changelog

Downgrade 'Closing scope' error to warning; prevent hammering sentry
Add metrics to track how the rate limiter is affecting requests (sleep/reject) (#13534 )
2022-08-18 11:58:38 +01:00 · 2022-08-18 11:46:54 +01:00 · 2022-08-17 16:10:07 -05:00 · 2022-08-17 18:08:23 +00:00 · 2022-08-17 16:46:09 +01:00 · 2022-08-17 11:02:38 +00:00
201 changed files with 8063 additions and 4675 deletions
@@ -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
@@ -53,10 +53,22 @@ jobs:
        env:
          PULL_REQUEST_NUMBER: ${{ github.event.number }}

+  lint-pydantic:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+        with:
+          ref: ${{ github.event.pull_request.head.sha }}
+          fetch-depth: 0
+      - uses: matrix-org/setup-python-poetry@v1
+        with:
+          extras: "all"
+      - run: poetry run scripts-dev/check_pydantic_models.py
+
  # Dummy step to gate other tests on without repeating the whole list
  linting-done:
    if: ${{ !cancelled() }} # Run this even if prior jobs were skipped
-    needs: [lint, lint-crlf, lint-newsfile, check-sampleconfig, check-schema-delta]
+    needs: [lint, lint-crlf, lint-newsfile, lint-pydantic, check-sampleconfig, check-schema-delta]
    runs-on: ubuntu-latest
    steps:
      - run: "true"
@@ -328,6 +340,9 @@ jobs:
          - arrangement: monolith
            database: Postgres

+          - arrangement: workers
+            database: Postgres
+
    steps:
      - name: Run actions/checkout@v2 for synapse
        uses: actions/checkout@v2
@@ -343,30 +358,6 @@ 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() }}
@@ -1,3 +1,92 @@
+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`) (this was reverted in v1.65.0rc2, see changelog notes above). ([\#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)
 ===========================

@@ -2,152 +2,70 @@
 Synapse |support| |development| |documentation| |license| |pypi| |python|
 =========================================================================

+Synapse is an open-source `Matrix <https://matrix.org/>`_ homeserver written and
+maintained by the Matrix.org Foundation. We began rapid development began in 2014,
+reaching v1.0.0 in 2019. Development on Synapse and the Matrix protocol itself continues
+in earnest today.
+
+Briefly, Matrix is an open standard for communications on the internet, supporting
+federation, encryption and VoIP. Matrix.org has more to say about the `goals of the
+Matrix project <https://matrix.org/docs/guides/introduction>`_, and the `formal specification
+<https://spec.matrix.org/>`_ describes the technical details.
+
 .. contents::

-Introduction
-============
+Installing and configuration
+============================

-Matrix is an ambitious new ecosystem for open federated Instant Messaging and
-VoIP.  The basics you need to know to get up and running are:
-
- Everything in Matrix happens in a room.  Rooms are distributed and do not
-  exist on any single server.  Rooms can be located using convenience aliases
-  like ``#matrix:matrix.org`` or ``#test:localhost:8448``.
-
- Matrix user IDs look like ``@matthew:matrix.org`` (although in the future
-  you will normally refer to yourself and others using a third party identifier
-  (3PID): email address, phone number, etc rather than manipulating Matrix user IDs)
-
-The overall architecture is::
-
-      client <----> homeserver <=====================> homeserver <----> client
-             https://somewhere.org/_matrix      https://elsewhere.net/_matrix
-
-``#matrix:matrix.org`` is the official support room for Matrix, and can be
-accessed by any client from https://matrix.org/docs/projects/try-matrix-now.html or
-via IRC bridge at irc://irc.libera.chat/matrix.
-
-Synapse is currently in rapid development, but as of version 0.5 we believe it
-is sufficiently stable to be run as an internet-facing service for real usage!
-
-About Matrix
-============
-
-Matrix specifies a set of pragmatic RESTful HTTP JSON APIs as an open standard,
-which handle:
-
- Creating and managing fully distributed chat rooms with no
-  single points of control or failure
- Eventually-consistent cryptographically secure synchronisation of room
-  state across a global open network of federated servers and services
- Sending and receiving extensible messages in a room with (optional)
-  end-to-end encryption
- Inviting, joining, leaving, kicking, banning room members
- Managing user accounts (registration, login, logout)
- Using 3rd Party IDs (3PIDs) such as email addresses, phone numbers,
-  Facebook accounts to authenticate, identify and discover users on Matrix.
- Placing 1:1 VoIP and Video calls
-
-These APIs are intended to be implemented on a wide range of servers, services
-and clients, letting developers build messaging and VoIP functionality on top
-of the entirely open Matrix ecosystem rather than using closed or proprietary
-solutions. The hope is for Matrix to act as the building blocks for a new
-generation of fully open and interoperable messaging and VoIP apps for the
-internet.
-
-Synapse is a Matrix "homeserver" implementation developed by the matrix.org core
-team, written in Python 3/Twisted.
-
-In Matrix, every user runs one or more Matrix clients, which connect through to
-a Matrix homeserver. The homeserver stores all their personal chat history and
-user account information - much as a mail client connects through to an
-IMAP/SMTP server. Just like email, you can either run your own Matrix
-homeserver and control and own your own communications and history or use one
-hosted by someone else (e.g. matrix.org) - there is no single point of control
-or mandatory service provider in Matrix, unlike WhatsApp, Facebook, Hangouts,
-etc.
-
-We'd like to invite you to join #matrix:matrix.org (via
-https://matrix.org/docs/projects/try-matrix-now.html), run a homeserver, take a look
-at the `Matrix spec <https://matrix.org/docs/spec>`_, and experiment with the
-`APIs <https://matrix.org/docs/api>`_ and `Client SDKs
-<https://matrix.org/docs/projects/try-matrix-now.html#client-sdks>`_.
-
-Thanks for using Matrix!
-
-Support
-=======
-
-For support installing or managing Synapse, please join |room|_ (from a matrix.org
-account if necessary) and ask questions there. We do not use GitHub issues for
-support requests, only for bug reports and feature requests.
-
-Synapse's documentation is `nicely rendered on GitHub Pages <https://matrix-org.github.io/synapse>`_,
-with its source available in |docs|_.
-
-.. |room| replace:: ``#synapse:matrix.org``
-.. _room: https://matrix.to/#/#synapse:matrix.org
-
-.. |docs| replace:: ``docs``
-.. _docs: docs
-
-Synapse Installation
-====================
+The Synapse documentation describes `how to install Synapse <https://matrix-org.github.io/synapse/latest/setup/installation.html>`_. We recommend using
+`Docker images <https://matrix-org.github.io/synapse/latest/setup/installation.html#docker-images-and-ansible-playbooks>`_ or `Debian packages from Matrix.org
+<https://matrix-org.github.io/synapse/latest/setup/installation.html#matrixorg-packages>`_.

 .. _federation:

-* For details on how to install synapse, see
-  `Installation Instructions <https://matrix-org.github.io/synapse/latest/setup/installation.html>`_.
-* For specific details on how to configure Synapse for federation see `docs/federate.md <docs/federate.md>`_
+Synapse has a variety of `config options
+<https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html>`_
+which can be used to customise its behaviour after installation.
+There are additional details on how to `configure Synapse for federation here
+<https://matrix-org.github.io/synapse/latest/federate.html>`_.
+
+.. _reverse-proxy:
+
+Using a reverse proxy with Synapse
+----------------------------------
+
+It is recommended to put a reverse proxy such as
+`nginx <https://nginx.org/en/docs/http/ngx_http_proxy_module.html>`_,
+`Apache <https://httpd.apache.org/docs/current/mod/mod_proxy_http.html>`_,
+`Caddy <https://caddyserver.com/docs/quick-starts/reverse-proxy>`_,
+`HAProxy <https://www.haproxy.org/>`_ or
+`relayd <https://man.openbsd.org/relayd.8>`_ in front of Synapse. One advantage of
+doing so is that it means that you can expose the default https port (443) to
+Matrix clients without needing to run Synapse with root privileges.
+For information on configuring one, see `the reverse proxy docs
+<https://matrix-org.github.io/synapse/latest/reverse_proxy.html>`_.
+
+Upgrading an existing Synapse
+-----------------------------
+
+The instructions for upgrading Synapse are in `the upgrade notes`_.
+Please check these instructions as upgrading may require extra steps for some
+versions of Synapse.
+
+.. _the upgrade notes: https://matrix-org.github.io/synapse/develop/upgrade.html


-Connecting to Synapse from a client
-===================================
+Platform dependencies
+---------------------

-The easiest way to try out your new Synapse installation is by connecting to it
-from a web client.
+Synapse uses a number of platform dependencies such as Python and PostgreSQL,
+and aims to follow supported upstream versions. See the
+`deprecation policy <https://matrix-org.github.io/synapse/latest/deprecation_policy.html>`_
+for more details.

-Unless you are running a test instance of Synapse on your local machine, in
-general, you will need to enable TLS support before you can successfully
-connect from a client: see
-`TLS certificates <https://matrix-org.github.io/synapse/latest/setup/installation.html#tls-certificates>`_.
-
-An easy way to get started is to login or register via Element at
-https://app.element.io/#/login or https://app.element.io/#/register respectively.
-You will need to change the server you are logging into from ``matrix.org``
-and instead specify a Homeserver URL of ``https://<server_name>:8448``
-(or just ``https://<server_name>`` if you are using a reverse proxy).
-If you prefer to use another client, refer to our
-`client breakdown <https://matrix.org/docs/projects/clients-matrix>`_.
-
-If all goes well you should at least be able to log in, create a room, and
-start sending messages.
-
-.. _`client-user-reg`:
-
-Registering a new user from a client
------------------------------------
-
-By default, registration of new users via Matrix clients is disabled. To enable
-it, specify ``enable_registration: true`` in ``homeserver.yaml``. (It is then
-recommended to also set up CAPTCHA - see `<docs/CAPTCHA_SETUP.md>`_.)
-
-Once ``enable_registration`` is set to ``true``, it is possible to register a
-user via a Matrix client.
-
-Your new user name will be formed partly from the ``server_name``, and partly
-from a localpart you specify when you create the account. Your name will take
-the form of::
-
-    @localpart:my.domain.name
-
-(pronounced "at localpart on my dot domain dot name").
-
-As when logging in, you will need to specify a "Custom server".  Specify your
-desired ``localpart`` in the 'User name' box.

 Security note
-=============
+-------------

 Matrix serves raw, user-supplied data in some APIs -- specifically the `content
 repository endpoints`_.
@@ -187,30 +105,76 @@ Following this advice ensures that even if an XSS is found in Synapse, the
 impact to other applications will be minimal.


-Upgrading an existing Synapse
-=============================
+Testing a new installation
+==========================

-The instructions for upgrading synapse are in `the upgrade notes`_.
-Please check these instructions as upgrading may require extra steps for some
-versions of synapse.
+The easiest way to try out your new Synapse installation is by connecting to it
+from a web client.

-.. _the upgrade notes: https://matrix-org.github.io/synapse/develop/upgrade.html
+Unless you are running a test instance of Synapse on your local machine, in
+general, you will need to enable TLS support before you can successfully
+connect from a client: see
+`TLS certificates <https://matrix-org.github.io/synapse/latest/setup/installation.html#tls-certificates>`_.

-.. _reverse-proxy:
+An easy way to get started is to login or register via Element at
+https://app.element.io/#/login or https://app.element.io/#/register respectively.
+You will need to change the server you are logging into from ``matrix.org``
+and instead specify a Homeserver URL of ``https://<server_name>:8448``
+(or just ``https://<server_name>`` if you are using a reverse proxy).
+If you prefer to use another client, refer to our
+`client breakdown <https://matrix.org/docs/projects/clients-matrix>`_.

-Using a reverse proxy with Synapse
-==================================
+If all goes well you should at least be able to log in, create a room, and
+start sending messages.

-It is recommended to put a reverse proxy such as
-`nginx <https://nginx.org/en/docs/http/ngx_http_proxy_module.html>`_,
-`Apache <https://httpd.apache.org/docs/current/mod/mod_proxy_http.html>`_,
-`Caddy <https://caddyserver.com/docs/quick-starts/reverse-proxy>`_,
-`HAProxy <https://www.haproxy.org/>`_ or
-`relayd <https://man.openbsd.org/relayd.8>`_ in front of Synapse. One advantage of
-doing so is that it means that you can expose the default https port (443) to
-Matrix clients without needing to run Synapse with root privileges.
+.. _`client-user-reg`:

-For information on configuring one, see `<docs/reverse_proxy.md>`_.
+Registering a new user from a client
+------------------------------------
+
+By default, registration of new users via Matrix clients is disabled. To enable
+it:
+
+1. In the
+   `registration config section <https://matrix-org.github.io/synapse/latest/usage/configuration/config_documentation.html#registration>`_
+   set ``enable_registration: true`` in ``homeserver.yaml``.
+2. Then **either**:
+
+   a. set up a `CAPTCHA <https://matrix-org.github.io/synapse/latest/CAPTCHA_SETUP.html>`_, or
+   b. set ``enable_registration_without_verification: true`` in ``homeserver.yaml``.
+
+We **strongly** recommend using a CAPTCHA, particularly if your homeserver is exposed to
+the public internet. Without it, anyone can freely register accounts on your homeserver.
+This can be exploited by attackers to create spambots targetting the rest of the Matrix
+federation.
+
+Your new user name will be formed partly from the ``server_name``, and partly
+from a localpart you specify when you create the account. Your name will take
+the form of::
+
+    @localpart:my.domain.name
+
+(pronounced "at localpart on my dot domain dot name").
+
+As when logging in, you will need to specify a "Custom server".  Specify your
+desired ``localpart`` in the 'User name' box.
+
+Troubleshooting and support
+===========================
+
+The `Admin FAQ <https://matrix-org.github.io/synapse/latest/usage/administration/admin_faq.html>`_
+includes tips on dealing with some common problems. For more details, see
+`Synapse's wider documentation <https://matrix-org.github.io/synapse/latest/>`_.
+
+For additional support installing or managing Synapse, please ask in the community
+support room |room|_ (from a matrix.org account if necessary). We do not use GitHub
+issues for support requests, only for bug reports and feature requests.
+
+.. |room| replace:: ``#synapse:matrix.org``
+.. _room: https://matrix.to/#/#synapse:matrix.org
+
+.. |docs| replace:: ``docs``
+.. _docs: docs

 Identity Servers
 ================
@@ -242,34 +206,15 @@ an email address with your account, or send an invite to another user via their
 email address.


-Password reset
-==============
-
-Users can reset their password through their client. Alternatively, a server admin
-can reset a users password using the `admin API <docs/admin_api/user_admin_api.md#reset-password>`_
-or by directly editing the database as shown below.
-
-First calculate the hash of the new password::
-
-    $ ~/synapse/env/bin/hash_password
-    Password:
-    Confirm password:
-    $2a$12$xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
-
-Then update the ``users`` table in the database::
-
-    UPDATE users SET password_hash='$2a$12$xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx'
-        WHERE name='@test:test.com';
-
-
-Synapse Development
-===================
+Development
+===========

+We welcome contributions to Synapse from the community!
 The best place to get started is our
 `guide for contributors <https://matrix-org.github.io/synapse/latest/development/contributing_guide.html>`_.
 This is part of our larger `documentation <https://matrix-org.github.io/synapse/latest>`_, which includes
-information for synapse developers as well as synapse administrators.

+information for Synapse developers as well as Synapse administrators.
 Developers might be particularly interested in:

 * `Synapse's database schema <https://matrix-org.github.io/synapse/latest/development/database_schema.html>`_,
@@ -280,187 +225,6 @@ Alongside all that, join our developer community on Matrix:
 `#synapse-dev:matrix.org <https://matrix.to/#/#synapse-dev:matrix.org>`_, featuring real humans!


-Quick start
-----------
-
-Before setting up a development environment for synapse, make sure you have the
-system dependencies (such as the python header files) installed - see
-`Platform-specific prerequisites <https://matrix-org.github.io/synapse/latest/setup/installation.html#platform-specific-prerequisites>`_.
-
-To check out a synapse for development, clone the git repo into a working
-directory of your choice::
-
-    git clone https://github.com/matrix-org/synapse.git
-    cd synapse
-
-Synapse has a number of external dependencies. We maintain a fixed development
-environment using `Poetry <https://python-poetry.org/>`_. First, install poetry. We recommend::
-
-    pip install --user pipx
-    pipx install poetry
-
-as described `here <https://python-poetry.org/docs/#installing-with-pipx>`_.
-(See `poetry's installation docs <https://python-poetry.org/docs/#installation>`_
-for other installation methods.) Then ask poetry to create a virtual environment
-from the project and install Synapse's dependencies::
-
-    poetry install --extras "all test"
-
-This will run a process of downloading and installing all the needed
-dependencies into a virtual env.
-
-We recommend using the demo which starts 3 federated instances running on ports `8080` - `8082`::
-
-    poetry run ./demo/start.sh
-
-(to stop, you can use ``poetry run ./demo/stop.sh``)
-
-See the `demo documentation <https://matrix-org.github.io/synapse/develop/development/demo.html>`_
-for more information.
-
-If you just want to start a single instance of the app and run it directly::
-
-    # Create the homeserver.yaml config once
-    poetry run synapse_homeserver \
-      --server-name my.domain.name \
-      --config-path homeserver.yaml \
-      --generate-config \
-      --report-stats=[yes|no]
-
-    # Start the app
-    poetry run synapse_homeserver --config-path homeserver.yaml
-
-
-Running the unit tests
----------------------
-
-After getting up and running, you may wish to run Synapse's unit tests to
-check that everything is installed correctly::
-
-    poetry run trial tests
-
-This should end with a 'PASSED' result (note that exact numbers will
-differ)::
-
-    Ran 1337 tests in 716.064s
-
-    PASSED (skips=15, successes=1322)
-
-For more tips on running the unit tests, like running a specific test or
-to see the logging output, see the `CONTRIBUTING doc <CONTRIBUTING.md#run-the-unit-tests>`_.
-
-
-Running the Integration Tests
-----------------------------
-
-Synapse is accompanied by `SyTest <https://github.com/matrix-org/sytest>`_,
-a Matrix homeserver integration testing suite, which uses HTTP requests to
-access the API as a Matrix client would. It is able to run Synapse directly from
-the source tree, so installation of the server is not required.
-
-Testing with SyTest is recommended for verifying that changes related to the
-Client-Server API are functioning correctly. See the `SyTest installation
-instructions <https://github.com/matrix-org/sytest#installing>`_ for details.
-
-
-Platform dependencies
-=====================
-
-Synapse uses a number of platform dependencies such as Python and PostgreSQL,
-and aims to follow supported upstream versions. See the
-`<docs/deprecation_policy.md>`_ document for more details.
-
-
-Troubleshooting
-===============
-
-Need help? Join our community support room on Matrix:
-`#synapse:matrix.org <https://matrix.to/#/#synapse:matrix.org>`_
-
-Running out of File Handles
---------------------------
-
-If synapse runs out of file handles, it typically fails badly - live-locking
-at 100% CPU, and/or failing to accept new TCP connections (blocking the
-connecting client).  Matrix currently can legitimately use a lot of file handles,
-thanks to busy rooms like #matrix:matrix.org containing hundreds of participating
-servers.  The first time a server talks in a room it will try to connect
-simultaneously to all participating servers, which could exhaust the available
-file descriptors between DNS queries & HTTPS sockets, especially if DNS is slow
-to respond. (We need to improve the routing algorithm used to be better than
-full mesh, but as of March 2019 this hasn't happened yet).
-
-If you hit this failure mode, we recommend increasing the maximum number of
-open file handles to be at least 4096 (assuming a default of 1024 or 256).
-This is typically done by editing ``/etc/security/limits.conf``
-
-Separately, Synapse may leak file handles if inbound HTTP requests get stuck
-during processing - e.g. blocked behind a lock or talking to a remote server etc.
-This is best diagnosed by matching up the 'Received request' and 'Processed request'
-log lines and looking for any 'Processed request' lines which take more than
-a few seconds to execute. Please let us know at #synapse:matrix.org if
-you see this failure mode so we can help debug it, however.
-
-Help!! Synapse is slow and eats all my RAM/CPU!
-----------------------------------------------
-
-First, ensure you are running the latest version of Synapse, using Python 3
-with a PostgreSQL database.
-
-Synapse's architecture is quite RAM hungry currently - we deliberately
-cache a lot of recent room data and metadata in RAM in order to speed up
-common requests. We'll improve this in the future, but for now the easiest
-way to either reduce the RAM usage (at the risk of slowing things down)
-is to set the almost-undocumented ``SYNAPSE_CACHE_FACTOR`` environment
-variable. The default is 0.5, which can be decreased to reduce RAM usage
-in memory constrained enviroments, or increased if performance starts to
-degrade.
-
-However, degraded performance due to a low cache factor, common on
-machines with slow disks, often leads to explosions in memory use due
-backlogged requests. In this case, reducing the cache factor will make
-things worse. Instead, try increasing it drastically. 2.0 is a good
-starting value.
-
-Using `libjemalloc <http://jemalloc.net/>`_ can also yield a significant
-improvement in overall memory use, and especially in terms of giving back
-RAM to the OS. To use it, the library must simply be put in the
-LD_PRELOAD environment variable when launching Synapse. On Debian, this
-can be done by installing the ``libjemalloc1`` package and adding this
-line to ``/etc/default/matrix-synapse``::
-
-    LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libjemalloc.so.1
-
-This can make a significant difference on Python 2.7 - it's unclear how
-much of an improvement it provides on Python 3.x.
-
-If you're encountering high CPU use by the Synapse process itself, you
-may be affected by a bug with presence tracking that leads to a
-massive excess of outgoing federation requests (see `discussion
-<https://github.com/matrix-org/synapse/issues/3971>`_). If metrics
-indicate that your server is also issuing far more outgoing federation
-requests than can be accounted for by your users' activity, this is a
-likely cause. The misbehavior can be worked around by setting
-the following in the Synapse config file:
-
-.. code-block:: yaml
-
-   presence:
-       enabled: false
-
-People can't accept room invitations from me
--------------------------------------------
-
-The typical failure mode here is that you send an invitation to someone
-to join a room or direct chat, but when they go to accept it, they get an
-error (typically along the lines of "Invalid signature"). They might see
-something like the following in their logs::
-
-    2019-09-11 19:32:04,271 - synapse.federation.transport.server - 288 - WARNING - GET-11752 - authenticate_request failed: 401: Invalid signature for server <server> with key ed25519:a_EqML: Unable to verify signature for <server>
-
-This is normally caused by a misconfiguration in your reverse-proxy. See
-`<docs/reverse_proxy.md>`_ and double-check that your settings are correct.
-
 .. |support| image:: https://img.shields.io/matrix/synapse:matrix.org?label=support&logo=matrix
  :alt: (get support on #synapse:matrix.org)
  :target: https://matrix.to/#/#synapse:matrix.org
@@ -0,0 +1 @@
+Improve validation of request bodies for the following client-server API endpoints: [`/account/password`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountpassword), [`/account/password/email/requestToken`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountpasswordemailrequesttoken), [`/account/deactivate`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountdeactivate) and [`/account/3pid/email/requestToken`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3account3pidemailrequesttoken).
@@ -0,0 +1 @@
+Allow use of both `@trace` and `@tag_args` stacked on the same function (tracing).
@@ -0,0 +1 @@
+Faster joins: update the rejected state of events during de-partial-stating.
@@ -0,0 +1 @@
+Clean-up tests for notifications.
@@ -0,0 +1 @@
+Add `openssl` example for generating registration HMAC digest.
@@ -0,0 +1 @@
+Add some miscellaneous comments to document sync, especially around `compute_state_delta`.
@@ -0,0 +1 @@
+Use literals in place of `HTTPStatus` constants in tests.
@@ -0,0 +1 @@
+Add comments about how event push actions are rotated.
@@ -0,0 +1 @@
+Use literals in place of `HTTPStatus` constants in tests.
@@ -0,0 +1 @@
+Instrument the federation/backfill part of `/messages` for understandable traces in Jaeger.
@@ -0,0 +1 @@
+Tidy up Synapse's README.
@@ -0,0 +1 @@
+Document that event purging related to the `redaction_retention_period` config option is executed only every 5 minutes.
@@ -0,0 +1 @@
+Modify HTML template content to better support mobile devices' screen sizes.
@@ -0,0 +1,2 @@
+Add a warning to retention documentation regarding the possibility of database corruption.
+
@@ -0,0 +1 @@
+Instrument `FederationStateIdsServlet` (`/state_ids`) for understandable traces in Jaeger.
@@ -0,0 +1 @@
+Add a linter script which will reject non-strict types in Pydantic models.
@@ -0,0 +1 @@
+Add forgotten status to Room Details API.
@@ -0,0 +1 @@
+Faster room joins: make `/joined_members` block whilst the room is partial stated.
@@ -0,0 +1 @@
+Document that the `DOCKER_BUILDKIT=1` flag is needed to build the docker image.
@@ -0,0 +1 @@
+Improve performance of sending messages in rooms with thousands of local users.
@@ -0,0 +1 @@
+Fix a bug in the `/event_reports` Admin API which meant that the total count could be larger than the number of results you can actually query for.
@@ -0,0 +1 @@
+Faster room joins: Refuse to start when faster joins is enabled on a deployment with workers, since worker configurations are not currently supported.
@@ -0,0 +1 @@
+Track HTTP response times over 10 seconds from `/messages` (`synapse_room_message_list_rest_servlet_response_time_seconds`).
@@ -0,0 +1 @@
+Add metrics to track how the rate limiter is affecting requests (sleep/reject).
@@ -0,0 +1 @@
+Add metrics to time how long it takes us to do backfill processing (`synapse_federation_backfill_processing_before_time_seconds`, `synapse_federation_backfill_processing_after_time_seconds`).
@@ -0,0 +1 @@
+Add missing links in `user_consent` section of configuration manual.
@@ -0,0 +1 @@
+Fix the doc and some warnings that were referring to the nonexistent `custom_templates_directory` setting (instead of `custom_template_directory`).
@@ -0,0 +1 @@
+Add metrics to track rate limiter queue timing (`synapse_rate_limit_queue_wait_time_seconds`).
@@ -0,0 +1 @@
+Improve performance of sending messages in rooms with thousands of local users.
@@ -0,0 +1 @@
+Downgrade an error that had the capacity to spam Sentry instances to a warning.
@@ -1,3 +1,21 @@
+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.
@@ -40,7 +40,8 @@ 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
@@ -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
@@ -1,39 +1,62 @@
 # syntax=docker/dockerfile:1
-# Inherit from the official Synapse docker image
+
 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
@@ -191,7 +191,7 @@ If you need to build the image from a Synapse checkout, use the following `docke
 build` command from the repo's root:

 ```
-docker build -t matrixdotorg/synapse -f docker/Dockerfile .
+DOCKER_BUILDKIT=1 docker build -t matrixdotorg/synapse -f docker/Dockerfile .
 ```

 You can choose to build a different docker image by changing the value of the `-f` flag to
@@ -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
@@ -46,7 +46,24 @@ As an example:
 The MAC is the hex digest output of the HMAC-SHA1 algorithm, with the key being
 the shared secret and the content being the nonce, user, password, either the
 string "admin" or "notadmin", and optionally the user_type
-each separated by NULs. For an example of generation in Python:
+each separated by NULs.
+
+Here is an easy way to generate the HMAC digest if you have Bash and OpenSSL:
+
+```bash
+# Update these values and then paste this code block into a bash terminal
+nonce='thisisanonce'
+username='pepper_roni'
+password='pizza'
+admin='admin'
+secret='shared_secret'
+
+printf '%s\0%s\0%s\0%s' "$nonce" "$username" "$password" "$admin" |
+  openssl sha1 -hmac "$secret" |
+  awk '{print $2}'
+```
+
+For an example of generation in Python:

 ```python
 import hmac, hashlib
@@ -70,4 +87,4 @@ def generate_mac(nonce, user, password, admin=False, user_type=None):
        mac.update(user_type.encode('utf8'))

    return mac.hexdigest()
-```
+```
@@ -302,6 +302,8 @@ The following fields are possible in the JSON response body:
 * `state_events` - Total number of state_events of a room. Complexity of the room.
 * `room_type` - The type of the room taken from the room's creation event; for example "m.space" if the room is a space.
  If the room does not define a type, the value will be `null`.
+* `forgotten` - Whether all local users have
+  [forgotten](https://spec.matrix.org/latest/client-server-api/#leaving-rooms) the room.

 The API is:

@@ -330,7 +332,8 @@ A response body like the following is returned:
  "guest_access": null,
  "history_visibility": "shared",
  "state_events": 93534,
-  "room_type": "m.space"
+  "room_type": "m.space",
+  "forgotten": false
 }
 ```

@@ -8,7 +8,8 @@ and allow server and room admins to configure how long messages should
 be kept in a homeserver's database before being purged from it.
 **Please note that, as this feature isn't part of the Matrix
 specification yet, this implementation is to be considered as
-experimental.** 
+experimental. There are known bugs which may cause database corruption.
+Proceed with caution.** 

 A message retention policy is mainly defined by its `max_lifetime`
 parameter, which defines how long a message can be kept around after
@@ -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)
 ```
@@ -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
@@ -9,7 +9,7 @@ in, allowing them to specify custom templates:

 ```yaml
 templates:
-  custom_templates_directory: /path/to/custom/templates/
+  custom_template_directory: /path/to/custom/templates/
 ```

 If this setting is not set, or the files named below are not found within the directory,
@@ -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';
 ```
@@ -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
@@ -2,9 +2,9 @@

 How do I become a server admin?
 ---
-If your server already has an admin account you should use the user admin API to promote other accounts to become admins. See [User Admin API](../../admin_api/user_admin_api.md#Change-whether-a-user-is-a-server-administrator-or-not)
+If your server already has an admin account you should use the [User Admin API](../../admin_api/user_admin_api.md#Change-whether-a-user-is-a-server-administrator-or-not) to promote other accounts to become admins.

-If you don't have any admin accounts yet you won't be able to use the admin API so you'll have to edit the database manually. Manually editing the database is generally not recommended so once you have an admin account, use the admin APIs to make further changes.
+If you don't have any admin accounts yet you won't be able to use the admin API, so you'll have to edit the database manually. Manually editing the database is generally not recommended so once you have an admin account: use the admin APIs to make further changes.

 ```sql
 UPDATE users SET admin = 1 WHERE name = '@foo:bar.com';
@@ -32,9 +32,11 @@ What users are registered on my server?
 SELECT NAME from users;
 ```

-Manually resetting passwords:
+Manually resetting passwords
 ---
-See https://github.com/matrix-org/synapse/blob/master/README.rst#password-reset
+Users can reset their password through their client. Alternatively, a server admin
+can reset a user's password using the [admin API](../../admin_api/user_admin_api.md#reset-password).
+

 I have a problem with my server. Can I just delete my database and start again?
 ---
@@ -101,3 +103,83 @@ LIMIT 10;

 You can also use the [List Room API](../../admin_api/rooms.md#list-room-api)
 and `order_by` `state_events`.
+
+
+People can't accept room invitations from me
+---
+
+The typical failure mode here is that you send an invitation to someone
+to join a room or direct chat, but when they go to accept it, they get an
+error (typically along the lines of "Invalid signature"). They might see
+something like the following in their logs:
+
+    2019-09-11 19:32:04,271 - synapse.federation.transport.server - 288 - WARNING - GET-11752 - authenticate_request failed: 401: Invalid signature for server <server> with key ed25519:a_EqML: Unable to verify signature for <server>
+
+This is normally caused by a misconfiguration in your reverse-proxy. See [the reverse proxy docs](docs/reverse_proxy.md) and double-check that your settings are correct.
+
+
+Help!! Synapse is slow and eats all my RAM/CPU!
+-----------------------------------------------
+
+First, ensure you are running the latest version of Synapse, using Python 3
+with a [PostgreSQL database](../../postgres.md).
+
+Synapse's architecture is quite RAM hungry currently - we deliberately
+cache a lot of recent room data and metadata in RAM in order to speed up
+common requests. We'll improve this in the future, but for now the easiest
+way to either reduce the RAM usage (at the risk of slowing things down)
+is to set the almost-undocumented ``SYNAPSE_CACHE_FACTOR`` environment
+variable. The default is 0.5, which can be decreased to reduce RAM usage
+in memory constrained environments, or increased if performance starts to
+degrade.
+
+However, degraded performance due to a low cache factor, common on
+machines with slow disks, often leads to explosions in memory use due
+backlogged requests. In this case, reducing the cache factor will make
+things worse. Instead, try increasing it drastically. 2.0 is a good
+starting value.
+
+Using [libjemalloc](https://jemalloc.net) can also yield a significant
+improvement in overall memory use, and especially in terms of giving back
+RAM to the OS. To use it, the library must simply be put in the
+LD_PRELOAD environment variable when launching Synapse. On Debian, this
+can be done by installing the `libjemalloc1` package and adding this
+line to `/etc/default/matrix-synapse`:
+
+    LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libjemalloc.so.1
+
+This made a significant difference on Python 2.7 - it's unclear how
+much of an improvement it provides on Python 3.x.
+
+If you're encountering high CPU use by the Synapse process itself, you
+may be affected by a bug with presence tracking that leads to a
+massive excess of outgoing federation requests (see [discussion](https://github.com/matrix-org/synapse/issues/3971)). If metrics
+indicate that your server is also issuing far more outgoing federation
+requests than can be accounted for by your users' activity, this is a
+likely cause. The misbehavior can be worked around by disabling presence
+in the Synapse config file: [see here](../configuration/config_documentation.md#presence).
+
+
+Running out of File Handles
+---------------------------
+
+If Synapse runs out of file handles, it typically fails badly - live-locking
+at 100% CPU, and/or failing to accept new TCP connections (blocking the
+connecting client).  Matrix currently can legitimately use a lot of file handles,
+thanks to busy rooms like `#matrix:matrix.org` containing hundreds of participating
+servers.  The first time a server talks in a room it will try to connect
+simultaneously to all participating servers, which could exhaust the available
+file descriptors between DNS queries & HTTPS sockets, especially if DNS is slow
+to respond. (We need to improve the routing algorithm used to be better than
+full mesh, but as of March 2019 this hasn't happened yet).
+
+If you hit this failure mode, we recommend increasing the maximum number of
+open file handles to be at least 4096 (assuming a default of 1024 or 256).
+This is typically done by editing ``/etc/security/limits.conf``
+
+Separately, Synapse may leak file handles if inbound HTTP requests get stuck
+during processing - e.g. blocked behind a lock or talking to a remote server etc.
+This is best diagnosed by matching up the 'Received request' and 'Processed request'
+log lines and looking for any 'Processed request' lines which take more than
+a few seconds to execute. Please let us know at [`#synapse:matrix.org`](https://matrix.to/#/#synapse-dev:matrix.org) if
+you see this failure mode so we can help debug it, however.
@@ -72,49 +72,6 @@ apply if you want your config file to be read properly. A few helpful things to
  In addition, each setting has an example of its usage, with the proper indentation
  shown.

-## Contents
-[Modules](#modules)
-
-[Server](#server)
-
-[Homeserver Blocking](#homeserver-blocking)
-
-[TLS](#tls)
-
-[Federation](#federation)
-
-[Caching](#caching)
-
-[Database](#database)
-
-[Logging](#logging)
-
-[Ratelimiting](#ratelimiting)
-
-[Media Store](#media-store)
-
-[Captcha](#captcha)
-
-[TURN](#turn)
-
-[Registration](#registration)
-
-[API Configuration](#api-configuration)
-
-[Signing Keys](#signing-keys)
-
-[Single Sign On Integration](#single-sign-on-integration)
-
-[Push](#push)
-
-[Rooms](#rooms)
-
-[Opentracing](#opentracing)
-
-[Workers](#workers)
-
-[Background Updates](#background-updates)
-
 ## Modules

 Server admins can expand Synapse's functionality with external modules.
@@ -486,7 +443,8 @@ Sub-options for each listener include:

   * `names`: a list of names of HTTP resources. See below for a list of valid resource names.

-   * `compress`: set to true to enable HTTP compression for this resource.
+   * `compress`: set to true to enable gzip compression on HTTP bodies for this resource. This is currently only supported with the
+     `client`, `consent` and `metrics` resources.

 * `additional_resources`: Only valid for an 'http' listener. A map of
   additional endpoints which should be loaded via dynamic modules.
@@ -801,6 +759,10 @@ allowed_avatar_mimetypes: ["image/png", "image/jpeg", "image/gif"]
 How long to keep redacted events in unredacted form in the database. After
 this period redacted events get replaced with their redacted form in the DB.

+Synapse will check whether the rentention period has concluded for redacted
+events every 5 minutes. Thus, even if this option is set to `0`, Synapse may
+still take up to 5 minutes to purge redacted events from the database.
+
 Defaults to `7d`. Set to `null` to disable.

 Example configuration:
@@ -887,7 +849,11 @@ which are older than the room's maximum retention period. Synapse will also
 filter events received over federation so that events that should have been
 purged are ignored and not stored again.

-The message retention policies feature is disabled by default.
+The message retention policies feature is disabled by default. Please be advised 
+that enabling this feature carries some risk. There are known bugs with the implementation
+which can cause database corruption. Setting retention to delete older history
+is less risky than deleting newer history but in general caution is advised when enabling this
+experimental feature. You can read more about this feature [here](../../message_retention_policies.md).

 This setting has the following sub-options:
 * `default_policy`: Default retention policy. If set, Synapse will apply it to rooms that lack the
@@ -1098,26 +1064,26 @@ allow_device_name_lookup_over_federation: true
 ---
 ## Caching ##

-Options related to caching
+Options related to caching.

 ---
 ### `event_cache_size`

 The number of events to cache in memory. Not affected by
-`caches.global_factor`. Defaults to 10K.
+`caches.global_factor` and is not part of the `caches` section. Defaults to 10K.

 Example configuration:
 ```yaml
 event_cache_size: 15K
 ```
 ---
-### `cache` and associated values
+### `caches` and associated values

 A cache 'factor' is a multiplier that can be applied to each of
 Synapse's caches in order to increase or decrease the maximum
 number of entries that can be stored.

-Caching can be configured through the following sub-options:
+`caches` can be configured through the following sub-options:

 * `global_factor`: Controls the global cache factor, which is the default cache factor
  for all caches if a specific factor for that cache is not otherwise
@@ -1179,6 +1145,7 @@ Caching can be configured through the following sub-options:

 Example configuration:
 ```yaml
+event_cache_size: 15K
 caches:
  global_factor: 1.0
  per_cache_factors:
@@ -1858,7 +1825,7 @@ Example configuration:
 max_spider_size: 8M
 ```
 ---
-### `url_preview_language`
+### `url_preview_accept_language`

 A list of values for the Accept-Language HTTP header used when
 downloading webpages during URL preview generation. This allows
@@ -2537,9 +2504,13 @@ track_appservice_user_ips: true
 ---
 ### `macaroon_secret_key`

-A secret which is used to sign access tokens. If none is specified,
-the `registration_shared_secret` is used, if one is given; otherwise,
-a secret key is derived from the signing key.
+A secret which is used to sign
+- access token for guest users,
+- short-term login token used during SSO logins (OIDC or SAML2) and
+- token used for unsubscribing from email notifications.
+
+If none is specified, the `registration_shared_secret` is used, if one is given;
+otherwise, a secret key is derived from the signing key.

 Example configuration:
 ```yaml
@@ -3381,7 +3352,7 @@ user_directory:
 For detailed instructions on user consent configuration, see [here](../../consent_tracking.md).

 Parts of this section are required if enabling the `consent` resource under
-`listeners`, in particular `template_dir` and `version`. # TODO: link `listeners`
+[`listeners`](#listeners), in particular `template_dir` and `version`.

 * `template_dir`: gives the location of the templates for the HTML forms.
  This directory should contain one subdirectory per language (eg, `en`, `fr`),
@@ -3393,7 +3364,7 @@ Parts of this section are required if enabling the `consent` resource under
   parameter.

 * `server_notice_content`: if enabled, will send a user a "Server Notice"
-   asking them to consent to the privacy policy. The `server_notices` section ##TODO: link
+   asking them to consent to the privacy policy. The [`server_notices` section](#server_notices)
   must also be configured for this to work. Notices will *not* be sent to
   guest users unless `send_server_notice_to_guests` is set to true.

@@ -1,6 +1,6 @@
 [mypy]
 namespace_packages = True
-plugins = mypy_zope:plugin, scripts-dev/mypy_synapse_plugin.py
+plugins = pydantic.mypy, mypy_zope:plugin, scripts-dev/mypy_synapse_plugin.py
 follow_imports = normal
 check_untyped_defs = True
 show_error_codes = True
@@ -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"
@@ -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"
@@ -778,6 +778,21 @@ category = "main"
 optional = false
 python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*"

+[[package]]
+name = "pydantic"
+version = "1.9.1"
+description = "Data validation and settings management using python type hints"
+category = "main"
+optional = false
+python-versions = ">=3.6.1"
+
+[package.dependencies]
+typing-extensions = ">=3.7.4.3"
+
+[package.extras]
+dotenv = ["python-dotenv (>=0.10.4)"]
+email = ["email-validator (>=1.0.3)"]
+
 [[package]]
 name = "pyflakes"
 version = "2.4.0"
@@ -820,10 +835,10 @@ 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 +1578,7 @@ url_preview = ["lxml"]
 [metadata]
 lock-version = "1.1"
 python-versions = "^3.7.1"
-content-hash = "c24bbcee7e86dbbe7cdbf49f91a25b310bf21095452641e7440129f59b077f78"
+content-hash = "7de518bf27967b3547eab8574342cfb67f87d6b47b4145c13de11112141dbf2d"

 [metadata.files]
 attrs = [
@@ -2055,8 +2070,8 @@ 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"},
@@ -2260,6 +2275,43 @@ pycparser = [
    {file = "pycparser-2.21-py2.py3-none-any.whl", hash = "sha256:8ee45429555515e1f6b185e78100aea234072576aa43ab53aefcae078162fca9"},
    {file = "pycparser-2.21.tar.gz", hash = "sha256:e644fdec12f7872f86c58ff790da456218b10f863970249516d60a5eaca77206"},
 ]
+pydantic = [
+    {file = "pydantic-1.9.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:c8098a724c2784bf03e8070993f6d46aa2eeca031f8d8a048dff277703e6e193"},
+    {file = "pydantic-1.9.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c320c64dd876e45254bdd350f0179da737463eea41c43bacbee9d8c9d1021f11"},
+    {file = "pydantic-1.9.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:18f3e912f9ad1bdec27fb06b8198a2ccc32f201e24174cec1b3424dda605a310"},
+    {file = "pydantic-1.9.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:c11951b404e08b01b151222a1cb1a9f0a860a8153ce8334149ab9199cd198131"},
+    {file = "pydantic-1.9.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:8bc541a405423ce0e51c19f637050acdbdf8feca34150e0d17f675e72d119580"},
+    {file = "pydantic-1.9.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:e565a785233c2d03724c4dc55464559639b1ba9ecf091288dd47ad9c629433bd"},
+    {file = "pydantic-1.9.1-cp310-cp310-win_amd64.whl", hash = "sha256:a4a88dcd6ff8fd47c18b3a3709a89adb39a6373f4482e04c1b765045c7e282fd"},
+    {file = "pydantic-1.9.1-cp36-cp36m-macosx_10_9_x86_64.whl", hash = "sha256:447d5521575f18e18240906beadc58551e97ec98142266e521c34968c76c8761"},
+    {file = "pydantic-1.9.1-cp36-cp36m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:985ceb5d0a86fcaa61e45781e567a59baa0da292d5ed2e490d612d0de5796918"},
+    {file = "pydantic-1.9.1-cp36-cp36m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:059b6c1795170809103a1538255883e1983e5b831faea6558ef873d4955b4a74"},
+    {file = "pydantic-1.9.1-cp36-cp36m-musllinux_1_1_i686.whl", hash = "sha256:d12f96b5b64bec3f43c8e82b4aab7599d0157f11c798c9f9c528a72b9e0b339a"},
+    {file = "pydantic-1.9.1-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:ae72f8098acb368d877b210ebe02ba12585e77bd0db78ac04a1ee9b9f5dd2166"},
+    {file = "pydantic-1.9.1-cp36-cp36m-win_amd64.whl", hash = "sha256:79b485767c13788ee314669008d01f9ef3bc05db9ea3298f6a50d3ef596a154b"},
+    {file = "pydantic-1.9.1-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:494f7c8537f0c02b740c229af4cb47c0d39840b829ecdcfc93d91dcbb0779892"},
+    {file = "pydantic-1.9.1-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f0f047e11febe5c3198ed346b507e1d010330d56ad615a7e0a89fae604065a0e"},
+    {file = "pydantic-1.9.1-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:969dd06110cb780da01336b281f53e2e7eb3a482831df441fb65dd30403f4608"},
+    {file = "pydantic-1.9.1-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:177071dfc0df6248fd22b43036f936cfe2508077a72af0933d0c1fa269b18537"},
+    {file = "pydantic-1.9.1-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:9bcf8b6e011be08fb729d110f3e22e654a50f8a826b0575c7196616780683380"},
+    {file = "pydantic-1.9.1-cp37-cp37m-win_amd64.whl", hash = "sha256:a955260d47f03df08acf45689bd163ed9df82c0e0124beb4251b1290fa7ae728"},
+    {file = "pydantic-1.9.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:9ce157d979f742a915b75f792dbd6aa63b8eccaf46a1005ba03aa8a986bde34a"},
+    {file = "pydantic-1.9.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:0bf07cab5b279859c253d26a9194a8906e6f4a210063b84b433cf90a569de0c1"},
+    {file = "pydantic-1.9.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5d93d4e95eacd313d2c765ebe40d49ca9dd2ed90e5b37d0d421c597af830c195"},
+    {file = "pydantic-1.9.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:1542636a39c4892c4f4fa6270696902acb186a9aaeac6f6cf92ce6ae2e88564b"},
+    {file = "pydantic-1.9.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:a9af62e9b5b9bc67b2a195ebc2c2662fdf498a822d62f902bf27cccb52dbbf49"},
+    {file = "pydantic-1.9.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:fe4670cb32ea98ffbf5a1262f14c3e102cccd92b1869df3bb09538158ba90fe6"},
+    {file = "pydantic-1.9.1-cp38-cp38-win_amd64.whl", hash = "sha256:9f659a5ee95c8baa2436d392267988fd0f43eb774e5eb8739252e5a7e9cf07e0"},
+    {file = "pydantic-1.9.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:b83ba3825bc91dfa989d4eed76865e71aea3a6ca1388b59fc801ee04c4d8d0d6"},
+    {file = "pydantic-1.9.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:1dd8fecbad028cd89d04a46688d2fcc14423e8a196d5b0a5c65105664901f810"},
+    {file = "pydantic-1.9.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:02eefd7087268b711a3ff4db528e9916ac9aa18616da7bca69c1871d0b7a091f"},
+    {file = "pydantic-1.9.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:7eb57ba90929bac0b6cc2af2373893d80ac559adda6933e562dcfb375029acee"},
+    {file = "pydantic-1.9.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:4ce9ae9e91f46c344bec3b03d6ee9612802682c1551aaf627ad24045ce090761"},
+    {file = "pydantic-1.9.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:72ccb318bf0c9ab97fc04c10c37683d9eea952ed526707fabf9ac5ae59b701fd"},
+    {file = "pydantic-1.9.1-cp39-cp39-win_amd64.whl", hash = "sha256:61b6760b08b7c395975d893e0b814a11cf011ebb24f7d869e7118f5a339a82e1"},
+    {file = "pydantic-1.9.1-py3-none-any.whl", hash = "sha256:4988c0f13c42bfa9ddd2fe2f569c9d54646ce84adc5de84228cfe83396f3bd58"},
+    {file = "pydantic-1.9.1.tar.gz", hash = "sha256:1ed987c3ff29fff7fd8c3ea3a3ea877ad310aae2ef9889a119e22d3f2db0691a"},
+]
 pyflakes = [
    {file = "pyflakes-2.4.0-py2.py3-none-any.whl", hash = "sha256:3bb3a3f256f4b7968c9c788781e4ff07dce46bdf12339dcda61053375426ee2e"},
    {file = "pyflakes-2.4.0.tar.gz", hash = "sha256:05a85c2872edf37a4ed30b0cce2f6093e1d0581f8c19d7393122da7e25b2b24c"},
@@ -54,7 +54,7 @@ skip_gitignore = true

 [tool.poetry]
 name = "matrix-synapse"
-version = "1.64.0"
+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"
@@ -158,6 +158,9 @@ packaging = ">=16.1"
 # At the time of writing, we only use functions from the version `importlib.metadata`
 # which shipped in Python 3.8. This corresponds to version 1.4 of the backport.
 importlib_metadata = { version = ">=1.4", python = "<3.8" }
+# This is the most recent version of Pydantic with available on common distros.
+pydantic = ">=1.7.4"
+


 # Optional Dependencies
@@ -0,0 +1,425 @@
+#! /usr/bin/env python
+# Copyright 2022 The Matrix.org Foundation C.I.C.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# 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.
+"""
+A script which enforces that Synapse always uses strict types when defining a Pydantic
+model.
+
+Pydantic does not yet offer a strict mode, but it is planned for pydantic v2. See
+
+    https://github.com/pydantic/pydantic/issues/1098
+    https://pydantic-docs.helpmanual.io/blog/pydantic-v2/#strict-mode
+
+until then, this script is a best effort to stop us from introducing type coersion bugs
+(like the infamous stringy power levels fixed in room version 10).
+"""
+import argparse
+import contextlib
+import functools
+import importlib
+import logging
+import os
+import pkgutil
+import sys
+import textwrap
+import traceback
+import unittest.mock
+from contextlib import contextmanager
+from typing import Any, Callable, Dict, Generator, List, Set, Type, TypeVar
+
+from parameterized import parameterized
+from pydantic import BaseModel as PydanticBaseModel, conbytes, confloat, conint, constr
+from pydantic.typing import get_args
+from typing_extensions import ParamSpec
+
+logger = logging.getLogger(__name__)
+
+CONSTRAINED_TYPE_FACTORIES_WITH_STRICT_FLAG: List[Callable] = [
+    constr,
+    conbytes,
+    conint,
+    confloat,
+]
+
+TYPES_THAT_PYDANTIC_WILL_COERCE_TO = [
+    str,
+    bytes,
+    int,
+    float,
+    bool,
+]
+
+
+P = ParamSpec("P")
+R = TypeVar("R")
+
+
+class ModelCheckerException(Exception):
+    """Dummy exception. Allows us to detect unwanted types during a module import."""
+
+
+class MissingStrictInConstrainedTypeException(ModelCheckerException):
+    factory_name: str
+
+    def __init__(self, factory_name: str):
+        self.factory_name = factory_name
+
+
+class FieldHasUnwantedTypeException(ModelCheckerException):
+    message: str
+
+    def __init__(self, message: str):
+        self.message = message
+
+
+def make_wrapper(factory: Callable[P, R]) -> Callable[P, R]:
+    """We patch `constr` and friends with wrappers that enforce strict=True."""
+
+    @functools.wraps(factory)
+    def wrapper(*args: P.args, **kwargs: P.kwargs) -> R:
+        # type-ignore: should be redundant once we can use https://github.com/python/mypy/pull/12668
+        if "strict" not in kwargs:  # type: ignore[attr-defined]
+            raise MissingStrictInConstrainedTypeException(factory.__name__)
+        if not kwargs["strict"]:  # type: ignore[index]
+            raise MissingStrictInConstrainedTypeException(factory.__name__)
+        return factory(*args, **kwargs)
+
+    return wrapper
+
+
+def field_type_unwanted(type_: Any) -> bool:
+    """Very rough attempt to detect if a type is unwanted as a Pydantic annotation.
+
+    At present, we exclude types which will coerce, or any generic type involving types
+    which will coerce."""
+    logger.debug("Is %s unwanted?")
+    if type_ in TYPES_THAT_PYDANTIC_WILL_COERCE_TO:
+        logger.debug("yes")
+        return True
+    logger.debug("Maybe. Subargs are %s", get_args(type_))
+    rv = any(field_type_unwanted(t) for t in get_args(type_))
+    logger.debug("Conclusion: %s %s unwanted", type_, "is" if rv else "is not")
+    return rv
+
+
+class PatchedBaseModel(PydanticBaseModel):
+    """A patched version of BaseModel that inspects fields after models are defined.
+
+    We complain loudly if we see an unwanted type.
+
+    Beware: ModelField.type_ is presumably private; this is likely to be very brittle.
+    """
+
+    @classmethod
+    def __init_subclass__(cls: Type[PydanticBaseModel], **kwargs: object):
+        for field in cls.__fields__.values():
+            # Note that field.type_ and field.outer_type are computed based on the
+            # annotation type, see pydantic.fields.ModelField._type_analysis
+            if field_type_unwanted(field.outer_type_):
+                # TODO: this only reports the first bad field. Can we find all bad ones
+                #  and report them all?
+                raise FieldHasUnwantedTypeException(
+                    f"{cls.__module__}.{cls.__qualname__} has field '{field.name}' "
+                    f"with unwanted type `{field.outer_type_}`"
+                )
+
+
+@contextmanager
+def monkeypatch_pydantic() -> Generator[None, None, None]:
+    """Patch pydantic with our snooping versions of BaseModel and the con* functions.
+
+    If the snooping functions see something they don't like, they'll raise a
+    ModelCheckingException instance.
+    """
+    with contextlib.ExitStack() as patches:
+        # Most Synapse code ought to import the patched objects directly from
+        # `pydantic`. But we also patch their containing modules `pydantic.main` and
+        # `pydantic.types` for completeness.
+        patch_basemodel1 = unittest.mock.patch(
+            "pydantic.BaseModel", new=PatchedBaseModel
+        )
+        patch_basemodel2 = unittest.mock.patch(
+            "pydantic.main.BaseModel", new=PatchedBaseModel
+        )
+        patches.enter_context(patch_basemodel1)
+        patches.enter_context(patch_basemodel2)
+        for factory in CONSTRAINED_TYPE_FACTORIES_WITH_STRICT_FLAG:
+            wrapper: Callable = make_wrapper(factory)
+            patch1 = unittest.mock.patch(f"pydantic.{factory.__name__}", new=wrapper)
+            patch2 = unittest.mock.patch(
+                f"pydantic.types.{factory.__name__}", new=wrapper
+            )
+            patches.enter_context(patch1)
+            patches.enter_context(patch2)
+        yield
+
+
+def format_model_checker_exception(e: ModelCheckerException) -> str:
+    """Work out which line of code caused e. Format the line in a human-friendly way."""
+    # TODO. FieldHasUnwantedTypeException gives better error messages. Can we ditch the
+    #   patches of constr() etc, and instead inspect fields to look for ConstrainedStr
+    #   with strict=False? There is some difficulty with the inheritance hierarchy
+    #   because StrictStr < ConstrainedStr < str.
+    if isinstance(e, FieldHasUnwantedTypeException):
+        return e.message
+    elif isinstance(e, MissingStrictInConstrainedTypeException):
+        frame_summary = traceback.extract_tb(e.__traceback__)[-2]
+        return (
+            f"Missing `strict=True` from {e.factory_name}() call \n"
+            + traceback.format_list([frame_summary])[0].lstrip()
+        )
+    else:
+        raise ValueError(f"Unknown exception {e}") from e
+
+
+def lint() -> int:
+    """Try to import all of Synapse and see if we spot any Pydantic type coercions.
+
+    Print any problems, then return a status code suitable for sys.exit."""
+    failures = do_lint()
+    if failures:
+        print(f"Found {len(failures)} problem(s)")
+    for failure in sorted(failures):
+        print(failure)
+    return os.EX_DATAERR if failures else os.EX_OK
+
+
+def do_lint() -> Set[str]:
+    """Try to import all of Synapse and see if we spot any Pydantic type coercions."""
+    failures = set()
+
+    with monkeypatch_pydantic():
+        logger.debug("Importing synapse")
+        try:
+            # TODO: make "synapse" an argument so we can target this script at
+            # a subpackage
+            module = importlib.import_module("synapse")
+        except ModelCheckerException as e:
+            logger.warning("Bad annotation found when importing synapse")
+            failures.add(format_model_checker_exception(e))
+            return failures
+
+        try:
+            logger.debug("Fetching subpackages")
+            module_infos = list(
+                pkgutil.walk_packages(module.__path__, f"{module.__name__}.")
+            )
+        except ModelCheckerException as e:
+            logger.warning("Bad annotation found when looking for modules to import")
+            failures.add(format_model_checker_exception(e))
+            return failures
+
+        for module_info in module_infos:
+            logger.debug("Importing %s", module_info.name)
+            try:
+                importlib.import_module(module_info.name)
+            except ModelCheckerException as e:
+                logger.warning(
+                    f"Bad annotation found when importing {module_info.name}"
+                )
+                failures.add(format_model_checker_exception(e))
+
+    return failures
+
+
+def run_test_snippet(source: str) -> None:
+    """Exec a snippet of source code in an isolated environment."""
+    # To emulate `source` being called at the top level of the module,
+    # the globals and locals we provide apparently have to be the same mapping.
+    #
+    # > Remember that at the module level, globals and locals are the same dictionary.
+    # > If exec gets two separate objects as globals and locals, the code will be
+    # > executed as if it were embedded in a class definition.
+    globals_: Dict[str, object]
+    locals_: Dict[str, object]
+    globals_ = locals_ = {}
+    exec(textwrap.dedent(source), globals_, locals_)
+
+
+class TestConstrainedTypesPatch(unittest.TestCase):
+    def test_expression_without_strict_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import constr
+                constr()
+                """
+            )
+
+    def test_called_as_module_attribute_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                import pydantic
+                pydantic.constr()
+                """
+            )
+
+    def test_wildcard_import_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import *
+                constr()
+                """
+            )
+
+    def test_alternative_import_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic.types import constr
+                constr()
+                """
+            )
+
+    def test_alternative_import_attribute_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                import pydantic.types
+                pydantic.types.constr()
+                """
+            )
+
+    def test_kwarg_but_no_strict_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import constr
+                constr(min_length=10)
+                """
+            )
+
+    def test_kwarg_strict_False_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import constr
+                constr(strict=False)
+                """
+            )
+
+    def test_kwarg_strict_True_doesnt_raise(self) -> None:
+        with monkeypatch_pydantic():
+            run_test_snippet(
+                """
+                from pydantic import constr
+                constr(strict=True)
+                """
+            )
+
+    def test_annotation_without_strict_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import constr
+                x: constr()
+                """
+            )
+
+    def test_field_annotation_without_strict_raises(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic import BaseModel, conint
+                class C:
+                    x: conint()
+                """
+            )
+
+
+class TestFieldTypeInspection(unittest.TestCase):
+    @parameterized.expand(
+        [
+            ("str",),
+            ("bytes"),
+            ("int",),
+            ("float",),
+            ("bool"),
+            ("Optional[str]",),
+            ("Union[None, str]",),
+            ("List[str]",),
+            ("List[List[str]]",),
+            ("Dict[StrictStr, str]",),
+            ("Dict[str, StrictStr]",),
+            ("TypedDict('D', x=int)",),
+        ]
+    )
+    def test_field_holding_unwanted_type_raises(self, annotation: str) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                f"""
+                from typing import *
+                from pydantic import *
+                class C(BaseModel):
+                    f: {annotation}
+                """
+            )
+
+    @parameterized.expand(
+        [
+            ("StrictStr",),
+            ("StrictBytes"),
+            ("StrictInt",),
+            ("StrictFloat",),
+            ("StrictBool"),
+            ("constr(strict=True, min_length=10)",),
+            ("Optional[StrictStr]",),
+            ("Union[None, StrictStr]",),
+            ("List[StrictStr]",),
+            ("List[List[StrictStr]]",),
+            ("Dict[StrictStr, StrictStr]",),
+            ("TypedDict('D', x=StrictInt)",),
+        ]
+    )
+    def test_field_holding_accepted_type_doesnt_raise(self, annotation: str) -> None:
+        with monkeypatch_pydantic():
+            run_test_snippet(
+                f"""
+                from typing import *
+                from pydantic import *
+                class C(BaseModel):
+                    f: {annotation}
+                """
+            )
+
+    def test_field_holding_str_raises_with_alternative_import(self) -> None:
+        with monkeypatch_pydantic(), self.assertRaises(ModelCheckerException):
+            run_test_snippet(
+                """
+                from pydantic.main import BaseModel
+                class C(BaseModel):
+                    f: str
+                """
+            )
+
+
+parser = argparse.ArgumentParser()
+parser.add_argument("mode", choices=["lint", "test"], default="lint", nargs="?")
+parser.add_argument("-v", "--verbose", action="store_true")
+
+
+if __name__ == "__main__":
+    args = parser.parse_args(sys.argv[1:])
+    logging.basicConfig(
+        format="%(asctime)s %(name)s:%(lineno)d %(levelname)s %(message)s",
+        level=logging.DEBUG if args.verbose else logging.INFO,
+    )
+    # suppress logs we don't care about
+    logging.getLogger("xmlschema").setLevel(logging.WARNING)
+    if args.mode == "lint":
+        sys.exit(lint())
+    elif args.mode == "test":
+        unittest.main(argv=sys.argv[:1])
@@ -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::"

@@ -106,4 +106,5 @@ isort "${files[@]}"
 python3 -m black "${files[@]}"
 ./scripts-dev/config-lint.sh
 flake8 "${files[@]}"
+./scripts-dev/check_pydantic_models.py lint
 mypy
@@ -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


@@ -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,
            )
@@ -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"
@@ -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}
@@ -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
@@ -441,6 +441,13 @@ def start(config_options: List[str]) -> None:
        "synapse.app.user_dir",
    )

+    if config.experimental.faster_joins_enabled:
+        raise ConfigError(
+            "You have enabled the experimental `faster_joins` config option, but it is "
+            "not compatible with worker deployments yet. Please disable `faster_joins` "
+            "or run Synapse as a single process deployment instead."
+        )
+
    synapse.events.USE_FROZEN_DICTS = config.server.use_frozen_dicts
    synapse.util.caches.TRACK_MEMORY_USAGE = config.caches.track_memory_usage

@@ -23,7 +23,7 @@ LEGACY_TEMPLATE_DIR_WARNING = """
 This server's configuration file is using the deprecated 'template_dir' setting in the
 'account_validity' section. Support for this setting has been deprecated and will be
 removed in a future version of Synapse. Server admins should instead use the new
-'custom_templates_directory' setting documented here:
+'custom_template_directory' setting documented here:
 https://matrix-org.github.io/synapse/latest/templates.html
 ---------------------------------------------------------------------------------------"""

@@ -53,7 +53,7 @@ LEGACY_TEMPLATE_DIR_WARNING = """
 This server's configuration file is using the deprecated 'template_dir' setting in the
 'email' section. Support for this setting has been deprecated and will be removed in a
 future version of Synapse. Server admins should instead use the new
-'custom_templates_directory' setting documented here:
+'custom_template_directory' setting documented here:
 https://matrix-org.github.io/synapse/latest/templates.html
 ---------------------------------------------------------------------------------------"""

@@ -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)
@@ -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", {})
        )

@@ -101,20 +101,20 @@ 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 = RateLimitConfig(
+        self.rc_joins_per_room = RatelimitSettings(
            config.get("rc_joins_per_room", {}),
            defaults={"per_second": 1, "burst_count": 10},
        )
@@ -124,31 +124,31 @@ class RatelimitConfig(Config):
        # * 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,
@@ -26,7 +26,7 @@ LEGACY_TEMPLATE_DIR_WARNING = """
 This server's configuration file is using the deprecated 'template_dir' setting in the
 'sso' section. Support for this setting has been deprecated and will be removed in a
 future version of Synapse. Server admins should instead use the new
-'custom_templates_directory' setting documented here:
+'custom_template_directory' setting documented here:
 https://matrix-org.github.io/synapse/latest/templates.html
 ---------------------------------------------------------------------------------------"""

@@ -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,
        )


@@ -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
@@ -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 SynapseTags, set_tag, tag_args, 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
@@ -233,6 +234,8 @@ class FederationClient(FederationBase):
            destination, content, timeout
        )

+    @trace
+    @tag_args
    async def backfill(
        self, dest: str, room_id: str, limit: int, extremities: Collection[str]
    ) -> Optional[List[EventBase]]:
@@ -335,6 +338,8 @@ class FederationClient(FederationBase):

        return None

+    @trace
+    @tag_args
    async def get_pdu(
        self,
        destinations: Iterable[str],
@@ -403,9 +408,9 @@ class FederationClient(FederationBase):
                        # Prime the cache
                        self._get_pdu_cache[event.event_id] = event

-                        # FIXME: We should add a `break` here to avoid calling every
-                        # destination after we already found a PDU (will follow-up
-                        # in a separate PR)
+                        # Now that we have an event, we can break out of this
+                        # loop and stop asking other destinations.
+                        break

                except SynapseError as e:
                    logger.info(
@@ -446,6 +451,8 @@ class FederationClient(FederationBase):

        return event_copy

+    @trace
+    @tag_args
    async def get_room_state_ids(
        self, destination: str, room_id: str, event_id: str
    ) -> Tuple[List[str], List[str]]:
@@ -465,6 +472,23 @@ class FederationClient(FederationBase):
        state_event_ids = result["pdu_ids"]
        auth_event_ids = result.get("auth_chain_ids", [])

+        set_tag(
+            SynapseTags.RESULT_PREFIX + "state_event_ids",
+            str(state_event_ids),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "state_event_ids.length",
+            str(len(state_event_ids)),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "auth_event_ids",
+            str(auth_event_ids),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "auth_event_ids.length",
+            str(len(auth_event_ids)),
+        )
+
        if not isinstance(state_event_ids, list) or not isinstance(
            auth_event_ids, list
        ):
@@ -472,6 +496,8 @@ class FederationClient(FederationBase):

        return state_event_ids, auth_event_ids

+    @trace
+    @tag_args
    async def get_room_state(
        self,
        destination: str,
@@ -531,6 +557,7 @@ class FederationClient(FederationBase):

        return valid_state_events, valid_auth_events

+    @trace
    async def _check_sigs_and_hash_and_fetch(
        self,
        origin: str,
@@ -725,6 +752,12 @@ 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
@@ -774,7 +807,7 @@ 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,
@@ -61,7 +61,12 @@ from synapse.logging.context import (
    nested_logging_context,
    run_in_background,
 )
-from synapse.logging.opentracing import log_kv, start_active_span_from_edu, trace
+from synapse.logging.opentracing import (
+    log_kv,
+    start_active_span_from_edu,
+    tag_args,
+    trace,
+)
 from synapse.metrics.background_process_metrics import wrap_as_background_process
 from synapse.replication.http.federation import (
    ReplicationFederationSendEduRestServlet,
@@ -469,7 +474,7 @@ 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]:
@@ -547,6 +552,8 @@ class FederationServer(FederationBase):

        return 200, resp

+    @trace
+    @tag_args
    async def on_state_ids_request(
        self, origin: str, room_id: str, event_id: str
    ) -> Tuple[int, JsonDict]:
@@ -569,6 +576,8 @@ class FederationServer(FederationBase):

        return 200, resp

+    @trace
+    @tag_args
    async def _on_state_ids_request_compute(
        self, room_id: str, event_id: str
    ) -> JsonDict:
@@ -843,8 +852,25 @@ 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,
@@ -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:
@@ -32,6 +32,7 @@ from typing import (
 )

 import attr
+from prometheus_client import Histogram
 from signedjson.key import decode_verify_key_bytes
 from signedjson.sign import verify_signed_json
 from unpaddedbase64 import decode_base64
@@ -59,6 +60,7 @@ 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 SynapseTags, set_tag, tag_args, 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 (
@@ -78,6 +80,24 @@ if TYPE_CHECKING:

 logger = logging.getLogger(__name__)

+# Added to debug performance and track progress on optimizations
+backfill_processing_before_timer = Histogram(
+    "synapse_federation_backfill_processing_before_time_seconds",
+    "sec",
+    [],
+    buckets=(
+        1.0,
+        5.0,
+        10.0,
+        20.0,
+        30.0,
+        40.0,
+        60.0,
+        80.0,
+        "+Inf",
+    ),
+)
+

 def get_domains_from_state(state: StateMap[EventBase]) -> List[Tuple[str, int]]:
    """Get joined domains from state
@@ -137,6 +157,7 @@ class FederationHandler:
    def __init__(self, hs: "HomeServer"):
        self.hs = hs

+        self.clock = hs.get_clock()
        self.store = hs.get_datastores().main
        self._storage_controllers = hs.get_storage_controllers()
        self._state_storage_controller = self._storage_controllers.state
@@ -180,6 +201,7 @@ 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:
@@ -195,12 +217,39 @@ class FederationHandler:
                return. This is used as part of the heuristic to decide if we
                should back paginate.
        """
+        # Starting the processing time here so we can include the room backfill
+        # linearizer lock queue in the timing
+        processing_start_time = self.clock.time_msec()
+
        async with self._room_backfill.queue(room_id):
-            return await self._maybe_backfill_inner(room_id, current_depth, limit)
+            return await self._maybe_backfill_inner(
+                room_id,
+                current_depth,
+                limit,
+                processing_start_time=processing_start_time,
+            )

    async def _maybe_backfill_inner(
-        self, room_id: str, current_depth: int, limit: int
+        self,
+        room_id: str,
+        current_depth: int,
+        limit: int,
+        *,
+        processing_start_time: int,
    ) -> bool:
+        """
+        Checks whether the `current_depth` is at or approaching any backfill
+        points in the room and if so, will backfill. We only care about
+        checking backfill points that happened before the `current_depth`
+        (meaning less than or equal to the `current_depth`).
+
+        Args:
+            room_id: The room to backfill in.
+            current_depth: The depth to check at for any upcoming backfill points.
+            limit: The max number of events to request from the remote federated server.
+            processing_start_time: The time when `maybe_backfill` started
+                processing. Only used for timing.
+        """
        backwards_extremities = [
            _BackfillPoint(event_id, depth, _BackfillPointType.BACKWARDS_EXTREMITY)
            for event_id, depth in await self.store.get_oldest_event_ids_with_depth_in_room(
@@ -368,6 +417,14 @@ class FederationHandler:
        logger.debug(
            "_maybe_backfill_inner: extremities_to_request %s", extremities_to_request
        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "extremities_to_request",
+            str(extremities_to_request),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "extremities_to_request.length",
+            str(len(extremities_to_request)),
+        )

        # Now we need to decide which hosts to hit first.

@@ -423,6 +480,11 @@ class FederationHandler:

            return False

+        processing_end_time = self.clock.time_msec()
+        backfill_processing_before_timer.observe(
+            (processing_start_time - processing_end_time) / 1000
+        )
+
        success = await try_backfill(likely_domains)
        if success:
            return True
@@ -546,9 +608,9 @@ 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 +636,21 @@ 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 +814,23 @@ 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
@@ -1058,6 +1141,8 @@ class FederationHandler:

        return event

+    @trace
+    @tag_args
    async def get_state_ids_for_pdu(self, room_id: str, event_id: str) -> List[str]:
        """Returns the state at the event. i.e. not including said event."""
        event = await self.store.get_event(event_id, check_room_id=room_id)
@@ -1539,15 +1624,16 @@ 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)
@@ -29,7 +29,7 @@ from typing import (
    Tuple,
 )

-from prometheus_client import Counter
+from prometheus_client import Counter, Histogram

 from synapse import event_auth
 from synapse.api.constants import (
@@ -59,6 +59,13 @@ 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 (
+    SynapseTags,
+    set_tag,
+    start_active_span,
+    tag_args,
+    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 (
@@ -91,6 +98,26 @@ soft_failed_event_counter = Counter(
    "Events received over federation that we marked as soft_failed",
 )

+# Added to debug performance and track progress on optimizations
+backfill_processing_after_timer = Histogram(
+    "synapse_federation_backfill_processing_after_time_seconds",
+    "sec",
+    [],
+    buckets=(
+        1.0,
+        5.0,
+        10.0,
+        20.0,
+        30.0,
+        40.0,
+        60.0,
+        80.0,
+        120.0,
+        180.0,
+        "+Inf",
+    ),
+)
+

 class FederationEventHandler:
    """Handles events that originated from federation.
@@ -278,7 +305,8 @@ 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 +314,8 @@ 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 +345,7 @@ 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 +406,7 @@ 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

@@ -406,6 +436,7 @@ class FederationEventHandler:
            prev_member_event,
        )

+    @trace
    async def process_remote_join(
        self,
        origin: str,
@@ -534,32 +565,36 @@ 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:
@@ -589,21 +624,23 @@ class FederationEventHandler:
        if not events:
            return

-        # if there are any events in the wrong room, the remote server is buggy and
-        # should not be trusted.
-        for ev in events:
-            if ev.room_id != room_id:
-                raise InvalidResponseError(
-                    f"Remote server {dest} returned event {ev.event_id} which is in "
-                    f"room {ev.room_id}, when we were backfilling in {room_id}"
-                )
+        with backfill_processing_after_timer.time():
+            # if there are any events in the wrong room, the remote server is buggy and
+            # should not be trusted.
+            for ev in events:
+                if ev.room_id != room_id:
+                    raise InvalidResponseError(
+                        f"Remote server {dest} returned event {ev.event_id} which is in "
+                        f"room {ev.room_id}, when we were backfilling in {room_id}"
+                    )

-        await self._process_pulled_events(
-            dest,
-            events,
-            backfilled=True,
-        )
+            await self._process_pulled_events(
+                dest,
+                events,
+                backfilled=True,
+            )

+    @trace
    async def _get_missing_events_for_pdu(
        self, origin: str, pdu: EventBase, prevs: Set[str], min_depth: int
    ) -> None:
@@ -704,8 +741,9 @@ 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
+        self, origin: str, events: Collection[EventBase], backfilled: bool
    ) -> None:
        """Process a batch of events we have pulled from a remote server

@@ -720,6 +758,15 @@ class FederationEventHandler:
            backfilled: True if this is part of a historical batch of events (inhibits
                notification to clients, and validation of device keys.)
        """
+        set_tag(
+            SynapseTags.FUNC_ARG_PREFIX + "event_ids",
+            str([event.event_id for event in events]),
+        )
+        set_tag(
+            SynapseTags.FUNC_ARG_PREFIX + "event_ids.length",
+            str(len(events)),
+        )
+        set_tag(SynapseTags.FUNC_ARG_PREFIX + "backfilled", str(backfilled))
        logger.debug(
            "processing pulled backfilled=%s events=%s",
            backfilled,
@@ -742,6 +789,8 @@ class FederationEventHandler:
            with nested_logging_context(ev.event_id):
                await self._process_pulled_event(origin, ev, backfilled=backfilled)

+    @trace
+    @tag_args
    async def _process_pulled_event(
        self, origin: str, event: EventBase, backfilled: bool
    ) -> None:
@@ -806,29 +855,56 @@ 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(
+    @trace
+    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`,
@@ -849,8 +925,7 @@ 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
@@ -864,7 +939,7 @@ 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 "
@@ -876,9 +951,15 @@ 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
@@ -924,8 +1005,12 @@ 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
+        )

+    @trace
+    @tag_args
    async def _get_state_ids_after_missing_prev_event(
        self,
        destination: str,
@@ -965,10 +1050,10 @@ class FederationEventHandler:
        logger.debug("Fetching %i events from cache/store", len(desired_events))
        have_events = await self._store.have_seen_events(room_id, desired_events)

-        missing_desired_events = desired_events - have_events
+        missing_desired_event_ids = desired_events - have_events
        logger.debug(
            "We are missing %i events (got %i)",
-            len(missing_desired_events),
+            len(missing_desired_event_ids),
            len(have_events),
        )

@@ -980,13 +1065,30 @@ class FederationEventHandler:
        #   already have a bunch of the state events. It would be nice if the
        #   federation api gave us a way of finding out which we actually need.

-        missing_auth_events = set(auth_event_ids) - have_events
-        missing_auth_events.difference_update(
-            await self._store.have_seen_events(room_id, missing_auth_events)
+        missing_auth_event_ids = set(auth_event_ids) - have_events
+        missing_auth_event_ids.difference_update(
+            await self._store.have_seen_events(room_id, missing_auth_event_ids)
        )
-        logger.debug("We are also missing %i auth events", len(missing_auth_events))
+        logger.debug("We are also missing %i auth events", len(missing_auth_event_ids))

-        missing_events = missing_desired_events | missing_auth_events
+        missing_event_ids = missing_desired_event_ids | missing_auth_event_ids
+
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "missing_auth_event_ids",
+            str(missing_auth_event_ids),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "missing_auth_event_ids.length",
+            str(len(missing_auth_event_ids)),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "missing_desired_event_ids",
+            str(missing_desired_event_ids),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "missing_desired_event_ids.length",
+            str(len(missing_desired_event_ids)),
+        )

        # Making an individual request for each of 1000s of events has a lot of
        # overhead. On the other hand, we don't really want to fetch all of the events
@@ -997,13 +1099,13 @@ class FederationEventHandler:
        #
        # TODO: might it be better to have an API which lets us do an aggregate event
        #   request
-        if (len(missing_events) * 10) >= len(auth_event_ids) + len(state_event_ids):
+        if (len(missing_event_ids) * 10) >= len(auth_event_ids) + len(state_event_ids):
            logger.debug("Requesting complete state from remote")
            await self._get_state_and_persist(destination, room_id, event_id)
        else:
-            logger.debug("Fetching %i events from remote", len(missing_events))
+            logger.debug("Fetching %i events from remote", len(missing_event_ids))
            await self._get_events_and_persist(
-                destination=destination, room_id=room_id, event_ids=missing_events
+                destination=destination, room_id=room_id, event_ids=missing_event_ids
            )

        # We now need to fill out the state map, which involves fetching the
@@ -1060,6 +1162,14 @@ class FederationEventHandler:
                event_id,
                failed_to_fetch,
            )
+            set_tag(
+                SynapseTags.RESULT_PREFIX + "failed_to_fetch",
+                str(failed_to_fetch),
+            )
+            set_tag(
+                SynapseTags.RESULT_PREFIX + "failed_to_fetch.length",
+                str(len(failed_to_fetch)),
+            )

        if remote_event.is_state() and remote_event.rejected_reason is None:
            state_map[
@@ -1068,6 +1178,8 @@ class FederationEventHandler:

        return state_map

+    @trace
+    @tag_args
    async def _get_state_and_persist(
        self, destination: str, room_id: str, event_id: str
    ) -> None:
@@ -1089,11 +1201,12 @@ class FederationEventHandler:
                destination=destination, room_id=room_id, event_ids=(event_id,)
            )

+    @trace
    async def _process_received_pdu(
        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.
@@ -1115,24 +1228,18 @@ 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:
@@ -1144,7 +1251,7 @@ 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)

@@ -1245,6 +1352,7 @@ class FederationEventHandler:
        except Exception:
            logger.exception("Failed to resync device for %s", sender)

+    @trace
    async def _handle_marker_event(self, origin: str, marker_event: EventBase) -> None:
        """Handles backfilling the insertion event when we receive a marker
        event that points to one.
@@ -1376,6 +1484,8 @@ class FederationEventHandler:

        return event_from_response

+    @trace
+    @tag_args
    async def _get_events_and_persist(
        self, destination: str, room_id: str, event_ids: Collection[str]
    ) -> None:
@@ -1421,6 +1531,7 @@ class FederationEventHandler:
        logger.info("Fetched %i events of %i requested", len(events), len(event_ids))
        await self._auth_and_persist_outliers(room_id, events)

+    @trace
    async def _auth_and_persist_outliers(
        self, room_id: str, events: Iterable[EventBase]
    ) -> None:
@@ -1439,6 +1550,16 @@ class FederationEventHandler:
        """
        event_map = {event.event_id: event for event in events}

+        event_ids = event_map.keys()
+        set_tag(
+            SynapseTags.FUNC_ARG_PREFIX + "event_ids",
+            str(event_ids),
+        )
+        set_tag(
+            SynapseTags.FUNC_ARG_PREFIX + "event_ids.length",
+            str(len(event_ids)),
+        )
+
        # filter out any events we have already seen. This might happen because
        # the events were eagerly pushed to us (eg, during a room join), or because
        # another thread has raced against us since we decided to request the event.
@@ -1555,14 +1676,17 @@ class FederationEventHandler:
            backfilled=True,
        )

+    @trace
    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.
@@ -1591,6 +1715,14 @@ class FederationEventHandler:
        claimed_auth_events = await self._load_or_fetch_auth_events_for_event(
            origin, event
        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "claimed_auth_events",
+            str([ev.event_id for ev in claimed_auth_events]),
+        )
+        set_tag(
+            SynapseTags.RESULT_PREFIX + "claimed_auth_events.length",
+            str(len(claimed_auth_events)),
+        )

        # ... and check that the event passes auth at those auth events.
        # https://spec.matrix.org/v1.3/server-server-api/#checks-performed-on-receipt-of-a-pdu:
@@ -1688,6 +1820,7 @@ class FederationEventHandler:
            )
            context.rejected = RejectedReason.AUTH_ERROR

+    @trace
    async def _maybe_kick_guest_users(self, event: EventBase) -> None:
        if event.type != EventTypes.GuestAccess:
            return
@@ -1705,7 +1838,7 @@ 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
@@ -1716,7 +1849,7 @@ class FederationEventHandler:

        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):
@@ -1742,11 +1875,15 @@ 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.
            #
@@ -1759,6 +1896,7 @@ 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(
@@ -1808,7 +1946,7 @@ 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

@@ -1824,12 +1962,19 @@ 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())
@@ -1841,6 +1986,13 @@ 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",
@@ -1876,6 +2028,8 @@ class FederationEventHandler:
        # instead we raise an AuthError, which will make the caller ignore it.
        raise AuthError(code=HTTPStatus.FORBIDDEN, msg="Auth events could not be found")

+    @trace
+    @tag_args
    async def _get_remote_auth_chain_for_event(
        self, destination: str, room_id: str, event_id: str
    ) -> None:
@@ -1904,6 +2058,7 @@ class FederationEventHandler:

        await self._auth_and_persist_outliers(room_id, remote_auth_events)

+    @trace
    async def _run_push_actions_and_persist_event(
        self, event: EventBase, context: EventContext, backfilled: bool = False
    ) -> None:
@@ -2012,8 +2167,17 @@ class FederationEventHandler:
                    self._message_handler.maybe_schedule_expiry(event)

            if not backfilled:  # Never notify for backfilled events
-                for event in events:
-                    await self._notify_persisted_event(event, max_stream_token)
+                with start_active_span("notify_persisted_events"):
+                    set_tag(
+                        SynapseTags.RESULT_PREFIX + "event_ids",
+                        str([ev.event_id for ev in events]),
+                    )
+                    set_tag(
+                        SynapseTags.RESULT_PREFIX + "event_ids.length",
+                        str(len(events)),
+                    )
+                    for event in events:
+                        await self._notify_persisted_event(event, max_stream_token)

            return max_stream_token.stream

@@ -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(
@@ -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

@@ -319,11 +325,17 @@ 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)
+        users_with_profile = (
+            await self._state_storage_controller.get_users_in_room_with_profiles(
+                room_id
+            )
+        )

        # If this is an AS, double check that they are allowed to see the members.
        # This can either be because the AS user is in the room or because there
@@ -334,7 +346,11 @@ 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: {
@@ -1135,6 +1151,10 @@ 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)
@@ -1359,9 +1379,10 @@ 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.
@@ -1448,9 +1469,10 @@ 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.
@@ -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,
@@ -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

@@ -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]:
@@ -182,7 +182,7 @@ class RoomListHandler:
                == HistoryVisibility.WORLD_READABLE,
                "guest_can_join": room["guest_access"] == "can_join",
                "join_rule": room["join_rules"],
-                "org.matrix.msc3827.room_type": room["room_type"],
+                "room_type": room["room_type"],
            }

            # Filter out Nones – rather omit the field altogether
@@ -32,6 +32,7 @@ from synapse.event_auth import get_named_level, get_power_level_event
 from synapse.events import EventBase
 from synapse.events.snapshot import EventContext
 from synapse.handlers.profile import MAX_AVATAR_URL_LEN, MAX_DISPLAYNAME_LEN
+from synapse.logging import opentracing
 from synapse.module_api import NOT_SPAM
 from synapse.storage.state import StateFilter
 from synapse.types import (
@@ -428,14 +429,14 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
                await self._join_rate_per_room_limiter.ratelimit(
                    requester, key=room_id, update=False
                )
-
-        result_event = await self.event_creation_handler.handle_new_client_event(
-            requester,
-            event,
-            context,
-            extra_users=[target],
-            ratelimit=ratelimit,
-        )
+        with opentracing.start_active_span("handle_new_client_event"):
+            result_event = await self.event_creation_handler.handle_new_client_event(
+                requester,
+                event,
+                context,
+                extra_users=[target],
+                ratelimit=ratelimit,
+            )

        if event.membership == Membership.LEAVE:
            if prev_member_event_id:
@@ -564,25 +565,26 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
        # by application services), and then by room ID.
        async with self.member_as_limiter.queue(as_id):
            async with self.member_linearizer.queue(key):
-                result = await self.update_membership_locked(
-                    requester,
-                    target,
-                    room_id,
-                    action,
-                    txn_id=txn_id,
-                    remote_room_hosts=remote_room_hosts,
-                    third_party_signed=third_party_signed,
-                    ratelimit=ratelimit,
-                    content=content,
-                    new_room=new_room,
-                    require_consent=require_consent,
-                    outlier=outlier,
-                    historical=historical,
-                    allow_no_prev_events=allow_no_prev_events,
-                    prev_event_ids=prev_event_ids,
-                    state_event_ids=state_event_ids,
-                    depth=depth,
-                )
+                with opentracing.start_active_span("update_membership_locked"):
+                    result = await self.update_membership_locked(
+                        requester,
+                        target,
+                        room_id,
+                        action,
+                        txn_id=txn_id,
+                        remote_room_hosts=remote_room_hosts,
+                        third_party_signed=third_party_signed,
+                        ratelimit=ratelimit,
+                        content=content,
+                        new_room=new_room,
+                        require_consent=require_consent,
+                        outlier=outlier,
+                        historical=historical,
+                        allow_no_prev_events=allow_no_prev_events,
+                        prev_event_ids=prev_event_ids,
+                        state_event_ids=state_event_ids,
+                        depth=depth,
+                    )

        return result

@@ -649,6 +651,7 @@ class RoomMemberHandler(metaclass=abc.ABCMeta):
        Returns:
            A tuple of the new event ID and stream ID.
        """
+
        content_specified = bool(content)
        if content is None:
            content = {}
@@ -1679,7 +1682,11 @@ class RoomMemberMasterHandler(RoomMemberHandler):
        ]

        if len(remote_room_hosts) == 0:
-            raise SynapseError(404, "No known servers")
+            raise SynapseError(
+                404,
+                "Can't join remote room because no servers "
+                "that are in the room have been provided.",
+            )

        check_complexity = self.hs.config.server.limit_remote_rooms.enabled
        if (
@@ -28,11 +28,11 @@ from synapse.api.constants import (
    RoomTypes,
 )
 from synapse.api.errors import (
-    AuthError,
    Codes,
    NotFoundError,
    StoreError,
    SynapseError,
+    UnstableSpecAuthError,
    UnsupportedRoomVersionError,
 )
 from synapse.api.ratelimiting import Ratelimiter
@@ -175,10 +175,11 @@ class RoomSummaryHandler:

        # First of all, check that the room is accessible.
        if not await self._is_local_room_accessible(requested_room_id, requester):
-            raise AuthError(
+            raise UnstableSpecAuthError(
                403,
                "User %s not in room %s, and room previews are disabled"
                % (requester, requested_room_id),
+                errcode=Codes.NOT_JOINED,
            )

        # If this is continuing a previous session, pull the persisted data.
@@ -13,7 +13,17 @@
 # limitations under the License.
 import itertools
 import logging
-from typing import TYPE_CHECKING, Any, Dict, FrozenSet, List, Optional, Set, Tuple
+from typing import (
+    TYPE_CHECKING,
+    Any,
+    Dict,
+    FrozenSet,
+    List,
+    Optional,
+    Sequence,
+    Set,
+    Tuple,
+)

 import attr
 from prometheus_client import Counter
@@ -89,7 +99,7 @@ class SyncConfig:
@attr.s(slots=True, frozen=True, auto_attribs=True)
 class TimelineBatch:
    prev_batch: StreamToken
-    events: List[EventBase]
+    events: Sequence[EventBase]
    limited: bool
    # A mapping of event ID to the bundled aggregations for the above events.
    # This is only calculated if limited is true.
@@ -852,16 +862,26 @@ class SyncHandler:
        now_token: StreamToken,
        full_state: bool,
    ) -> MutableStateMap[EventBase]:
-        """Works out the difference in state between the start of the timeline
-        and the previous sync.
+        """Works out the difference in state between the end of the previous sync and
+        the start of the timeline.

        Args:
            room_id:
            batch: The timeline batch for the room that will be sent to the user.
            sync_config:
-            since_token: Token of the end of the previous batch. May be None.
+            since_token: Token of the end of the previous batch. May be `None`.
            now_token: Token of the end of the current batch.
            full_state: Whether to force returning the full state.
+                `lazy_load_members` still applies when `full_state` is `True`.
+
+        Returns:
+            The state to return in the sync response for the room.
+
+            Clients will overlay this onto the state at the end of the previous sync to
+            arrive at the state at the start of the timeline.
+
+            Clients will then overlay state events in the timeline to arrive at the
+            state at the end of the timeline, in preparation for the next sync.
        """
        # TODO(mjark) Check if the state events were received by the server
        # after the previous sync, since we need to include those state
@@ -869,7 +889,8 @@ class SyncHandler:
        # TODO(mjark) Check for new redactions in the state events.

        with Measure(self.clock, "compute_state_delta"):
-
+            # The memberships needed for events in the timeline.
+            # Only calculated when `lazy_load_members` is on.
            members_to_fetch = None

            lazy_load_members = sync_config.filter_collection.lazy_load_members()
@@ -897,38 +918,46 @@ class SyncHandler:
            else:
                state_filter = StateFilter.all()

+            # The contribution to the room state from state events in the timeline.
+            # Only contains the last event for any given state key.
            timeline_state = {
                (event.type, event.state_key): event.event_id
                for event in batch.events
                if event.is_state()
            }

+            # Now calculate the state to return in the sync response for the room.
+            # This is more or less the change in state between the end of the previous
+            # sync's timeline and the start of the current sync's timeline.
+            # See the docstring above for details.
+            state_ids: StateMap[str]
+
            if full_state:
                if batch:
-                    current_state_ids = (
+                    state_at_timeline_end = (
                        await self._state_storage_controller.get_state_ids_for_event(
                            batch.events[-1].event_id, state_filter=state_filter
                        )
                    )

-                    state_ids = (
+                    state_at_timeline_start = (
                        await self._state_storage_controller.get_state_ids_for_event(
                            batch.events[0].event_id, state_filter=state_filter
                        )
                    )

                else:
-                    current_state_ids = await self.get_state_at(
+                    state_at_timeline_end = await self.get_state_at(
                        room_id, stream_position=now_token, state_filter=state_filter
                    )

-                    state_ids = current_state_ids
+                    state_at_timeline_start = state_at_timeline_end

                state_ids = _calculate_state(
                    timeline_contains=timeline_state,
-                    timeline_start=state_ids,
-                    previous={},
-                    current=current_state_ids,
+                    timeline_start=state_at_timeline_start,
+                    timeline_end=state_at_timeline_end,
+                    previous_timeline_end={},
                    lazy_load_members=lazy_load_members,
                )
            elif batch.limited:
@@ -968,24 +997,23 @@ class SyncHandler:
                )

                if batch:
-                    current_state_ids = (
+                    state_at_timeline_end = (
                        await self._state_storage_controller.get_state_ids_for_event(
                            batch.events[-1].event_id, state_filter=state_filter
                        )
                    )
                else:
-                    # Its not clear how we get here, but empirically we do
-                    # (#5407). Logging has been added elsewhere to try and
-                    # figure out where this state comes from.
-                    current_state_ids = await self.get_state_at(
+                    # We can get here if the user has ignored the senders of all
+                    # the recent events.
+                    state_at_timeline_end = await self.get_state_at(
                        room_id, stream_position=now_token, state_filter=state_filter
                    )

                state_ids = _calculate_state(
                    timeline_contains=timeline_state,
                    timeline_start=state_at_timeline_start,
-                    previous=state_at_previous_sync,
-                    current=current_state_ids,
+                    timeline_end=state_at_timeline_end,
+                    previous_timeline_end=state_at_previous_sync,
                    # we have to include LL members in case LL initial sync missed them
                    lazy_load_members=lazy_load_members,
                )
@@ -1010,6 +1038,13 @@ class SyncHandler:
                            ),
                        )

+            # At this point, if `lazy_load_members` is enabled, `state_ids` includes
+            # the memberships of all event senders in the timeline. This is because we
+            # may not have sent the memberships in a previous sync.
+
+            # When `include_redundant_members` is on, we send all the lazy-loaded
+            # memberships of event senders. Otherwise we make an effort to limit the set
+            # of memberships we send to those that we have not already sent to this client.
            if lazy_load_members and not include_redundant_members:
                cache_key = (sync_config.user.to_string(), sync_config.device_id)
                cache = self.get_lazy_loaded_members_cache(cache_key)
@@ -1536,15 +1571,13 @@ class SyncHandler:
        ignored_users = await self.store.ignored_users(user_id)
        if since_token:
            room_changes = await self._get_rooms_changed(
-                sync_result_builder, ignored_users, self.rooms_to_exclude
+                sync_result_builder, ignored_users
            )
            tags_by_room = await self.store.get_updated_tags(
                user_id, since_token.account_data_key
            )
        else:
-            room_changes = await self._get_all_rooms(
-                sync_result_builder, ignored_users, self.rooms_to_exclude
-            )
+            room_changes = await self._get_all_rooms(sync_result_builder, ignored_users)
            tags_by_room = await self.store.get_tags_for_user(user_id)

        log_kv({"rooms_changed": len(room_changes.room_entries)})
@@ -1623,13 +1656,14 @@ class SyncHandler:
        self,
        sync_result_builder: "SyncResultBuilder",
        ignored_users: FrozenSet[str],
-        excluded_rooms: List[str],
    ) -> _RoomChanges:
        """Determine the changes in rooms to report to the user.

        This function is a first pass at generating the rooms part of the sync response.
        It determines which rooms have changed during the sync period, and categorises
-        them into four buckets: "knock", "invite", "join" and "leave".
+        them into four buckets: "knock", "invite", "join" and "leave". It also excludes
+        from that list any room that appears in the list of rooms to exclude from sync
+        results in the server configuration.

        1. Finds all membership changes for the user in the sync period (from
           `since_token` up to `now_token`).
@@ -1655,7 +1689,7 @@ class SyncHandler:
        #       _have_rooms_changed. We could keep the results in memory to avoid a
        #       second query, at the cost of more complicated source code.
        membership_change_events = await self.store.get_membership_changes_for_user(
-            user_id, since_token.room_key, now_token.room_key, excluded_rooms
+            user_id, since_token.room_key, now_token.room_key, self.rooms_to_exclude
        )

        mem_change_events_by_room_id: Dict[str, List[EventBase]] = {}
@@ -1862,7 +1896,6 @@ class SyncHandler:
        self,
        sync_result_builder: "SyncResultBuilder",
        ignored_users: FrozenSet[str],
-        ignored_rooms: List[str],
    ) -> _RoomChanges:
        """Returns entries for all rooms for the user.

@@ -1884,7 +1917,7 @@ class SyncHandler:
        room_list = await self.store.get_rooms_for_local_user_where_membership_is(
            user_id=user_id,
            membership_list=Membership.LIST,
-            excluded_rooms=ignored_rooms,
+            excluded_rooms=self.rooms_to_exclude,
        )

        room_entries = []
@@ -2150,7 +2183,9 @@ class SyncHandler:
                raise Exception("Unrecognized rtype: %r", room_builder.rtype)

    async def get_rooms_for_user_at(
-        self, user_id: str, room_key: RoomStreamToken
+        self,
+        user_id: str,
+        room_key: RoomStreamToken,
    ) -> FrozenSet[str]:
        """Get set of joined rooms for a user at the given stream ordering.

@@ -2176,7 +2211,12 @@ class SyncHandler:
        # If the membership's stream ordering is after the given stream
        # ordering, we need to go and work out if the user was in the room
        # before.
+        # We also need to check whether the room should be excluded from sync
+        # responses as per the homeserver config.
        for joined_room in joined_rooms:
+            if joined_room.room_id in self.rooms_to_exclude:
+                continue
+
            if not joined_room.event_pos.persisted_after(room_key):
                joined_room_ids.add(joined_room.room_id)
                continue
@@ -2211,8 +2251,8 @@ def _action_has_highlight(actions: List[JsonDict]) -> bool:
 def _calculate_state(
    timeline_contains: StateMap[str],
    timeline_start: StateMap[str],
-    previous: StateMap[str],
-    current: StateMap[str],
+    timeline_end: StateMap[str],
+    previous_timeline_end: StateMap[str],
    lazy_load_members: bool,
 ) -> StateMap[str]:
    """Works out what state to include in a sync response.
@@ -2220,45 +2260,50 @@ def _calculate_state(
    Args:
        timeline_contains: state in the timeline
        timeline_start: state at the start of the timeline
-        previous: state at the end of the previous sync (or empty dict
+        timeline_end: state at the end of the timeline
+        previous_timeline_end: state at the end of the previous sync (or empty dict
            if this is an initial sync)
-        current: state at the end of the timeline
        lazy_load_members: whether to return members from timeline_start
            or not.  assumes that timeline_start has already been filtered to
            include only the members the client needs to know about.
    """
-    event_id_to_key = {
-        e: key
-        for key, e in itertools.chain(
+    event_id_to_state_key = {
+        event_id: state_key
+        for state_key, event_id in itertools.chain(
            timeline_contains.items(),
-            previous.items(),
            timeline_start.items(),
-            current.items(),
+            timeline_end.items(),
+            previous_timeline_end.items(),
        )
    }

-    c_ids = set(current.values())
-    ts_ids = set(timeline_start.values())
-    p_ids = set(previous.values())
-    tc_ids = set(timeline_contains.values())
+    timeline_end_ids = set(timeline_end.values())
+    timeline_start_ids = set(timeline_start.values())
+    previous_timeline_end_ids = set(previous_timeline_end.values())
+    timeline_contains_ids = set(timeline_contains.values())

    # If we are lazyloading room members, we explicitly add the membership events
    # for the senders in the timeline into the state block returned by /sync,
    # as we may not have sent them to the client before.  We find these membership
    # events by filtering them out of timeline_start, which has already been filtered
    # to only include membership events for the senders in the timeline.
-    # In practice, we can do this by removing them from the p_ids list,
-    # which is the list of relevant state we know we have already sent to the client.
+    # In practice, we can do this by removing them from the previous_timeline_end_ids
+    # list, which is the list of relevant state we know we have already sent to the
+    # client.
    # see https://github.com/matrix-org/synapse/pull/2970/files/efcdacad7d1b7f52f879179701c7e0d9b763511f#r204732809

    if lazy_load_members:
-        p_ids.difference_update(
+        previous_timeline_end_ids.difference_update(
            e for t, e in timeline_start.items() if t[0] == EventTypes.Member
        )

-    state_ids = ((c_ids | ts_ids) - p_ids) - tc_ids
+    state_ids = (
+        (timeline_end_ids | timeline_start_ids)
+        - previous_timeline_end_ids
+        - timeline_contains_ids
+    )

-    return {event_id_to_key[e]: e for e in state_ids}
+    return {event_id_to_state_key[e]: e for e in state_ids}


@attr.s(slots=True, auto_attribs=True)
@@ -489,8 +489,15 @@ class TypingNotificationEventSource(EventSource[int, JsonDict]):
            handler = self.get_typing_handler()

            events = []
-            for room_id in handler._room_serials.keys():
-                if handler._room_serials[room_id] <= from_key:
+
+            # Work on a copy of things here as these may change in the handler while
+            # waiting for the AS `is_interested_in_room` call to complete.
+            # Shallow copy is safe as no nested data is present.
+            latest_room_serial = handler._latest_room_serial
+            room_serials = handler._room_serials.copy()
+
+            for room_id, serial in room_serials.items():
+                if serial <= from_key:
                    continue

                if not await service.is_interested_in_room(room_id, self._main_store):
@@ -498,7 +505,7 @@ class TypingNotificationEventSource(EventSource[int, JsonDict]):

                events.append(self._make_event_for(room_id))

-            return events, handler._latest_room_serial
+            return events, latest_room_serial

    async def get_new_events(
        self,
@@ -58,6 +58,7 @@ from synapse.api.errors import (
    SynapseError,
    UnrecognizedRequestError,
 )
+from synapse.config.homeserver import HomeServerConfig
 from synapse.http.site import SynapseRequest
 from synapse.logging.context import defer_to_thread, preserve_fn, run_in_background
 from synapse.logging.opentracing import active_span, start_active_span, trace_servlet
@@ -155,15 +156,16 @@ def is_method_cancellable(method: Callable[..., Any]) -> bool:
    return getattr(method, "cancellable", False)


-def return_json_error(f: failure.Failure, request: SynapseRequest) -> None:
+def return_json_error(
+    f: failure.Failure, request: SynapseRequest, config: Optional[HomeServerConfig]
+) -> None:
    """Sends a JSON error response to clients."""

    if f.check(SynapseError):
        # mypy doesn't understand that f.check asserts the type.
        exc: SynapseError = f.value  # type: ignore
        error_code = exc.code
-        error_dict = exc.error_dict()
-
+        error_dict = exc.error_dict(config)
        logger.info("%s SynapseError: %s - %s", request, error_code, exc.msg)
    elif f.check(CancelledError):
        error_code = HTTP_STATUS_REQUEST_CANCELLED
@@ -450,7 +452,7 @@ class DirectServeJsonResource(_AsyncResource):
        request: SynapseRequest,
    ) -> None:
        """Implements _AsyncResource._send_error_response"""
-        return_json_error(f, request)
+        return_json_error(f, request, None)


@attr.s(slots=True, frozen=True, auto_attribs=True)
@@ -575,6 +577,14 @@ class JsonResource(DirectServeJsonResource):

        return callback_return

+    def _send_error_response(
+        self,
+        f: failure.Failure,
+        request: SynapseRequest,
+    ) -> None:
+        """Implements _AsyncResource._send_error_response"""
+        return_json_error(f, request, self.hs.config)
+

 class DirectServeHtmlResource(_AsyncResource):
    """A resource that will call `self._async_on_<METHOD>` on new requests,
@@ -23,9 +23,12 @@ from typing import (
    Optional,
    Sequence,
    Tuple,
+    Type,
+    TypeVar,
    overload,
 )

+from pydantic import BaseModel, ValidationError
 from typing_extensions import Literal

 from twisted.web.server import Request
@@ -694,6 +697,28 @@ def parse_json_object_from_request(
    return content


+Model = TypeVar("Model", bound=BaseModel)
+
+
+def parse_and_validate_json_object_from_request(
+    request: Request, model_type: Type[Model]
+) -> Model:
+    """Parse a JSON object from the body of a twisted HTTP request, then deserialise and
+    validate using the given pydantic model.
+
+    Raises:
+        SynapseError if the request body couldn't be decoded as JSON or
+            if it wasn't a JSON object.
+    """
+    content = parse_json_object_from_request(request, allow_empty_body=False)
+    try:
+        instance = model_type.parse_obj(content)
+    except ValidationError as e:
+        raise SynapseError(HTTPStatus.BAD_REQUEST, str(e), errcode=Codes.BAD_JSON)
+
+    return instance
+
+
 def assert_params_in_dict(body: JsonDict, required: Iterable[str]) -> None:
    absent = []
    for k in required:
@@ -173,6 +173,7 @@ from typing import (
    Any,
    Callable,
    Collection,
+    ContextManager,
    Dict,
    Generator,
    Iterable,
@@ -309,6 +310,19 @@ class SynapseTags:
    # The name of the external cache
    CACHE_NAME = "cache.name"

+    # Used to tag function arguments
+    #
+    # Tag a named arg. The name of the argument should be appended to this prefix.
+    FUNC_ARG_PREFIX = "ARG."
+    # Tag extra variadic number of positional arguments (`def foo(first, second, *extras)`)
+    FUNC_ARGS = "args"
+    # Tag keyword args
+    FUNC_KWARGS = "kwargs"
+
+    # Some intermediate result that's interesting to the function. The label for
+    # the result should be appended to this prefix.
+    RESULT_PREFIX = "RESULT."
+

 class SynapseBaggage:
    FORCE_TRACING = "synapse-force-tracing"
@@ -823,75 +837,117 @@ def extract_text_map(carrier: Dict[str, str]) -> Optional["opentracing.SpanConte
 # Tracing decorators


-def trace_with_opname(opname: str) -> Callable[[Callable[P, R]], Callable[P, R]]:
+def _custom_sync_async_decorator(
+    func: Callable[P, R],
+    wrapping_logic: Callable[[Callable[P, R], Any, Any], ContextManager[None]],
+) -> Callable[P, R]:
+    """
+    Decorates a function that is sync or async (coroutines), or that returns a Twisted
+    `Deferred`. The custom business logic of the decorator goes in `wrapping_logic`.
+
+    Example usage:
+    ```py
+    # Decorator to time the function and log it out
+    def duration(func: Callable[P, R]) -> Callable[P, R]:
+        @contextlib.contextmanager
+        def _wrapping_logic(func: Callable[P, R], *args: P.args, **kwargs: P.kwargs) -> Generator[None, None, None]:
+            start_ts = time.time()
+            try:
+                yield
+            finally:
+                end_ts = time.time()
+                duration = end_ts - start_ts
+                logger.info("%s took %s seconds", func.__name__, duration)
+        return _custom_sync_async_decorator(func, _wrapping_logic)
+    ```
+
+    Args:
+        func: The function to be decorated
+        wrapping_logic: The business logic of your custom decorator.
+            This should be a ContextManager so you are able to run your logic
+            before/after the function as desired.
+    """
+
+    if inspect.iscoroutinefunction(func):
+
+        @wraps(func)
+        async def _wrapper(*args: P.args, **kwargs: P.kwargs) -> R:
+            with wrapping_logic(func, *args, **kwargs):
+                return await func(*args, **kwargs)  # type: ignore[misc]
+
+    else:
+        # The other case here handles both sync functions and those
+        # decorated with inlineDeferred.
+        @wraps(func)
+        def _wrapper(*args: P.args, **kwargs: P.kwargs) -> R:
+            scope = wrapping_logic(func, *args, **kwargs)
+            scope.__enter__()
+
+            try:
+                result = func(*args, **kwargs)
+                if isinstance(result, defer.Deferred):
+
+                    def call_back(result: R) -> R:
+                        scope.__exit__(None, None, None)
+                        return result
+
+                    def err_back(result: R) -> R:
+                        scope.__exit__(None, None, None)
+                        return result
+
+                    result.addCallbacks(call_back, err_back)
+
+                else:
+                    if inspect.isawaitable(result):
+                        logger.error(
+                            "@trace may not have wrapped %s correctly! "
+                            "The function is not async but returned a %s.",
+                            func.__qualname__,
+                            type(result).__name__,
+                        )
+
+                    scope.__exit__(None, None, None)
+
+                return result
+
+            except Exception as e:
+                scope.__exit__(type(e), None, e.__traceback__)
+                raise
+
+    return _wrapper  # type: ignore[return-value]
+
+
+def trace_with_opname(
+    opname: str,
+    *,
+    tracer: Optional["opentracing.Tracer"] = None,
+) -> Callable[[Callable[P, R]], Callable[P, R]]:
    """
    Decorator to trace a function with a custom opname.
-
    See the module's doc string for usage examples.
-
    """

-    def decorator(func: Callable[P, R]) -> Callable[P, R]:
-        if opentracing is None:
-            return func  # type: ignore[unreachable]
+    # type-ignore: mypy bug, see https://github.com/python/mypy/issues/12909
+    @contextlib.contextmanager  # type: ignore[arg-type]
+    def _wrapping_logic(
+        func: Callable[P, R], *args: P.args, **kwargs: P.kwargs
+    ) -> Generator[None, None, None]:
+        with start_active_span(opname, tracer=tracer):
+            yield

-        if inspect.iscoroutinefunction(func):
+    def _decorator(func: Callable[P, R]) -> Callable[P, R]:
+        if not opentracing:
+            return func

-            @wraps(func)
-            async def _trace_inner(*args: P.args, **kwargs: P.kwargs) -> R:
-                with start_active_span(opname):
-                    return await func(*args, **kwargs)  # type: ignore[misc]
+        return _custom_sync_async_decorator(func, _wrapping_logic)

-        else:
-            # The other case here handles both sync functions and those
-            # decorated with inlineDeferred.
-            @wraps(func)
-            def _trace_inner(*args: P.args, **kwargs: P.kwargs) -> R:
-                scope = start_active_span(opname)
-                scope.__enter__()
-
-                try:
-                    result = func(*args, **kwargs)
-                    if isinstance(result, defer.Deferred):
-
-                        def call_back(result: R) -> R:
-                            scope.__exit__(None, None, None)
-                            return result
-
-                        def err_back(result: R) -> R:
-                            scope.__exit__(None, None, None)
-                            return result
-
-                        result.addCallbacks(call_back, err_back)
-
-                    else:
-                        if inspect.isawaitable(result):
-                            logger.error(
-                                "@trace may not have wrapped %s correctly! "
-                                "The function is not async but returned a %s.",
-                                func.__qualname__,
-                                type(result).__name__,
-                            )
-
-                        scope.__exit__(None, None, None)
-
-                    return result
-
-                except Exception as e:
-                    scope.__exit__(type(e), None, e.__traceback__)
-                    raise
-
-        return _trace_inner  # type: ignore[return-value]
-
-    return decorator
+    return _decorator


 def trace(func: Callable[P, R]) -> Callable[P, R]:
    """
    Decorator to trace a function.
-
    Sets the operation name to that of the function's name.
-
    See the module's doc string for usage examples.
    """

@@ -900,22 +956,36 @@ def trace(func: Callable[P, R]) -> Callable[P, R]:

 def tag_args(func: Callable[P, R]) -> Callable[P, R]:
    """
-    Tags all of the args to the active span.
+    Decorator to tag all of the args to the active span.
+
+    Args:
+        func: `func` is assumed to be a method taking a `self` parameter, or a
+            `classmethod` taking a `cls` parameter. In either case, a tag is not
+            created for this parameter.
    """

    if not opentracing:
        return func

-    @wraps(func)
-    def _tag_args_inner(*args: P.args, **kwargs: P.kwargs) -> R:
+    # type-ignore: mypy bug, see https://github.com/python/mypy/issues/12909
+    @contextlib.contextmanager  # type: ignore[arg-type]
+    def _wrapping_logic(
+        func: Callable[P, R], *args: P.args, **kwargs: P.kwargs
+    ) -> Generator[None, None, None]:
        argspec = inspect.getfullargspec(func)
-        for i, arg in enumerate(argspec.args[1:]):
-            set_tag("ARG_" + arg, str(args[i]))  # type: ignore[index]
-        set_tag("args", str(args[len(argspec.args) :]))  # type: ignore[index]
-        set_tag("kwargs", str(kwargs))
-        return func(*args, **kwargs)
+        # We use `[1:]` to skip the `self` object reference and `start=1` to
+        # make the index line up with `argspec.args`.
+        #
+        # FIXME: We could update this to handle any type of function by ignoring the
+        #   first argument only if it's named `self` or `cls`. This isn't fool-proof
+        #   but handles the idiomatic cases.
+        for i, arg in enumerate(args[1:], start=1):  # type: ignore[index]
+            set_tag(SynapseTags.FUNC_ARG_PREFIX + argspec.args[i], str(arg))
+        set_tag(SynapseTags.FUNC_ARGS, str(args[len(argspec.args) :]))  # type: ignore[index]
+        set_tag(SynapseTags.FUNC_KWARGS, str(kwargs))
+        yield

-    return _tag_args_inner
+    return _custom_sync_async_decorator(func, _wrapping_logic)


@contextlib.contextmanager
@@ -156,7 +156,7 @@ class _LogContextScope(Scope):
    def close(self) -> None:
        active_scope = self.manager.active
        if active_scope is not self:
-            logger.error(
+            logger.warning(
                "Closing scope %s which is not the currently-active one %s",
                self,
                active_scope,
@@ -929,10 +929,12 @@ class ModuleApi:
        room_id: str,
        new_membership: str,
        content: Optional[JsonDict] = None,
+        remote_room_hosts: Optional[List[str]] = None,
    ) -> EventBase:
        """Updates the membership of a user to the given value.

        Added in Synapse v1.46.0.
+        Changed in Synapse v1.65.0: Added the 'remote_room_hosts' parameter.

        Args:
            sender: The user performing the membership change. Must be a user local to
@@ -946,6 +948,7 @@ class ModuleApi:
                https://spec.matrix.org/unstable/client-server-api/#mroommember for the
                list of allowed values.
            content: Additional values to include in the resulting event's content.
+            remote_room_hosts: Remote servers to use for remote joins/knocks/etc.

        Returns:
            The newly created membership event.
@@ -1005,15 +1008,12 @@ class ModuleApi:
            room_id=room_id,
            action=new_membership,
            content=content,
+            remote_room_hosts=remote_room_hosts,
        )

        # Try to retrieve the resulting event.
        event = await self._hs.get_datastores().main.get_event(event_id)

-        # update_membership is supposed to always return after the event has been
-        # successfully persisted.
-        assert event is not None
-
        return event

    async def create_and_send_event_into_room(self, event_dict: JsonDict) -> EventBase:
@@ -1452,6 +1452,81 @@ class ModuleApi:
            start_timestamp, end_timestamp
        )

+    async def lookup_room_alias(self, room_alias: str) -> Tuple[str, List[str]]:
+        """
+        Get the room ID associated with a room alias.
+
+        Added in Synapse v1.65.0.
+
+        Args:
+            room_alias: The alias to look up.
+
+        Returns:
+            A tuple of:
+                The room ID (str).
+                Hosts likely to be participating in the room ([str]).
+
+        Raises:
+            SynapseError if room alias is invalid or could not be found.
+        """
+        alias = RoomAlias.from_string(room_alias)
+        (room_id, hosts) = await self._hs.get_room_member_handler().lookup_room_alias(
+            alias
+        )
+
+        return room_id.to_string(), hosts
+
+    async def create_room(
+        self,
+        user_id: str,
+        config: JsonDict,
+        ratelimit: bool = True,
+        creator_join_profile: Optional[JsonDict] = None,
+    ) -> Tuple[str, Optional[str]]:
+        """Creates a new room.
+
+        Added in Synapse v1.65.0.
+
+        Args:
+            user_id:
+                The user who requested the room creation.
+            config : A dict of configuration options. See "Request body" of:
+                https://spec.matrix.org/latest/client-server-api/#post_matrixclientv3createroom
+            ratelimit: set to False to disable the rate limiter for this specific operation.
+
+            creator_join_profile:
+                Set to override the displayname and avatar for the creating
+                user in this room. If unset, displayname and avatar will be
+                derived from the user's profile. If set, should contain the
+                values to go in the body of the 'join' event (typically
+                `avatar_url` and/or `displayname`.
+
+        Returns:
+                A tuple containing: 1) the room ID (str), 2) if an alias was requested,
+                the room alias (str), otherwise None if no alias was requested.
+
+        Raises:
+            ResourceLimitError if server is blocked to some resource being
+            exceeded.
+            RuntimeError if the user_id does not refer to a local user.
+            SynapseError if the user_id is invalid, room ID couldn't be stored, or
+            something went horribly wrong.
+        """
+        if not self.is_mine(user_id):
+            raise RuntimeError(
+                "Tried to create a room as a user that isn't local to this homeserver",
+            )
+
+        requester = create_requester(user_id)
+        room_id_and_alias, _ = await self._hs.get_room_creation_handler().create_room(
+            requester=requester,
+            config=config,
+            ratelimit=ratelimit,
+            creator_join_profile=creator_join_profile,
+        )
+
+        return room_id_and_alias["room_id"], room_id_and_alias.get("room_alias", None)
+

 class PublicRoomListManager:
    """Contains methods for adding to, removing from and querying whether a room
@@ -14,128 +14,235 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.

-import copy
-from typing import Any, Dict, List
+"""
+Push rules is the system used to determine which events trigger a push (and a
+bump in notification counts).

-from synapse.push.rulekinds import PRIORITY_CLASS_INVERSE_MAP, PRIORITY_CLASS_MAP
+This consists of a list of "push rules" for each user, where a push rule is a
+pair of "conditions" and "actions". When a user receives an event Synapse
+iterates over the list of push rules until it finds one where all the conditions
+match the event, at which point "actions" describe the outcome (e.g. notify,
+highlight, etc).
+
+Push rules are split up into 5 different "kinds" (aka "priority classes"), which
+are run in order:
+    1. Override — highest priority rules, e.g. always ignore notices
+    2. Content — content specific rules, e.g. @ notifications
+    3. Room — per room rules, e.g. enable/disable notifications for all messages
+       in a room
+    4. Sender — per sender rules, e.g. never notify for messages from a given
+       user
+    5. Underride — the lowest priority "default" rules, e.g. notify for every
+       message.
+
+The set of "base rules" are the list of rules that every user has by default. A
+user can modify their copy of the push rules in one of three ways:
+
+    1. Adding a new push rule of a certain kind
+    2. Changing the actions of a base rule
+    3. Enabling/disabling a base rule.
+
+The base rules are split into whether they come before or after a particular
+kind, so the order of push rule evaluation would be: base rules for before
+"override" kind, user defined "override" rules, base rules after "override"
+kind, etc, etc.
+"""
+
+import itertools
+import logging
+from typing import Dict, Iterator, List, Mapping, Sequence, Tuple, Union
+
+import attr
+
+from synapse.config.experimental import ExperimentalConfig
+from synapse.push.rulekinds import PRIORITY_CLASS_MAP
+
+logger = logging.getLogger(__name__)


-def list_with_base_rules(rawrules: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
-    """Combine the list of rules set by the user with the default push rules
+@attr.s(auto_attribs=True, slots=True, frozen=True)
+class PushRule:
+    """A push rule

-    Args:
-        rawrules: The rules the user has modified or set.
-
-    Returns:
-        A new list with the rules set by the user combined with the defaults.
+    Attributes:
+        rule_id: a unique ID for this rule
+        priority_class: what "kind" of push rule this is (see
+            `PRIORITY_CLASS_MAP` for mapping between int and kind)
+        conditions: the sequence of conditions that all need to match
+        actions: the actions to apply if all conditions are met
+        default: is this a base rule?
+        default_enabled: is this enabled by default?
    """
-    ruleslist = []

-    # Grab the base rules that the user has modified.
-    # The modified base rules have a priority_class of -1.
-    modified_base_rules = {r["rule_id"]: r for r in rawrules if r["priority_class"] < 0}
+    rule_id: str
+    priority_class: int
+    conditions: Sequence[Mapping[str, str]]
+    actions: Sequence[Union[str, Mapping]]
+    default: bool = False
+    default_enabled: bool = True

-    # Remove the modified base rules from the list, They'll be added back
-    # in the default positions in the list.
-    rawrules = [r for r in rawrules if r["priority_class"] >= 0]

-    # shove the server default rules for each kind onto the end of each
-    current_prio_class = list(PRIORITY_CLASS_INVERSE_MAP)[-1]
+@attr.s(auto_attribs=True, slots=True, frozen=True, weakref_slot=False)
+class PushRules:
+    """A collection of push rules for an account.

-    ruleslist.extend(
-        make_base_prepend_rules(
-            PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules
+    Can be iterated over, producing push rules in priority order.
+    """
+
+    # A mapping from rule ID to push rule that overrides a base rule. These will
+    # be returned instead of the base rule.
+    overriden_base_rules: Dict[str, PushRule] = attr.Factory(dict)
+
+    # The following stores the custom push rules at each priority class.
+    #
+    # We keep these separate (rather than combining into one big list) to avoid
+    # copying the base rules around all the time.
+    override: List[PushRule] = attr.Factory(list)
+    content: List[PushRule] = attr.Factory(list)
+    room: List[PushRule] = attr.Factory(list)
+    sender: List[PushRule] = attr.Factory(list)
+    underride: List[PushRule] = attr.Factory(list)
+
+    def __iter__(self) -> Iterator[PushRule]:
+        # When iterating over the push rules we need to return the base rules
+        # interspersed at the correct spots.
+        for rule in itertools.chain(
+            BASE_PREPEND_OVERRIDE_RULES,
+            self.override,
+            BASE_APPEND_OVERRIDE_RULES,
+            self.content,
+            BASE_APPEND_CONTENT_RULES,
+            self.room,
+            self.sender,
+            self.underride,
+            BASE_APPEND_UNDERRIDE_RULES,
+        ):
+            # Check if a base rule has been overriden by a custom rule. If so
+            # return that instead.
+            override_rule = self.overriden_base_rules.get(rule.rule_id)
+            if override_rule:
+                yield override_rule
+            else:
+                yield rule
+
+    def __len__(self) -> int:
+        # The length is mostly used by caches to get a sense of "size" / amount
+        # of memory this object is using, so we only count the number of custom
+        # rules.
+        return (
+            len(self.overriden_base_rules)
+            + len(self.override)
+            + len(self.content)
+            + len(self.room)
+            + len(self.sender)
+            + len(self.underride)
        )
-    )

-    for r in rawrules:
-        if r["priority_class"] < current_prio_class:
-            while r["priority_class"] < current_prio_class:
-                ruleslist.extend(
-                    make_base_append_rules(
-                        PRIORITY_CLASS_INVERSE_MAP[current_prio_class],
-                        modified_base_rules,
-                    )
-                )
-                current_prio_class -= 1
-                if current_prio_class > 0:
-                    ruleslist.extend(
-                        make_base_prepend_rules(
-                            PRIORITY_CLASS_INVERSE_MAP[current_prio_class],
-                            modified_base_rules,
-                        )
-                    )

-        ruleslist.append(r)
+@attr.s(auto_attribs=True, slots=True, frozen=True, weakref_slot=False)
+class FilteredPushRules:
+    """A wrapper around `PushRules` that filters out disabled experimental push
+    rules, and includes the "enabled" state for each rule when iterated over.
+    """

-    while current_prio_class > 0:
-        ruleslist.extend(
-            make_base_append_rules(
-                PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules
+    push_rules: PushRules
+    enabled_map: Dict[str, bool]
+    experimental_config: ExperimentalConfig
+
+    def __iter__(self) -> Iterator[Tuple[PushRule, bool]]:
+        for rule in self.push_rules:
+            if not _is_experimental_rule_enabled(
+                rule.rule_id, self.experimental_config
+            ):
+                continue
+
+            enabled = self.enabled_map.get(rule.rule_id, rule.default_enabled)
+
+            yield rule, enabled
+
+    def __len__(self) -> int:
+        return len(self.push_rules)
+
+
+DEFAULT_EMPTY_PUSH_RULES = PushRules()
+
+
+def compile_push_rules(rawrules: List[PushRule]) -> PushRules:
+    """Given a set of custom push rules return a `PushRules` instance (which
+    includes the base rules).
+    """
+
+    if not rawrules:
+        # Fast path to avoid allocating empty lists when there are no custom
+        # rules for the user.
+        return DEFAULT_EMPTY_PUSH_RULES
+
+    rules = PushRules()
+
+    for rule in rawrules:
+        # We need to decide which bucket each custom push rule goes into.
+
+        # If it has the same ID as a base rule then it overrides that...
+        overriden_base_rule = BASE_RULES_BY_ID.get(rule.rule_id)
+        if overriden_base_rule:
+            rules.overriden_base_rules[rule.rule_id] = attr.evolve(
+                overriden_base_rule, actions=rule.actions
            )
-        )
-        current_prio_class -= 1
-        if current_prio_class > 0:
-            ruleslist.extend(
-                make_base_prepend_rules(
-                    PRIORITY_CLASS_INVERSE_MAP[current_prio_class], modified_base_rules
-                )
+            continue
+
+        # ... otherwise it gets added to the appropriate priority class bucket
+        collection: List[PushRule]
+        if rule.priority_class == 5:
+            collection = rules.override
+        elif rule.priority_class == 4:
+            collection = rules.content
+        elif rule.priority_class == 3:
+            collection = rules.room
+        elif rule.priority_class == 2:
+            collection = rules.sender
+        elif rule.priority_class == 1:
+            collection = rules.underride
+        elif rule.priority_class <= 0:
+            logger.info(
+                "Got rule with priority class less than zero, but doesn't override a base rule: %s",
+                rule,
            )
+            continue
+        else:
+            # We log and continue here so as not to break event sending
+            logger.error("Unknown priority class: %", rule.priority_class)
+            continue

-    return ruleslist
-
-
-def make_base_append_rules(
-    kind: str, modified_base_rules: Dict[str, Dict[str, Any]]
-) -> List[Dict[str, Any]]:
-    rules = []
-
-    if kind == "override":
-        rules = BASE_APPEND_OVERRIDE_RULES
-    elif kind == "underride":
-        rules = BASE_APPEND_UNDERRIDE_RULES
-    elif kind == "content":
-        rules = BASE_APPEND_CONTENT_RULES
-
-    # Copy the rules before modifying them
-    rules = copy.deepcopy(rules)
-    for r in rules:
-        # Only modify the actions, keep the conditions the same.
-        assert isinstance(r["rule_id"], str)
-        modified = modified_base_rules.get(r["rule_id"])
-        if modified:
-            r["actions"] = modified["actions"]
+        collection.append(rule)

    return rules


-def make_base_prepend_rules(
-    kind: str,
-    modified_base_rules: Dict[str, Dict[str, Any]],
-) -> List[Dict[str, Any]]:
-    rules = []
-
-    if kind == "override":
-        rules = BASE_PREPEND_OVERRIDE_RULES
-
-    # Copy the rules before modifying them
-    rules = copy.deepcopy(rules)
-    for r in rules:
-        # Only modify the actions, keep the conditions the same.
-        assert isinstance(r["rule_id"], str)
-        modified = modified_base_rules.get(r["rule_id"])
-        if modified:
-            r["actions"] = modified["actions"]
-
-    return rules
+def _is_experimental_rule_enabled(
+    rule_id: str, experimental_config: ExperimentalConfig
+) -> bool:
+    """Used by `FilteredPushRules` to filter out experimental rules when they
+    have not been enabled.
+    """
+    if (
+        rule_id == "global/override/.org.matrix.msc3786.rule.room.server_acl"
+        and not experimental_config.msc3786_enabled
+    ):
+        return False
+    if (
+        rule_id == "global/underride/.org.matrix.msc3772.thread_reply"
+        and not experimental_config.msc3772_enabled
+    ):
+        return False
+    return True


-# We have to annotate these types, otherwise mypy infers them as
-# `List[Dict[str, Sequence[Collection[str]]]]`.
-BASE_APPEND_CONTENT_RULES: List[Dict[str, Any]] = [
-    {
-        "rule_id": "global/content/.m.rule.contains_user_name",
-        "conditions": [
+BASE_APPEND_CONTENT_RULES = [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["content"],
+        rule_id="global/content/.m.rule.contains_user_name",
+        conditions=[
            {
                "kind": "event_match",
                "key": "content.body",
@@ -143,29 +250,33 @@ BASE_APPEND_CONTENT_RULES: List[Dict[str, Any]] = [
                "pattern_type": "user_localpart",
            }
        ],
-        "actions": [
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "default"},
            {"set_tweak": "highlight"},
        ],
-    }
+    )
 ]


-BASE_PREPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
-    {
-        "rule_id": "global/override/.m.rule.master",
-        "enabled": False,
-        "conditions": [],
-        "actions": ["dont_notify"],
-    }
+BASE_PREPEND_OVERRIDE_RULES = [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.master",
+        default_enabled=False,
+        conditions=[],
+        actions=["dont_notify"],
+    )
 ]


-BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
-    {
-        "rule_id": "global/override/.m.rule.suppress_notices",
-        "conditions": [
+BASE_APPEND_OVERRIDE_RULES = [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.suppress_notices",
+        conditions=[
            {
                "kind": "event_match",
                "key": "content.msgtype",
@@ -173,13 +284,15 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_suppress_notices",
            }
        ],
-        "actions": ["dont_notify"],
-    },
+        actions=["dont_notify"],
+    ),
    # NB. .m.rule.invite_for_me must be higher prio than .m.rule.member_event
    # otherwise invites will be matched by .m.rule.member_event
-    {
-        "rule_id": "global/override/.m.rule.invite_for_me",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.invite_for_me",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -195,21 +308,23 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
            # Match the requester's MXID.
            {"kind": "event_match", "key": "state_key", "pattern_type": "user_id"},
        ],
-        "actions": [
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "default"},
            {"set_tweak": "highlight", "value": False},
        ],
-    },
+    ),
    # Will we sometimes want to know about people joining and leaving?
    # Perhaps: if so, this could be expanded upon. Seems the most usual case
    # is that we don't though. We add this override rule so that even if
    # the room rule is set to notify, we don't get notifications about
    # join/leave/avatar/displayname events.
    # See also: https://matrix.org/jira/browse/SYN-607
-    {
-        "rule_id": "global/override/.m.rule.member_event",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.member_event",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -217,24 +332,28 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_member",
            }
        ],
-        "actions": ["dont_notify"],
-    },
+        actions=["dont_notify"],
+    ),
    # This was changed from underride to override so it's closer in priority
    # to the content rules where the user name highlight rule lives. This
    # way a room rule is lower priority than both but a custom override rule
    # is higher priority than both.
-    {
-        "rule_id": "global/override/.m.rule.contains_display_name",
-        "conditions": [{"kind": "contains_display_name"}],
-        "actions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.contains_display_name",
+        conditions=[{"kind": "contains_display_name"}],
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "default"},
            {"set_tweak": "highlight"},
        ],
-    },
-    {
-        "rule_id": "global/override/.m.rule.roomnotif",
-        "conditions": [
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.roomnotif",
+        conditions=[
            {
                "kind": "event_match",
                "key": "content.body",
@@ -247,11 +366,13 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_roomnotif_pl",
            },
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": True}],
-    },
-    {
-        "rule_id": "global/override/.m.rule.tombstone",
-        "conditions": [
+        actions=["notify", {"set_tweak": "highlight", "value": True}],
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.tombstone",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -265,11 +386,13 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_tombstone_statekey",
            },
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": True}],
-    },
-    {
-        "rule_id": "global/override/.m.rule.reaction",
-        "conditions": [
+        actions=["notify", {"set_tweak": "highlight", "value": True}],
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.m.rule.reaction",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -277,14 +400,16 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_reaction",
            }
        ],
-        "actions": ["dont_notify"],
-    },
+        actions=["dont_notify"],
+    ),
    # XXX: This is an experimental rule that is only enabled if msc3786_enabled
    # is enabled, if it is not the rule gets filtered out in _load_rules() in
    # PushRulesWorkerStore
-    {
-        "rule_id": "global/override/.org.matrix.msc3786.rule.room.server_acl",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["override"],
+        rule_id="global/override/.org.matrix.msc3786.rule.room.server_acl",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -298,15 +423,17 @@ BASE_APPEND_OVERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_room_server_acl_state_key",
            },
        ],
-        "actions": [],
-    },
+        actions=[],
+    ),
 ]


-BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
-    {
-        "rule_id": "global/underride/.m.rule.call",
-        "conditions": [
+BASE_APPEND_UNDERRIDE_RULES = [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.m.rule.call",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -314,17 +441,19 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_call",
            }
        ],
-        "actions": [
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "ring"},
            {"set_tweak": "highlight", "value": False},
        ],
-    },
+    ),
    # XXX: once m.direct is standardised everywhere, we should use it to detect
    # a DM from the user's perspective rather than this heuristic.
-    {
-        "rule_id": "global/underride/.m.rule.room_one_to_one",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.m.rule.room_one_to_one",
+        conditions=[
            {"kind": "room_member_count", "is": "2", "_cache_key": "member_count"},
            {
                "kind": "event_match",
@@ -333,17 +462,19 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_message",
            },
        ],
-        "actions": [
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "default"},
            {"set_tweak": "highlight", "value": False},
        ],
-    },
+    ),
    # XXX: this is going to fire for events which aren't m.room.messages
    # but are encrypted (e.g. m.call.*)...
-    {
-        "rule_id": "global/underride/.m.rule.encrypted_room_one_to_one",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.m.rule.encrypted_room_one_to_one",
+        conditions=[
            {"kind": "room_member_count", "is": "2", "_cache_key": "member_count"},
            {
                "kind": "event_match",
@@ -352,15 +483,17 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_encrypted",
            },
        ],
-        "actions": [
+        actions=[
            "notify",
            {"set_tweak": "sound", "value": "default"},
            {"set_tweak": "highlight", "value": False},
        ],
-    },
-    {
-        "rule_id": "global/underride/.org.matrix.msc3772.thread_reply",
-        "conditions": [
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.org.matrix.msc3772.thread_reply",
+        conditions=[
            {
                "kind": "org.matrix.msc3772.relation_match",
                "rel_type": "m.thread",
@@ -368,11 +501,13 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "sender_type": "user_id",
            }
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": False}],
-    },
-    {
-        "rule_id": "global/underride/.m.rule.message",
-        "conditions": [
+        actions=["notify", {"set_tweak": "highlight", "value": False}],
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.m.rule.message",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -380,13 +515,15 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_message",
            }
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": False}],
-    },
+        actions=["notify", {"set_tweak": "highlight", "value": False}],
+    ),
    # XXX: this is going to fire for events which aren't m.room.messages
    # but are encrypted (e.g. m.call.*)...
-    {
-        "rule_id": "global/underride/.m.rule.encrypted",
-        "conditions": [
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.m.rule.encrypted",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -394,11 +531,13 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_encrypted",
            }
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": False}],
-    },
-    {
-        "rule_id": "global/underride/.im.vector.jitsi",
-        "conditions": [
+        actions=["notify", {"set_tweak": "highlight", "value": False}],
+    ),
+    PushRule(
+        default=True,
+        priority_class=PRIORITY_CLASS_MAP["underride"],
+        rule_id="global/underride/.im.vector.jitsi",
+        conditions=[
            {
                "kind": "event_match",
                "key": "type",
@@ -418,29 +557,27 @@ BASE_APPEND_UNDERRIDE_RULES: List[Dict[str, Any]] = [
                "_cache_key": "_is_state_event",
            },
        ],
-        "actions": ["notify", {"set_tweak": "highlight", "value": False}],
-    },
+        actions=["notify", {"set_tweak": "highlight", "value": False}],
+    ),
 ]


 BASE_RULE_IDS = set()

+BASE_RULES_BY_ID: Dict[str, PushRule] = {}
+
 for r in BASE_APPEND_CONTENT_RULES:
-    r["priority_class"] = PRIORITY_CLASS_MAP["content"]
-    r["default"] = True
-    BASE_RULE_IDS.add(r["rule_id"])
+    BASE_RULE_IDS.add(r.rule_id)
+    BASE_RULES_BY_ID[r.rule_id] = r

 for r in BASE_PREPEND_OVERRIDE_RULES:
-    r["priority_class"] = PRIORITY_CLASS_MAP["override"]
-    r["default"] = True
-    BASE_RULE_IDS.add(r["rule_id"])
+    BASE_RULE_IDS.add(r.rule_id)
+    BASE_RULES_BY_ID[r.rule_id] = r

 for r in BASE_APPEND_OVERRIDE_RULES:
-    r["priority_class"] = PRIORITY_CLASS_MAP["override"]
-    r["default"] = True
-    BASE_RULE_IDS.add(r["rule_id"])
+    BASE_RULE_IDS.add(r.rule_id)
+    BASE_RULES_BY_ID[r.rule_id] = r

 for r in BASE_APPEND_UNDERRIDE_RULES:
-    r["priority_class"] = PRIORITY_CLASS_MAP["underride"]
-    r["default"] = True
-    BASE_RULE_IDS.add(r["rule_id"])
+    BASE_RULE_IDS.add(r.rule_id)
+    BASE_RULES_BY_ID[r.rule_id] = r
@@ -15,7 +15,18 @@

 import itertools
 import logging
-from typing import TYPE_CHECKING, Any, Dict, Iterable, List, Optional, Set, Tuple, Union
+from typing import (
+    TYPE_CHECKING,
+    Collection,
+    Dict,
+    Iterable,
+    List,
+    Mapping,
+    Optional,
+    Set,
+    Tuple,
+    Union,
+)

 from prometheus_client import Counter

@@ -30,6 +41,7 @@ from synapse.util.caches import register_cache
 from synapse.util.metrics import measure_func
 from synapse.visibility import filter_event_for_clients_with_state

+from .baserules import FilteredPushRules, PushRule
 from .push_rule_evaluator import PushRuleEvaluatorForEvent

 if TYPE_CHECKING:
@@ -112,7 +124,7 @@ class BulkPushRuleEvaluator:
    async def _get_rules_for_event(
        self,
        event: EventBase,
-    ) -> Dict[str, List[Dict[str, Any]]]:
+    ) -> Dict[str, FilteredPushRules]:
        """Get the push rules for all users who may need to be notified about
        the event.

@@ -186,7 +198,7 @@ class BulkPushRuleEvaluator:
        return pl_event.content if pl_event else {}, sender_level

    async def _get_mutual_relations(
-        self, event: EventBase, rules: Iterable[Dict[str, Any]]
+        self, event: EventBase, rules: Iterable[Tuple[PushRule, bool]]
    ) -> Dict[str, Set[Tuple[str, str]]]:
        """
        Fetch event metadata for events which related to the same event as the given event.
@@ -216,12 +228,11 @@ class BulkPushRuleEvaluator:

        # Pre-filter to figure out which relation types are interesting.
        rel_types = set()
-        for rule in rules:
-            # Skip disabled rules.
-            if "enabled" in rule and not rule["enabled"]:
+        for rule, enabled in rules:
+            if not enabled:
                continue

-            for condition in rule["conditions"]:
+            for condition in rule.conditions:
                if condition["kind"] != "org.matrix.msc3772.relation_match":
                    continue

@@ -254,7 +265,7 @@ class BulkPushRuleEvaluator:
        count_as_unread = _should_count_as_unread(event, context)

        rules_by_user = await self._get_rules_for_event(event)
-        actions_by_user: Dict[str, List[Union[dict, str]]] = {}
+        actions_by_user: Dict[str, Collection[Union[Mapping, str]]] = {}

        room_member_count = await self.store.get_number_joined_users_in_room(
            event.room_id
@@ -317,15 +328,13 @@ class BulkPushRuleEvaluator:
                # current user, it'll be added to the dict later.
                actions_by_user[uid] = []

-            for rule in rules:
-                if "enabled" in rule and not rule["enabled"]:
+            for rule, enabled in rules:
+                if not enabled:
                    continue

-                matches = evaluator.check_conditions(
-                    rule["conditions"], uid, display_name
-                )
+                matches = evaluator.check_conditions(rule.conditions, uid, display_name)
                if matches:
-                    actions = [x for x in rule["actions"] if x != "dont_notify"]
+                    actions = [x for x in rule.actions if x != "dont_notify"]
                    if actions and "notify" in actions:
                        # Push rules say we should notify the user of this event
                        actions_by_user[uid] = actions
@@ -18,16 +18,15 @@ from typing import Any, Dict, List, Optional
 from synapse.push.rulekinds import PRIORITY_CLASS_INVERSE_MAP, PRIORITY_CLASS_MAP
 from synapse.types import UserID

+from .baserules import FilteredPushRules, PushRule
+

 def format_push_rules_for_user(
-    user: UserID, ruleslist: List
+    user: UserID, ruleslist: FilteredPushRules
 ) -> Dict[str, Dict[str, list]]:
    """Converts a list of rawrules and a enabled map into nested dictionaries
    to match the Matrix client-server format for push rules"""

-    # We're going to be mutating this a lot, so do a deep copy
-    ruleslist = copy.deepcopy(ruleslist)
-
    rules: Dict[str, Dict[str, List[Dict[str, Any]]]] = {
        "global": {},
        "device": {},
@@ -35,11 +34,30 @@ def format_push_rules_for_user(

    rules["global"] = _add_empty_priority_class_arrays(rules["global"])

-    for r in ruleslist:
-        template_name = _priority_class_to_template_name(r["priority_class"])
+    for r, enabled in ruleslist:
+        template_name = _priority_class_to_template_name(r.priority_class)
+
+        rulearray = rules["global"][template_name]
+
+        template_rule = _rule_to_template(r)
+        if not template_rule:
+            continue
+
+        rulearray.append(template_rule)
+
+        template_rule["enabled"] = enabled
+
+        if "conditions" not in template_rule:
+            # Not all formatted rules have explicit conditions, e.g. "room"
+            # rules omit them as they can be derived from the kind and rule ID.
+            #
+            # If the formatted rule has no conditions then we can skip the
+            # formatting of conditions.
+            continue

        # Remove internal stuff.
-        for c in r["conditions"]:
+        template_rule["conditions"] = copy.deepcopy(template_rule["conditions"])
+        for c in template_rule["conditions"]:
            c.pop("_cache_key", None)

            pattern_type = c.pop("pattern_type", None)
@@ -52,16 +70,6 @@ def format_push_rules_for_user(
            if sender_type == "user_id":
                c["sender"] = user.to_string()

-        rulearray = rules["global"][template_name]
-
-        template_rule = _rule_to_template(r)
-        if template_rule:
-            if "enabled" in r:
-                template_rule["enabled"] = r["enabled"]
-            else:
-                template_rule["enabled"] = True
-            rulearray.append(template_rule)
-
    return rules


@@ -71,24 +79,24 @@ def _add_empty_priority_class_arrays(d: Dict[str, list]) -> Dict[str, list]:
    return d


-def _rule_to_template(rule: Dict[str, Any]) -> Optional[Dict[str, Any]]:
-    unscoped_rule_id = None
-    if "rule_id" in rule:
-        unscoped_rule_id = _rule_id_from_namespaced(rule["rule_id"])
+def _rule_to_template(rule: PushRule) -> Optional[Dict[str, Any]]:
+    templaterule: Dict[str, Any]

-    template_name = _priority_class_to_template_name(rule["priority_class"])
+    unscoped_rule_id = _rule_id_from_namespaced(rule.rule_id)
+
+    template_name = _priority_class_to_template_name(rule.priority_class)
    if template_name in ["override", "underride"]:
-        templaterule = {k: rule[k] for k in ["conditions", "actions"]}
+        templaterule = {"conditions": rule.conditions, "actions": rule.actions}
    elif template_name in ["sender", "room"]:
-        templaterule = {"actions": rule["actions"]}
-        unscoped_rule_id = rule["conditions"][0]["pattern"]
+        templaterule = {"actions": rule.actions}
+        unscoped_rule_id = rule.conditions[0]["pattern"]
    elif template_name == "content":
-        if len(rule["conditions"]) != 1:
+        if len(rule.conditions) != 1:
            return None
-        thecond = rule["conditions"][0]
+        thecond = rule.conditions[0]
        if "pattern" not in thecond:
            return None
-        templaterule = {"actions": rule["actions"]}
+        templaterule = {"actions": rule.actions}
        templaterule["pattern"] = thecond["pattern"]
    else:
        # This should not be reached unless this function is not kept in sync
@@ -97,8 +105,8 @@ def _rule_to_template(rule: Dict[str, Any]) -> Optional[Dict[str, Any]]:

    if unscoped_rule_id:
        templaterule["rule_id"] = unscoped_rule_id
-    if "default" in rule:
-        templaterule["default"] = rule["default"]
+    if rule.default:
+        templaterule["default"] = True
    return templaterule


@@ -15,7 +15,18 @@

 import logging
 import re
-from typing import Any, Dict, List, Mapping, Optional, Pattern, Set, Tuple, Union
+from typing import (
+    Any,
+    Dict,
+    List,
+    Mapping,
+    Optional,
+    Pattern,
+    Sequence,
+    Set,
+    Tuple,
+    Union,
+)

 from matrix_common.regex import glob_to_regex, to_word_pattern

@@ -32,14 +43,14 @@ INEQUALITY_EXPR = re.compile("^([=<>]*)([0-9]*)$")


 def _room_member_count(
-    ev: EventBase, condition: Dict[str, Any], room_member_count: int
+    ev: EventBase, condition: Mapping[str, Any], room_member_count: int
 ) -> bool:
    return _test_ineq_condition(condition, room_member_count)


 def _sender_notification_permission(
    ev: EventBase,
-    condition: Dict[str, Any],
+    condition: Mapping[str, Any],
    sender_power_level: int,
    power_levels: Dict[str, Union[int, Dict[str, int]]],
 ) -> bool:
@@ -54,7 +65,7 @@ def _sender_notification_permission(
    return sender_power_level >= room_notif_level


-def _test_ineq_condition(condition: Dict[str, Any], number: int) -> bool:
+def _test_ineq_condition(condition: Mapping[str, Any], number: int) -> bool:
    if "is" not in condition:
        return False
    m = INEQUALITY_EXPR.match(condition["is"])
@@ -137,7 +148,7 @@ class PushRuleEvaluatorForEvent:
        self._condition_cache: Dict[str, bool] = {}

    def check_conditions(
-        self, conditions: List[dict], uid: str, display_name: Optional[str]
+        self, conditions: Sequence[Mapping], uid: str, display_name: Optional[str]
    ) -> bool:
        """
        Returns true if a user's conditions/user ID/display name match the event.
@@ -169,7 +180,7 @@ class PushRuleEvaluatorForEvent:
        return True

    def matches(
-        self, condition: Dict[str, Any], user_id: str, display_name: Optional[str]
+        self, condition: Mapping[str, Any], user_id: str, display_name: Optional[str]
    ) -> bool:
        """
        Returns true if a user's condition/user ID/display name match the event.
@@ -204,7 +215,7 @@ class PushRuleEvaluatorForEvent:
            #     endpoint with an unknown kind, see _rule_tuple_from_request_object.
            return True

-    def _event_match(self, condition: dict, user_id: str) -> bool:
+    def _event_match(self, condition: Mapping, user_id: str) -> bool:
        """
        Check an "event_match" push rule condition.

@@ -269,7 +280,7 @@ class PushRuleEvaluatorForEvent:

        return bool(r.search(body))

-    def _relation_match(self, condition: dict, user_id: str) -> bool:
+    def _relation_match(self, condition: Mapping, user_id: str) -> bool:
        """
        Check an "relation_match" push rule condition.

@@ -416,7 +416,10 @@ class FederationSenderHandler:
            if not self._is_mine_id(receipt.user_id):
                continue
            # Private read receipts never get sent over federation.
-            if receipt.receipt_type == ReceiptTypes.READ_PRIVATE:
+            if receipt.receipt_type in (
+                ReceiptTypes.READ_PRIVATE,
+                ReceiptTypes.UNSTABLE_READ_PRIVATE,
+            ):
                continue
            receipt_info = ReadReceipt(
                receipt.room_id,
@@ -1 +1,12 @@
-<html><body>Your account is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.</body><html>
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Your account is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.</title>
+</head>
+<body>
+    Your account is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.
+</body>
+</html>
@@ -1 +1,12 @@
-<html><body>Your account has been successfully renewed and is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.</body><html>
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Your account has been successfully renewed and is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.</title>
+</head>
+<body>
+    Your account has been successfully renewed and is valid until {{ expiration_ts|format_ts("%d-%m-%Y") }}.
+</body>
+</html>
@@ -1,9 +1,14 @@
-<html>
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Request to add an email address to your Matrix account</title>
+</head>
 <body>
    <p>A request to add an email address to your Matrix account has been received. If this was you, please click the link below to confirm adding this email:</p>
-
    <a href="{{ link }}">{{ link }}</a>
-
    <p>If this was not you, you can safely ignore this email. Thank you.</p>
 </body>
 </html>
@@ -1,8 +1,13 @@
-<html>
-<head></head>
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Request failed</title>
+</head>
 <body>
-<p>The request failed for the following reason: {{ failure_reason }}.</p>
-
-<p>No changes have been made to your account.</p>
+    <p>The request failed for the following reason: {{ failure_reason }}.</p>
+    <p>No changes have been made to your account.</p>
 </body>
 </html>
@@ -1,6 +1,12 @@
-<html>
-<head></head>
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+    <title>Your email has now been validated</title>
+</head>
 <body>
-<p>Your email has now been validated, please return to your client. You may now close this window.</p>
+    <p>Your email has now been validated, please return to your client. You may now close this window.</p>
 </body>
-</html>
+</html>
@@ -1,8 +1,8 @@
 <html>
 <head>
 <title>Success!</title>
-<meta name='viewport' content='width=device-width, initial-scale=1,
-    user-scalable=no, minimum-scale=1.0, maximum-scale=1.0'>
+    <meta http-equiv="X-UA-Compatible" content="IE=edge">
+    <meta name="viewport" content="width=device-width, initial-scale=1.0">
 <link rel="stylesheet" href="/_matrix/static/client/register/style.css">
 <script>
 if (window.onAuthDone) {
--- a/Show More
+++ b/Show More
				`@@ -0,0 +1 @@`
				Improve validation of request bodies for the following client-server API endpoints: [`/account/password`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountpassword), [`/account/password/email/requestToken`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountpasswordemailrequesttoken), [`/account/deactivate`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3accountdeactivate) and [`/account/3pid/email/requestToken`](https://spec.matrix.org/v1.3/client-server-api/#post_matrixclientv3account3pidemailrequesttoken).
				`@@ -0,0 +1 @@`
				Allow use of both `@trace` and `@tag_args` stacked on the same function (tracing).
				`@@ -0,0 +1 @@`
				`Faster joins: update the rejected state of events during de-partial-stating.`
				`@@ -0,0 +1 @@`
				Add `openssl` example for generating registration HMAC digest.
				`@@ -0,0 +1 @@`
				Add some miscellaneous comments to document sync, especially around `compute_state_delta`.
				`@@ -0,0 +1 @@`
				Use literals in place of `HTTPStatus` constants in tests.
				`@@ -0,0 +1 @@`
				`Add comments about how event push actions are rotated.`
				`@@ -0,0 +1 @@`
				Instrument the federation/backfill part of `/messages` for understandable traces in Jaeger.
				`@@ -0,0 +1 @@`
				Document that event purging related to the `redaction_retention_period` config option is executed only every 5 minutes.
				`@@ -0,0 +1 @@`
				`Modify HTML template content to better support mobile devices' screen sizes.`