by_user(
-                event, context
-            )
+            with opentracing.start_active_span("calculate_push_actions"):
+                await self._bulk_push_rule_evaluator.action_for_event_by_user(
+                    event, context
+                )
 
         try:
             # If we're a worker we need to hit out to the master.
@@ -1447,9 +1465,10 @@ class EventCreationHandler:
             state = await state_entry.get_state(
                 self._storage_controllers.state, StateFilter.all()
             )
-            joined_hosts = await self.store.get_joined_hosts(
-                event.room_id, state, state_entry
-            )
+            with opentracing.start_active_span("get_joined_hosts"):
+                joined_hosts = await self.store.get_joined_hosts(
+                    event.room_id, state, state_entry
+                )
 
             # Note that the expiry times must be larger than the expiry time in
             # _external_cache_joined_hosts_updates.
@@ -1550,6 +1569,16 @@ class EventCreationHandler:
                 requester, is_admin_redaction=is_admin_redaction
             )
 
+        if event.type == EventTypes.Member and event.membership == Membership.JOIN:
+            (
+                current_membership,
+                _,
+            ) = await self.store.get_local_current_membership_for_user_in_room(
+                event.state_key, event.room_id
+            )
+            if current_membership != Membership.JOIN:
+                self._notifier.notify_user_joined_room(event.event_id, event.room_id)
+
         await self._maybe_kick_guest_users(event, context)
 
         if event.type == EventTypes.CanonicalAlias:
@@ -1849,13 +1878,8 @@ class EventCreationHandler:
 
         # For each room we need to find a joined member we can use to send
         # the dummy event with.
-        latest_event_ids = await self.store.get_prev_events_for_room(room_id)
-        members = await self.state.get_current_users_in_room(
-            room_id, latest_event_ids=latest_event_ids
-        )
+        members = await self.store.get_local_users_in_room(room_id)
         for user_id in members:
-            if not self.hs.is_mine_id(user_id):
-                continue
             requester = create_requester(user_id, authenticated_entity=self.server_name)
             try:
                 event, context = await self.create_event(
@@ -1866,7 +1890,6 @@ class EventCreationHandler:
                         "room_id": room_id,
                         "sender": user_id,
                     },
-                    prev_event_ids=latest_event_ids,
                 )
 
                 event.internal_metadata.proactively_send = False

synapse/handlers/pagination.py

@@ -24,6 +24,7 @@ from synapse.api.errors import SynapseError
 from synapse.api.filtering import Filter
 from synapse.events.utils import SerializeEventConfig
 from synapse.handlers.room import ShutdownRoomResponse
+from synapse.logging.opentracing import trace
 from synapse.metrics.background_process_metrics import run_as_background_process
 from synapse.storage.state import StateFilter
 from synapse.streams.config import PaginationConfig
@@ -416,6 +417,7 @@ class PaginationHandler:
 
             await self._storage_controllers.purge_events.purge_room(room_id)
 
+    @trace
     async def get_messages(
         self,
         requester: Requester,

synapse/handlers/presence.py

@@ -34,7 +34,6 @@ from typing import (
     Callable,
     Collection,
     Dict,
-    FrozenSet,
     Generator,
     Iterable,
     List,
@@ -42,7 +41,6 @@ from typing import (
     Set,
     Tuple,
     Type,
-    Union,
 )
 
 from prometheus_client import Counter
@@ -68,7 +66,6 @@ from synapse.storage.databases.main import DataStore
 from synapse.streams import EventSource
 from synapse.types import JsonDict, StreamKeyType, UserID, get_domain_from_id
 from synapse.util.async_helpers import Linearizer
-from synapse.util.caches.descriptors import _CacheContext, cached
 from synapse.util.metrics import Measure
 from synapse.util.wheel_timer import WheelTimer
 
@@ -1656,15 +1653,18 @@ class PresenceEventSource(EventSource[int, UserPresenceState]):
                 # doesn't return. C.f. #5503.
                 return [], max_token
 
-            # Figure out which other users this user should receive updates for
-            users_interested_in = await self._get_interested_in(user, explicit_room_id)
+            # Figure out which other users this user should explicitly receive
+            # updates for
+            additional_users_interested_in = (
+                await self.get_presence_router().get_interested_users(user.to_string())
+            )
 
             # We have a set of users that we're interested in the presence of. We want to
             # cross-reference that with the users that have actually changed their presence.
 
             # Check whether this user should see all user updates
 
-            if users_interested_in == PresenceRouter.ALL_USERS:
+            if additional_users_interested_in == PresenceRouter.ALL_USERS:
                 # Provide presence state for all users
                 presence_updates = await self._filter_all_presence_updates_for_user(
                     user_id, include_offline, from_key
@@ -1673,34 +1673,47 @@ class PresenceEventSource(EventSource[int, UserPresenceState]):
                 return presence_updates, max_token
 
             # Make mypy happy. users_interested_in should now be a set
-            assert not isinstance(users_interested_in, str)
+            assert not isinstance(additional_users_interested_in, str)
+
+            # We always care about our own presence.
+            additional_users_interested_in.add(user_id)
+
+            if explicit_room_id:
+                user_ids = await self.store.get_users_in_room(explicit_room_id)
+                additional_users_interested_in.update(user_ids)
 
             # The set of users that we're interested in and that have had a presence update.
             # We'll actually pull the presence updates for these users at the end.
-            interested_and_updated_users: Union[Set[str], FrozenSet[str]] = set()
+            interested_and_updated_users: Collection[str]
 
             if from_key is not None:
                 # First get all users that have had a presence update
                 updated_users = stream_change_cache.get_all_entities_changed(from_key)
 
                 # Cross-reference users we're interested in with those that have had updates.
-                # Use a slightly-optimised method for processing smaller sets of updates.
-                if updated_users is not None and len(updated_users) < 500:
-                    # For small deltas, it's quicker to get all changes and then
-                    # cross-reference with the users we're interested in
+                if updated_users is not None:
+                    # If we have the full list of changes for presence we can
+                    # simply check which ones share a room with the user.
                     get_updates_counter.labels("stream").inc()
-                    for other_user_id in updated_users:
-                        if other_user_id in users_interested_in:
-                            # mypy thinks this variable could be a FrozenSet as it's possibly set
-                            # to one in the `get_entities_changed` call below, and `add()` is not
-                            # method on a FrozenSet. That doesn't affect us here though, as
-                            # `interested_and_updated_users` is clearly a set() above.
-                            interested_and_updated_users.add(other_user_id)  # type: ignore
+
+                    sharing_users = await self.store.do_users_share_a_room(
+                        user_id, updated_users
+                    )
+
+                    interested_and_updated_users = (
+                        sharing_users.union(additional_users_interested_in)
+                    ).intersection(updated_users)
+
                 else:
                     # Too many possible updates. Find all users we can see and check
                     # if any of them have changed.
                     get_updates_counter.labels("full").inc()
 
+                    users_interested_in = (
+                        await self.store.get_users_who_share_room_with_user(user_id)
+                    )
+                    users_interested_in.update(additional_users_interested_in)
+
                     interested_and_updated_users = (
                         stream_change_cache.get_entities_changed(
                             users_interested_in, from_key
@@ -1709,7 +1722,10 @@ class PresenceEventSource(EventSource[int, UserPresenceState]):
             else:
                 # No from_key has been specified. Return the presence for all users
                 # this user is interested in
-                interested_and_updated_users = users_interested_in
+                interested_and_updated_users = (
+                    await self.store.get_users_who_share_room_with_user(user_id)
+                )
+                interested_and_updated_users.update(additional_users_interested_in)
 
             # Retrieve the current presence state for each user
             users_to_state = await self.get_presence_handler().current_state_for_users(
@@ -1804,62 +1820,6 @@ class PresenceEventSource(EventSource[int, UserPresenceState]):
     def get_current_key(self) -> int:
         return self.store.get_current_presence_token()
 
-    @cached(num_args=2, cache_context=True)
-    async def _get_interested_in(
-        self,
-        user: UserID,
-        explicit_room_id: Optional[str] = None,
-        cache_context: Optional[_CacheContext] = None,
-    ) -> Union[Set[str], str]:
-        """Returns the set of users that the given user should see presence
-        updates for.
-
-        Args:
-            user: The user to retrieve presence updates for.
-            explicit_room_id: The users that are in the room will be returned.
-
-        Returns:
-            A set of user IDs to return presence updates for, or "ALL" to return all
-            known updates.
-        """
-        user_id = user.to_string()
-        users_interested_in = set()
-        users_interested_in.add(user_id)  # So that we receive our own presence
-
-        # cache_context isn't likely to ever be None due to the @cached decorator,
-        # but we can't have a non-optional argument after the optional argument
-        # explicit_room_id either. Assert cache_context is not None so we can use it
-        # without mypy complaining.
-        assert cache_context
-
-        # Check with the presence router whether we should poll additional users for
-        # their presence information
-        additional_users = await self.get_presence_router().get_interested_users(
-            user.to_string()
-        )
-        if additional_users == PresenceRouter.ALL_USERS:
-            # If the module requested that this user see the presence updates of *all*
-            # users, then simply return that instead of calculating what rooms this
-            # user shares
-            return PresenceRouter.ALL_USERS
-
-        # Add the additional users from the router
-        users_interested_in.update(additional_users)
-
-        # Find the users who share a room with this user
-        users_who_share_room = await self.store.get_users_who_share_room_with_user(
-            user_id, on_invalidate=cache_context.invalidate
-        )
-        users_interested_in.update(users_who_share_room)
-
-        if explicit_room_id:
-            user_ids = await self.store.get_users_in_room(
-                explicit_room_id, on_invalidate=cache_context.invalidate
-            )
-            users_interested_in.update(user_ids)
-
-        return users_interested_in
-
 
 def handle_timeouts(
     user_states: List[UserPresenceState],

synapse/handlers/receipts.py

@@ -163,7 +163,10 @@ class ReceiptsHandler:
         if not is_new:
             return
 
-        if self.federation_sender and receipt_type != ReceiptTypes.READ_PRIVATE:
+        if self.federation_sender and receipt_type not in (
+            ReceiptTypes.READ_PRIVATE,
+            ReceiptTypes.UNSTABLE_READ_PRIVATE,
+        ):
             await self.federation_sender.send_read_receipt(receipt)
 
 
@@ -203,24 +206,38 @@ class ReceiptEventSource(EventSource[int, JsonDict]):
             for event_id, orig_event_content in room.get("content", {}).items():
                 event_content = orig_event_content
                 # If there are private read receipts, additional logic is necessary.
-                if ReceiptTypes.READ_PRIVATE in event_content:
+                if (
+                    ReceiptTypes.READ_PRIVATE in event_content
+                    or ReceiptTypes.UNSTABLE_READ_PRIVATE in event_content
+                ):
                     # Make a copy without private read receipts to avoid leaking
                     # other user's private read receipts..
                     event_content = {
                         receipt_type: receipt_value
                         for receipt_type, receipt_value in event_content.items()
-                        if receipt_type != ReceiptTypes.READ_PRIVATE
+                        if receipt_type
+                        not in (
+                            ReceiptTypes.READ_PRIVATE,
+                            ReceiptTypes.UNSTABLE_READ_PRIVATE,
+                        )
                     }
 
                     # Copy the current user's private read receipt from the
                     # original content, if it exists.
-                    user_private_read_receipt = orig_event_content[
-                        ReceiptTypes.READ_PRIVATE
-                    ].get(user_id, None)
+                    user_private_read_receipt = orig_event_content.get(
+                        ReceiptTypes.READ_PRIVATE, {}
+                    ).get(user_id, None)
                     if user_private_read_receipt:
                         event_content[ReceiptTypes.READ_PRIVATE] = {
                             user_id: user_private_read_receipt
                         }
+                    user_unstable_private_read_receipt = orig_event_content.get(
+                        ReceiptTypes.UNSTABLE_READ_PRIVATE, {}
+                    ).get(user_id, None)
+                    if user_unstable_private_read_receipt:
+                        event_content[ReceiptTypes.UNSTABLE_READ_PRIVATE] = {
+                            user_id: user_unstable_private_read_receipt
+                        }
 
                 # Include the event if there is at least one non-private read
                 # receipt or the current user has a private read receipt.
@@ -256,10 +273,9 @@ class ReceiptEventSource(EventSource[int, JsonDict]):
             room_ids, from_key=from_key, to_key=to_key
         )
 
-        if self.config.experimental.msc2285_enabled:
-            events = ReceiptEventSource.filter_out_private_receipts(
-                events, user.to_string()
-            )
+        events = ReceiptEventSource.filter_out_private_receipts(
+            events, user.to_string()
+        )
 
         return events, to_key
 

synapse/handlers/relations.py

@@ -19,6 +19,7 @@ import attr
 from synapse.api.constants import RelationTypes
 from synapse.api.errors import SynapseError
 from synapse.events import EventBase, relation_from_event
+from synapse.logging.opentracing import trace
 from synapse.storage.databases.main.relations import _RelatedEvent
 from synapse.types import JsonDict, Requester, StreamToken, UserID
 from synapse.visibility import filter_events_for_client
@@ -73,7 +74,6 @@ class RelationsHandler:
         room_id: str,
         relation_type: Optional[str] = None,
         event_type: Optional[str] = None,
-        aggregation_key: Optional[str] = None,
         limit: int = 5,
         direction: str = "b",
         from_token: Optional[StreamToken] = None,
@@ -89,7 +89,6 @@ class RelationsHandler:
             room_id: The room the event belongs to.
             relation_type: Only fetch events with this relation type, if given.
             event_type: Only fetch events with this event type, if given.
-            aggregation_key: Only fetch events with this aggregation key, if given.
             limit: Only fetch the most recent `limit` events.
             direction: Whether to fetch the most recent first (`"b"`) or the
                 oldest first (`"f"`).
@@ -122,7 +121,6 @@ class RelationsHandler:
             room_id=room_id,
             relation_type=relation_type,
             event_type=event_type,
-            aggregation_key=aggregation_key,
             limit=limit,
             direction=direction,
             from_token=from_token,
@@ -364,6 +362,7 @@ class RelationsHandler:
 
         return results
 
+    @trace
     async def get_bundled_aggregations(
         self, events: Iterable[EventBase], user_id: str
     ) -> Dict[str, BundledAggregations]:

synapse/handlers/room.py

@@ -1384,6 +1384,7 @@ class TimestampLookupHandler:
         self.store = hs.get_datastores().main
         self.state_handler = hs.get_state_handler()
         self.federation_client = hs.get_federation_client()
+        self.federation_event_handler = hs.get_federation_event_handler()
         self._storage_controllers = hs.get_storage_controllers()
 
     async def get_event_for_timestamp(
@@ -1479,38 +1480,68 @@ class TimestampLookupHandler:
                         remote_response,
                     )
 
-                    # TODO: Do we want to persist this as an extremity?
-                    # TODO: I think ideally, we would try to backfill from
-                    # this event and run this whole
-                    # `get_event_for_timestamp` function again to make sure
-                    # they didn't give us an event from their gappy history.
                     remote_event_id = remote_response.event_id
-                    origin_server_ts = remote_response.origin_server_ts
+                    remote_origin_server_ts = remote_response.origin_server_ts
+
+                    # Backfill this event so we can get a pagination token for
+                    # it with `/context` and paginate `/messages` from this
+                    # point.
+                    #
+                    # TODO: The requested timestamp may lie in a part of the
+                    #   event graph that the remote server *also* didn't have,
+                    #   in which case they will have returned another event
+                    #   which may be nowhere near the requested timestamp. In
+                    #   the future, we may need to reconcile that gap and ask
+                    #   other homeservers, and/or extend `/timestamp_to_event`
+                    #   to return events on *both* sides of the timestamp to
+                    #   help reconcile the gap faster.
+                    remote_event = (
+                        await self.federation_event_handler.backfill_event_id(
+                            domain, room_id, remote_event_id
+                        )
+                    )
+
+                    # XXX: When we see that the remote server is not trustworthy,
+                    # maybe we should not ask them first in the future.
+                    if remote_origin_server_ts != remote_event.origin_server_ts:
+                        logger.info(
+                            "get_event_for_timestamp: Remote server (%s) claimed that remote_event_id=%s occured at remote_origin_server_ts=%s but that isn't true (actually occured at %s). Their claims are dubious and we should consider not trusting them.",
+                            domain,
+                            remote_event_id,
+                            remote_origin_server_ts,
+                            remote_event.origin_server_ts,
+                        )
 
                     # Only return the remote event if it's closer than the local event
                     if not local_event or (
-                        abs(origin_server_ts - timestamp)
+                        abs(remote_event.origin_server_ts - timestamp)
                         < abs(local_event.origin_server_ts - timestamp)
                     ):
-                        return remote_event_id, origin_server_ts
+                        logger.info(
+                            "get_event_for_timestamp: returning remote_event_id=%s (%s) since it's closer to timestamp=%s than local_event=%s (%s)",
+                            remote_event_id,
+                            remote_event.origin_server_ts,
+                            timestamp,
+                            local_event.event_id if local_event else None,
+                            local_event.origin_server_ts if local_event else None,
+                        )
+                        return remote_event_id, remote_origin_server_ts
                 except (HttpResponseException, InvalidResponseError) as ex:
                     # Let's not put a high priority on some other homeserver
                     # failing to respond or giving a random response
                     logger.debug(
-                        "Failed to fetch /timestamp_to_event from %s because of exception(%s) %s args=%s",
+                        "get_event_for_timestamp: Failed to fetch /timestamp_to_event from %s because of exception(%s) %s args=%s",
                         domain,
                         type(ex).__name__,
                         ex,
                         ex.args,
                     )
-                except Exception as ex:
+                except Exception:
                     # But we do want to see some exceptions in our code
                     logger.warning(
-                        "Failed to fetch /timestamp_to_event from %s because of exception(%s) %s args=%s",
+                        "get_event_for_timestamp: Failed to fetch /timestamp_to_event from %s because of exception",
                         domain,
-                        type(ex).__name__,
-                        ex,
-                        ex.args,
+                        exc_info=True,
                     )
 
         # To appease mypy, we have to add both of these conditions to check for