diff --git a/.buildkite/worker-blacklist b/.buildkite/worker-blacklist index 094b6c94d..fd98cbbaf 100644 --- a/.buildkite/worker-blacklist +++ b/.buildkite/worker-blacklist @@ -5,8 +5,6 @@ Message history can be paginated Can re-join room if re-invited -/upgrade creates a new room - The only membership state included in an initial sync is for all the senders in the timeline Local device key changes get to remote servers diff --git a/CHANGES.md b/CHANGES.md index a8a45e49a..225fced28 100644 --- a/CHANGES.md +++ b/CHANGES.md @@ -1,3 +1,206 @@ +Synapse 1.13.0 (2020-05-19) +=========================== + +This release brings some potential changes necessary for certain +configurations of Synapse: + +* If your Synapse is configured to use SSO and have a custom + `sso_redirect_confirm_template_dir` configuration option set, you will need + to duplicate the new `sso_auth_confirm.html`, `sso_auth_success.html` and + `sso_account_deactivated.html` templates into that directory. +* Synapse plugins using the `complete_sso_login` method of + `synapse.module_api.ModuleApi` should instead switch to the async/await + version, `complete_sso_login_async`, which includes additional checks. The + former version is now deprecated. +* A bug was introduced in Synapse 1.4.0 which could cause the room directory + to be incomplete or empty if Synapse was upgraded directly from v1.2.1 or + earlier, to versions between v1.4.0 and v1.12.x. + +Please review [UPGRADE.rst](UPGRADE.rst) for more details on these changes +and for general upgrade guidance. + + +Notice of change to the default `git` branch for Synapse +-------------------------------------------------------- + +With the release of Synapse 1.13.0, the default `git` branch for Synapse has +changed to `develop`, which is the development tip. This is more consistent with +common practice and modern `git` usage. + +The `master` branch, which tracks the latest release, is still available. It is +recommended that developers and distributors who have scripts which run builds +using the default branch of Synapse should therefore consider pinning their +scripts to `master`. + + +Internal Changes +---------------- + +- Update the version of dh-virtualenv we use to build debs, and add focal to the list of target distributions. ([\#7526](https://github.com/matrix-org/synapse/issues/7526)) + + +Synapse 1.13.0rc3 (2020-05-18) +============================== + +Bugfixes +-------- + +- Hash passwords as early as possible during registration. ([\#7523](https://github.com/matrix-org/synapse/issues/7523)) + + +Synapse 1.13.0rc2 (2020-05-14) +============================== + +Bugfixes +-------- + +- Fix a long-standing bug which could cause messages not to be sent over federation, when state events with state keys matching user IDs (such as custom user statuses) were received. ([\#7376](https://github.com/matrix-org/synapse/issues/7376)) +- Restore compatibility with non-compliant clients during the user interactive authentication process, fixing a problem introduced in v1.13.0rc1. ([\#7483](https://github.com/matrix-org/synapse/issues/7483)) + +Internal Changes +---------------- + +- Fix linting errors in new version of Flake8. ([\#7470](https://github.com/matrix-org/synapse/issues/7470)) + + +Synapse 1.13.0rc1 (2020-05-11) +============================== + +Features +-------- + +- Extend the `web_client_location` option to accept an absolute URL to use as a redirect. Adds a warning when running the web client on the same hostname as homeserver. Contributed by Martin Milata. ([\#7006](https://github.com/matrix-org/synapse/issues/7006)) +- Set `Referrer-Policy` header to `no-referrer` on media downloads. ([\#7009](https://github.com/matrix-org/synapse/issues/7009)) +- Add support for running replication over Redis when using workers. ([\#7040](https://github.com/matrix-org/synapse/issues/7040), [\#7325](https://github.com/matrix-org/synapse/issues/7325), [\#7352](https://github.com/matrix-org/synapse/issues/7352), [\#7401](https://github.com/matrix-org/synapse/issues/7401), [\#7427](https://github.com/matrix-org/synapse/issues/7427), [\#7439](https://github.com/matrix-org/synapse/issues/7439), [\#7446](https://github.com/matrix-org/synapse/issues/7446), [\#7450](https://github.com/matrix-org/synapse/issues/7450), [\#7454](https://github.com/matrix-org/synapse/issues/7454)) +- Admin API `POST /_synapse/admin/v1/join/` to join users to a room like `auto_join_rooms` for creation of users. ([\#7051](https://github.com/matrix-org/synapse/issues/7051)) +- Add options to prevent users from changing their profile or associated 3PIDs. ([\#7096](https://github.com/matrix-org/synapse/issues/7096)) +- Support SSO in the user interactive authentication workflow. ([\#7102](https://github.com/matrix-org/synapse/issues/7102), [\#7186](https://github.com/matrix-org/synapse/issues/7186), [\#7279](https://github.com/matrix-org/synapse/issues/7279), [\#7343](https://github.com/matrix-org/synapse/issues/7343)) +- Allow server admins to define and enforce a password policy ([MSC2000](https://github.com/matrix-org/matrix-doc/issues/2000)). ([\#7118](https://github.com/matrix-org/synapse/issues/7118)) +- Improve the support for SSO authentication on the login fallback page. ([\#7152](https://github.com/matrix-org/synapse/issues/7152), [\#7235](https://github.com/matrix-org/synapse/issues/7235)) +- Always whitelist the login fallback in the SSO configuration if `public_baseurl` is set. ([\#7153](https://github.com/matrix-org/synapse/issues/7153)) +- Admin users are no longer required to be in a room to create an alias for it. ([\#7191](https://github.com/matrix-org/synapse/issues/7191)) +- Require admin privileges to enable room encryption by default. This does not affect existing rooms. ([\#7230](https://github.com/matrix-org/synapse/issues/7230)) +- Add a config option for specifying the value of the Accept-Language HTTP header when generating URL previews. ([\#7265](https://github.com/matrix-org/synapse/issues/7265)) +- Allow `/requestToken` endpoints to hide the existence (or lack thereof) of 3PID associations on the homeserver. ([\#7315](https://github.com/matrix-org/synapse/issues/7315)) +- Add a configuration setting to tweak the threshold for dummy events. ([\#7422](https://github.com/matrix-org/synapse/issues/7422)) + + +Bugfixes +-------- + +- Don't attempt to use an invalid sqlite config if no database configuration is provided. Contributed by @nekatak. ([\#6573](https://github.com/matrix-org/synapse/issues/6573)) +- Fix single-sign on with CAS systems: pass the same service URL when requesting the CAS ticket and when calling the `proxyValidate` URL. Contributed by @Naugrimm. ([\#6634](https://github.com/matrix-org/synapse/issues/6634)) +- Fix missing field `default` when fetching user-defined push rules. ([\#6639](https://github.com/matrix-org/synapse/issues/6639)) +- Improve error responses when accessing remote public room lists. ([\#6899](https://github.com/matrix-org/synapse/issues/6899), [\#7368](https://github.com/matrix-org/synapse/issues/7368)) +- Transfer alias mappings on room upgrade. ([\#6946](https://github.com/matrix-org/synapse/issues/6946)) +- Ensure that a user interactive authentication session is tied to a single request. ([\#7068](https://github.com/matrix-org/synapse/issues/7068), [\#7455](https://github.com/matrix-org/synapse/issues/7455)) +- Fix a bug in the federation API which could cause occasional "Failed to get PDU" errors. ([\#7089](https://github.com/matrix-org/synapse/issues/7089)) +- Return the proper error (`M_BAD_ALIAS`) when a non-existant canonical alias is provided. ([\#7109](https://github.com/matrix-org/synapse/issues/7109)) +- Fix a bug which meant that groups updates were not correctly replicated between workers. ([\#7117](https://github.com/matrix-org/synapse/issues/7117)) +- Fix starting workers when federation sending not split out. ([\#7133](https://github.com/matrix-org/synapse/issues/7133)) +- Ensure `is_verified` is a boolean in responses to `GET /_matrix/client/r0/room_keys/keys`. Also warn the user if they forgot the `version` query param. ([\#7150](https://github.com/matrix-org/synapse/issues/7150)) +- Fix error page being shown when a custom SAML handler attempted to redirect when processing an auth response. ([\#7151](https://github.com/matrix-org/synapse/issues/7151)) +- Avoid importing `sqlite3` when using the postgres backend. Contributed by David Vo. ([\#7155](https://github.com/matrix-org/synapse/issues/7155)) +- Fix excessive CPU usage by `prune_old_outbound_device_pokes` job. ([\#7159](https://github.com/matrix-org/synapse/issues/7159)) +- Fix a bug which could cause outbound federation traffic to stop working if a client uploaded an incorrect e2e device signature. ([\#7177](https://github.com/matrix-org/synapse/issues/7177)) +- Fix a bug which could cause incorrect 'cyclic dependency' error. ([\#7178](https://github.com/matrix-org/synapse/issues/7178)) +- Fix a bug that could cause a user to be invited to a server notices (aka System Alerts) room without any notice being sent. ([\#7199](https://github.com/matrix-org/synapse/issues/7199)) +- Fix some worker-mode replication handling not being correctly recorded in CPU usage stats. ([\#7203](https://github.com/matrix-org/synapse/issues/7203)) +- Do not allow a deactivated user to login via SSO. ([\#7240](https://github.com/matrix-org/synapse/issues/7240), [\#7259](https://github.com/matrix-org/synapse/issues/7259)) +- Fix --help command-line argument. ([\#7249](https://github.com/matrix-org/synapse/issues/7249)) +- Fix room publish permissions not being checked on room creation. ([\#7260](https://github.com/matrix-org/synapse/issues/7260)) +- Reject unknown session IDs during user interactive authentication instead of silently creating a new session. ([\#7268](https://github.com/matrix-org/synapse/issues/7268)) +- Fix a SQL query introduced in Synapse 1.12.0 which could cause large amounts of logging to the postgres slow-query log. ([\#7274](https://github.com/matrix-org/synapse/issues/7274)) +- Persist user interactive authentication sessions across workers and Synapse restarts. ([\#7302](https://github.com/matrix-org/synapse/issues/7302)) +- Fixed backwards compatibility logic of the first value of `trusted_third_party_id_servers` being used for `account_threepid_delegates.email`, which occurs when the former, deprecated option is set and the latter is not. ([\#7316](https://github.com/matrix-org/synapse/issues/7316)) +- Fix a bug where event updates might not be sent over replication to worker processes after the stream falls behind. ([\#7337](https://github.com/matrix-org/synapse/issues/7337), [\#7358](https://github.com/matrix-org/synapse/issues/7358)) +- Fix bad error handling that would cause Synapse to crash if it's provided with a YAML configuration file that's either empty or doesn't parse into a key-value map. ([\#7341](https://github.com/matrix-org/synapse/issues/7341)) +- Fix incorrect metrics reporting for `renew_attestations` background task. ([\#7344](https://github.com/matrix-org/synapse/issues/7344)) +- Prevent non-federating rooms from appearing in responses to federated `POST /publicRoom` requests when a filter was included. ([\#7367](https://github.com/matrix-org/synapse/issues/7367)) +- Fix a bug which would cause the room durectory to be incorrectly populated if Synapse was upgraded directly from v1.2.1 or earlier to v1.4.0 or later. Note that this fix does not apply retrospectively; see the [upgrade notes](UPGRADE.rst#upgrading-to-v1130) for more information. ([\#7387](https://github.com/matrix-org/synapse/issues/7387)) +- Fix bug in `EventContext.deserialize`. ([\#7393](https://github.com/matrix-org/synapse/issues/7393)) + + +Improved Documentation +---------------------- + +- Update Debian installation instructions to recommend installing the `virtualenv` package instead of `python3-virtualenv`. ([\#6892](https://github.com/matrix-org/synapse/issues/6892)) +- Improve the documentation for database configuration. ([\#6988](https://github.com/matrix-org/synapse/issues/6988)) +- Improve the documentation of application service configuration files. ([\#7091](https://github.com/matrix-org/synapse/issues/7091)) +- Update pre-built package name for FreeBSD. ([\#7107](https://github.com/matrix-org/synapse/issues/7107)) +- Update postgres docs with login troubleshooting information. ([\#7119](https://github.com/matrix-org/synapse/issues/7119)) +- Clean up INSTALL.md a bit. ([\#7141](https://github.com/matrix-org/synapse/issues/7141)) +- Add documentation for running a local CAS server for testing. ([\#7147](https://github.com/matrix-org/synapse/issues/7147)) +- Improve README.md by being explicit about public IP recommendation for TURN relaying. ([\#7167](https://github.com/matrix-org/synapse/issues/7167)) +- Fix a small typo in the `metrics_flags` config option. ([\#7171](https://github.com/matrix-org/synapse/issues/7171)) +- Update the contributed documentation on managing synapse workers with systemd, and bring it into the core distribution. ([\#7234](https://github.com/matrix-org/synapse/issues/7234)) +- Add documentation to the `password_providers` config option. Add known password provider implementations to docs. ([\#7238](https://github.com/matrix-org/synapse/issues/7238), [\#7248](https://github.com/matrix-org/synapse/issues/7248)) +- Modify suggested nginx reverse proxy configuration to match Synapse's default file upload size. Contributed by @ProCycleDev. ([\#7251](https://github.com/matrix-org/synapse/issues/7251)) +- Documentation of media_storage_providers options updated to avoid misunderstandings. Contributed by Tristan Lins. ([\#7272](https://github.com/matrix-org/synapse/issues/7272)) +- Add documentation on monitoring workers with Prometheus. ([\#7357](https://github.com/matrix-org/synapse/issues/7357)) +- Clarify endpoint usage in the users admin api documentation. ([\#7361](https://github.com/matrix-org/synapse/issues/7361)) + + +Deprecations and Removals +------------------------- + +- Remove nonfunctional `captcha_bypass_secret` option from `homeserver.yaml`. ([\#7137](https://github.com/matrix-org/synapse/issues/7137)) + + +Internal Changes +---------------- + +- Add benchmarks for LruCache. ([\#6446](https://github.com/matrix-org/synapse/issues/6446)) +- Return total number of users and profile attributes in admin users endpoint. Contributed by Awesome Technologies Innovationslabor GmbH. ([\#6881](https://github.com/matrix-org/synapse/issues/6881)) +- Change device list streams to have one row per ID. ([\#7010](https://github.com/matrix-org/synapse/issues/7010)) +- Remove concept of a non-limited stream. ([\#7011](https://github.com/matrix-org/synapse/issues/7011)) +- Move catchup of replication streams logic to worker. ([\#7024](https://github.com/matrix-org/synapse/issues/7024), [\#7195](https://github.com/matrix-org/synapse/issues/7195), [\#7226](https://github.com/matrix-org/synapse/issues/7226), [\#7239](https://github.com/matrix-org/synapse/issues/7239), [\#7286](https://github.com/matrix-org/synapse/issues/7286), [\#7290](https://github.com/matrix-org/synapse/issues/7290), [\#7318](https://github.com/matrix-org/synapse/issues/7318), [\#7326](https://github.com/matrix-org/synapse/issues/7326), [\#7378](https://github.com/matrix-org/synapse/issues/7378), [\#7421](https://github.com/matrix-org/synapse/issues/7421)) +- Convert some of synapse.rest.media to async/await. ([\#7110](https://github.com/matrix-org/synapse/issues/7110), [\#7184](https://github.com/matrix-org/synapse/issues/7184), [\#7241](https://github.com/matrix-org/synapse/issues/7241)) +- De-duplicate / remove unused REST code for login and auth. ([\#7115](https://github.com/matrix-org/synapse/issues/7115)) +- Convert `*StreamRow` classes to inner classes. ([\#7116](https://github.com/matrix-org/synapse/issues/7116)) +- Clean up some LoggingContext code. ([\#7120](https://github.com/matrix-org/synapse/issues/7120), [\#7181](https://github.com/matrix-org/synapse/issues/7181), [\#7183](https://github.com/matrix-org/synapse/issues/7183), [\#7408](https://github.com/matrix-org/synapse/issues/7408), [\#7426](https://github.com/matrix-org/synapse/issues/7426)) +- Add explicit `instance_id` for USER_SYNC commands and remove implicit `conn_id` usage. ([\#7128](https://github.com/matrix-org/synapse/issues/7128)) +- Refactored the CAS authentication logic to a separate class. ([\#7136](https://github.com/matrix-org/synapse/issues/7136)) +- Run replication streamers on workers. ([\#7146](https://github.com/matrix-org/synapse/issues/7146)) +- Add tests for outbound device pokes. ([\#7157](https://github.com/matrix-org/synapse/issues/7157)) +- Fix device list update stream ids going backward. ([\#7158](https://github.com/matrix-org/synapse/issues/7158)) +- Use `stream.current_token()` and remove `stream_positions()`. ([\#7172](https://github.com/matrix-org/synapse/issues/7172)) +- Move client command handling out of TCP protocol. ([\#7185](https://github.com/matrix-org/synapse/issues/7185)) +- Move server command handling out of TCP protocol. ([\#7187](https://github.com/matrix-org/synapse/issues/7187)) +- Fix consistency of HTTP status codes reported in log lines. ([\#7188](https://github.com/matrix-org/synapse/issues/7188)) +- Only run one background database update at a time. ([\#7190](https://github.com/matrix-org/synapse/issues/7190)) +- Remove sent outbound device list pokes from the database. ([\#7192](https://github.com/matrix-org/synapse/issues/7192)) +- Add a background database update job to clear out duplicate `device_lists_outbound_pokes`. ([\#7193](https://github.com/matrix-org/synapse/issues/7193)) +- Remove some extraneous debugging log lines. ([\#7207](https://github.com/matrix-org/synapse/issues/7207)) +- Add explicit Python build tooling as dependencies for the snapcraft build. ([\#7213](https://github.com/matrix-org/synapse/issues/7213)) +- Add typing information to federation server code. ([\#7219](https://github.com/matrix-org/synapse/issues/7219)) +- Extend room admin api (`GET /_synapse/admin/v1/rooms`) with additional attributes. ([\#7225](https://github.com/matrix-org/synapse/issues/7225)) +- Unblacklist '/upgrade creates a new room' sytest for workers. ([\#7228](https://github.com/matrix-org/synapse/issues/7228)) +- Remove redundant checks on `daemonize` from synctl. ([\#7233](https://github.com/matrix-org/synapse/issues/7233)) +- Upgrade jQuery to v3.4.1 on fallback login/registration pages. ([\#7236](https://github.com/matrix-org/synapse/issues/7236)) +- Change log line that told user to implement onLogin/onRegister fallback js functions to a warning, instead of an info, so it's more visible. ([\#7237](https://github.com/matrix-org/synapse/issues/7237)) +- Correct the parameters of a test fixture. Contributed by Isaiah Singletary. ([\#7243](https://github.com/matrix-org/synapse/issues/7243)) +- Convert auth handler to async/await. ([\#7261](https://github.com/matrix-org/synapse/issues/7261)) +- Add some unit tests for replication. ([\#7278](https://github.com/matrix-org/synapse/issues/7278)) +- Improve typing annotations in `synapse.replication.tcp.streams.Stream`. ([\#7291](https://github.com/matrix-org/synapse/issues/7291)) +- Reduce log verbosity of url cache cleanup tasks. ([\#7295](https://github.com/matrix-org/synapse/issues/7295)) +- Fix sample SAML Service Provider configuration. Contributed by @frcl. ([\#7300](https://github.com/matrix-org/synapse/issues/7300)) +- Fix StreamChangeCache to work with multiple entities changing on the same stream id. ([\#7303](https://github.com/matrix-org/synapse/issues/7303)) +- Fix an incorrect import in IdentityHandler. ([\#7319](https://github.com/matrix-org/synapse/issues/7319)) +- Reduce logging verbosity for successful federation requests. ([\#7321](https://github.com/matrix-org/synapse/issues/7321)) +- Convert some federation handler code to async/await. ([\#7338](https://github.com/matrix-org/synapse/issues/7338)) +- Fix collation for postgres for unit tests. ([\#7359](https://github.com/matrix-org/synapse/issues/7359)) +- Convert RegistrationWorkerStore.is_server_admin and dependent code to async/await. ([\#7363](https://github.com/matrix-org/synapse/issues/7363)) +- Add an `instance_name` to `RDATA` and `POSITION` replication commands. ([\#7364](https://github.com/matrix-org/synapse/issues/7364)) +- Thread through instance name to replication client. ([\#7369](https://github.com/matrix-org/synapse/issues/7369)) +- Convert synapse.server_notices to async/await. ([\#7394](https://github.com/matrix-org/synapse/issues/7394)) +- Convert synapse.notifier to async/await. ([\#7395](https://github.com/matrix-org/synapse/issues/7395)) +- Fix issues with the Python package manifest. ([\#7404](https://github.com/matrix-org/synapse/issues/7404)) +- Prevent methods in `synapse.handlers.auth` from polling the homeserver config every request. ([\#7420](https://github.com/matrix-org/synapse/issues/7420)) +- Speed up fetching device lists changes when handling `/sync` requests. ([\#7423](https://github.com/matrix-org/synapse/issues/7423)) +- Run group attestation renewal in series rather than parallel for performance. ([\#7442](https://github.com/matrix-org/synapse/issues/7442)) + + Synapse 1.12.4 (2020-04-23) =========================== @@ -21,7 +224,6 @@ Bugfixes - Do not treat display names as globs in push rules. ([\#7271](https://github.com/matrix-org/synapse/issues/7271)) - Fix a bug with cross-signing devices belonging to remote users who did not share a room with any user on the local homeserver. ([\#7289](https://github.com/matrix-org/synapse/issues/7289)) - Synapse 1.12.3 (2020-04-03) =========================== @@ -31,13 +233,10 @@ correctly fix the issue with building the Debian packages. ([\#7212](https://git Synapse 1.12.2 (2020-04-02) =========================== -This release works around [an -issue](https://github.com/matrix-org/synapse/issues/7208) with building the -debian packages. +This release works around [an issue](https://github.com/matrix-org/synapse/issues/7208) with building the debian packages. No other significant changes since 1.12.1. - Synapse 1.12.1 (2020-04-02) =========================== @@ -54,7 +253,6 @@ Bugfixes - Avoid importing `sqlite3` when using the postgres backend. Contributed by David Vo. ([\#7155](https://github.com/matrix-org/synapse/issues/7155)). Introduced in v1.12.0rc1. - Fix a bug which could cause outbound federation traffic to stop working if a client uploaded an incorrect e2e device signature. ([\#7177](https://github.com/matrix-org/synapse/issues/7177)). Introduced in v1.11.0. - Synapse 1.12.0 (2020-03-23) =========================== diff --git a/INSTALL.md b/INSTALL.md index 8ded6e909..b8f8a6732 100644 --- a/INSTALL.md +++ b/INSTALL.md @@ -2,7 +2,6 @@ - [Installing Synapse](#installing-synapse) - [Installing from source](#installing-from-source) - [Platform-Specific Instructions](#platform-specific-instructions) - - [Troubleshooting Installation](#troubleshooting-installation) - [Prebuilt packages](#prebuilt-packages) - [Setting up Synapse](#setting-up-synapse) - [TLS certificates](#tls-certificates) @@ -10,6 +9,7 @@ - [Registering a user](#registering-a-user) - [Setting up a TURN server](#setting-up-a-turn-server) - [URL previews](#url-previews) +- [Troubleshooting Installation](#troubleshooting-installation) # Choosing your server name @@ -70,7 +70,7 @@ pip install -U matrix-synapse ``` Before you can start Synapse, you will need to generate a configuration -file. To do this, run (in your virtualenv, as before):: +file. To do this, run (in your virtualenv, as before): ``` cd ~/synapse @@ -84,22 +84,24 @@ python -m synapse.app.homeserver \ ... substituting an appropriate value for `--server-name`. This command will generate you a config file that you can then customise, but it will -also generate a set of keys for you. These keys will allow your Home Server to -identify itself to other Home Servers, so don't lose or delete them. It would be +also generate a set of keys for you. These keys will allow your homeserver to +identify itself to other homeserver, so don't lose or delete them. It would be wise to back them up somewhere safe. (If, for whatever reason, you do need to -change your Home Server's keys, you may find that other Home Servers have the +change your homeserver's keys, you may find that other homeserver have the old key cached. If you update the signing key, you should change the name of the key in the `.signing.key` file (the second word) to something different. See the [spec](https://matrix.org/docs/spec/server_server/latest.html#retrieving-server-keys) -for more information on key management.) +for more information on key management). To actually run your new homeserver, pick a working directory for Synapse to -run (e.g. `~/synapse`), and:: +run (e.g. `~/synapse`), and: - cd ~/synapse - source env/bin/activate - synctl start +``` +cd ~/synapse +source env/bin/activate +synctl start +``` ### Platform-Specific Instructions @@ -110,7 +112,7 @@ Installing prerequisites on Ubuntu or Debian: ``` sudo apt-get install build-essential python3-dev libffi-dev \ python3-pip python3-setuptools sqlite3 \ - libssl-dev python3-virtualenv libjpeg-dev libxslt1-dev + libssl-dev virtualenv libjpeg-dev libxslt1-dev ``` #### ArchLinux @@ -188,7 +190,7 @@ doas pkg_add python libffi py-pip py-setuptools sqlite3 py-virtualenv \ There is currently no port for OpenBSD. Additionally, OpenBSD's security settings require a slightly more difficult installation process. -XXX: I suspect this is out of date. +(XXX: I suspect this is out of date) 1. Create a new directory in `/usr/local` called `_synapse`. Also, create a new user called `_synapse` and set that directory as the new user's home. @@ -196,7 +198,7 @@ XXX: I suspect this is out of date. write and execute permissions on the same memory space to be run from `/usr/local`. 2. `su` to the new `_synapse` user and change to their home directory. -3. Create a new virtualenv: `virtualenv -p python2.7 ~/.synapse` +3. Create a new virtualenv: `virtualenv -p python3 ~/.synapse` 4. Source the virtualenv configuration located at `/usr/local/_synapse/.synapse/bin/activate`. This is done in `ksh` by using the `.` command, rather than `bash`'s `source`. @@ -217,45 +219,6 @@ be found at https://docs.microsoft.com/en-us/windows/wsl/install-win10 for Windows 10 and https://docs.microsoft.com/en-us/windows/wsl/install-on-server for Windows Server. -### Troubleshooting Installation - -XXX a bunch of this is no longer relevant. - -Synapse requires pip 8 or later, so if your OS provides too old a version you -may need to manually upgrade it:: - - sudo pip install --upgrade pip - -Installing may fail with `Could not find any downloads that satisfy the requirement pymacaroons-pynacl (from matrix-synapse==0.12.0)`. -You can fix this by manually upgrading pip and virtualenv:: - - sudo pip install --upgrade virtualenv - -You can next rerun `virtualenv -p python3 synapse` to update the virtual env. - -Installing may fail during installing virtualenv with `InsecurePlatformWarning: A true SSLContext object is not available. This prevents urllib3 from configuring SSL appropriately and may cause certain SSL connections to fail. For more information, see https://urllib3.readthedocs.org/en/latest/security.html#insecureplatformwarning.` -You can fix this by manually installing ndg-httpsclient:: - - pip install --upgrade ndg-httpsclient - -Installing may fail with `mock requires setuptools>=17.1. Aborting installation`. -You can fix this by upgrading setuptools:: - - pip install --upgrade setuptools - -If pip crashes mid-installation for reason (e.g. lost terminal), pip may -refuse to run until you remove the temporary installation directory it -created. To reset the installation:: - - rm -rf /tmp/pip_install_matrix - -pip seems to leak *lots* of memory during installation. For instance, a Linux -host with 512MB of RAM may run out of memory whilst installing Twisted. If this -happens, you will have to individually install the dependencies which are -failing, e.g.:: - - pip install twisted - ## Prebuilt packages As an alternative to installing from source, prebuilt packages are available @@ -314,7 +277,7 @@ For `buster` and `sid`, Synapse is available in the Debian repositories and it should be possible to install it with simply: ``` - sudo apt install matrix-synapse +sudo apt install matrix-synapse ``` There is also a version of `matrix-synapse` in `stretch-backports`. Please see @@ -375,15 +338,17 @@ sudo pip install py-bcrypt Synapse can be found in the void repositories as 'synapse': - xbps-install -Su - xbps-install -S synapse +``` +xbps-install -Su +xbps-install -S synapse +``` ### FreeBSD Synapse can be installed via FreeBSD Ports or Packages contributed by Brendan Molloy from: - Ports: `cd /usr/ports/net-im/py-matrix-synapse && make install clean` - - Packages: `pkg install py27-matrix-synapse` + - Packages: `pkg install py37-matrix-synapse` ### NixOS @@ -420,6 +385,7 @@ so, you will need to edit `homeserver.yaml`, as follows: resources: - names: [client, federation] ``` + * You will also need to uncomment the `tls_certificate_path` and `tls_private_key_path` lines under the `TLS` section. You can either point these settings at an existing certificate and key, or you can @@ -435,7 +401,7 @@ so, you will need to edit `homeserver.yaml`, as follows: `cert.pem`). For a more detailed guide to configuring your server for federation, see -[federate.md](docs/federate.md) +[federate.md](docs/federate.md). ## Email @@ -482,7 +448,7 @@ on your server even if `enable_registration` is `false`. ## Setting up a TURN server For reliable VoIP calls to be routed via this homeserver, you MUST configure -a TURN server. See [docs/turn-howto.md](docs/turn-howto.md) for details. +a TURN server. See [docs/turn-howto.md](docs/turn-howto.md) for details. ## URL previews @@ -491,10 +457,24 @@ turn it on you must enable the `url_preview_enabled: True` config parameter and explicitly specify the IP ranges that Synapse is not allowed to spider for previewing in the `url_preview_ip_range_blacklist` configuration parameter. This is critical from a security perspective to stop arbitrary Matrix users -spidering 'internal' URLs on your network. At the very least we recommend that +spidering 'internal' URLs on your network. At the very least we recommend that your loopback and RFC1918 IP addresses are blacklisted. -This also requires the optional lxml and netaddr python dependencies to be -installed. This in turn requires the libxml2 library to be available - on +This also requires the optional `lxml` and `netaddr` python dependencies to be +installed. This in turn requires the `libxml2` library to be available - on Debian/Ubuntu this means `apt-get install libxml2-dev`, or equivalent for your OS. + +# Troubleshooting Installation + +`pip` seems to leak *lots* of memory during installation. For instance, a Linux +host with 512MB of RAM may run out of memory whilst installing Twisted. If this +happens, you will have to individually install the dependencies which are +failing, e.g.: + +``` +pip install twisted +``` + +If you have any other problems, feel free to ask in +[#synapse:matrix.org](https://matrix.to/#/#synapse:matrix.org). diff --git a/MANIFEST.in b/MANIFEST.in index 156d6f04f..120ce5b77 100644 --- a/MANIFEST.in +++ b/MANIFEST.in @@ -30,23 +30,24 @@ recursive-include synapse/static *.gif recursive-include synapse/static *.html recursive-include synapse/static *.js -exclude Dockerfile +exclude .codecov.yml +exclude .coveragerc exclude .dockerignore -exclude test_postgresql.sh exclude .editorconfig +exclude Dockerfile +exclude mypy.ini exclude sytest-blacklist +exclude test_postgresql.sh include pyproject.toml recursive-include changelog.d * prune .buildkite prune .circleci -prune .codecov.yml -prune .coveragerc prune .github +prune contrib prune debian prune demo/etc prune docker -prune mypy.ini prune snap prune stubs diff --git a/UPGRADE.rst b/UPGRADE.rst index 768d94a39..41c47e964 100644 --- a/UPGRADE.rst +++ b/UPGRADE.rst @@ -75,6 +75,74 @@ for example: wget https://packages.matrix.org/debian/pool/main/m/matrix-synapse-py3/matrix-synapse-py3_1.3.0+stretch1_amd64.deb dpkg -i matrix-synapse-py3_1.3.0+stretch1_amd64.deb +Upgrading to v1.13.0 +==================== + + +Incorrect database migration in old synapse versions +---------------------------------------------------- + +A bug was introduced in Synapse 1.4.0 which could cause the room directory to +be incomplete or empty if Synapse was upgraded directly from v1.2.1 or +earlier, to versions between v1.4.0 and v1.12.x. + +This will *not* be a problem for Synapse installations which were: + * created at v1.4.0 or later, + * upgraded via v1.3.x, or + * upgraded straight from v1.2.1 or earlier to v1.13.0 or later. + +If completeness of the room directory is a concern, installations which are +affected can be repaired as follows: + +1. Run the following sql from a `psql` or `sqlite3` console: + + .. code:: sql + + INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES + ('populate_stats_process_rooms', '{}', 'current_state_events_membership'); + + INSERT INTO background_updates (update_name, progress_json, depends_on) VALUES + ('populate_stats_process_users', '{}', 'populate_stats_process_rooms'); + +2. Restart synapse. + +New Single Sign-on HTML Templates +--------------------------------- + +New templates (``sso_auth_confirm.html``, ``sso_auth_success.html``, and +``sso_account_deactivated.html``) were added to Synapse. If your Synapse is +configured to use SSO and a custom ``sso_redirect_confirm_template_dir`` +configuration then these templates will need to be copied from +`synapse/res/templates `_ into that directory. + +Synapse SSO Plugins Method Deprecation +-------------------------------------- + +Plugins using the ``complete_sso_login`` method of +``synapse.module_api.ModuleApi`` should update to using the async/await +version ``complete_sso_login_async`` which includes additional checks. The +non-async version is considered deprecated. + +Rolling back to v1.12.4 after a failed upgrade +---------------------------------------------- + +v1.13.0 includes a lot of large changes. If something problematic occurs, you +may want to roll-back to a previous version of Synapse. Because v1.13.0 also +includes a new database schema version, reverting that version is also required +alongside the generic rollback instructions mentioned above. In short, to roll +back to v1.12.4 you need to: + +1. Stop the server +2. Decrease the schema version in the database: + + .. code:: sql + + UPDATE schema_version SET version = 57; + +3. Downgrade Synapse by following the instructions for your installation method + in the "Rolling back to older versions" section above. + + Upgrading to v1.12.0 ==================== diff --git a/contrib/systemd-with-workers/README.md b/contrib/systemd-with-workers/README.md index 74b261e9f..8d21d532b 100644 --- a/contrib/systemd-with-workers/README.md +++ b/contrib/systemd-with-workers/README.md @@ -1,150 +1,2 @@ -# Setup Synapse with Workers and Systemd - -This is a setup for managing synapse with systemd including support for -managing workers. It provides a `matrix-synapse`, as well as a -`matrix-synapse-worker@` service for any workers you require. Additionally to -group the required services it sets up a `matrix.target`. You can use this to -automatically start any bot- or bridge-services. More on this in -[Bots and Bridges](#bots-and-bridges). - -See the folder [system](system) for any service and target files. - -The folder [workers](workers) contains an example configuration for the -`federation_reader` worker. Pay special attention to the name of the -configuration file. In order to work with the `matrix-synapse-worker@.service` -service, it needs to have the exact same name as the worker app. - -This setup expects neither the homeserver nor any workers to fork. Forking is -handled by systemd. - -## Setup - -1. Adjust your matrix configs. Make sure that the worker config files have the -exact same name as the worker app. Compare `matrix-synapse-worker@.service` for -why. You can find an example worker config in the [workers](workers) folder. See -below for relevant settings in the `homeserver.yaml`. -2. Copy the `*.service` and `*.target` files in [system](system) to -`/etc/systemd/system`. -3. `systemctl enable matrix-synapse.service` this adds the homeserver -app to the `matrix.target` -4. *Optional.* `systemctl enable -matrix-synapse-worker@federation_reader.service` this adds the federation_reader -app to the `matrix-synapse.service` -5. *Optional.* Repeat step 4 for any additional workers you require. -6. *Optional.* Add any bots or bridges by enabling them. -7. Start all matrix related services via `systemctl start matrix.target` -8. *Optional.* Enable autostart of all matrix related services on system boot -via `systemctl enable matrix.target` - -## Usage - -After you have setup you can use the following commands to manage your synapse -installation: - -``` -# Start matrix-synapse, all workers and any enabled bots or bridges. -systemctl start matrix.target - -# Restart matrix-synapse and all workers (not necessarily restarting bots -# or bridges, see "Bots and Bridges") -systemctl restart matrix-synapse.service - -# Stop matrix-synapse and all workers (not necessarily restarting bots -# or bridges, see "Bots and Bridges") -systemctl stop matrix-synapse.service - -# Restart a specific worker (i. e. federation_reader), the homeserver is -# unaffected by this. -systemctl restart matrix-synapse-worker@federation_reader.service - -# Add a new worker (assuming all configs are setup already) -systemctl enable matrix-synapse-worker@federation_writer.service -systemctl restart matrix-synapse.service -``` - -## The Configs - -Make sure the `worker_app` is set in the `homeserver.yaml` and it does not fork. - -``` -worker_app: synapse.app.homeserver -daemonize: false -``` - -None of the workers should fork, as forking is handled by systemd. Hence make -sure this is present in all worker config files. - -``` -worker_daemonize: false -``` - -The config files of all workers are expected to be located in -`/etc/matrix-synapse/workers`. If you want to use a different location you have -to edit the provided `*.service` files accordingly. - -## Bots and Bridges - -Most bots and bridges do not care if the homeserver goes down or is restarted. -Depending on the implementation this may crash them though. So look up the docs -or ask the community of the specific bridge or bot you want to run to make sure -you choose the correct setup. - -Whichever configuration you choose, after the setup the following will enable -automatically starting (and potentially restarting) your bot/bridge with the -`matrix.target`. - -``` -systemctl enable .service -``` - -**Note** that from an inactive synapse the bots/bridges will only be started with -synapse if you start the `matrix.target`, not if you start the -`matrix-synapse.service`. This is on purpose. Think of `matrix-synapse.service` -as *just* synapse, but `matrix.target` being anything matrix related, including -synapse and any and all enabled bots and bridges. - -### Start with synapse but ignore synapse going down - -If the bridge can handle shutdowns of the homeserver you'll want to install the -service in the `matrix.target` and optionally add a -`After=matrix-synapse.service` dependency to have the bot/bridge start after -synapse on starting everything. - -In this case the service file should look like this. - -``` -[Unit] -# ... -# Optional, this will only ensure that if you start everything, synapse will -# be started before the bot/bridge will be started. -After=matrix-synapse.service - -[Service] -# ... - -[Install] -WantedBy=matrix.target -``` - -### Stop/restart when synapse stops/restarts - -If the bridge can't handle shutdowns of the homeserver you'll still want to -install the service in the `matrix.target` but also have to specify the -`After=matrix-synapse.service` *and* `BindsTo=matrix-synapse.service` -dependencies to have the bot/bridge stop/restart with synapse. - -In this case the service file should look like this. - -``` -[Unit] -# ... -# Mandatory -After=matrix-synapse.service -BindsTo=matrix-synapse.service - -[Service] -# ... - -[Install] -WantedBy=matrix.target -``` +The documentation for using systemd to manage synapse workers is now part of +the main synapse distribution. See [docs/systemd-with-workers](../../docs/systemd-with-workers). diff --git a/contrib/systemd-with-workers/system/matrix-synapse-worker@.service b/contrib/systemd-with-workers/system/matrix-synapse-worker@.service deleted file mode 100644 index 3507e2e98..000000000 --- a/contrib/systemd-with-workers/system/matrix-synapse-worker@.service +++ /dev/null @@ -1,19 +0,0 @@ -[Unit] -Description=Synapse Matrix Worker -After=matrix-synapse.service -BindsTo=matrix-synapse.service - -[Service] -Type=notify -NotifyAccess=main -User=matrix-synapse -WorkingDirectory=/var/lib/matrix-synapse -EnvironmentFile=/etc/default/matrix-synapse -ExecStart=/opt/venvs/matrix-synapse/bin/python -m synapse.app.%i --config-path=/etc/matrix-synapse/homeserver.yaml --config-path=/etc/matrix-synapse/conf.d/ --config-path=/etc/matrix-synapse/workers/%i.yaml -ExecReload=/bin/kill -HUP $MAINPID -Restart=always -RestartSec=3 -SyslogIdentifier=matrix-synapse-%i - -[Install] -WantedBy=matrix-synapse.service diff --git a/contrib/systemd-with-workers/system/matrix.target b/contrib/systemd-with-workers/system/matrix.target deleted file mode 100644 index aff97d03e..000000000 --- a/contrib/systemd-with-workers/system/matrix.target +++ /dev/null @@ -1,7 +0,0 @@ -[Unit] -Description=Contains matrix services like synapse, bridges and bots -After=network.target -AllowIsolate=no - -[Install] -WantedBy=multi-user.target diff --git a/debian/build_virtualenv b/debian/build_virtualenv index d892fd5c9..4c9aabcac 100755 --- a/debian/build_virtualenv +++ b/debian/build_virtualenv @@ -36,7 +36,6 @@ esac dh_virtualenv \ --install-suffix "matrix-synapse" \ --builtin-venv \ - --setuptools \ --python "$SNAKE" \ --upgrade-pip \ --preinstall="lxml" \ diff --git a/debian/changelog b/debian/changelog index 8eaca8523..e7842d417 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,13 @@ +matrix-synapse-py3 (1.13.0) stable; urgency=medium + + [ Patrick Cloke ] + * Add information about .well-known files to Debian installation scripts. + + [ Synapse Packaging team ] + * New synapse release 1.13.0. + + -- Synapse Packaging team Tue, 19 May 2020 09:16:56 -0400 + matrix-synapse-py3 (1.12.4) stable; urgency=medium * New synapse release 1.12.4. diff --git a/debian/po/templates.pot b/debian/po/templates.pot index 84d960761..f0af9e70f 100644 --- a/debian/po/templates.pot +++ b/debian/po/templates.pot @@ -1,14 +1,14 @@ # SOME DESCRIPTIVE TITLE. # Copyright (C) YEAR THE PACKAGE'S COPYRIGHT HOLDER -# This file is distributed under the same license as the matrix-synapse package. +# This file is distributed under the same license as the matrix-synapse-py3 package. # FIRST AUTHOR , YEAR. # #, fuzzy msgid "" msgstr "" -"Project-Id-Version: matrix-synapse\n" -"Report-Msgid-Bugs-To: matrix-synapse@packages.debian.org\n" -"POT-Creation-Date: 2017-02-21 07:51+0000\n" +"Project-Id-Version: matrix-synapse-py3\n" +"Report-Msgid-Bugs-To: matrix-synapse-py3@packages.debian.org\n" +"POT-Creation-Date: 2020-04-06 16:39-0400\n" "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n" "Last-Translator: FULL NAME \n" "Language-Team: LANGUAGE \n" @@ -28,7 +28,10 @@ msgstr "" #: ../templates:1001 msgid "" "The name that this homeserver will appear as, to clients and other servers " -"via federation. This name should match the SRV record published in DNS." +"via federation. This is normally the public hostname of the server running " +"synapse, but can be different if you set up delegation. Please refer to the " +"delegation documentation in this case: https://github.com/matrix-org/synapse/" +"blob/master/docs/delegate.md." msgstr "" #. Type: boolean diff --git a/debian/templates b/debian/templates index 647358731..458fe8bbe 100644 --- a/debian/templates +++ b/debian/templates @@ -2,8 +2,10 @@ Template: matrix-synapse/server-name Type: string _Description: Name of the server: The name that this homeserver will appear as, to clients and other - servers via federation. This name should match the SRV record - published in DNS. + servers via federation. This is normally the public hostname of the + server running synapse, but can be different if you set up delegation. + Please refer to the delegation documentation in this case: + https://github.com/matrix-org/synapse/blob/master/docs/delegate.md. Template: matrix-synapse/report-stats Type: boolean diff --git a/docker/Dockerfile-dhvirtualenv b/docker/Dockerfile-dhvirtualenv index ac9ebcfd8..579724685 100644 --- a/docker/Dockerfile-dhvirtualenv +++ b/docker/Dockerfile-dhvirtualenv @@ -27,15 +27,16 @@ RUN env DEBIAN_FRONTEND=noninteractive apt-get install \ wget # fetch and unpack the package -RUN wget -q -O /dh-virtuenv-1.1.tar.gz https://github.com/spotify/dh-virtualenv/archive/1.1.tar.gz -RUN tar xvf /dh-virtuenv-1.1.tar.gz +RUN mkdir /dh-virtualenv +RUN wget -q -O /dh-virtualenv.tar.gz https://github.com/matrix-org/dh-virtualenv/archive/matrixorg-20200519.tar.gz +RUN tar -xv --strip-components=1 -C /dh-virtualenv -f /dh-virtualenv.tar.gz # install its build deps -RUN cd dh-virtualenv-1.1/ \ - && env DEBIAN_FRONTEND=noninteractive mk-build-deps -ri -t "apt-get -yqq --no-install-recommends" +RUN cd /dh-virtualenv \ + && env DEBIAN_FRONTEND=noninteractive mk-build-deps -ri -t "apt-get -y --no-install-recommends" # build it -RUN cd dh-virtualenv-1.1 && dpkg-buildpackage -us -uc -b +RUN cd /dh-virtualenv && dpkg-buildpackage -us -uc -b ### ### Stage 1 @@ -68,12 +69,12 @@ RUN apt-get update -qq -o Acquire::Languages=none \ sqlite3 \ libpq-dev -COPY --from=builder /dh-virtualenv_1.1-1_all.deb / +COPY --from=builder /dh-virtualenv_1.2~dev-1_all.deb / # install dhvirtualenv. Update the apt cache again first, in case we got a # cached cache from docker the first time. RUN apt-get update -qq -o Acquire::Languages=none \ - && apt-get install -yq /dh-virtualenv_1.1-1_all.deb + && apt-get install -yq /dh-virtualenv_1.2~dev-1_all.deb WORKDIR /synapse/source ENTRYPOINT ["bash","/synapse/source/docker/build_debian.sh"] diff --git a/docs/admin_api/room_membership.md b/docs/admin_api/room_membership.md new file mode 100644 index 000000000..16736d3d3 --- /dev/null +++ b/docs/admin_api/room_membership.md @@ -0,0 +1,34 @@ +# Edit Room Membership API + +This API allows an administrator to join an user account with a given `user_id` +to a room with a given `room_id_or_alias`. You can only modify the membership of +local users. The server administrator must be in the room and have permission to +invite users. + +## Parameters + +The following parameters are available: + +* `user_id` - Fully qualified user: for example, `@user:server.com`. +* `room_id_or_alias` - The room identifier or alias to join: for example, + `!636q39766251:server.com`. + +## Usage + +``` +POST /_synapse/admin/v1/join/ + +{ + "user_id": "@user:server.com" +} +``` + +Including an `access_token` of a server admin. + +Response: + +``` +{ + "room_id": "!636q39766251:server.com" +} +``` diff --git a/docs/admin_api/rooms.md b/docs/admin_api/rooms.md index 2db457c1b..26fe8b867 100644 --- a/docs/admin_api/rooms.md +++ b/docs/admin_api/rooms.md @@ -11,8 +11,21 @@ The following query parameters are available: * `from` - Offset in the returned list. Defaults to `0`. * `limit` - Maximum amount of rooms to return. Defaults to `100`. * `order_by` - The method in which to sort the returned list of rooms. Valid values are: - - `alphabetical` - Rooms are ordered alphabetically by room name. This is the default. - - `size` - Rooms are ordered by the number of members. Largest to smallest. + - `alphabetical` - Same as `name`. This is deprecated. + - `size` - Same as `joined_members`. This is deprecated. + - `name` - Rooms are ordered alphabetically by room name. This is the default. + - `canonical_alias` - Rooms are ordered alphabetically by main alias address of the room. + - `joined_members` - Rooms are ordered by the number of members. Largest to smallest. + - `joined_local_members` - Rooms are ordered by the number of local members. Largest to smallest. + - `version` - Rooms are ordered by room version. Largest to smallest. + - `creator` - Rooms are ordered alphabetically by creator of the room. + - `encryption` - Rooms are ordered alphabetically by the end-to-end encryption algorithm. + - `federatable` - Rooms are ordered by whether the room is federatable. + - `public` - Rooms are ordered by visibility in room list. + - `join_rules` - Rooms are ordered alphabetically by join rules of the room. + - `guest_access` - Rooms are ordered alphabetically by guest access option of the room. + - `history_visibility` - Rooms are ordered alphabetically by visibility of history of the room. + - `state_events` - Rooms are ordered by number of state events. Largest to smallest. * `dir` - Direction of room order. Either `f` for forwards or `b` for backwards. Setting this value to `b` will reverse the above sort order. Defaults to `f`. * `search_term` - Filter rooms by their room name. Search term can be contained in any @@ -26,6 +39,16 @@ The following fields are possible in the JSON response body: - `name` - The name of the room. - `canonical_alias` - The canonical (main) alias address of the room. - `joined_members` - How many users are currently in the room. + - `joined_local_members` - How many local users are currently in the room. + - `version` - The version of the room as a string. + - `creator` - The `user_id` of the room creator. + - `encryption` - Algorithm of end-to-end encryption of messages. Is `null` if encryption is not active. + - `federatable` - Whether users on other servers can join this room. + - `public` - Whether the room is visible in room directory. + - `join_rules` - The type of rules used for users wishing to join this room. One of: ["public", "knock", "invite", "private"]. + - `guest_access` - Whether guests can join the room. One of: ["can_join", "forbidden"]. + - `history_visibility` - Who can see the room history. One of: ["invited", "joined", "shared", "world_readable"]. + - `state_events` - Total number of state_events of a room. Complexity of the room. * `offset` - The current pagination offset in rooms. This parameter should be used instead of `next_token` for room offset as `next_token` is not intended to be parsed. @@ -60,14 +83,34 @@ Response: "room_id": "!OGEhHVWSdvArJzumhm:matrix.org", "name": "Matrix HQ", "canonical_alias": "#matrix:matrix.org", - "joined_members": 8326 + "joined_members": 8326, + "joined_local_members": 2, + "version": "1", + "creator": "@foo:matrix.org", + "encryption": null, + "federatable": true, + "public": true, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 93534 }, ... (8 hidden items) ... { "room_id": "!xYvNcQPhnkrdUmYczI:matrix.org", "name": "This Week In Matrix (TWIM)", "canonical_alias": "#twim:matrix.org", - "joined_members": 314 + "joined_members": 314, + "joined_local_members": 20, + "version": "4", + "creator": "@foo:matrix.org", + "encryption": "m.megolm.v1.aes-sha2", + "federatable": true, + "public": false, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 8345 } ], "offset": 0, @@ -92,7 +135,17 @@ Response: "room_id": "!xYvNcQPhnkrdUmYczI:matrix.org", "name": "This Week In Matrix (TWIM)", "canonical_alias": "#twim:matrix.org", - "joined_members": 314 + "joined_members": 314, + "joined_local_members": 20, + "version": "4", + "creator": "@foo:matrix.org", + "encryption": "m.megolm.v1.aes-sha2", + "federatable": true, + "public": false, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 8 } ], "offset": 0, @@ -117,14 +170,34 @@ Response: "room_id": "!OGEhHVWSdvArJzumhm:matrix.org", "name": "Matrix HQ", "canonical_alias": "#matrix:matrix.org", - "joined_members": 8326 + "joined_members": 8326, + "joined_local_members": 2, + "version": "1", + "creator": "@foo:matrix.org", + "encryption": null, + "federatable": true, + "public": true, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 93534 }, ... (98 hidden items) ... { "room_id": "!xYvNcQPhnkrdUmYczI:matrix.org", "name": "This Week In Matrix (TWIM)", "canonical_alias": "#twim:matrix.org", - "joined_members": 314 + "joined_members": 314, + "joined_local_members": 20, + "version": "4", + "creator": "@foo:matrix.org", + "encryption": "m.megolm.v1.aes-sha2", + "federatable": true, + "public": false, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 8345 } ], "offset": 0, @@ -154,6 +227,16 @@ Response: "name": "Music Theory", "canonical_alias": "#musictheory:matrix.org", "joined_members": 127 + "joined_local_members": 2, + "version": "1", + "creator": "@foo:matrix.org", + "encryption": null, + "federatable": true, + "public": true, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 93534 }, ... (48 hidden items) ... { @@ -161,6 +244,16 @@ Response: "name": "weechat-matrix", "canonical_alias": "#weechat-matrix:termina.org.uk", "joined_members": 137 + "joined_local_members": 20, + "version": "4", + "creator": "@foo:termina.org.uk", + "encryption": null, + "federatable": true, + "public": true, + "join_rules": "invite", + "guest_access": null, + "history_visibility": "shared", + "state_events": 8345 } ], "offset": 100, diff --git a/docs/admin_api/user_admin_api.rst b/docs/admin_api/user_admin_api.rst index 9ce10119f..859d7f99e 100644 --- a/docs/admin_api/user_admin_api.rst +++ b/docs/admin_api/user_admin_api.rst @@ -33,12 +33,22 @@ with a body of: including an ``access_token`` of a server admin. -The parameter ``displayname`` is optional and defaults to ``user_id``. -The parameter ``threepids`` is optional. -The parameter ``avatar_url`` is optional. -The parameter ``admin`` is optional and defaults to 'false'. -The parameter ``deactivated`` is optional and defaults to 'false'. -The parameter ``password`` is optional. If provided the user's password is updated and all devices are logged out. +The parameter ``displayname`` is optional and defaults to the value of +``user_id``. + +The parameter ``threepids`` is optional and allows setting the third-party IDs +(email, msisdn) belonging to a user. + +The parameter ``avatar_url`` is optional. Must be a [MXC +URI](https://matrix.org/docs/spec/client_server/r0.6.0#matrix-content-mxc-uris). + +The parameter ``admin`` is optional and defaults to ``false``. + +The parameter ``deactivated`` is optional and defaults to ``false``. + +The parameter ``password`` is optional. If provided, the user's password is +updated and all devices are logged out. + If the user already exists then optional parameters default to the current value. List Accounts @@ -51,16 +61,25 @@ The api is:: GET /_synapse/admin/v2/users?from=0&limit=10&guests=false including an ``access_token`` of a server admin. -The parameters ``from`` and ``limit`` are required only for pagination. -By default, a ``limit`` of 100 is used. -The parameter ``user_id`` can be used to select only users with user ids that -contain this value. -The parameter ``guests=false`` can be used to exclude guest users, -default is to include guest users. -The parameter ``deactivated=true`` can be used to include deactivated users, -default is to exclude deactivated users. -If the endpoint does not return a ``next_token`` then there are no more users left. -It returns a JSON body like the following: + +The parameter ``from`` is optional but used for pagination, denoting the +offset in the returned results. This should be treated as an opaque value and +not explicitly set to anything other than the return value of ``next_token`` +from a previous call. + +The parameter ``limit`` is optional but is used for pagination, denoting the +maximum number of items to return in this call. Defaults to ``100``. + +The parameter ``user_id`` is optional and filters to only users with user IDs +that contain this value. + +The parameter ``guests`` is optional and if ``false`` will **exclude** guest users. +Defaults to ``true`` to include guest users. + +The parameter ``deactivated`` is optional and if ``true`` will **include** deactivated users. +Defaults to ``false`` to exclude deactivated users. + +A JSON body is returned with the following shape: .. code:: json @@ -72,19 +91,29 @@ It returns a JSON body like the following: "is_guest": 0, "admin": 0, "user_type": null, - "deactivated": 0 + "deactivated": 0, + "displayname": "", + "avatar_url": null }, { "name": "", "password_hash": "", "is_guest": 0, "admin": 1, "user_type": null, - "deactivated": 0 + "deactivated": 0, + "displayname": "", + "avatar_url": "" } ], - "next_token": "100" + "next_token": "100", + "total": 200 } +To paginate, check for ``next_token`` and if present, call the endpoint again +with ``from`` set to the value of ``next_token``. This will return a new page. + +If the endpoint does not return a ``next_token`` then there are no more users +to paginate through. Query Account ============= diff --git a/docs/application_services.md b/docs/application_services.md index 06cb79f1f..e4592010a 100644 --- a/docs/application_services.md +++ b/docs/application_services.md @@ -23,9 +23,13 @@ namespaces: users: # List of users we're interested in - exclusive: regex: + group_id: - ... aliases: [] # List of aliases we're interested in rooms: [] # List of room ids we're interested in ``` +`exclusive`: If enabled, only this application service is allowed to register users in its namespace(s). +`group_id`: All users of this application service are dynamically joined to this group. This is useful for e.g user organisation or flairs. + See the [spec](https://matrix.org/docs/spec/application_service/unstable.html) for further details on how application services work. diff --git a/docs/dev/cas.md b/docs/dev/cas.md new file mode 100644 index 000000000..f8d02cc82 --- /dev/null +++ b/docs/dev/cas.md @@ -0,0 +1,64 @@ +# How to test CAS as a developer without a server + +The [django-mama-cas](https://github.com/jbittel/django-mama-cas) project is an +easy to run CAS implementation built on top of Django. + +## Prerequisites + +1. Create a new virtualenv: `python3 -m venv ` +2. Activate your virtualenv: `source /path/to/your/virtualenv/bin/activate` +3. Install Django and django-mama-cas: + ``` + python -m pip install "django<3" "django-mama-cas==2.4.0" + ``` +4. Create a Django project in the current directory: + ``` + django-admin startproject cas_test . + ``` +5. Follow the [install directions](https://django-mama-cas.readthedocs.io/en/latest/installation.html#configuring) for django-mama-cas +6. Setup the SQLite database: `python manage.py migrate` +7. Create a user: + ``` + python manage.py createsuperuser + ``` + 1. Use whatever you want as the username and password. + 2. Leave the other fields blank. +8. Use the built-in Django test server to serve the CAS endpoints on port 8000: + ``` + python manage.py runserver + ``` + +You should now have a Django project configured to serve CAS authentication with +a single user created. + +## Configure Synapse (and Riot) to use CAS + +1. Modify your `homeserver.yaml` to enable CAS and point it to your locally + running Django test server: + ```yaml + cas_config: + enabled: true + server_url: "http://localhost:8000" + service_url: "http://localhost:8081" + #displayname_attribute: name + #required_attributes: + # name: value + ``` +2. Restart Synapse. + +Note that the above configuration assumes the homeserver is running on port 8081 +and that the CAS server is on port 8000, both on localhost. + +## Testing the configuration + +Then in Riot: + +1. Visit the login page with a Riot pointing at your homeserver. +2. Click the Single Sign-On button. +3. Login using the credentials created with `createsuperuser`. +4. You should be logged in. + +If you want to repeat this process you'll need to manually logout first: + +1. http://localhost:8000/admin/ +2. Click "logout" in the top right. diff --git a/docs/dev/saml.md b/docs/dev/saml.md index f41aadce4..a9bfd2dc0 100644 --- a/docs/dev/saml.md +++ b/docs/dev/saml.md @@ -18,9 +18,13 @@ To make Synapse (and therefore Riot) use it: metadata: local: ["samling.xml"] ``` -5. Run `apt-get install xmlsec1` and `pip install --upgrade --force 'pysaml2>=4.5.0'` to ensure +5. Ensure that your `homeserver.yaml` has a setting for `public_baseurl`: + ```yaml + public_baseurl: http://localhost:8080/ + ``` +6. Run `apt-get install xmlsec1` and `pip install --upgrade --force 'pysaml2>=4.5.0'` to ensure the dependencies are installed and ready to go. -6. Restart Synapse. +7. Restart Synapse. Then in Riot: diff --git a/docs/log_contexts.md b/docs/log_contexts.md index 5331e8c88..fe30ca279 100644 --- a/docs/log_contexts.md +++ b/docs/log_contexts.md @@ -29,14 +29,13 @@ from synapse.logging import context # omitted from future snippets def handle_request(request_id): request_context = context.LoggingContext() - calling_context = context.LoggingContext.current_context() - context.LoggingContext.set_current_context(request_context) + calling_context = context.set_current_context(request_context) try: request_context.request = request_id do_request_handling() logger.debug("finished") finally: - context.LoggingContext.set_current_context(calling_context) + context.set_current_context(calling_context) def do_request_handling(): logger.debug("phew") # this will be logged against request_id diff --git a/docs/metrics-howto.md b/docs/metrics-howto.md index 32abb9f44..cf69938a2 100644 --- a/docs/metrics-howto.md +++ b/docs/metrics-howto.md @@ -60,6 +60,31 @@ 1. Restart Prometheus. +## Monitoring workers + +To monitor a Synapse installation using +[workers](https://github.com/matrix-org/synapse/blob/master/docs/workers.md), +every worker needs to be monitored independently, in addition to +the main homeserver process. This is because workers don't send +their metrics to the main homeserver process, but expose them +directly (if they are configured to do so). + +To allow collecting metrics from a worker, you need to add a +`metrics` listener to its configuration, by adding the following +under `worker_listeners`: + +```yaml + - type: metrics + bind_address: '' + port: 9101 +``` + +The `bind_address` and `port` parameters should be set so that +the resulting listener can be reached by prometheus, and they +don't clash with an existing worker. +With this example, the worker's metrics would then be available +on `http://127.0.0.1:9101`. + ## Renaming of metrics & deprecation of old names in 1.2 Synapse 1.2 updates the Prometheus metrics to match the naming diff --git a/docs/password_auth_providers.md b/docs/password_auth_providers.md index 0db1a3804..5d9ae6704 100644 --- a/docs/password_auth_providers.md +++ b/docs/password_auth_providers.md @@ -9,7 +9,11 @@ into Synapse, and provides a number of methods by which it can integrate with the authentication system. This document serves as a reference for those looking to implement their -own password auth providers. +own password auth providers. Additionally, here is a list of known +password auth provider module implementations: + +* [matrix-synapse-ldap3](https://github.com/matrix-org/matrix-synapse-ldap3/) +* [matrix-synapse-shared-secret-auth](https://github.com/devture/matrix-synapse-shared-secret-auth) ## Required methods diff --git a/docs/postgres.md b/docs/postgres.md index ca7ef1cf3..70fe29cdc 100644 --- a/docs/postgres.md +++ b/docs/postgres.md @@ -61,7 +61,33 @@ Note that the PostgreSQL database *must* have the correct encoding set You may need to enable password authentication so `synapse_user` can connect to the database. See -. +. + +If you get an error along the lines of `FATAL: Ident authentication failed for +user "synapse_user"`, you may need to use an authentication method other than +`ident`: + +* If the `synapse_user` user has a password, add the password to the `database:` + section of `homeserver.yaml`. Then add the following to `pg_hba.conf`: + + ``` + host synapse synapse_user ::1/128 md5 # or `scram-sha-256` instead of `md5` if you use that + ``` + +* If the `synapse_user` user does not have a password, then a password doesn't + have to be added to `homeserver.yaml`. But the following does need to be added + to `pg_hba.conf`: + + ``` + host synapse synapse_user ::1/128 trust + ``` + +Note that line order matters in `pg_hba.conf`, so make sure that if you do add a +new line, it is inserted before: + +``` +host all all ::1/128 ident +``` ### Fixing incorrect `COLLATE` or `CTYPE` @@ -104,19 +130,41 @@ of free memory the database host has available. When you are ready to start using PostgreSQL, edit the `database` section in your config file to match the following lines: - database: - name: psycopg2 - args: - user: - password: - database: - host: - cp_min: 5 - cp_max: 10 +```yaml +database: + name: psycopg2 + args: + user: + password: + database: + host: + cp_min: 5 + cp_max: 10 +``` All key, values in `args` are passed to the `psycopg2.connect(..)` function, except keys beginning with `cp_`, which are consumed by the -twisted adbapi connection pool. +twisted adbapi connection pool. See the [libpq +documentation](https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-PARAMKEYWORDS) +for a list of options which can be passed. + +You should consider tuning the `args.keepalives_*` options if there is any danger of +the connection between your homeserver and database dropping, otherwise Synapse +may block for an extended period while it waits for a response from the +database server. Example values might be: + +```yaml +# seconds of inactivity after which TCP should send a keepalive message to the server +keepalives_idle: 10 + +# the number of seconds after which a TCP keepalive message that is not +# acknowledged by the server should be retransmitted +keepalives_interval: 10 + +# the number of TCP keepalives that can be lost before the client's connection +# to the server is considered dead +keepalives_count: 3 +``` ## Porting from SQLite diff --git a/docs/reverse_proxy.md b/docs/reverse_proxy.md index 50f4a0f48..7c300023c 100644 --- a/docs/reverse_proxy.md +++ b/docs/reverse_proxy.md @@ -42,6 +42,9 @@ the reverse proxy and the homeserver. location /_matrix { proxy_pass http://localhost:8008; proxy_set_header X-Forwarded-For $remote_addr; + # Nginx by default only allows file uploads up to 1M in size + # Increase client_max_body_size to match max_upload_size defined in homeserver.yaml + client_max_body_size 10M; } } diff --git a/docs/sample_config.yaml b/docs/sample_config.yaml index 2ff0dd05a..98ead7dc0 100644 --- a/docs/sample_config.yaml +++ b/docs/sample_config.yaml @@ -33,10 +33,15 @@ server_name: "SERVERNAME" # pid_file: DATADIR/homeserver.pid -# The path to the web client which will be served at /_matrix/client/ -# if 'webclient' is configured under the 'listeners' configuration. +# The absolute URL to the web client which /_matrix/client will redirect +# to if 'webclient' is configured under the 'listeners' configuration. # -#web_client_location: "/path/to/web/root" +# This option can be also set to the filesystem path to the web client +# which will be served at /_matrix/client/ if 'webclient' is configured +# under the 'listeners' configuration, however this is a security risk: +# https://github.com/matrix-org/synapse#security-note +# +#web_client_location: https://riot.example.com/ # The public-facing base URL that clients use to access this HS # (not including _matrix/...). This is the same URL a user would @@ -248,6 +253,18 @@ listeners: # bind_addresses: ['::1', '127.0.0.1'] # type: manhole +# Forward extremities can build up in a room due to networking delays between +# homeservers. Once this happens in a large room, calculation of the state of +# that room can become quite expensive. To mitigate this, once the number of +# forward extremities reaches a given threshold, Synapse will send an +# org.matrix.dummy_event event, which will reduce the forward extremities +# in the room. +# +# This setting defines the threshold (i.e. number of forward extremities in the +# room) at which dummy events are sent. The default value is 10. +# +#dummy_events_threshold: 5 + ## Homeserver blocking ## @@ -409,6 +426,16 @@ retention: # longest_max_lifetime: 1y # interval: 1d +# Inhibits the /requestToken endpoints from returning an error that might leak +# information about whether an e-mail address is in use or not on this +# homeserver. +# Note that for some endpoints the error situation is the e-mail already being +# used, and for others the error is entering the e-mail being unused. +# If this option is enabled, instead of returning an error, these endpoints will +# act as if no error happened and return a fake session ID ('sid') to clients. +# +#request_token_inhibit_3pid_errors: true + ## TLS ## @@ -578,13 +605,46 @@ acme: ## Database ## +# The 'database' setting defines the database that synapse uses to store all of +# its data. +# +# 'name' gives the database engine to use: either 'sqlite3' (for SQLite) or +# 'psycopg2' (for PostgreSQL). +# +# 'args' gives options which are passed through to the database engine, +# except for options starting 'cp_', which are used to configure the Twisted +# connection pool. For a reference to valid arguments, see: +# * for sqlite: https://docs.python.org/3/library/sqlite3.html#sqlite3.connect +# * for postgres: https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-PARAMKEYWORDS +# * for the connection pool: https://twistedmatrix.com/documents/current/api/twisted.enterprise.adbapi.ConnectionPool.html#__init__ +# +# +# Example SQLite configuration: +# +#database: +# name: sqlite3 +# args: +# database: /path/to/homeserver.db +# +# +# Example Postgres configuration: +# +#database: +# name: psycopg2 +# args: +# user: synapse +# password: secretpassword +# database: synapse +# host: localhost +# cp_min: 5 +# cp_max: 10 +# +# For more information on using Synapse with Postgres, see `docs/postgres.md`. +# database: - # The database engine name - name: "sqlite3" - # Arguments to pass to the engine + name: sqlite3 args: - # Path to the database - database: "DATADIR/homeserver.db" + database: DATADIR/homeserver.db # Number of events to cache in memory. # @@ -697,12 +757,11 @@ media_store_path: "DATADIR/media_store" # #media_storage_providers: # - module: file_system -# # Whether to write new local files. +# # Whether to store newly uploaded local files # store_local: false -# # Whether to write new remote media +# # Whether to store newly downloaded remote files # store_remote: false -# # Whether to block upload requests waiting for write to this -# # provider to complete +# # Whether to wait for successful storage for local uploads # store_synchronous: false # config: # directory: /mnt/some/other/directory @@ -821,6 +880,31 @@ media_store_path: "DATADIR/media_store" # #max_spider_size: 10M +# A list of values for the Accept-Language HTTP header used when +# downloading webpages during URL preview generation. This allows +# Synapse to specify the preferred languages that URL previews should +# be in when communicating with remote servers. +# +# Each value is a IETF language tag; a 2-3 letter identifier for a +# language, optionally followed by subtags separated by '-', specifying +# a country or region variant. +# +# Multiple values can be provided, and a weight can be added to each by +# using quality value syntax (;q=). '*' translates to any language. +# +# Defaults to "en". +# +# Example: +# +# url_preview_accept_language: +# - en-UK +# - en-US;q=0.9 +# - fr;q=0.8 +# - *;q=0.7 +# +url_preview_accept_language: +# - en + ## Captcha ## # See docs/CAPTCHA_SETUP for full details of configuring this. @@ -839,10 +923,6 @@ media_store_path: "DATADIR/media_store" # #enable_registration_captcha: false -# A secret key used to bypass the captcha test entirely. -# -#captcha_bypass_secret: "YOUR_SECRET_HERE" - # The API endpoint to use for verifying m.login.recaptcha responses. # #recaptcha_siteverify_api: "https://www.recaptcha.net/recaptcha/api/siteverify" @@ -1057,6 +1137,29 @@ account_threepid_delegates: #email: https://example.com # Delegate email sending to example.com #msisdn: http://localhost:8090 # Delegate SMS sending to this local process +# Whether users are allowed to change their displayname after it has +# been initially set. Useful when provisioning users based on the +# contents of a third-party directory. +# +# Does not apply to server administrators. Defaults to 'true' +# +#enable_set_displayname: false + +# Whether users are allowed to change their avatar after it has been +# initially set. Useful when provisioning users based on the contents +# of a third-party directory. +# +# Does not apply to server administrators. Defaults to 'true' +# +#enable_set_avatar_url: false + +# Whether users can change the 3PIDs associated with their accounts +# (email address and msisdn). +# +# Defaults to 'true' +# +#enable_3pid_changes: false + # Users who register on this homeserver will automatically be joined # to these rooms # @@ -1092,7 +1195,7 @@ account_threepid_delegates: # enabled by default, either for performance reasons or limited use. # metrics_flags: - # Publish synapse_federation_known_servers, a g auge of the number of + # Publish synapse_federation_known_servers, a gauge of the number of # servers this homeserver knows about, including itself. May cause # performance problems on large homeservers. # @@ -1258,32 +1361,32 @@ saml2_config: # remote: # - url: https://our_idp/metadata.xml # - # # By default, the user has to go to our login page first. If you'd like - # # to allow IdP-initiated login, set 'allow_unsolicited: true' in a - # # 'service.sp' section: - # # - # #service: - # # sp: - # # allow_unsolicited: true + # # By default, the user has to go to our login page first. If you'd like + # # to allow IdP-initiated login, set 'allow_unsolicited: true' in a + # # 'service.sp' section: + # # + # #service: + # # sp: + # # allow_unsolicited: true # - # # The examples below are just used to generate our metadata xml, and you - # # may well not need them, depending on your setup. Alternatively you - # # may need a whole lot more detail - see the pysaml2 docs! + # # The examples below are just used to generate our metadata xml, and you + # # may well not need them, depending on your setup. Alternatively you + # # may need a whole lot more detail - see the pysaml2 docs! # - # description: ["My awesome SP", "en"] - # name: ["Test SP", "en"] + # description: ["My awesome SP", "en"] + # name: ["Test SP", "en"] # - # organization: - # name: Example com - # display_name: - # - ["Example co", "en"] - # url: "http://example.com" + # organization: + # name: Example com + # display_name: + # - ["Example co", "en"] + # url: "http://example.com" # - # contact_person: - # - given_name: Bob - # sur_name: "the Sysadmin" - # email_address": ["admin@example.com"] - # contact_type": technical + # contact_person: + # - given_name: Bob + # sur_name: "the Sysadmin" + # email_address": ["admin@example.com"] + # contact_type": technical # Instead of putting the config inline as above, you can specify a # separate pysaml2 configuration file: @@ -1392,6 +1495,10 @@ sso: # phishing attacks from evil.site. To avoid this, include a slash after the # hostname: "https://my.client/". # + # If public_baseurl is set, then the login fallback page (used by clients + # that don't natively support the required login flows) is whitelisted in + # addition to any URLs in this list. + # # By default, this list is empty. # #client_whitelist: @@ -1423,6 +1530,30 @@ sso: # # * server_name: the homeserver's name. # + # * HTML page which notifies the user that they are authenticating to confirm + # an operation on their account during the user interactive authentication + # process: 'sso_auth_confirm.html'. + # + # When rendering, this template is given the following variables: + # * redirect_url: the URL the user is about to be redirected to. Needs + # manual escaping (see + # https://jinja.palletsprojects.com/en/2.11.x/templates/#html-escaping). + # + # * description: the operation which the user is being asked to confirm + # + # * HTML page shown after a successful user interactive authentication session: + # 'sso_auth_success.html'. + # + # Note that this page must include the JavaScript which notifies of a successful authentication + # (see https://matrix.org/docs/spec/client_server/r0.6.0#fallback). + # + # This template has no additional variables. + # + # * HTML page shown during single sign-on if a deactivated user (according to Synapse's database) + # attempts to login: 'sso_account_deactivated.html'. + # + # This template has no additional variables. + # # You can see the default templates at: # https://github.com/matrix-org/synapse/tree/master/synapse/res/templates # @@ -1453,6 +1584,41 @@ password_config: # #pepper: "EVEN_MORE_SECRET" + # Define and enforce a password policy. Each parameter is optional. + # This is an implementation of MSC2000. + # + policy: + # Whether to enforce the password policy. + # Defaults to 'false'. + # + #enabled: true + + # Minimum accepted length for a password. + # Defaults to 0. + # + #minimum_length: 15 + + # Whether a password must contain at least one digit. + # Defaults to 'false'. + # + #require_digit: true + + # Whether a password must contain at least one symbol. + # A symbol is any character that's not a number or a letter. + # Defaults to 'false'. + # + #require_symbol: true + + # Whether a password must contain at least one lowercase letter. + # Defaults to 'false'. + # + #require_lowercase: true + + # Whether a password must contain at least one lowercase letter. + # Defaults to 'false'. + # + #require_uppercase: true + # Configuration for sending emails from Synapse. # @@ -1561,7 +1727,19 @@ email: #template_dir: "res/templates" -#password_providers: +# Password providers allow homeserver administrators to integrate +# their Synapse installation with existing authentication methods +# ex. LDAP, external tokens, etc. +# +# For more information and known implementations, please see +# https://github.com/matrix-org/synapse/blob/master/docs/password_auth_providers.md +# +# Note: instances wishing to use SAML or CAS authentication should +# instead use the `saml2_config` or `cas_config` options, +# respectively. +# +password_providers: +# # Example config for an LDAP auth provider # - module: "ldap_auth_provider.LdapAuthProvider" # config: # enabled: true diff --git a/docs/systemd-with-workers/README.md b/docs/systemd-with-workers/README.md new file mode 100644 index 000000000..257c09446 --- /dev/null +++ b/docs/systemd-with-workers/README.md @@ -0,0 +1,67 @@ +# Setting up Synapse with Workers and Systemd + +This is a setup for managing synapse with systemd, including support for +managing workers. It provides a `matrix-synapse` service for the master, as +well as a `matrix-synapse-worker@` service template for any workers you +require. Additionally, to group the required services, it sets up a +`matrix-synapse.target`. + +See the folder [system](system) for the systemd unit files. + +The folder [workers](workers) contains an example configuration for the +`federation_reader` worker. + +## Synapse configuration files + +See [workers.md](../workers.md) for information on how to set up the +configuration files and reverse-proxy correctly. You can find an example worker +config in the [workers](workers) folder. + +Systemd manages daemonization itself, so ensure that none of the configuration +files set either `daemonize` or `worker_daemonize`. + +The config files of all workers are expected to be located in +`/etc/matrix-synapse/workers`. If you want to use a different location, edit +the provided `*.service` files accordingly. + +There is no need for a separate configuration file for the master process. + +## Set up + +1. Adjust synapse configuration files as above. +1. Copy the `*.service` and `*.target` files in [system](system) to +`/etc/systemd/system`. +1. Run `systemctl deamon-reload` to tell systemd to load the new unit files. +1. Run `systemctl enable matrix-synapse.service`. This will configure the +synapse master process to be started as part of the `matrix-synapse.target` +target. +1. For each worker process to be enabled, run `systemctl enable +matrix-synapse-worker@.service`. For each ``, there +should be a corresponding configuration file +`/etc/matrix-synapse/workers/.yaml`. +1. Start all the synapse processes with `systemctl start matrix-synapse.target`. +1. Tell systemd to start synapse on boot with `systemctl enable matrix-synapse.target`/ + +## Usage + +Once the services are correctly set up, you can use the following commands +to manage your synapse installation: + +```sh +# Restart Synapse master and all workers +systemctl restart matrix-synapse.target + +# Stop Synapse and all workers +systemctl stop matrix-synapse.target + +# Restart the master alone +systemctl start matrix-synapse.service + +# Restart a specific worker (eg. federation_reader); the master is +# unaffected by this. +systemctl restart matrix-synapse-worker@federation_reader.service + +# Add a new worker (assuming all configs are set up already) +systemctl enable matrix-synapse-worker@federation_writer.service +systemctl restart matrix-synapse.target +``` diff --git a/docs/systemd-with-workers/system/matrix-synapse-worker@.service b/docs/systemd-with-workers/system/matrix-synapse-worker@.service new file mode 100644 index 000000000..70589a7a5 --- /dev/null +++ b/docs/systemd-with-workers/system/matrix-synapse-worker@.service @@ -0,0 +1,20 @@ +[Unit] +Description=Synapse %i + +# This service should be restarted when the synapse target is restarted. +PartOf=matrix-synapse.target + +[Service] +Type=notify +NotifyAccess=main +User=matrix-synapse +WorkingDirectory=/var/lib/matrix-synapse +EnvironmentFile=/etc/default/matrix-synapse +ExecStart=/opt/venvs/matrix-synapse/bin/python -m synapse.app.generic_worker --config-path=/etc/matrix-synapse/homeserver.yaml --config-path=/etc/matrix-synapse/conf.d/ --config-path=/etc/matrix-synapse/workers/%i.yaml +ExecReload=/bin/kill -HUP $MAINPID +Restart=always +RestartSec=3 +SyslogIdentifier=matrix-synapse-%i + +[Install] +WantedBy=matrix-synapse.target diff --git a/contrib/systemd-with-workers/system/matrix-synapse.service b/docs/systemd-with-workers/system/matrix-synapse.service similarity index 79% rename from contrib/systemd-with-workers/system/matrix-synapse.service rename to docs/systemd-with-workers/system/matrix-synapse.service index 68e8991f1..c7b5ddfa4 100644 --- a/contrib/systemd-with-workers/system/matrix-synapse.service +++ b/docs/systemd-with-workers/system/matrix-synapse.service @@ -1,5 +1,8 @@ [Unit] -Description=Synapse Matrix Homeserver +Description=Synapse master + +# This service should be restarted when the synapse target is restarted. +PartOf=matrix-synapse.target [Service] Type=notify @@ -15,4 +18,4 @@ RestartSec=3 SyslogIdentifier=matrix-synapse [Install] -WantedBy=matrix.target +WantedBy=matrix-synapse.target diff --git a/docs/systemd-with-workers/system/matrix-synapse.target b/docs/systemd-with-workers/system/matrix-synapse.target new file mode 100644 index 000000000..e0eba1b34 --- /dev/null +++ b/docs/systemd-with-workers/system/matrix-synapse.target @@ -0,0 +1,6 @@ +[Unit] +Description=Synapse parent target +After=network.target + +[Install] +WantedBy=multi-user.target diff --git a/contrib/systemd-with-workers/workers/federation_reader.yaml b/docs/systemd-with-workers/workers/federation_reader.yaml similarity index 92% rename from contrib/systemd-with-workers/workers/federation_reader.yaml rename to docs/systemd-with-workers/workers/federation_reader.yaml index 47c54ec0d..5b65c7040 100644 --- a/contrib/systemd-with-workers/workers/federation_reader.yaml +++ b/docs/systemd-with-workers/workers/federation_reader.yaml @@ -10,5 +10,4 @@ worker_listeners: resources: - names: [federation] -worker_daemonize: false worker_log_config: /etc/matrix-synapse/federation-reader-log.yaml diff --git a/docs/tcp_replication.md b/docs/tcp_replication.md index e3a4634b1..ab2fffbfe 100644 --- a/docs/tcp_replication.md +++ b/docs/tcp_replication.md @@ -14,16 +14,18 @@ example flow would be (where '>' indicates master to worker and '<' worker to master flows): > SERVER example.com - < REPLICATE events 53 - > RDATA events 54 ["$foo1:bar.com", ...] - > RDATA events 55 ["$foo4:bar.com", ...] + < REPLICATE + > POSITION events master 53 + > RDATA events master 54 ["$foo1:bar.com", ...] + > RDATA events master 55 ["$foo4:bar.com", ...] -The example shows the server accepting a new connection and sending its -identity with the `SERVER` command, followed by the client asking to -subscribe to the `events` stream from the token `53`. The server then -periodically sends `RDATA` commands which have the format -`RDATA `, where the format of `` is -defined by the individual streams. +The example shows the server accepting a new connection and sending its identity +with the `SERVER` command, followed by the client server to respond with the +position of all streams. The server then periodically sends `RDATA` commands +which have the format `RDATA `, where +the format of `` is defined by the individual streams. The +`` is the name of the Synapse process that generated the data +(usually "master"). Error reporting happens by either the client or server sending an ERROR command, and usually the connection will be closed. @@ -32,9 +34,6 @@ Since the protocol is a simple line based, its possible to manually connect to the server using a tool like netcat. A few things should be noted when manually using the protocol: -- When subscribing to a stream using `REPLICATE`, the special token - `NOW` can be used to get all future updates. The special stream name - `ALL` can be used with `NOW` to subscribe to all available streams. - The federation stream is only available if federation sending has been disabled on the main process. - The server will only time connections out that have sent a `PING` @@ -55,7 +54,7 @@ The basic structure of the protocol is line based, where the initial word of each line specifies the command. The rest of the line is parsed based on the command. For example, the RDATA command is defined as: - RDATA + RDATA (Note that may contains spaces, but cannot contain newlines.) @@ -91,9 +90,7 @@ The client: - Sends a `NAME` command, allowing the server to associate a human friendly name with the connection. This is optional. - Sends a `PING` as above -- For each stream the client wishes to subscribe to it sends a - `REPLICATE` with the `stream_name` and token it wants to subscribe - from. +- Sends a `REPLICATE` to get the current position of all streams. - On receipt of a `SERVER` command, checks that the server name matches the expected server name. @@ -140,14 +137,12 @@ the wire: > PING 1490197665618 < NAME synapse.app.appservice < PING 1490197665618 - < REPLICATE events 1 - < REPLICATE backfill 1 - < REPLICATE caches 1 - > POSITION events 1 - > POSITION backfill 1 - > POSITION caches 1 - > RDATA caches 2 ["get_user_by_id",["@01register-user:localhost:8823"],1490197670513] - > RDATA events 14 ["$149019767112vOHxz:localhost:8823", + < REPLICATE + > POSITION events master 1 + > POSITION backfill master 1 + > POSITION caches master 1 + > RDATA caches master 2 ["get_user_by_id",["@01register-user:localhost:8823"],1490197670513] + > RDATA events master 14 ["$149019767112vOHxz:localhost:8823", "!AFDCvgApUmpdfVjIXm:localhost:8823","m.room.guest_access","",null] < PING 1490197675618 > ERROR server stopping @@ -158,10 +153,10 @@ position without needing to send data with the `RDATA` command. An example of a batched set of `RDATA` is: - > RDATA caches batch ["get_user_by_id",["@test:localhost:8823"],1490197670513] - > RDATA caches batch ["get_user_by_id",["@test2:localhost:8823"],1490197670513] - > RDATA caches batch ["get_user_by_id",["@test3:localhost:8823"],1490197670513] - > RDATA caches 54 ["get_user_by_id",["@test4:localhost:8823"],1490197670513] + > RDATA caches master batch ["get_user_by_id",["@test:localhost:8823"],1490197670513] + > RDATA caches master batch ["get_user_by_id",["@test2:localhost:8823"],1490197670513] + > RDATA caches master batch ["get_user_by_id",["@test3:localhost:8823"],1490197670513] + > RDATA caches master 54 ["get_user_by_id",["@test4:localhost:8823"],1490197670513] In this case the client shouldn't advance their caches token until it sees the the last `RDATA`. @@ -181,9 +176,14 @@ client (C): #### POSITION (S) - The position of the stream has been updated. Sent to the client - after all missing updates for a stream have been sent to the client - and they're now up to date. + On receipt of a POSITION command clients should check if they have missed any + updates, and if so then fetch them out of band. Sent in response to a + REPLICATE command (but can happen at any time). + + The POSITION command includes the source of the stream. Currently all streams + are written by a single process (usually "master"). If fetching missing + updates via HTTP API, rather than via the DB, then processes should make the + request to the appropriate process. #### ERROR (S, C) @@ -199,24 +199,17 @@ client (C): #### REPLICATE (C) -Asks the server to replicate a given stream. The syntax is: - -``` - REPLICATE -``` - -Where `` may be either: - * a numeric stream_id to stream updates since (exclusive) - * `NOW` to stream all subsequent updates. - -The `` is the name of a replication stream to subscribe -to (see [here](../synapse/replication/tcp/streams/_base.py) for a list -of streams). It can also be `ALL` to subscribe to all known streams, -in which case the `` must be set to `NOW`. +Asks the server for the current position of all streams. #### USER_SYNC (C) - A user has started or stopped syncing + A user has started or stopped syncing on this process. + +#### CLEAR_USER_SYNC (C) + + The server should clear all associated user sync data from the worker. + + This is used when a worker is shutting down. #### FEDERATION_ACK (C) @@ -230,10 +223,6 @@ in which case the `` must be set to `NOW`. Inform the server a cache should be invalidated -#### SYNC (S, C) - - Used exclusively in tests - ### REMOTE_SERVER_UP (S, C) Inform other processes that a remote server may have come back online. @@ -252,12 +241,12 @@ Each individual cache invalidation results in a row being sent down replication, which includes the cache name (the name of the function) and they key to invalidate. For example: - > RDATA caches 550953771 ["get_user_by_id", ["@bob:example.com"], 1550574873251] + > RDATA caches master 550953771 ["get_user_by_id", ["@bob:example.com"], 1550574873251] Alternatively, an entire cache can be invalidated by sending down a `null` instead of the key. For example: - > RDATA caches 550953772 ["get_user_by_id", null, 1550574873252] + > RDATA caches master 550953772 ["get_user_by_id", null, 1550574873252] However, there are times when a number of caches need to be invalidated at the same time with the same key. To reduce traffic we batch those diff --git a/docs/turn-howto.md b/docs/turn-howto.md index 1e121ead6..b8a2ba3e8 100644 --- a/docs/turn-howto.md +++ b/docs/turn-howto.md @@ -11,6 +11,13 @@ TURN server. The following sections describe how to install [coturn]() (which implements the TURN REST API) and integrate it with synapse. +## Requirements + +For TURN relaying with `coturn` to work, it must be hosted on a server/endpoint with a public IP. + +Hosting TURN behind a NAT (even with appropriate port forwarding) is known to cause issues +and to often not work. + ## `coturn` Setup ### Initial installation diff --git a/docs/workers.md b/docs/workers.md index cb3b9f8e6..7512eff43 100644 --- a/docs/workers.md +++ b/docs/workers.md @@ -1,23 +1,31 @@ # Scaling synapse via workers -Synapse has experimental support for splitting out functionality into -multiple separate python processes, helping greatly with scalability. These +For small instances it recommended to run Synapse in monolith mode (the +default). For larger instances where performance is a concern it can be helpful +to split out functionality into multiple separate python processes. These processes are called 'workers', and are (eventually) intended to scale horizontally independently. -All of the below is highly experimental and subject to change as Synapse evolves, -but documenting it here to help folks needing highly scalable Synapses similar -to the one running matrix.org! +Synapse's worker support is under active development and subject to change as +we attempt to rapidly scale ever larger Synapse instances. However we are +documenting it here to help admins needing a highly scalable Synapse instance +similar to the one running `matrix.org`. -All processes continue to share the same database instance, and as such, workers -only work with postgres based synapse deployments (sharing a single sqlite -across multiple processes is a recipe for disaster, plus you should be using -postgres anyway if you care about scalability). +All processes continue to share the same database instance, and as such, +workers only work with PostgreSQL-based Synapse deployments. SQLite should only +be used for demo purposes and any admin considering workers should already be +running PostgreSQL. -The workers communicate with the master synapse process via a synapse-specific -TCP protocol called 'replication' - analogous to MySQL or Postgres style -database replication; feeding a stream of relevant data to the workers so they -can be kept in sync with the main synapse process and database state. +## Master/worker communication + +The workers communicate with the master process via a Synapse-specific protocol +called 'replication' (analogous to MySQL- or Postgres-style database +replication) which feeds a stream of relevant data from the master to the +workers so they can be kept in sync with the master process and database state. + +Additionally, workers may make HTTP requests to the master, to send information +in the other direction. Typically this is used for operations which need to +wait for a reply - such as sending an event. ## Configuration @@ -27,72 +35,61 @@ the correct worker, or to the main synapse instance. Note that this includes requests made to the federation port. See [reverse_proxy.md](reverse_proxy.md) for information on setting up a reverse proxy. -To enable workers, you need to add two replication listeners to the master -synapse, e.g.: +To enable workers, you need to add *two* replication listeners to the +main Synapse configuration file (`homeserver.yaml`). For example: - listeners: - # The TCP replication port - - port: 9092 - bind_address: '127.0.0.1' - type: replication - # The HTTP replication port - - port: 9093 - bind_address: '127.0.0.1' - type: http - resources: - - names: [replication] +```yaml +listeners: + # The TCP replication port + - port: 9092 + bind_address: '127.0.0.1' + type: replication + + # The HTTP replication port + - port: 9093 + bind_address: '127.0.0.1' + type: http + resources: + - names: [replication] +``` Under **no circumstances** should these replication API listeners be exposed to -the public internet; it currently implements no authentication whatsoever and is -unencrypted. +the public internet; they have no authentication and are unencrypted. -(Roughly, the TCP port is used for streaming data from the master to the -workers, and the HTTP port for the workers to send data to the main -synapse process.) +You should then create a set of configs for the various worker processes. Each +worker configuration file inherits the configuration of the main homeserver +configuration file. You can then override configuration specific to that +worker, e.g. the HTTP listener that it provides (if any); logging +configuration; etc. You should minimise the number of overrides though to +maintain a usable config. -You then create a set of configs for the various worker processes. These -should be worker configuration files, and should be stored in a dedicated -subdirectory, to allow synctl to manipulate them. An additional configuration -for the master synapse process will need to be created because the process will -not be started automatically. That configuration should look like this: +In the config file for each worker, you must specify the type of worker +application (`worker_app`). The currently available worker applications are +listed below. You must also specify the replication endpoints that it should +talk to on the main synapse process. `worker_replication_host` should specify +the host of the main synapse, `worker_replication_port` should point to the TCP +replication listener port and `worker_replication_http_port` should point to +the HTTP replication port. - worker_app: synapse.app.homeserver - daemonize: true +For example: -Each worker configuration file inherits the configuration of the main homeserver -configuration file. You can then override configuration specific to that worker, -e.g. the HTTP listener that it provides (if any); logging configuration; etc. -You should minimise the number of overrides though to maintain a usable config. +```yaml +worker_app: synapse.app.synchrotron -You must specify the type of worker application (`worker_app`). The currently -available worker applications are listed below. You must also specify the -replication endpoints that it's talking to on the main synapse process. -`worker_replication_host` should specify the host of the main synapse, -`worker_replication_port` should point to the TCP replication listener port and -`worker_replication_http_port` should point to the HTTP replication port. +# The replication listener on the synapse to talk to. +worker_replication_host: 127.0.0.1 +worker_replication_port: 9092 +worker_replication_http_port: 9093 -Currently, the `event_creator` and `federation_reader` workers require specifying -`worker_replication_http_port`. +worker_listeners: + - type: http + port: 8083 + resources: + - names: + - client -For instance: - - worker_app: synapse.app.synchrotron - - # The replication listener on the synapse to talk to. - worker_replication_host: 127.0.0.1 - worker_replication_port: 9092 - worker_replication_http_port: 9093 - - worker_listeners: - - type: http - port: 8083 - resources: - - names: - - client - - worker_daemonize: True - worker_pid_file: /home/matrix/synapse/synchrotron.pid - worker_log_config: /home/matrix/synapse/config/synchrotron_log_config.yaml +worker_log_config: /home/matrix/synapse/config/synchrotron_log_config.yaml +``` ...is a full configuration for a synchrotron worker instance, which will expose a plain HTTP `/sync` endpoint on port 8083 separately from the `/sync` endpoint provided @@ -101,7 +98,75 @@ by the main synapse. Obviously you should configure your reverse-proxy to route the relevant endpoints to the worker (`localhost:8083` in the above example). -Finally, to actually run your worker-based synapse, you must pass synctl the -a +Finally, you need to start your worker processes. This can be done with either +`synctl` or your distribution's preferred service manager such as `systemd`. We +recommend the use of `systemd` where available: for information on setting up +`systemd` to start synapse workers, see +[systemd-with-workers](systemd-with-workers). To use `synctl`, see below. + +### **Experimental** support for replication over redis + +As of Synapse v1.13.0, it is possible to configure Synapse to send replication +via a [Redis pub/sub channel](https://redis.io/topics/pubsub). This is an +alternative to direct TCP connections to the master: rather than all the +workers connecting to the master, all the workers and the master connect to +Redis, which relays replication commands between processes. This can give a +significant cpu saving on the master and will be a prerequisite for upcoming +performance improvements. + +Note that this support is currently experimental; you may experience lost +messages and similar problems! It is strongly recommended that admins setting +up workers for the first time use direct TCP replication as above. + +To configure Synapse to use Redis: + +1. Install Redis following the normal procedure for your distribution - for + example, on Debian, `apt install redis-server`. (It is safe to use an + existing Redis deployment if you have one: we use a pub/sub stream named + according to the `server_name` of your synapse server.) +2. Check Redis is running and accessible: you should be able to `echo PING | nc -q1 + localhost 6379` and get a response of `+PONG`. +3. Install the python prerequisites. If you installed synapse into a + virtualenv, this can be done with: + ```sh + pip install matrix-synapse[redis] + ``` + The debian packages from matrix.org already include the required + dependencies. +4. Add config to the shared configuration (`homeserver.yaml`): + ```yaml + redis: + enabled: true + ``` + Optional parameters which can go alongside `enabled` are `host`, `port`, + `password`. Normally none of these are required. +5. Restart master and all workers. + +Once redis replication is in use, `worker_replication_port` is redundant and +can be removed from the worker configuration files. Similarly, the +configuration for the `listener` for the TCP replication port can be removed +from the main configuration file. Note that the HTTP replication port is +still required. + +### Using synctl + +If you want to use `synctl` to manage your synapse processes, you will need to +create an an additional configuration file for the master synapse process. That +configuration should look like this: + +```yaml +worker_app: synapse.app.homeserver +``` + +Additionally, each worker app must be configured with the name of a "pid file", +to which it will write its process ID when it starts. For example, for a +synchrotron, you might write: + +```yaml +worker_pid_file: /home/matrix/synapse/synchrotron.pid +``` + +Finally, to actually run your worker-based synapse, you must pass synctl the `-a` commandline option to tell it to operate on all the worker configurations found in the given directory, e.g.: diff --git a/scripts-dev/build_debian_packages b/scripts-dev/build_debian_packages index 84eaec6a9..ae2145d71 100755 --- a/scripts-dev/build_debian_packages +++ b/scripts-dev/build_debian_packages @@ -27,6 +27,7 @@ DISTS = ( "ubuntu:cosmic", "ubuntu:disco", "ubuntu:eoan", + "ubuntu:focal", ) DESC = '''\ diff --git a/snap/snapcraft.yaml b/snap/snapcraft.yaml index 6b62b7911..9a01152c1 100644 --- a/snap/snapcraft.yaml +++ b/snap/snapcraft.yaml @@ -33,6 +33,10 @@ parts: python-version: python3 python-packages: - '.[all]' + - pip + - setuptools + - setuptools-scm + - wheel build-packages: - libffi-dev - libturbojpeg0-dev diff --git a/stubs/sortedcontainers/__init__.pyi b/stubs/sortedcontainers/__init__.pyi new file mode 100644 index 000000000..073b806d3 --- /dev/null +++ b/stubs/sortedcontainers/__init__.pyi @@ -0,0 +1,13 @@ +from .sorteddict import ( + SortedDict, + SortedKeysView, + SortedItemsView, + SortedValuesView, +) + +__all__ = [ + "SortedDict", + "SortedKeysView", + "SortedItemsView", + "SortedValuesView", +] diff --git a/stubs/sortedcontainers/sorteddict.pyi b/stubs/sortedcontainers/sorteddict.pyi new file mode 100644 index 000000000..68779f968 --- /dev/null +++ b/stubs/sortedcontainers/sorteddict.pyi @@ -0,0 +1,124 @@ +# stub for SortedDict. This is a lightly edited copy of +# https://github.com/grantjenks/python-sortedcontainers/blob/eea42df1f7bad2792e8da77335ff888f04b9e5ae/sortedcontainers/sorteddict.pyi +# (from https://github.com/grantjenks/python-sortedcontainers/pull/107) + +from typing import ( + Any, + Callable, + Dict, + Hashable, + Iterator, + Iterable, + ItemsView, + KeysView, + List, + Mapping, + Optional, + Sequence, + Type, + TypeVar, + Tuple, + Union, + ValuesView, + overload, +) + +_T = TypeVar("_T") +_S = TypeVar("_S") +_T_h = TypeVar("_T_h", bound=Hashable) +_KT = TypeVar("_KT", bound=Hashable) # Key type. +_VT = TypeVar("_VT") # Value type. +_KT_co = TypeVar("_KT_co", covariant=True, bound=Hashable) +_VT_co = TypeVar("_VT_co", covariant=True) +_SD = TypeVar("_SD", bound=SortedDict) +_Key = Callable[[_T], Any] + +class SortedDict(Dict[_KT, _VT]): + @overload + def __init__(self, **kwargs: _VT) -> None: ... + @overload + def __init__(self, __map: Mapping[_KT, _VT], **kwargs: _VT) -> None: ... + @overload + def __init__( + self, __iterable: Iterable[Tuple[_KT, _VT]], **kwargs: _VT + ) -> None: ... + @overload + def __init__(self, __key: _Key[_KT], **kwargs: _VT) -> None: ... + @overload + def __init__( + self, __key: _Key[_KT], __map: Mapping[_KT, _VT], **kwargs: _VT + ) -> None: ... + @overload + def __init__( + self, __key: _Key[_KT], __iterable: Iterable[Tuple[_KT, _VT]], **kwargs: _VT + ) -> None: ... + @property + def key(self) -> Optional[_Key[_KT]]: ... + @property + def iloc(self) -> SortedKeysView[_KT]: ... + def clear(self) -> None: ... + def __delitem__(self, key: _KT) -> None: ... + def __iter__(self) -> Iterator[_KT]: ... + def __reversed__(self) -> Iterator[_KT]: ... + def __setitem__(self, key: _KT, value: _VT) -> None: ... + def _setitem(self, key: _KT, value: _VT) -> None: ... + def copy(self: _SD) -> _SD: ... + def __copy__(self: _SD) -> _SD: ... + @classmethod + @overload + def fromkeys(cls, seq: Iterable[_T_h]) -> SortedDict[_T_h, None]: ... + @classmethod + @overload + def fromkeys(cls, seq: Iterable[_T_h], value: _S) -> SortedDict[_T_h, _S]: ... + def keys(self) -> SortedKeysView[_KT]: ... + def items(self) -> SortedItemsView[_KT, _VT]: ... + def values(self) -> SortedValuesView[_VT]: ... + @overload + def pop(self, key: _KT) -> _VT: ... + @overload + def pop(self, key: _KT, default: _T = ...) -> Union[_VT, _T]: ... + def popitem(self, index: int = ...) -> Tuple[_KT, _VT]: ... + def peekitem(self, index: int = ...) -> Tuple[_KT, _VT]: ... + def setdefault(self, key: _KT, default: Optional[_VT] = ...) -> _VT: ... + @overload + def update(self, __map: Mapping[_KT, _VT], **kwargs: _VT) -> None: ... + @overload + def update(self, __iterable: Iterable[Tuple[_KT, _VT]], **kwargs: _VT) -> None: ... + @overload + def update(self, **kwargs: _VT) -> None: ... + def __reduce__( + self, + ) -> Tuple[ + Type[SortedDict[_KT, _VT]], Tuple[Callable[[_KT], Any], List[Tuple[_KT, _VT]]], + ]: ... + def __repr__(self) -> str: ... + def _check(self) -> None: ... + def islice( + self, start: Optional[int] = ..., stop: Optional[int] = ..., reverse=bool, + ) -> Iterator[_KT]: ... + def bisect_left(self, value: _KT) -> int: ... + def bisect_right(self, value: _KT) -> int: ... + +class SortedKeysView(KeysView[_KT_co], Sequence[_KT_co]): + @overload + def __getitem__(self, index: int) -> _KT_co: ... + @overload + def __getitem__(self, index: slice) -> List[_KT_co]: ... + def __delitem__(self, index: Union[int, slice]) -> None: ... + +class SortedItemsView( # type: ignore + ItemsView[_KT_co, _VT_co], Sequence[Tuple[_KT_co, _VT_co]] +): + def __iter__(self) -> Iterator[Tuple[_KT_co, _VT_co]]: ... + @overload + def __getitem__(self, index: int) -> Tuple[_KT_co, _VT_co]: ... + @overload + def __getitem__(self, index: slice) -> List[Tuple[_KT_co, _VT_co]]: ... + def __delitem__(self, index: Union[int, slice]) -> None: ... + +class SortedValuesView(ValuesView[_VT_co], Sequence[_VT_co]): + @overload + def __getitem__(self, index: int) -> _VT_co: ... + @overload + def __getitem__(self, index: slice) -> List[_VT_co]: ... + def __delitem__(self, index: Union[int, slice]) -> None: ... diff --git a/stubs/txredisapi.pyi b/stubs/txredisapi.pyi new file mode 100644 index 000000000..cac689d4f --- /dev/null +++ b/stubs/txredisapi.pyi @@ -0,0 +1,43 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +"""Contains *incomplete* type hints for txredisapi. +""" + +from typing import List, Optional, Union + +class RedisProtocol: + def publish(self, channel: str, message: bytes): ... + +class SubscriberProtocol: + password: Optional[str] + def subscribe(self, channels: Union[str, List[str]]): ... + def connectionMade(self): ... + def connectionLost(self, reason): ... + +def lazyConnection( + host: str = ..., + port: int = ..., + dbid: Optional[int] = ..., + reconnect: bool = ..., + charset: str = ..., + password: Optional[str] = ..., + connectTimeout: Optional[int] = ..., + replyTimeout: Optional[int] = ..., + convertNumbers: bool = ..., +) -> RedisProtocol: ... + +class SubscriberFactory: + def buildProtocol(self, addr): ... diff --git a/synapse/__init__.py b/synapse/__init__.py index d8d340f42..0abf49117 100644 --- a/synapse/__init__.py +++ b/synapse/__init__.py @@ -36,7 +36,7 @@ try: except ImportError: pass -__version__ = "1.12.4" +__version__ = "1.13.0" if bool(os.environ.get("SYNAPSE_TEST_PATCH_LOG_CONTEXTS", False)): # We import here so that we don't have to install a bunch of deps when diff --git a/synapse/api/auth.py b/synapse/api/auth.py index c1ade1333..1ad5ff941 100644 --- a/synapse/api/auth.py +++ b/synapse/api/auth.py @@ -26,16 +26,15 @@ from twisted.internet import defer import synapse.logging.opentracing as opentracing import synapse.types from synapse import event_auth -from synapse.api.constants import EventTypes, LimitBlockingTypes, Membership, UserTypes +from synapse.api.auth_blocking import AuthBlocking +from synapse.api.constants import EventTypes, Membership from synapse.api.errors import ( AuthError, Codes, InvalidClientTokenError, MissingClientTokenError, - ResourceLimitError, ) from synapse.api.room_versions import KNOWN_ROOM_VERSIONS -from synapse.config.server import is_threepid_reserved from synapse.events import EventBase from synapse.types import StateMap, UserID from synapse.util.caches import CACHE_SIZE_FACTOR, register_cache @@ -77,7 +76,11 @@ class Auth(object): self.token_cache = LruCache(CACHE_SIZE_FACTOR * 10000) register_cache("cache", "token_cache", self.token_cache) + self._auth_blocking = AuthBlocking(self.hs) + self._account_validity = hs.config.account_validity + self._track_appservice_user_ips = hs.config.track_appservice_user_ips + self._macaroon_secret_key = hs.config.macaroon_secret_key @defer.inlineCallbacks def check_from_context(self, room_version: str, event, context, do_sig_check=True): @@ -191,7 +194,7 @@ class Auth(object): opentracing.set_tag("authenticated_entity", user_id) opentracing.set_tag("appservice_id", app_service.id) - if ip_addr and self.hs.config.track_appservice_user_ips: + if ip_addr and self._track_appservice_user_ips: yield self.store.insert_client_ip( user_id=user_id, access_token=access_token, @@ -454,7 +457,7 @@ class Auth(object): # access_tokens include a nonce for uniqueness: any value is acceptable v.satisfy_general(lambda c: c.startswith("nonce = ")) - v.verify(macaroon, self.hs.config.macaroon_secret_key) + v.verify(macaroon, self._macaroon_secret_key) def _verify_expiry(self, caveat): prefix = "time < " @@ -537,8 +540,7 @@ class Auth(object): return defer.succeed(auth_ids) - @defer.inlineCallbacks - def check_can_change_room_list(self, room_id: str, user: UserID): + async def check_can_change_room_list(self, room_id: str, user: UserID): """Determine whether the user is allowed to edit the room's entry in the published room list. @@ -547,17 +549,17 @@ class Auth(object): user """ - is_admin = yield self.is_server_admin(user) + is_admin = await self.is_server_admin(user) if is_admin: return True user_id = user.to_string() - yield self.check_user_in_room(room_id, user_id) + await self.check_user_in_room(room_id, user_id) # We currently require the user is a "moderator" in the room. We do this # by checking if they would (theoretically) be able to change the # m.room.canonical_alias events - power_level_event = yield self.state.get_current_state( + power_level_event = await self.state.get_current_state( room_id, EventTypes.PowerLevels, "" ) @@ -664,71 +666,5 @@ class Auth(object): % (user_id, room_id), ) - @defer.inlineCallbacks - def check_auth_blocking(self, user_id=None, threepid=None, user_type=None): - """Checks if the user should be rejected for some external reason, - such as monthly active user limiting or global disable flag - - Args: - user_id(str|None): If present, checks for presence against existing - MAU cohort - - threepid(dict|None): If present, checks for presence against configured - reserved threepid. Used in cases where the user is trying register - with a MAU blocked server, normally they would be rejected but their - threepid is on the reserved list. user_id and - threepid should never be set at the same time. - - user_type(str|None): If present, is used to decide whether to check against - certain blocking reasons like MAU. - """ - - # Never fail an auth check for the server notices users or support user - # This can be a problem where event creation is prohibited due to blocking - if user_id is not None: - if user_id == self.hs.config.server_notices_mxid: - return - if (yield self.store.is_support_user(user_id)): - return - - if self.hs.config.hs_disabled: - raise ResourceLimitError( - 403, - self.hs.config.hs_disabled_message, - errcode=Codes.RESOURCE_LIMIT_EXCEEDED, - admin_contact=self.hs.config.admin_contact, - limit_type=LimitBlockingTypes.HS_DISABLED, - ) - if self.hs.config.limit_usage_by_mau is True: - assert not (user_id and threepid) - - # If the user is already part of the MAU cohort or a trial user - if user_id: - timestamp = yield self.store.user_last_seen_monthly_active(user_id) - if timestamp: - return - - is_trial = yield self.store.is_trial_user(user_id) - if is_trial: - return - elif threepid: - # If the user does not exist yet, but is signing up with a - # reserved threepid then pass auth check - if is_threepid_reserved( - self.hs.config.mau_limits_reserved_threepids, threepid - ): - return - elif user_type == UserTypes.SUPPORT: - # If the user does not exist yet and is of type "support", - # allow registration. Support users are excluded from MAU checks. - return - # Else if there is no room in the MAU bucket, bail - current_mau = yield self.store.get_monthly_active_count() - if current_mau >= self.hs.config.max_mau_value: - raise ResourceLimitError( - 403, - "Monthly Active User Limit Exceeded", - admin_contact=self.hs.config.admin_contact, - errcode=Codes.RESOURCE_LIMIT_EXCEEDED, - limit_type=LimitBlockingTypes.MONTHLY_ACTIVE_USER, - ) + def check_auth_blocking(self, *args, **kwargs): + return self._auth_blocking.check_auth_blocking(*args, **kwargs) diff --git a/synapse/api/auth_blocking.py b/synapse/api/auth_blocking.py new file mode 100644 index 000000000..5c499b6b4 --- /dev/null +++ b/synapse/api/auth_blocking.py @@ -0,0 +1,104 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +import logging + +from twisted.internet import defer + +from synapse.api.constants import LimitBlockingTypes, UserTypes +from synapse.api.errors import Codes, ResourceLimitError +from synapse.config.server import is_threepid_reserved + +logger = logging.getLogger(__name__) + + +class AuthBlocking(object): + def __init__(self, hs): + self.store = hs.get_datastore() + + self._server_notices_mxid = hs.config.server_notices_mxid + self._hs_disabled = hs.config.hs_disabled + self._hs_disabled_message = hs.config.hs_disabled_message + self._admin_contact = hs.config.admin_contact + self._max_mau_value = hs.config.max_mau_value + self._limit_usage_by_mau = hs.config.limit_usage_by_mau + self._mau_limits_reserved_threepids = hs.config.mau_limits_reserved_threepids + + @defer.inlineCallbacks + def check_auth_blocking(self, user_id=None, threepid=None, user_type=None): + """Checks if the user should be rejected for some external reason, + such as monthly active user limiting or global disable flag + + Args: + user_id(str|None): If present, checks for presence against existing + MAU cohort + + threepid(dict|None): If present, checks for presence against configured + reserved threepid. Used in cases where the user is trying register + with a MAU blocked server, normally they would be rejected but their + threepid is on the reserved list. user_id and + threepid should never be set at the same time. + + user_type(str|None): If present, is used to decide whether to check against + certain blocking reasons like MAU. + """ + + # Never fail an auth check for the server notices users or support user + # This can be a problem where event creation is prohibited due to blocking + if user_id is not None: + if user_id == self._server_notices_mxid: + return + if (yield self.store.is_support_user(user_id)): + return + + if self._hs_disabled: + raise ResourceLimitError( + 403, + self._hs_disabled_message, + errcode=Codes.RESOURCE_LIMIT_EXCEEDED, + admin_contact=self._admin_contact, + limit_type=LimitBlockingTypes.HS_DISABLED, + ) + if self._limit_usage_by_mau is True: + assert not (user_id and threepid) + + # If the user is already part of the MAU cohort or a trial user + if user_id: + timestamp = yield self.store.user_last_seen_monthly_active(user_id) + if timestamp: + return + + is_trial = yield self.store.is_trial_user(user_id) + if is_trial: + return + elif threepid: + # If the user does not exist yet, but is signing up with a + # reserved threepid then pass auth check + if is_threepid_reserved(self._mau_limits_reserved_threepids, threepid): + return + elif user_type == UserTypes.SUPPORT: + # If the user does not exist yet and is of type "support", + # allow registration. Support users are excluded from MAU checks. + return + # Else if there is no room in the MAU bucket, bail + current_mau = yield self.store.get_monthly_active_count() + if current_mau >= self._max_mau_value: + raise ResourceLimitError( + 403, + "Monthly Active User Limit Exceeded", + admin_contact=self._admin_contact, + errcode=Codes.RESOURCE_LIMIT_EXCEEDED, + limit_type=LimitBlockingTypes.MONTHLY_ACTIVE_USER, + ) diff --git a/synapse/api/constants.py b/synapse/api/constants.py index cc8577552..bcaf2c360 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py @@ -61,6 +61,7 @@ class LoginType(object): MSISDN = "m.login.msisdn" RECAPTCHA = "m.login.recaptcha" TERMS = "m.login.terms" + SSO = "org.matrix.login.sso" DUMMY = "m.login.dummy" # Only for C/S API v1 @@ -96,6 +97,8 @@ class EventTypes(object): Retention = "m.room.retention" + Presence = "m.presence" + class RejectedReason(object): AUTH_ERROR = "auth_error" diff --git a/synapse/api/errors.py b/synapse/api/errors.py index 616942b05..d54dfb385 100644 --- a/synapse/api/errors.py +++ b/synapse/api/errors.py @@ -64,6 +64,13 @@ class Codes(object): INCOMPATIBLE_ROOM_VERSION = "M_INCOMPATIBLE_ROOM_VERSION" WRONG_ROOM_KEYS_VERSION = "M_WRONG_ROOM_KEYS_VERSION" EXPIRED_ACCOUNT = "ORG_MATRIX_EXPIRED_ACCOUNT" + PASSWORD_TOO_SHORT = "M_PASSWORD_TOO_SHORT" + PASSWORD_NO_DIGIT = "M_PASSWORD_NO_DIGIT" + PASSWORD_NO_UPPERCASE = "M_PASSWORD_NO_UPPERCASE" + PASSWORD_NO_LOWERCASE = "M_PASSWORD_NO_LOWERCASE" + PASSWORD_NO_SYMBOL = "M_PASSWORD_NO_SYMBOL" + PASSWORD_IN_DICTIONARY = "M_PASSWORD_IN_DICTIONARY" + WEAK_PASSWORD = "M_WEAK_PASSWORD" INVALID_SIGNATURE = "M_INVALID_SIGNATURE" USER_DEACTIVATED = "M_USER_DEACTIVATED" BAD_ALIAS = "M_BAD_ALIAS" @@ -79,7 +86,14 @@ class CodeMessageException(RuntimeError): def __init__(self, code, msg): super(CodeMessageException, self).__init__("%d: %s" % (code, msg)) - self.code = code + + # Some calls to this method pass instances of http.HTTPStatus for `code`. + # While HTTPStatus is a subclass of int, it has magic __str__ methods + # which emit `HTTPStatus.FORBIDDEN` when converted to a str, instead of `403`. + # This causes inconsistency in our log lines. + # + # To eliminate this behaviour, we convert them to their integer equivalents here. + self.code = int(code) self.msg = msg @@ -439,6 +453,20 @@ class IncompatibleRoomVersionError(SynapseError): return cs_error(self.msg, self.errcode, room_version=self._room_version) +class PasswordRefusedError(SynapseError): + """A password has been refused, either during password reset/change or registration. + """ + + def __init__( + self, + msg="This password doesn't comply with the server's policy", + errcode=Codes.WEAK_PASSWORD, + ): + super(PasswordRefusedError, self).__init__( + code=400, msg=msg, errcode=errcode, + ) + + class RequestSendFailed(RuntimeError): """Sending a HTTP request over federation failed due to not being able to talk to the remote server for some reason. diff --git a/synapse/app/_base.py b/synapse/app/_base.py index 4d84f4595..dedff81af 100644 --- a/synapse/app/_base.py +++ b/synapse/app/_base.py @@ -22,6 +22,7 @@ import sys import traceback from daemonize import Daemonize +from typing_extensions import NoReturn from twisted.internet import defer, error, reactor from twisted.protocols.tls import TLSMemoryBIOFactory @@ -139,9 +140,9 @@ def start_reactor( run() -def quit_with_error(error_string): +def quit_with_error(error_string: str) -> NoReturn: message_lines = error_string.split("\n") - line_length = max(len(l) for l in message_lines if len(l) < 80) + 2 + line_length = max(len(line) for line in message_lines if len(line) < 80) + 2 sys.stderr.write("*" * line_length + "\n") for line in message_lines: sys.stderr.write(" %s\n" % (line.rstrip(),)) @@ -270,7 +271,7 @@ def start(hs, listeners=None): # Start the tracer synapse.logging.opentracing.init_tracer( # type: ignore[attr-defined] # noqa - hs.config + hs ) # It is now safe to start your Synapse. @@ -316,7 +317,7 @@ def setup_sentry(hs): scope.set_tag("matrix_server_name", hs.config.server_name) app = hs.config.worker_app if hs.config.worker_app else "synapse.app.homeserver" - name = hs.config.worker_name if hs.config.worker_name else "master" + name = hs.get_instance_name() scope.set_tag("worker_app", app) scope.set_tag("worker_name", name) diff --git a/synapse/app/admin_cmd.py b/synapse/app/admin_cmd.py index 1c7c6ec0c..a37818fe9 100644 --- a/synapse/app/admin_cmd.py +++ b/synapse/app/admin_cmd.py @@ -43,7 +43,6 @@ from synapse.replication.slave.storage.push_rule import SlavedPushRuleStore from synapse.replication.slave.storage.receipts import SlavedReceiptsStore from synapse.replication.slave.storage.registration import SlavedRegistrationStore from synapse.replication.slave.storage.room import RoomStore -from synapse.replication.tcp.client import ReplicationClientHandler from synapse.server import HomeServer from synapse.util.logcontext import LoggingContext from synapse.util.versionstring import get_version_string @@ -79,17 +78,6 @@ class AdminCmdServer(HomeServer): def start_listening(self, listeners): pass - def build_tcp_replication(self): - return AdminCmdReplicationHandler(self) - - -class AdminCmdReplicationHandler(ReplicationClientHandler): - async def on_rdata(self, stream_name, token, rows): - pass - - def get_streams_to_replicate(self): - return {} - @defer.inlineCallbacks def export_data_command(hs, args): diff --git a/synapse/app/generic_worker.py b/synapse/app/generic_worker.py index 66be6ea2e..667ad2042 100644 --- a/synapse/app/generic_worker.py +++ b/synapse/app/generic_worker.py @@ -17,6 +17,9 @@ import contextlib import logging import sys +from typing import Dict, Iterable + +from typing_extensions import ContextManager from twisted.internet import defer, reactor from twisted.web.resource import NoResource @@ -38,14 +41,14 @@ from synapse.config.homeserver import HomeServerConfig from synapse.config.logger import setup_logging from synapse.federation import send_queue from synapse.federation.transport.server import TransportLayerServer -from synapse.handlers.presence import PresenceHandler, get_interested_parties +from synapse.handlers.presence import BasePresenceHandler, get_interested_parties from synapse.http.server import JsonResource from synapse.http.servlet import RestServlet, parse_json_object_from_request from synapse.http.site import SynapseSite -from synapse.logging.context import LoggingContext, run_in_background +from synapse.logging.context import LoggingContext from synapse.metrics import METRICS_PREFIX, MetricsResource, RegistryProxy from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.replication.slave.storage._base import BaseSlavedStore, __func__ +from synapse.replication.slave.storage._base import BaseSlavedStore from synapse.replication.slave.storage.account_data import SlavedAccountDataStore from synapse.replication.slave.storage.appservice import SlavedApplicationServiceStore from synapse.replication.slave.storage.client_ips import SlavedClientIpStore @@ -64,13 +67,25 @@ from synapse.replication.slave.storage.receipts import SlavedReceiptsStore from synapse.replication.slave.storage.registration import SlavedRegistrationStore from synapse.replication.slave.storage.room import RoomStore from synapse.replication.slave.storage.transactions import SlavedTransactionStore -from synapse.replication.tcp.client import ReplicationClientHandler -from synapse.replication.tcp.streams._base import ( +from synapse.replication.tcp.client import ReplicationDataHandler +from synapse.replication.tcp.commands import ClearUserSyncsCommand +from synapse.replication.tcp.streams import ( + AccountDataStream, DeviceListsStream, + GroupServerStream, + PresenceStream, + PushersStream, + PushRulesStream, ReceiptsStream, + TagAccountDataStream, ToDeviceStream, + TypingStream, +) +from synapse.replication.tcp.streams.events import ( + EventsStream, + EventsStreamEventRow, + EventsStreamRow, ) -from synapse.replication.tcp.streams.events import EventsStreamEventRow, EventsStreamRow from synapse.rest.admin import register_servlets_for_media_repo from synapse.rest.client.v1 import events from synapse.rest.client.v1.initial_sync import InitialSyncRestServlet @@ -112,12 +127,12 @@ from synapse.storage.data_stores.main.monthly_active_users import ( MonthlyActiveUsersWorkerStore, ) from synapse.storage.data_stores.main.presence import UserPresenceState +from synapse.storage.data_stores.main.ui_auth import UIAuthWorkerStore from synapse.storage.data_stores.main.user_directory import UserDirectoryStore from synapse.types import ReadReceipt from synapse.util.async_helpers import Linearizer from synapse.util.httpresourcetree import create_resource_tree from synapse.util.manhole import manhole -from synapse.util.stringutils import random_string from synapse.util.versionstring import get_version_string logger = logging.getLogger("synapse.app.generic_worker") @@ -214,21 +229,31 @@ class KeyUploadServlet(RestServlet): return 200, {"one_time_key_counts": result} +class _NullContextManager(ContextManager[None]): + """A context manager which does nothing.""" + + def __exit__(self, exc_type, exc_val, exc_tb): + pass + + UPDATE_SYNCING_USERS_MS = 10 * 1000 -class GenericWorkerPresence(object): +class GenericWorkerPresence(BasePresenceHandler): def __init__(self, hs): + super().__init__(hs) self.hs = hs self.is_mine_id = hs.is_mine_id self.http_client = hs.get_simple_http_client() - self.store = hs.get_datastore() - self.user_to_num_current_syncs = {} - self.clock = hs.get_clock() - self.notifier = hs.get_notifier() - active_presence = self.store.take_presence_startup_info() - self.user_to_current_state = {state.user_id: state for state in active_presence} + self._presence_enabled = hs.config.use_presence + + # The number of ongoing syncs on this process, by user id. + # Empty if _presence_enabled is false. + self._user_to_num_current_syncs = {} # type: Dict[str, int] + + self.notifier = hs.get_notifier() + self.instance_id = hs.get_instance_id() # user_id -> last_sync_ms. Lists the users that have stopped syncing # but we haven't notified the master of that yet @@ -238,13 +263,24 @@ class GenericWorkerPresence(object): self.send_stop_syncing, UPDATE_SYNCING_USERS_MS ) - self.process_id = random_string(16) - logger.info("Presence process_id is %r", self.process_id) + hs.get_reactor().addSystemEventTrigger( + "before", + "shutdown", + run_as_background_process, + "generic_presence.on_shutdown", + self._on_shutdown, + ) + + def _on_shutdown(self): + if self._presence_enabled: + self.hs.get_tcp_replication().send_command( + ClearUserSyncsCommand(self.instance_id) + ) def send_user_sync(self, user_id, is_syncing, last_sync_ms): - if self.hs.config.use_presence: + if self._presence_enabled: self.hs.get_tcp_replication().send_user_sync( - user_id, is_syncing, last_sync_ms + self.instance_id, user_id, is_syncing, last_sync_ms ) def mark_as_coming_online(self, user_id): @@ -284,28 +320,33 @@ class GenericWorkerPresence(object): # TODO Hows this supposed to work? return defer.succeed(None) - get_states = __func__(PresenceHandler.get_states) - get_state = __func__(PresenceHandler.get_state) - current_state_for_users = __func__(PresenceHandler.current_state_for_users) + async def user_syncing( + self, user_id: str, affect_presence: bool + ) -> ContextManager[None]: + """Record that a user is syncing. - def user_syncing(self, user_id, affect_presence): - if affect_presence: - curr_sync = self.user_to_num_current_syncs.get(user_id, 0) - self.user_to_num_current_syncs[user_id] = curr_sync + 1 + Called by the sync and events servlets to record that a user has connected to + this worker and is waiting for some events. + """ + if not affect_presence or not self._presence_enabled: + return _NullContextManager() - # If we went from no in flight sync to some, notify replication - if self.user_to_num_current_syncs[user_id] == 1: - self.mark_as_coming_online(user_id) + curr_sync = self._user_to_num_current_syncs.get(user_id, 0) + self._user_to_num_current_syncs[user_id] = curr_sync + 1 + + # If we went from no in flight sync to some, notify replication + if self._user_to_num_current_syncs[user_id] == 1: + self.mark_as_coming_online(user_id) def _end(): # We check that the user_id is in user_to_num_current_syncs because # user_to_num_current_syncs may have been cleared if we are # shutting down. - if affect_presence and user_id in self.user_to_num_current_syncs: - self.user_to_num_current_syncs[user_id] -= 1 + if user_id in self._user_to_num_current_syncs: + self._user_to_num_current_syncs[user_id] -= 1 # If we went from one in flight sync to non, notify replication - if self.user_to_num_current_syncs[user_id] == 0: + if self._user_to_num_current_syncs[user_id] == 0: self.mark_as_going_offline(user_id) @contextlib.contextmanager @@ -315,7 +356,7 @@ class GenericWorkerPresence(object): finally: _end() - return defer.succeed(_user_syncing()) + return _user_syncing() @defer.inlineCallbacks def notify_from_replication(self, states, stream_id): @@ -350,15 +391,12 @@ class GenericWorkerPresence(object): stream_id = token yield self.notify_from_replication(states, stream_id) - def get_currently_syncing_users(self): - if self.hs.config.use_presence: - return [ - user_id - for user_id, count in self.user_to_num_current_syncs.items() - if count > 0 - ] - else: - return set() + def get_currently_syncing_users_for_replication(self) -> Iterable[str]: + return [ + user_id + for user_id, count in self._user_to_num_current_syncs.items() + if count > 0 + ] class GenericWorkerTyping(object): @@ -375,12 +413,6 @@ class GenericWorkerTyping(object): # map room IDs to sets of users currently typing self._room_typing = {} - def stream_positions(self): - # We must update this typing token from the response of the previous - # sync. In particular, the stream id may "reset" back to zero/a low - # value which we *must* use for the next replication request. - return {"typing": self._latest_room_serial} - def process_replication_rows(self, token, rows): if self._latest_room_serial > token: # The master has gone backwards. To prevent inconsistent data, just @@ -394,11 +426,15 @@ class GenericWorkerTyping(object): self._room_serials[row.room_id] = token self._room_typing[row.room_id] = row.user_ids + def get_current_token(self) -> int: + return self._latest_room_serial + class GenericWorkerSlavedStore( # FIXME(#3714): We need to add UserDirectoryStore as we write directly # rather than going via the correct worker. UserDirectoryStore, + UIAuthWorkerStore, SlavedDeviceInboxStore, SlavedDeviceStore, SlavedReceiptsStore, @@ -583,7 +619,7 @@ class GenericWorkerServer(HomeServer): def remove_pusher(self, app_id, push_key, user_id): self.get_tcp_replication().send_remove_pusher(app_id, push_key, user_id) - def build_tcp_replication(self): + def build_replication_data_handler(self): return GenericWorkerReplicationHandler(self) def build_presence_handler(self): @@ -593,14 +629,13 @@ class GenericWorkerServer(HomeServer): return GenericWorkerTyping(self) -class GenericWorkerReplicationHandler(ReplicationClientHandler): +class GenericWorkerReplicationHandler(ReplicationDataHandler): def __init__(self, hs): super(GenericWorkerReplicationHandler, self).__init__(hs.get_datastore()) self.store = hs.get_datastore() self.typing_handler = hs.get_typing_handler() - # NB this is a SynchrotronPresence, not a normal PresenceHandler - self.presence_handler = hs.get_presence_handler() + self.presence_handler = hs.get_presence_handler() # type: GenericWorkerPresence self.notifier = hs.get_notifier() self.notify_pushers = hs.config.start_pushers @@ -611,28 +646,18 @@ class GenericWorkerReplicationHandler(ReplicationClientHandler): else: self.send_handler = None - async def on_rdata(self, stream_name, token, rows): - await super(GenericWorkerReplicationHandler, self).on_rdata( - stream_name, token, rows - ) - run_in_background(self.process_and_notify, stream_name, token, rows) + async def on_rdata(self, stream_name, instance_name, token, rows): + await super().on_rdata(stream_name, instance_name, token, rows) + await self._process_and_notify(stream_name, instance_name, token, rows) - def get_streams_to_replicate(self): - args = super(GenericWorkerReplicationHandler, self).get_streams_to_replicate() - args.update(self.typing_handler.stream_positions()) - if self.send_handler: - args.update(self.send_handler.stream_positions()) - return args - - def get_currently_syncing_users(self): - return self.presence_handler.get_currently_syncing_users() - - async def process_and_notify(self, stream_name, token, rows): + async def _process_and_notify(self, stream_name, instance_name, token, rows): try: if self.send_handler: - self.send_handler.process_replication_rows(stream_name, token, rows) + await self.send_handler.process_replication_rows( + stream_name, token, rows + ) - if stream_name == "events": + if stream_name == EventsStream.NAME: # We shouldn't get multiple rows per token for events stream, so # we don't need to optimise this for multiple rows. for row in rows: @@ -655,43 +680,44 @@ class GenericWorkerReplicationHandler(ReplicationClientHandler): ) await self.pusher_pool.on_new_notifications(token, token) - elif stream_name == "push_rules": + elif stream_name == PushRulesStream.NAME: self.notifier.on_new_event( "push_rules_key", token, users=[row.user_id for row in rows] ) - elif stream_name in ("account_data", "tag_account_data"): + elif stream_name in (AccountDataStream.NAME, TagAccountDataStream.NAME): self.notifier.on_new_event( "account_data_key", token, users=[row.user_id for row in rows] ) - elif stream_name == "receipts": + elif stream_name == ReceiptsStream.NAME: self.notifier.on_new_event( "receipt_key", token, rooms=[row.room_id for row in rows] ) await self.pusher_pool.on_new_receipts( token, token, {row.room_id for row in rows} ) - elif stream_name == "typing": + elif stream_name == TypingStream.NAME: self.typing_handler.process_replication_rows(token, rows) self.notifier.on_new_event( "typing_key", token, rooms=[row.room_id for row in rows] ) - elif stream_name == "to_device": + elif stream_name == ToDeviceStream.NAME: entities = [row.entity for row in rows if row.entity.startswith("@")] if entities: self.notifier.on_new_event("to_device_key", token, users=entities) - elif stream_name == "device_lists": + elif stream_name == DeviceListsStream.NAME: all_room_ids = set() for row in rows: - room_ids = await self.store.get_rooms_for_user(row.user_id) - all_room_ids.update(room_ids) + if row.entity.startswith("@"): + room_ids = await self.store.get_rooms_for_user(row.entity) + all_room_ids.update(room_ids) self.notifier.on_new_event("device_list_key", token, rooms=all_room_ids) - elif stream_name == "presence": + elif stream_name == PresenceStream.NAME: await self.presence_handler.process_replication_rows(token, rows) - elif stream_name == "receipts": + elif stream_name == GroupServerStream.NAME: self.notifier.on_new_event( "groups_key", token, users=[row.user_id for row in rows] ) - elif stream_name == "pushers": + elif stream_name == PushersStream.NAME: for row in rows: if row.deleted: self.stop_pusher(row.user_id, row.app_id, row.pushkey) @@ -758,15 +784,12 @@ class FederationSenderHandler(object): def wake_destination(self, server: str): self.federation_sender.wake_destination(server) - def stream_positions(self): - return {"federation": self.federation_position} - - def process_replication_rows(self, stream_name, token, rows): + async def process_replication_rows(self, stream_name, token, rows): # The federation stream contains things that we want to send out, e.g. # presence, typing, etc. if stream_name == "federation": send_queue.process_rows_for_federation(self.federation_sender, rows) - run_in_background(self.update_token, token) + await self.update_token(token) # We also need to poke the federation sender when new events happen elif stream_name == "events": @@ -774,13 +797,14 @@ class FederationSenderHandler(object): # ... and when new receipts happen elif stream_name == ReceiptsStream.NAME: - run_as_background_process( - "process_receipts_for_federation", self._on_new_receipts, rows - ) + await self._on_new_receipts(rows) # ... as well as device updates and messages elif stream_name == DeviceListsStream.NAME: - hosts = {row.destination for row in rows} + # The entities are either user IDs (starting with '@') whose devices + # have changed, or remote servers that we need to tell about + # changes. + hosts = {row.entity for row in rows if not row.entity.startswith("@")} for host in hosts: self.federation_sender.send_device_messages(host) @@ -795,7 +819,7 @@ class FederationSenderHandler(object): async def _on_new_receipts(self, rows): """ Args: - rows (iterable[synapse.replication.tcp.streams.ReceiptsStreamRow]): + rows (Iterable[synapse.replication.tcp.streams.ReceiptsStream.ReceiptsStreamRow]): new receipts to be processed """ for receipt in rows: @@ -920,17 +944,22 @@ def start(config_options): synapse.events.USE_FROZEN_DICTS = config.use_frozen_dicts - ss = GenericWorkerServer( + hs = GenericWorkerServer( config.server_name, config=config, version_string="Synapse/" + get_version_string(synapse), ) - setup_logging(ss, config, use_worker_options=True) + setup_logging(hs, config, use_worker_options=True) + + hs.setup() + + # Ensure the replication streamer is always started in case we write to any + # streams. Will no-op if no streams can be written to by this worker. + hs.get_replication_streamer() - ss.setup() reactor.addSystemEventTrigger( - "before", "startup", _base.start, ss, config.worker_listeners + "before", "startup", _base.start, hs, config.worker_listeners ) _base.start_worker_reactor("synapse-generic-worker", config) diff --git a/synapse/app/homeserver.py b/synapse/app/homeserver.py index f2b56a636..cbd1ea475 100644 --- a/synapse/app/homeserver.py +++ b/synapse/app/homeserver.py @@ -241,16 +241,26 @@ class SynapseHomeServer(HomeServer): resources[SERVER_KEY_V2_PREFIX] = KeyApiV2Resource(self) if name == "webclient": - webclient_path = self.get_config().web_client_location + webclient_loc = self.get_config().web_client_location - if webclient_path is None: + if webclient_loc is None: logger.warning( "Not enabling webclient resource, as web_client_location is unset." ) + elif webclient_loc.startswith("http://") or webclient_loc.startswith( + "https://" + ): + resources[WEB_CLIENT_PREFIX] = RootRedirect(webclient_loc) else: + logger.warning( + "Running webclient on the same domain is not recommended: " + "https://github.com/matrix-org/synapse#security-note - " + "after you move webclient to different host you can set " + "web_client_location to its full URL to enable redirection." + ) # GZip is disabled here due to # https://twistedmatrix.com/trac/ticket/7678 - resources[WEB_CLIENT_PREFIX] = File(webclient_path) + resources[WEB_CLIENT_PREFIX] = File(webclient_loc) if name == "metrics" and self.get_config().enable_metrics: resources[METRICS_PREFIX] = MetricsResource(RegistryProxy) @@ -263,6 +273,12 @@ class SynapseHomeServer(HomeServer): def start_listening(self, listeners): config = self.get_config() + if config.redis_enabled: + # If redis is enabled we connect via the replication command handler + # in the same way as the workers (since we're effectively a client + # rather than a server). + self.get_tcp_replication().start_replication(self) + for listener in listeners: if listener["type"] == "http": self._listening_services.extend(self._listener_http(config, listener)) diff --git a/synapse/config/_base.py b/synapse/config/_base.py index ba846042c..30d1050a9 100644 --- a/synapse/config/_base.py +++ b/synapse/config/_base.py @@ -294,7 +294,6 @@ class RootConfig(object): report_stats=None, open_private_ports=False, listeners=None, - database_conf=None, tls_certificate_path=None, tls_private_key_path=None, acme_domain=None, @@ -367,7 +366,6 @@ class RootConfig(object): report_stats=report_stats, open_private_ports=open_private_ports, listeners=listeners, - database_conf=database_conf, tls_certificate_path=tls_certificate_path, tls_private_key_path=tls_private_key_path, acme_domain=acme_domain, @@ -470,8 +468,8 @@ class RootConfig(object): Returns: Config object, or None if --generate-config or --generate-keys was set """ - config_parser = argparse.ArgumentParser(add_help=False) - config_parser.add_argument( + parser = argparse.ArgumentParser(description=description) + parser.add_argument( "-c", "--config-path", action="append", @@ -480,7 +478,7 @@ class RootConfig(object): " may specify directories containing *.yaml files.", ) - generate_group = config_parser.add_argument_group("Config generation") + generate_group = parser.add_argument_group("Config generation") generate_group.add_argument( "--generate-config", action="store_true", @@ -528,12 +526,13 @@ class RootConfig(object): ), ) - config_args, remaining_args = config_parser.parse_known_args(argv) + cls.invoke_all_static("add_arguments", parser) + config_args = parser.parse_args(argv) config_files = find_config_files(search_paths=config_args.config_path) if not config_files: - config_parser.error( + parser.error( "Must supply a config file.\nA config file can be automatically" ' generated using "--generate-config -H SERVER_NAME' ' -c CONFIG-FILE"' @@ -552,7 +551,7 @@ class RootConfig(object): if config_args.generate_config: if config_args.report_stats is None: - config_parser.error( + parser.error( "Please specify either --report-stats=yes or --report-stats=no\n\n" + MISSING_REPORT_STATS_SPIEL ) @@ -611,15 +610,6 @@ class RootConfig(object): ) generate_missing_configs = True - parser = argparse.ArgumentParser( - parents=[config_parser], - description=description, - formatter_class=argparse.RawDescriptionHelpFormatter, - ) - - obj.invoke_all_static("add_arguments", parser) - args = parser.parse_args(remaining_args) - config_dict = read_config_files(config_files) if generate_missing_configs: obj.generate_missing_files(config_dict, config_dir_path) @@ -628,7 +618,7 @@ class RootConfig(object): obj.parse_config_dict( config_dict, config_dir_path=config_dir_path, data_dir_path=data_dir_path ) - obj.invoke_all("read_arguments", args) + obj.invoke_all("read_arguments", config_args) return obj @@ -667,6 +657,12 @@ def read_config_files(config_files): for config_file in config_files: with open(config_file) as file_stream: yaml_config = yaml.safe_load(file_stream) + + if not isinstance(yaml_config, dict): + err = "File %r is empty or doesn't parse into a key-value map. IGNORING." + print(err % (config_file,)) + continue + specified_config.update(yaml_config) if "server_name" not in specified_config: diff --git a/synapse/config/captcha.py b/synapse/config/captcha.py index f0171bb5b..56c87fa29 100644 --- a/synapse/config/captcha.py +++ b/synapse/config/captcha.py @@ -24,7 +24,6 @@ class CaptchaConfig(Config): self.enable_registration_captcha = config.get( "enable_registration_captcha", False ) - self.captcha_bypass_secret = config.get("captcha_bypass_secret") self.recaptcha_siteverify_api = config.get( "recaptcha_siteverify_api", "https://www.recaptcha.net/recaptcha/api/siteverify", @@ -49,10 +48,6 @@ class CaptchaConfig(Config): # #enable_registration_captcha: false - # A secret key used to bypass the captcha test entirely. - # - #captcha_bypass_secret: "YOUR_SECRET_HERE" - # The API endpoint to use for verifying m.login.recaptcha responses. # #recaptcha_siteverify_api: "https://www.recaptcha.net/recaptcha/api/siteverify" diff --git a/synapse/config/database.py b/synapse/config/database.py index 219b32f67..5b662d1b0 100644 --- a/synapse/config/database.py +++ b/synapse/config/database.py @@ -1,5 +1,6 @@ # -*- coding: utf-8 -*- # Copyright 2014-2016 OpenMarket Ltd +# Copyright 2020 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. @@ -14,14 +15,65 @@ # limitations under the License. import logging import os -from textwrap import indent - -import yaml from synapse.config._base import Config, ConfigError logger = logging.getLogger(__name__) +NON_SQLITE_DATABASE_PATH_WARNING = """\ +Ignoring 'database_path' setting: not using a sqlite3 database. +-------------------------------------------------------------------------------- +""" + +DEFAULT_CONFIG = """\ +## Database ## + +# The 'database' setting defines the database that synapse uses to store all of +# its data. +# +# 'name' gives the database engine to use: either 'sqlite3' (for SQLite) or +# 'psycopg2' (for PostgreSQL). +# +# 'args' gives options which are passed through to the database engine, +# except for options starting 'cp_', which are used to configure the Twisted +# connection pool. For a reference to valid arguments, see: +# * for sqlite: https://docs.python.org/3/library/sqlite3.html#sqlite3.connect +# * for postgres: https://www.postgresql.org/docs/current/libpq-connect.html#LIBPQ-PARAMKEYWORDS +# * for the connection pool: https://twistedmatrix.com/documents/current/api/twisted.enterprise.adbapi.ConnectionPool.html#__init__ +# +# +# Example SQLite configuration: +# +#database: +# name: sqlite3 +# args: +# database: /path/to/homeserver.db +# +# +# Example Postgres configuration: +# +#database: +# name: psycopg2 +# args: +# user: synapse +# password: secretpassword +# database: synapse +# host: localhost +# cp_min: 5 +# cp_max: 10 +# +# For more information on using Synapse with Postgres, see `docs/postgres.md`. +# +database: + name: sqlite3 + args: + database: %(database_path)s + +# Number of events to cache in memory. +# +#event_cache_size: 10K +""" + class DatabaseConnectionConfig: """Contains the connection config for a particular database. @@ -36,10 +88,12 @@ class DatabaseConnectionConfig: """ def __init__(self, name: str, db_config: dict): - if db_config["name"] not in ("sqlite3", "psycopg2"): - raise ConfigError("Unsupported database type %r" % (db_config["name"],)) + db_engine = db_config.get("name", "sqlite3") - if db_config["name"] == "sqlite3": + if db_engine not in ("sqlite3", "psycopg2"): + raise ConfigError("Unsupported database type %r" % (db_engine,)) + + if db_engine == "sqlite3": db_config.setdefault("args", {}).update( {"cp_min": 1, "cp_max": 1, "check_same_thread": False} ) @@ -56,6 +110,11 @@ class DatabaseConnectionConfig: class DatabaseConfig(Config): section = "database" + def __init__(self, *args, **kwargs): + super().__init__(*args, **kwargs) + + self.databases = [] + def read_config(self, config, **kwargs): self.event_cache_size = self.parse_size(config.get("event_cache_size", "10K")) @@ -76,12 +135,13 @@ class DatabaseConfig(Config): multi_database_config = config.get("databases") database_config = config.get("database") + database_path = config.get("database_path") if multi_database_config and database_config: - raise ConfigError("Can't specify both 'database' and 'datbases' in config") + raise ConfigError("Can't specify both 'database' and 'databases' in config") if multi_database_config: - if config.get("database_path"): + if database_path: raise ConfigError("Can't specify 'database_path' with 'databases'") self.databases = [ @@ -89,65 +149,55 @@ class DatabaseConfig(Config): for name, db_conf in multi_database_config.items() ] - else: - if database_config is None: - database_config = {"name": "sqlite3", "args": {}} - + if database_config: self.databases = [DatabaseConnectionConfig("master", database_config)] - self.set_databasepath(config.get("database_path")) + if database_path: + if self.databases and self.databases[0].name != "sqlite3": + logger.warning(NON_SQLITE_DATABASE_PATH_WARNING) + return - def generate_config_section(self, data_dir_path, database_conf, **kwargs): - if not database_conf: - database_path = os.path.join(data_dir_path, "homeserver.db") - database_conf = ( - """# The database engine name - name: "sqlite3" - # Arguments to pass to the engine - args: - # Path to the database - database: "%(database_path)s" - """ - % locals() - ) - else: - database_conf = indent(yaml.dump(database_conf), " " * 10).lstrip() + database_config = {"name": "sqlite3", "args": {}} + self.databases = [DatabaseConnectionConfig("master", database_config)] + self.set_databasepath(database_path) - return ( - """\ - ## Database ## - - database: - %(database_conf)s - # Number of events to cache in memory. - # - #event_cache_size: 10K - """ - % locals() - ) + def generate_config_section(self, data_dir_path, **kwargs): + return DEFAULT_CONFIG % { + "database_path": os.path.join(data_dir_path, "homeserver.db") + } def read_arguments(self, args): - self.set_databasepath(args.database_path) + """ + Cases for the cli input: + - If no databases are configured and no database_path is set, raise. + - No databases and only database_path available ==> sqlite3 db. + - If there are multiple databases and a database_path raise an error. + - If the database set in the config file is sqlite then + overwrite with the command line argument. + """ + + if args.database_path is None: + if not self.databases: + raise ConfigError("No database config provided") + return + + if len(self.databases) == 0: + database_config = {"name": "sqlite3", "args": {}} + self.databases = [DatabaseConnectionConfig("master", database_config)] + self.set_databasepath(args.database_path) + return + + if self.get_single_database().name == "sqlite3": + self.set_databasepath(args.database_path) + else: + logger.warning(NON_SQLITE_DATABASE_PATH_WARNING) def set_databasepath(self, database_path): - if database_path is None: - return if database_path != ":memory:": database_path = self.abspath(database_path) - # We only support setting a database path if we have a single sqlite3 - # database. - if len(self.databases) != 1: - raise ConfigError("Cannot specify 'database_path' with multiple databases") - - database = self.get_single_database() - if database.config["name"] != "sqlite3": - # We don't raise here as we haven't done so before for this case. - logger.warn("Ignoring 'database_path' for non-sqlite3 database") - return - - database.config["args"]["database"] = database_path + self.databases[0].config["args"]["database"] = database_path @staticmethod def add_arguments(parser): @@ -162,7 +212,7 @@ class DatabaseConfig(Config): def get_single_database(self) -> DatabaseConnectionConfig: """Returns the database if there is only one, useful for e.g. tests """ - if len(self.databases) != 1: + if not self.databases: raise Exception("More than one database exists") return self.databases[0] diff --git a/synapse/config/emailconfig.py b/synapse/config/emailconfig.py index f31fc85ec..76b8957ea 100644 --- a/synapse/config/emailconfig.py +++ b/synapse/config/emailconfig.py @@ -108,9 +108,14 @@ class EmailConfig(Config): if self.trusted_third_party_id_servers: # XXX: It's a little confusing that account_threepid_delegate_email is modified # both in RegistrationConfig and here. We should factor this bit out - self.account_threepid_delegate_email = self.trusted_third_party_id_servers[ - 0 - ] # type: Optional[str] + + first_trusted_identity_server = self.trusted_third_party_id_servers[0] + + # trusted_third_party_id_servers does not contain a scheme whereas + # account_threepid_delegate_email is expected to. Presume https + self.account_threepid_delegate_email = ( + "https://" + first_trusted_identity_server + ) # type: Optional[str] self.using_identity_server_from_trusted_list = True else: raise ConfigError( diff --git a/synapse/config/homeserver.py b/synapse/config/homeserver.py index b4bca08b2..be6c6afa7 100644 --- a/synapse/config/homeserver.py +++ b/synapse/config/homeserver.py @@ -31,6 +31,7 @@ from .password import PasswordConfig from .password_auth_providers import PasswordAuthProviderConfig from .push import PushConfig from .ratelimiting import RatelimitConfig +from .redis import RedisConfig from .registration import RegistrationConfig from .repository import ContentRepositoryConfig from .room_directory import RoomDirectoryConfig @@ -82,4 +83,5 @@ class HomeServerConfig(RootConfig): RoomDirectoryConfig, ThirdPartyRulesConfig, TracerConfig, + RedisConfig, ] diff --git a/synapse/config/metrics.py b/synapse/config/metrics.py index 22538153e..6f517a71d 100644 --- a/synapse/config/metrics.py +++ b/synapse/config/metrics.py @@ -86,7 +86,7 @@ class MetricsConfig(Config): # enabled by default, either for performance reasons or limited use. # metrics_flags: - # Publish synapse_federation_known_servers, a g auge of the number of + # Publish synapse_federation_known_servers, a gauge of the number of # servers this homeserver knows about, including itself. May cause # performance problems on large homeservers. # diff --git a/synapse/config/password.py b/synapse/config/password.py index 2a634ac75..9c0ea8c30 100644 --- a/synapse/config/password.py +++ b/synapse/config/password.py @@ -31,6 +31,10 @@ class PasswordConfig(Config): self.password_localdb_enabled = password_config.get("localdb_enabled", True) self.password_pepper = password_config.get("pepper", "") + # Password policy + self.password_policy = password_config.get("policy") or {} + self.password_policy_enabled = self.password_policy.get("enabled", False) + def generate_config_section(self, config_dir_path, server_name, **kwargs): return """\ password_config: @@ -48,4 +52,39 @@ class PasswordConfig(Config): # DO NOT CHANGE THIS AFTER INITIAL SETUP! # #pepper: "EVEN_MORE_SECRET" + + # Define and enforce a password policy. Each parameter is optional. + # This is an implementation of MSC2000. + # + policy: + # Whether to enforce the password policy. + # Defaults to 'false'. + # + #enabled: true + + # Minimum accepted length for a password. + # Defaults to 0. + # + #minimum_length: 15 + + # Whether a password must contain at least one digit. + # Defaults to 'false'. + # + #require_digit: true + + # Whether a password must contain at least one symbol. + # A symbol is any character that's not a number or a letter. + # Defaults to 'false'. + # + #require_symbol: true + + # Whether a password must contain at least one lowercase letter. + # Defaults to 'false'. + # + #require_lowercase: true + + # Whether a password must contain at least one lowercase letter. + # Defaults to 'false'. + # + #require_uppercase: true """ diff --git a/synapse/config/password_auth_providers.py b/synapse/config/password_auth_providers.py index 9746bbc68..4fda8ae98 100644 --- a/synapse/config/password_auth_providers.py +++ b/synapse/config/password_auth_providers.py @@ -35,7 +35,7 @@ class PasswordAuthProviderConfig(Config): if ldap_config.get("enabled", False): providers.append({"module": LDAP_PROVIDER, "config": ldap_config}) - providers.extend(config.get("password_providers", [])) + providers.extend(config.get("password_providers") or []) for provider in providers: mod_name = provider["module"] @@ -52,7 +52,19 @@ class PasswordAuthProviderConfig(Config): def generate_config_section(self, **kwargs): return """\ - #password_providers: + # Password providers allow homeserver administrators to integrate + # their Synapse installation with existing authentication methods + # ex. LDAP, external tokens, etc. + # + # For more information and known implementations, please see + # https://github.com/matrix-org/synapse/blob/master/docs/password_auth_providers.md + # + # Note: instances wishing to use SAML or CAS authentication should + # instead use the `saml2_config` or `cas_config` options, + # respectively. + # + password_providers: + # # Example config for an LDAP auth provider # - module: "ldap_auth_provider.LdapAuthProvider" # config: # enabled: true diff --git a/synapse/config/redis.py b/synapse/config/redis.py new file mode 100644 index 000000000..d5d3ca1c9 --- /dev/null +++ b/synapse/config/redis.py @@ -0,0 +1,34 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +from synapse.config._base import Config +from synapse.python_dependencies import check_requirements + + +class RedisConfig(Config): + section = "redis" + + def read_config(self, config, **kwargs): + redis_config = config.get("redis", {}) + self.redis_enabled = redis_config.get("enabled", False) + + if not self.redis_enabled: + return + + check_requirements("redis") + + self.redis_host = redis_config.get("host", "localhost") + self.redis_port = redis_config.get("port", 6379) + self.redis_password = redis_config.get("password") diff --git a/synapse/config/registration.py b/synapse/config/registration.py index 9bb3beedb..e7ea3a01c 100644 --- a/synapse/config/registration.py +++ b/synapse/config/registration.py @@ -129,6 +129,10 @@ class RegistrationConfig(Config): raise ConfigError("Invalid auto_join_rooms entry %s" % (room_alias,)) self.autocreate_auto_join_rooms = config.get("autocreate_auto_join_rooms", True) + self.enable_set_displayname = config.get("enable_set_displayname", True) + self.enable_set_avatar_url = config.get("enable_set_avatar_url", True) + self.enable_3pid_changes = config.get("enable_3pid_changes", True) + self.disable_msisdn_registration = config.get( "disable_msisdn_registration", False ) @@ -330,6 +334,29 @@ class RegistrationConfig(Config): #email: https://example.com # Delegate email sending to example.com #msisdn: http://localhost:8090 # Delegate SMS sending to this local process + # Whether users are allowed to change their displayname after it has + # been initially set. Useful when provisioning users based on the + # contents of a third-party directory. + # + # Does not apply to server administrators. Defaults to 'true' + # + #enable_set_displayname: false + + # Whether users are allowed to change their avatar after it has been + # initially set. Useful when provisioning users based on the contents + # of a third-party directory. + # + # Does not apply to server administrators. Defaults to 'true' + # + #enable_set_avatar_url: false + + # Whether users can change the 3PIDs associated with their accounts + # (email address and msisdn). + # + # Defaults to 'true' + # + #enable_3pid_changes: false + # Users who register on this homeserver will automatically be joined # to these rooms # diff --git a/synapse/config/repository.py b/synapse/config/repository.py index 7d2dd27fd..9d2ce2022 100644 --- a/synapse/config/repository.py +++ b/synapse/config/repository.py @@ -192,6 +192,10 @@ class ContentRepositoryConfig(Config): self.url_preview_url_blacklist = config.get("url_preview_url_blacklist", ()) + self.url_preview_accept_language = config.get( + "url_preview_accept_language" + ) or ["en"] + def generate_config_section(self, data_dir_path, **kwargs): media_store = os.path.join(data_dir_path, "media_store") uploads_path = os.path.join(data_dir_path, "uploads") @@ -220,12 +224,11 @@ class ContentRepositoryConfig(Config): # #media_storage_providers: # - module: file_system - # # Whether to write new local files. + # # Whether to store newly uploaded local files # store_local: false - # # Whether to write new remote media + # # Whether to store newly downloaded remote files # store_remote: false - # # Whether to block upload requests waiting for write to this - # # provider to complete + # # Whether to wait for successful storage for local uploads # store_synchronous: false # config: # directory: /mnt/some/other/directory @@ -329,6 +332,31 @@ class ContentRepositoryConfig(Config): # The largest allowed URL preview spidering size in bytes # #max_spider_size: 10M + + # A list of values for the Accept-Language HTTP header used when + # downloading webpages during URL preview generation. This allows + # Synapse to specify the preferred languages that URL previews should + # be in when communicating with remote servers. + # + # Each value is a IETF language tag; a 2-3 letter identifier for a + # language, optionally followed by subtags separated by '-', specifying + # a country or region variant. + # + # Multiple values can be provided, and a weight can be added to each by + # using quality value syntax (;q=). '*' translates to any language. + # + # Defaults to "en". + # + # Example: + # + # url_preview_accept_language: + # - en-UK + # - en-US;q=0.9 + # - fr;q=0.8 + # - *;q=0.7 + # + url_preview_accept_language: + # - en """ % locals() ) diff --git a/synapse/config/saml2_config.py b/synapse/config/saml2_config.py index 8fe64d90f..726a27d7b 100644 --- a/synapse/config/saml2_config.py +++ b/synapse/config/saml2_config.py @@ -248,32 +248,32 @@ class SAML2Config(Config): # remote: # - url: https://our_idp/metadata.xml # - # # By default, the user has to go to our login page first. If you'd like - # # to allow IdP-initiated login, set 'allow_unsolicited: true' in a - # # 'service.sp' section: - # # - # #service: - # # sp: - # # allow_unsolicited: true + # # By default, the user has to go to our login page first. If you'd like + # # to allow IdP-initiated login, set 'allow_unsolicited: true' in a + # # 'service.sp' section: + # # + # #service: + # # sp: + # # allow_unsolicited: true # - # # The examples below are just used to generate our metadata xml, and you - # # may well not need them, depending on your setup. Alternatively you - # # may need a whole lot more detail - see the pysaml2 docs! + # # The examples below are just used to generate our metadata xml, and you + # # may well not need them, depending on your setup. Alternatively you + # # may need a whole lot more detail - see the pysaml2 docs! # - # description: ["My awesome SP", "en"] - # name: ["Test SP", "en"] + # description: ["My awesome SP", "en"] + # name: ["Test SP", "en"] # - # organization: - # name: Example com - # display_name: - # - ["Example co", "en"] - # url: "http://example.com" + # organization: + # name: Example com + # display_name: + # - ["Example co", "en"] + # url: "http://example.com" # - # contact_person: - # - given_name: Bob - # sur_name: "the Sysadmin" - # email_address": ["admin@example.com"] - # contact_type": technical + # contact_person: + # - given_name: Bob + # sur_name: "the Sysadmin" + # email_address": ["admin@example.com"] + # contact_type": technical # Instead of putting the config inline as above, you can specify a # separate pysaml2 configuration file: diff --git a/synapse/config/server.py b/synapse/config/server.py index 7525765fe..ed28da3de 100644 --- a/synapse/config/server.py +++ b/synapse/config/server.py @@ -505,10 +505,24 @@ class ServerConfig(Config): "cleanup_extremities_with_dummy_events", True ) + # The number of forward extremities in a room needed to send a dummy event. + self.dummy_events_threshold = config.get("dummy_events_threshold", 10) + self.enable_ephemeral_messages = config.get("enable_ephemeral_messages", False) + # Inhibits the /requestToken endpoints from returning an error that might leak + # information about whether an e-mail address is in use or not on this + # homeserver, and instead return a 200 with a fake sid if this kind of error is + # met, without sending anything. + # This is a compromise between sending an email, which could be a spam vector, + # and letting the client know which email address is bound to an account and + # which one isn't. + self.request_token_inhibit_3pid_errors = config.get( + "request_token_inhibit_3pid_errors", False, + ) + def has_tls_listener(self) -> bool: - return any(l["tls"] for l in self.listeners) + return any(listener["tls"] for listener in self.listeners) def generate_config_section( self, server_name, data_dir_path, open_private_ports, listeners, **kwargs @@ -604,10 +618,15 @@ class ServerConfig(Config): # pid_file: %(pid_file)s - # The path to the web client which will be served at /_matrix/client/ - # if 'webclient' is configured under the 'listeners' configuration. + # The absolute URL to the web client which /_matrix/client will redirect + # to if 'webclient' is configured under the 'listeners' configuration. # - #web_client_location: "/path/to/web/root" + # This option can be also set to the filesystem path to the web client + # which will be served at /_matrix/client/ if 'webclient' is configured + # under the 'listeners' configuration, however this is a security risk: + # https://github.com/matrix-org/synapse#security-note + # + #web_client_location: https://riot.example.com/ # The public-facing base URL that clients use to access this HS # (not including _matrix/...). This is the same URL a user would @@ -807,6 +826,18 @@ class ServerConfig(Config): # bind_addresses: ['::1', '127.0.0.1'] # type: manhole + # Forward extremities can build up in a room due to networking delays between + # homeservers. Once this happens in a large room, calculation of the state of + # that room can become quite expensive. To mitigate this, once the number of + # forward extremities reaches a given threshold, Synapse will send an + # org.matrix.dummy_event event, which will reduce the forward extremities + # in the room. + # + # This setting defines the threshold (i.e. number of forward extremities in the + # room) at which dummy events are sent. The default value is 10. + # + #dummy_events_threshold: 5 + ## Homeserver blocking ## @@ -967,6 +998,16 @@ class ServerConfig(Config): # - shortest_max_lifetime: 3d # longest_max_lifetime: 1y # interval: 1d + + # Inhibits the /requestToken endpoints from returning an error that might leak + # information about whether an e-mail address is in use or not on this + # homeserver. + # Note that for some endpoints the error situation is the e-mail already being + # used, and for others the error is entering the e-mail being unused. + # If this option is enabled, instead of returning an error, these endpoints will + # act as if no error happened and return a fake session ID ('sid') to clients. + # + #request_token_inhibit_3pid_errors: true """ % locals() ) diff --git a/synapse/config/sso.py b/synapse/config/sso.py index 95762689b..cac6bc013 100644 --- a/synapse/config/sso.py +++ b/synapse/config/sso.py @@ -12,6 +12,7 @@ # 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. +import os from typing import Any, Dict import pkg_resources @@ -36,9 +37,32 @@ class SSOConfig(Config): template_dir = pkg_resources.resource_filename("synapse", "res/templates",) self.sso_redirect_confirm_template_dir = template_dir + self.sso_account_deactivated_template = self.read_file( + os.path.join( + self.sso_redirect_confirm_template_dir, "sso_account_deactivated.html" + ), + "sso_account_deactivated_template", + ) + self.sso_auth_success_template = self.read_file( + os.path.join( + self.sso_redirect_confirm_template_dir, "sso_auth_success.html" + ), + "sso_auth_success_template", + ) self.sso_client_whitelist = sso_config.get("client_whitelist") or [] + # Attempt to also whitelist the server's login fallback, since that fallback sets + # the redirect URL to itself (so it can process the login token then return + # gracefully to the client). This would make it pointless to ask the user for + # confirmation, since the URL the confirmation page would be showing wouldn't be + # the client's. + # public_baseurl is an optional setting, so we only add the fallback's URL to the + # list if it's provided (because we can't figure out what that URL is otherwise). + if self.public_baseurl: + login_fallback_url = self.public_baseurl + "_matrix/static/client/login" + self.sso_client_whitelist.append(login_fallback_url) + def generate_config_section(self, **kwargs): return """\ # Additional settings to use with single-sign on systems such as SAML2 and CAS. @@ -54,6 +78,10 @@ class SSOConfig(Config): # phishing attacks from evil.site. To avoid this, include a slash after the # hostname: "https://my.client/". # + # If public_baseurl is set, then the login fallback page (used by clients + # that don't natively support the required login flows) is whitelisted in + # addition to any URLs in this list. + # # By default, this list is empty. # #client_whitelist: @@ -85,6 +113,30 @@ class SSOConfig(Config): # # * server_name: the homeserver's name. # + # * HTML page which notifies the user that they are authenticating to confirm + # an operation on their account during the user interactive authentication + # process: 'sso_auth_confirm.html'. + # + # When rendering, this template is given the following variables: + # * redirect_url: the URL the user is about to be redirected to. Needs + # manual escaping (see + # https://jinja.palletsprojects.com/en/2.11.x/templates/#html-escaping). + # + # * description: the operation which the user is being asked to confirm + # + # * HTML page shown after a successful user interactive authentication session: + # 'sso_auth_success.html'. + # + # Note that this page must include the JavaScript which notifies of a successful authentication + # (see https://matrix.org/docs/spec/client_server/r0.6.0#fallback). + # + # This template has no additional variables. + # + # * HTML page shown during single sign-on if a deactivated user (according to Synapse's database) + # attempts to login: 'sso_account_deactivated.html'. + # + # This template has no additional variables. + # # You can see the default templates at: # https://github.com/matrix-org/synapse/tree/master/synapse/res/templates # diff --git a/synapse/crypto/keyring.py b/synapse/crypto/keyring.py index 983f0ead8..a9f4025bf 100644 --- a/synapse/crypto/keyring.py +++ b/synapse/crypto/keyring.py @@ -43,8 +43,8 @@ from synapse.api.errors import ( SynapseError, ) from synapse.logging.context import ( - LoggingContext, PreserveLoggingContext, + current_context, make_deferred_yieldable, preserve_fn, run_in_background, @@ -236,7 +236,7 @@ class Keyring(object): """ try: - ctx = LoggingContext.current_context() + ctx = current_context() # map from server name to a set of outstanding request ids server_to_request_ids = {} diff --git a/synapse/events/snapshot.py b/synapse/events/snapshot.py index 9ea85e93e..7c5f620d0 100644 --- a/synapse/events/snapshot.py +++ b/synapse/events/snapshot.py @@ -322,11 +322,14 @@ class _AsyncEventContextImpl(EventContext): self._current_state_ids = yield self._storage.state.get_state_ids_for_group( self.state_group ) - if self._prev_state_id and self._event_state_key is not None: + if self._event_state_key is not None: self._prev_state_ids = dict(self._current_state_ids) key = (self._event_type, self._event_state_key) - self._prev_state_ids[key] = self._prev_state_id + if self._prev_state_id: + self._prev_state_ids[key] = self._prev_state_id + else: + self._prev_state_ids.pop(key, None) else: self._prev_state_ids = self._current_state_ids diff --git a/synapse/federation/federation_base.py b/synapse/federation/federation_base.py index 5c991e541..4b115aac0 100644 --- a/synapse/federation/federation_base.py +++ b/synapse/federation/federation_base.py @@ -25,19 +25,15 @@ from twisted.python.failure import Failure from synapse.api.constants import MAX_DEPTH, EventTypes, Membership from synapse.api.errors import Codes, SynapseError -from synapse.api.room_versions import ( - KNOWN_ROOM_VERSIONS, - EventFormatVersions, - RoomVersion, -) +from synapse.api.room_versions import EventFormatVersions, RoomVersion from synapse.crypto.event_signing import check_event_content_hash from synapse.crypto.keyring import Keyring from synapse.events import EventBase, make_event_from_dict from synapse.events.utils import prune_event from synapse.http.servlet import assert_params_in_dict from synapse.logging.context import ( - LoggingContext, PreserveLoggingContext, + current_context, make_deferred_yieldable, ) from synapse.types import JsonDict, get_domain_from_id @@ -55,13 +51,15 @@ class FederationBase(object): self.store = hs.get_datastore() self._clock = hs.get_clock() - def _check_sigs_and_hash(self, room_version: str, pdu: EventBase) -> Deferred: + def _check_sigs_and_hash( + self, room_version: RoomVersion, pdu: EventBase + ) -> Deferred: return make_deferred_yieldable( self._check_sigs_and_hashes(room_version, [pdu])[0] ) def _check_sigs_and_hashes( - self, room_version: str, pdus: List[EventBase] + self, room_version: RoomVersion, pdus: List[EventBase] ) -> List[Deferred]: """Checks that each of the received events is correctly signed by the sending server. @@ -80,7 +78,7 @@ class FederationBase(object): """ deferreds = _check_sigs_on_pdus(self.keyring, room_version, pdus) - ctx = LoggingContext.current_context() + ctx = current_context() def callback(_, pdu: EventBase): with PreserveLoggingContext(ctx): @@ -146,7 +144,7 @@ class PduToCheckSig( def _check_sigs_on_pdus( - keyring: Keyring, room_version: str, pdus: Iterable[EventBase] + keyring: Keyring, room_version: RoomVersion, pdus: Iterable[EventBase] ) -> List[Deferred]: """Check that the given events are correctly signed @@ -191,10 +189,6 @@ def _check_sigs_on_pdus( for p in pdus ] - v = KNOWN_ROOM_VERSIONS.get(room_version) - if not v: - raise RuntimeError("Unrecognized room version %s" % (room_version,)) - # First we check that the sender event is signed by the sender's domain # (except if its a 3pid invite, in which case it may be sent by any server) pdus_to_check_sender = [p for p in pdus_to_check if not _is_invite_via_3pid(p.pdu)] @@ -204,7 +198,7 @@ def _check_sigs_on_pdus( ( p.sender_domain, p.redacted_pdu_json, - p.pdu.origin_server_ts if v.enforce_key_validity else 0, + p.pdu.origin_server_ts if room_version.enforce_key_validity else 0, p.pdu.event_id, ) for p in pdus_to_check_sender @@ -227,7 +221,7 @@ def _check_sigs_on_pdus( # event id's domain (normally only the case for joins/leaves), and add additional # checks. Only do this if the room version has a concept of event ID domain # (ie, the room version uses old-style non-hash event IDs). - if v.event_format == EventFormatVersions.V1: + if room_version.event_format == EventFormatVersions.V1: pdus_to_check_event_id = [ p for p in pdus_to_check @@ -239,7 +233,7 @@ def _check_sigs_on_pdus( ( get_domain_from_id(p.pdu.event_id), p.redacted_pdu_json, - p.pdu.origin_server_ts if v.enforce_key_validity else 0, + p.pdu.origin_server_ts if room_version.enforce_key_validity else 0, p.pdu.event_id, ) for p in pdus_to_check_event_id diff --git a/synapse/federation/federation_client.py b/synapse/federation/federation_client.py index 8c6b83947..687cd841a 100644 --- a/synapse/federation/federation_client.py +++ b/synapse/federation/federation_client.py @@ -220,8 +220,7 @@ class FederationClient(FederationBase): # FIXME: We should handle signature failures more gracefully. pdus[:] = await make_deferred_yieldable( defer.gatherResults( - self._check_sigs_and_hashes(room_version.identifier, pdus), - consumeErrors=True, + self._check_sigs_and_hashes(room_version, pdus), consumeErrors=True, ).addErrback(unwrapFirstError) ) @@ -291,9 +290,7 @@ class FederationClient(FederationBase): pdu = pdu_list[0] # Check signatures are correct. - signed_pdu = await self._check_sigs_and_hash( - room_version.identifier, pdu - ) + signed_pdu = await self._check_sigs_and_hash(room_version, pdu) break @@ -350,7 +347,7 @@ class FederationClient(FederationBase): self, origin: str, pdus: List[EventBase], - room_version: str, + room_version: RoomVersion, outlier: bool = False, include_none: bool = False, ) -> List[EventBase]: @@ -396,7 +393,7 @@ class FederationClient(FederationBase): self.get_pdu( destinations=[pdu.origin], event_id=pdu.event_id, - room_version=room_version, # type: ignore + room_version=room_version, outlier=outlier, timeout=10000, ) @@ -434,7 +431,7 @@ class FederationClient(FederationBase): ] signed_auth = await self._check_sigs_and_hash_and_fetch( - destination, auth_chain, outlier=True, room_version=room_version.identifier + destination, auth_chain, outlier=True, room_version=room_version ) signed_auth.sort(key=lambda e: e.depth) @@ -661,7 +658,7 @@ class FederationClient(FederationBase): destination, list(pdus.values()), outlier=True, - room_version=room_version.identifier, + room_version=room_version, ) valid_pdus_map = {p.event_id: p for p in valid_pdus} @@ -756,7 +753,7 @@ class FederationClient(FederationBase): pdu = event_from_pdu_json(pdu_dict, room_version) # Check signatures are correct. - pdu = await self._check_sigs_and_hash(room_version.identifier, pdu) + pdu = await self._check_sigs_and_hash(room_version, pdu) # FIXME: We should handle signature failures more gracefully. @@ -886,18 +883,37 @@ class FederationClient(FederationBase): def get_public_rooms( self, - destination, - limit=None, - since_token=None, - search_filter=None, - include_all_networks=False, - third_party_instance_id=None, + remote_server: str, + limit: Optional[int] = None, + since_token: Optional[str] = None, + search_filter: Optional[Dict] = None, + include_all_networks: bool = False, + third_party_instance_id: Optional[str] = None, ): - if destination == self.server_name: - return + """Get the list of public rooms from a remote homeserver + Args: + remote_server: The name of the remote server + limit: Maximum amount of rooms to return + since_token: Used for result pagination + search_filter: A filter dictionary to send the remote homeserver + and filter the result set + include_all_networks: Whether to include results from all third party instances + third_party_instance_id: Whether to only include results from a specific third + party instance + + Returns: + Deferred[Dict[str, Any]]: The response from the remote server, or None if + `remote_server` is the same as the local server_name + + Raises: + HttpResponseException: There was an exception returned from the remote server + SynapseException: M_FORBIDDEN when the remote server has disallowed publicRoom + requests over federation + + """ return self.transport_layer.get_public_rooms( - destination, + remote_server, limit, since_token, search_filter, @@ -948,7 +964,7 @@ class FederationClient(FederationBase): ] signed_events = await self._check_sigs_and_hash_and_fetch( - destination, events, outlier=False, room_version=room_version.identifier + destination, events, outlier=False, room_version=room_version ) except HttpResponseException as e: if not e.code == 400: @@ -960,14 +976,13 @@ class FederationClient(FederationBase): return signed_events - @defer.inlineCallbacks - def forward_third_party_invite(self, destinations, room_id, event_dict): + async def forward_third_party_invite(self, destinations, room_id, event_dict): for destination in destinations: if destination == self.server_name: continue try: - yield self.transport_layer.exchange_third_party_invite( + await self.transport_layer.exchange_third_party_invite( destination=destination, room_id=room_id, event_dict=event_dict ) return None diff --git a/synapse/federation/federation_server.py b/synapse/federation/federation_server.py index 275b9c99d..32a8a2ee4 100644 --- a/synapse/federation/federation_server.py +++ b/synapse/federation/federation_server.py @@ -15,7 +15,7 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging -from typing import Dict +from typing import Any, Callable, Dict, List, Match, Optional, Tuple, Union import six from six import iteritems @@ -38,6 +38,7 @@ from synapse.api.errors import ( UnsupportedRoomVersionError, ) from synapse.api.room_versions import KNOWN_ROOM_VERSIONS +from synapse.events import EventBase from synapse.federation.federation_base import FederationBase, event_from_pdu_json from synapse.federation.persistence import TransactionActions from synapse.federation.units import Edu, Transaction @@ -94,7 +95,9 @@ class FederationServer(FederationBase): # come in waves. self._state_resp_cache = ResponseCache(hs, "state_resp", timeout_ms=30000) - async def on_backfill_request(self, origin, room_id, versions, limit): + async def on_backfill_request( + self, origin: str, room_id: str, versions: List[str], limit: int + ) -> Tuple[int, Dict[str, Any]]: with (await self._server_linearizer.queue((origin, room_id))): origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) @@ -107,23 +110,25 @@ class FederationServer(FederationBase): return 200, res - async def on_incoming_transaction(self, origin, transaction_data): + async def on_incoming_transaction( + self, origin: str, transaction_data: JsonDict + ) -> Tuple[int, Dict[str, Any]]: # keep this as early as possible to make the calculated origin ts as # accurate as possible. request_time = self._clock.time_msec() transaction = Transaction(**transaction_data) - if not transaction.transaction_id: + if not transaction.transaction_id: # type: ignore raise Exception("Transaction missing transaction_id") - logger.debug("[%s] Got transaction", transaction.transaction_id) + logger.debug("[%s] Got transaction", transaction.transaction_id) # type: ignore # use a linearizer to ensure that we don't process the same transaction # multiple times in parallel. with ( await self._transaction_linearizer.queue( - (origin, transaction.transaction_id) + (origin, transaction.transaction_id) # type: ignore ) ): result = await self._handle_incoming_transaction( @@ -132,31 +137,33 @@ class FederationServer(FederationBase): return result - async def _handle_incoming_transaction(self, origin, transaction, request_time): + async def _handle_incoming_transaction( + self, origin: str, transaction: Transaction, request_time: int + ) -> Tuple[int, Dict[str, Any]]: """ Process an incoming transaction and return the HTTP response Args: - origin (unicode): the server making the request - transaction (Transaction): incoming transaction - request_time (int): timestamp that the HTTP request arrived at + origin: the server making the request + transaction: incoming transaction + request_time: timestamp that the HTTP request arrived at Returns: - Deferred[(int, object)]: http response code and body + HTTP response code and body """ response = await self.transaction_actions.have_responded(origin, transaction) if response: logger.debug( "[%s] We've already responded to this request", - transaction.transaction_id, + transaction.transaction_id, # type: ignore ) return response - logger.debug("[%s] Transaction is new", transaction.transaction_id) + logger.debug("[%s] Transaction is new", transaction.transaction_id) # type: ignore # Reject if PDU count > 50 or EDU count > 100 - if len(transaction.pdus) > 50 or ( - hasattr(transaction, "edus") and len(transaction.edus) > 100 + if len(transaction.pdus) > 50 or ( # type: ignore + hasattr(transaction, "edus") and len(transaction.edus) > 100 # type: ignore ): logger.info("Transaction PDU or EDU count too large. Returning 400") @@ -204,13 +211,13 @@ class FederationServer(FederationBase): report back to the sending server. """ - received_pdus_counter.inc(len(transaction.pdus)) + received_pdus_counter.inc(len(transaction.pdus)) # type: ignore origin_host, _ = parse_server_name(origin) - pdus_by_room = {} + pdus_by_room = {} # type: Dict[str, List[EventBase]] - for p in transaction.pdus: + for p in transaction.pdus: # type: ignore if "unsigned" in p: unsigned = p["unsigned"] if "age" in unsigned: @@ -254,7 +261,7 @@ class FederationServer(FederationBase): # require callouts to other servers to fetch missing events), but # impose a limit to avoid going too crazy with ram/cpu. - async def process_pdus_for_room(room_id): + async def process_pdus_for_room(room_id: str): logger.debug("Processing PDUs for %s", room_id) try: await self.check_server_matches_acl(origin_host, room_id) @@ -310,7 +317,9 @@ class FederationServer(FederationBase): TRANSACTION_CONCURRENCY_LIMIT, ) - async def on_context_state_request(self, origin, room_id, event_id): + async def on_context_state_request( + self, origin: str, room_id: str, event_id: str + ) -> Tuple[int, Dict[str, Any]]: origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) @@ -338,7 +347,9 @@ class FederationServer(FederationBase): return 200, resp - async def on_state_ids_request(self, origin, room_id, event_id): + async def on_state_ids_request( + self, origin: str, room_id: str, event_id: str + ) -> Tuple[int, Dict[str, Any]]: if not event_id: raise NotImplementedError("Specify an event") @@ -354,7 +365,9 @@ class FederationServer(FederationBase): return 200, {"pdu_ids": state_ids, "auth_chain_ids": auth_chain_ids} - async def _on_context_state_request_compute(self, room_id, event_id): + async def _on_context_state_request_compute( + self, room_id: str, event_id: str + ) -> Dict[str, list]: if event_id: pdus = await self.handler.get_state_for_pdu(room_id, event_id) else: @@ -367,7 +380,9 @@ class FederationServer(FederationBase): "auth_chain": [pdu.get_pdu_json() for pdu in auth_chain], } - async def on_pdu_request(self, origin, event_id): + async def on_pdu_request( + self, origin: str, event_id: str + ) -> Tuple[int, Union[JsonDict, str]]: pdu = await self.handler.get_persisted_pdu(origin, event_id) if pdu: @@ -375,12 +390,16 @@ class FederationServer(FederationBase): else: return 404, "" - async def on_query_request(self, query_type, args): + async def on_query_request( + self, query_type: str, args: Dict[str, str] + ) -> Tuple[int, Dict[str, Any]]: received_queries_counter.labels(query_type).inc() resp = await self.registry.on_query(query_type, args) return 200, resp - async def on_make_join_request(self, origin, room_id, user_id, supported_versions): + async def on_make_join_request( + self, origin: str, room_id: str, user_id: str, supported_versions: List[str] + ) -> Dict[str, Any]: origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) @@ -397,7 +416,7 @@ class FederationServer(FederationBase): async def on_invite_request( self, origin: str, content: JsonDict, room_version_id: str - ): + ) -> Dict[str, Any]: room_version = KNOWN_ROOM_VERSIONS.get(room_version_id) if not room_version: raise SynapseError( @@ -409,12 +428,14 @@ class FederationServer(FederationBase): pdu = event_from_pdu_json(content, room_version) origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, pdu.room_id) - pdu = await self._check_sigs_and_hash(room_version.identifier, pdu) + pdu = await self._check_sigs_and_hash(room_version, pdu) ret_pdu = await self.handler.on_invite_request(origin, pdu, room_version) time_now = self._clock.time_msec() return {"event": ret_pdu.get_pdu_json(time_now)} - async def on_send_join_request(self, origin, content, room_id): + async def on_send_join_request( + self, origin: str, content: JsonDict, room_id: str + ) -> Dict[str, Any]: logger.debug("on_send_join_request: content: %s", content) room_version = await self.store.get_room_version(room_id) @@ -425,7 +446,7 @@ class FederationServer(FederationBase): logger.debug("on_send_join_request: pdu sigs: %s", pdu.signatures) - pdu = await self._check_sigs_and_hash(room_version.identifier, pdu) + pdu = await self._check_sigs_and_hash(room_version, pdu) res_pdus = await self.handler.on_send_join_request(origin, pdu) time_now = self._clock.time_msec() @@ -434,7 +455,9 @@ class FederationServer(FederationBase): "auth_chain": [p.get_pdu_json(time_now) for p in res_pdus["auth_chain"]], } - async def on_make_leave_request(self, origin, room_id, user_id): + async def on_make_leave_request( + self, origin: str, room_id: str, user_id: str + ) -> Dict[str, Any]: origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) pdu = await self.handler.on_make_leave_request(origin, room_id, user_id) @@ -444,7 +467,9 @@ class FederationServer(FederationBase): time_now = self._clock.time_msec() return {"event": pdu.get_pdu_json(time_now), "room_version": room_version} - async def on_send_leave_request(self, origin, content, room_id): + async def on_send_leave_request( + self, origin: str, content: JsonDict, room_id: str + ) -> dict: logger.debug("on_send_leave_request: content: %s", content) room_version = await self.store.get_room_version(room_id) @@ -455,12 +480,14 @@ class FederationServer(FederationBase): logger.debug("on_send_leave_request: pdu sigs: %s", pdu.signatures) - pdu = await self._check_sigs_and_hash(room_version.identifier, pdu) + pdu = await self._check_sigs_and_hash(room_version, pdu) await self.handler.on_send_leave_request(origin, pdu) return {} - async def on_event_auth(self, origin, room_id, event_id): + async def on_event_auth( + self, origin: str, room_id: str, event_id: str + ) -> Tuple[int, Dict[str, Any]]: with (await self._server_linearizer.queue((origin, room_id))): origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) @@ -471,15 +498,21 @@ class FederationServer(FederationBase): return 200, res @log_function - def on_query_client_keys(self, origin, content): - return self.on_query_request("client_keys", content) + async def on_query_client_keys( + self, origin: str, content: Dict[str, str] + ) -> Tuple[int, Dict[str, Any]]: + return await self.on_query_request("client_keys", content) - async def on_query_user_devices(self, origin: str, user_id: str): + async def on_query_user_devices( + self, origin: str, user_id: str + ) -> Tuple[int, Dict[str, Any]]: keys = await self.device_handler.on_federation_query_user_devices(user_id) return 200, keys @trace - async def on_claim_client_keys(self, origin, content): + async def on_claim_client_keys( + self, origin: str, content: JsonDict + ) -> Dict[str, Any]: query = [] for user_id, device_keys in content.get("one_time_keys", {}).items(): for device_id, algorithm in device_keys.items(): @@ -488,7 +521,7 @@ class FederationServer(FederationBase): log_kv({"message": "Claiming one time keys.", "user, device pairs": query}) results = await self.store.claim_e2e_one_time_keys(query) - json_result = {} + json_result = {} # type: Dict[str, Dict[str, dict]] for user_id, device_keys in results.items(): for device_id, keys in device_keys.items(): for key_id, json_bytes in keys.items(): @@ -511,8 +544,13 @@ class FederationServer(FederationBase): return {"one_time_keys": json_result} async def on_get_missing_events( - self, origin, room_id, earliest_events, latest_events, limit - ): + self, + origin: str, + room_id: str, + earliest_events: List[str], + latest_events: List[str], + limit: int, + ) -> Dict[str, list]: with (await self._server_linearizer.queue((origin, room_id))): origin_host, _ = parse_server_name(origin) await self.check_server_matches_acl(origin_host, room_id) @@ -541,11 +579,11 @@ class FederationServer(FederationBase): return {"events": [ev.get_pdu_json(time_now) for ev in missing_events]} @log_function - def on_openid_userinfo(self, token): + async def on_openid_userinfo(self, token: str) -> Optional[str]: ts_now_ms = self._clock.time_msec() - return self.store.get_user_id_for_open_id_token(token, ts_now_ms) + return await self.store.get_user_id_for_open_id_token(token, ts_now_ms) - def _transaction_from_pdus(self, pdu_list): + def _transaction_from_pdus(self, pdu_list: List[EventBase]) -> Transaction: """Returns a new Transaction containing the given PDUs suitable for transmission. """ @@ -558,7 +596,7 @@ class FederationServer(FederationBase): destination=None, ) - async def _handle_received_pdu(self, origin, pdu): + async def _handle_received_pdu(self, origin: str, pdu: EventBase) -> None: """ Process a PDU received in a federation /send/ transaction. If the event is invalid, then this method throws a FederationError. @@ -579,10 +617,8 @@ class FederationServer(FederationBase): until we try to backfill across the discontinuity. Args: - origin (str): server which sent the pdu - pdu (FrozenEvent): received pdu - - Returns (Deferred): completes with None + origin: server which sent the pdu + pdu: received pdu Raises: FederationError if the signatures / hash do not match, or if the event was unacceptable for any other reason (eg, too large, @@ -611,7 +647,7 @@ class FederationServer(FederationBase): logger.info("Accepting join PDU %s from %s", pdu.event_id, origin) # We've already checked that we know the room version by this point - room_version = await self.store.get_room_version_id(pdu.room_id) + room_version = await self.store.get_room_version(pdu.room_id) # Check signature. try: @@ -625,25 +661,27 @@ class FederationServer(FederationBase): return "" % self.server_name async def exchange_third_party_invite( - self, sender_user_id, target_user_id, room_id, signed + self, sender_user_id: str, target_user_id: str, room_id: str, signed: Dict ): ret = await self.handler.exchange_third_party_invite( sender_user_id, target_user_id, room_id, signed ) return ret - async def on_exchange_third_party_invite_request(self, room_id, event_dict): + async def on_exchange_third_party_invite_request( + self, room_id: str, event_dict: Dict + ): ret = await self.handler.on_exchange_third_party_invite_request( room_id, event_dict ) return ret - async def check_server_matches_acl(self, server_name, room_id): + async def check_server_matches_acl(self, server_name: str, room_id: str): """Check if the given server is allowed by the server ACLs in the room Args: - server_name (str): name of server, *without any port part* - room_id (str): ID of the room to check + server_name: name of server, *without any port part* + room_id: ID of the room to check Raises: AuthError if the server does not match the ACL @@ -661,15 +699,15 @@ class FederationServer(FederationBase): raise AuthError(code=403, msg="Server is banned from room") -def server_matches_acl_event(server_name, acl_event): +def server_matches_acl_event(server_name: str, acl_event: EventBase) -> bool: """Check if the given server is allowed by the ACL event Args: - server_name (str): name of server, without any port part - acl_event (EventBase): m.room.server_acl event + server_name: name of server, without any port part + acl_event: m.room.server_acl event Returns: - bool: True if this server is allowed by the ACLs + True if this server is allowed by the ACLs """ logger.debug("Checking %s against acl %s", server_name, acl_event.content) @@ -713,7 +751,7 @@ def server_matches_acl_event(server_name, acl_event): return False -def _acl_entry_matches(server_name, acl_entry): +def _acl_entry_matches(server_name: str, acl_entry: str) -> Match: if not isinstance(acl_entry, six.string_types): logger.warning( "Ignoring non-str ACL entry '%s' (is %s)", acl_entry, type(acl_entry) @@ -732,13 +770,13 @@ class FederationHandlerRegistry(object): self.edu_handlers = {} self.query_handlers = {} - def register_edu_handler(self, edu_type, handler): + def register_edu_handler(self, edu_type: str, handler: Callable[[str, dict], None]): """Sets the handler callable that will be used to handle an incoming federation EDU of the given type. Args: - edu_type (str): The type of the incoming EDU to register handler for - handler (Callable[[str, dict]]): A callable invoked on incoming EDU + edu_type: The type of the incoming EDU to register handler for + handler: A callable invoked on incoming EDU of the given type. The arguments are the origin server name and the EDU contents. """ @@ -749,14 +787,16 @@ class FederationHandlerRegistry(object): self.edu_handlers[edu_type] = handler - def register_query_handler(self, query_type, handler): + def register_query_handler( + self, query_type: str, handler: Callable[[dict], defer.Deferred] + ): """Sets the handler callable that will be used to handle an incoming federation query of the given type. Args: - query_type (str): Category name of the query, which should match + query_type: Category name of the query, which should match the string used by make_query. - handler (Callable[[dict], Deferred[dict]]): Invoked to handle + handler: Invoked to handle incoming queries of this type. The return will be yielded on and the result used as the response to the query request. """ @@ -767,10 +807,11 @@ class FederationHandlerRegistry(object): self.query_handlers[query_type] = handler - async def on_edu(self, edu_type, origin, content): + async def on_edu(self, edu_type: str, origin: str, content: dict): handler = self.edu_handlers.get(edu_type) if not handler: logger.warning("No handler registered for EDU type %s", edu_type) + return with start_active_span_from_edu(content, "handle_edu"): try: @@ -780,7 +821,7 @@ class FederationHandlerRegistry(object): except Exception: logger.exception("Failed to handle edu %r", edu_type) - def on_query(self, query_type, args): + def on_query(self, query_type: str, args: dict) -> defer.Deferred: handler = self.query_handlers.get(query_type) if not handler: logger.warning("No handler registered for query type %s", query_type) @@ -807,7 +848,7 @@ class ReplicationFederationHandlerRegistry(FederationHandlerRegistry): super(ReplicationFederationHandlerRegistry, self).__init__() - async def on_edu(self, edu_type, origin, content): + async def on_edu(self, edu_type: str, origin: str, content: dict): """Overrides FederationHandlerRegistry """ if not self.config.use_presence and edu_type == "m.presence": @@ -821,7 +862,7 @@ class ReplicationFederationHandlerRegistry(FederationHandlerRegistry): return await self._send_edu(edu_type=edu_type, origin=origin, content=content) - async def on_query(self, query_type, args): + async def on_query(self, query_type: str, args: dict): """Overrides FederationHandlerRegistry """ handler = self.query_handlers.get(query_type) diff --git a/synapse/federation/send_queue.py b/synapse/federation/send_queue.py index 876fb0e24..e1700ca8a 100644 --- a/synapse/federation/send_queue.py +++ b/synapse/federation/send_queue.py @@ -477,7 +477,7 @@ def process_rows_for_federation(transaction_queue, rows): Args: transaction_queue (FederationSender) - rows (list(synapse.replication.tcp.streams.FederationStreamRow)) + rows (list(synapse.replication.tcp.streams.federation.FederationStream.FederationStreamRow)) """ # The federation stream contains a bunch of different types of diff --git a/synapse/federation/sender/__init__.py b/synapse/federation/sender/__init__.py index 233cb33da..a477578e4 100644 --- a/synapse/federation/sender/__init__.py +++ b/synapse/federation/sender/__init__.py @@ -499,4 +499,13 @@ class FederationSender(object): self._get_per_destination_queue(destination).attempt_new_transaction() def get_current_token(self) -> int: + # Dummy implementation for case where federation sender isn't offloaded + # to a worker. return 0 + + async def get_replication_rows( + self, from_token, to_token, limit, federation_ack=None + ): + # Dummy implementation for case where federation sender isn't offloaded + # to a worker. + return [] diff --git a/synapse/federation/transport/client.py b/synapse/federation/transport/client.py index 383e3fdc8..060bf0719 100644 --- a/synapse/federation/transport/client.py +++ b/synapse/federation/transport/client.py @@ -15,13 +15,14 @@ # limitations under the License. import logging -from typing import Any, Dict +from typing import Any, Dict, Optional from six.moves import urllib from twisted.internet import defer from synapse.api.constants import Membership +from synapse.api.errors import Codes, HttpResponseException, SynapseError from synapse.api.urls import ( FEDERATION_UNSTABLE_PREFIX, FEDERATION_V1_PREFIX, @@ -326,18 +327,25 @@ class TransportLayerClient(object): @log_function def get_public_rooms( self, - remote_server, - limit, - since_token, - search_filter=None, - include_all_networks=False, - third_party_instance_id=None, + remote_server: str, + limit: Optional[int] = None, + since_token: Optional[str] = None, + search_filter: Optional[Dict] = None, + include_all_networks: bool = False, + third_party_instance_id: Optional[str] = None, ): + """Get the list of public rooms from a remote homeserver + + See synapse.federation.federation_client.FederationClient.get_public_rooms for + more information. + """ if search_filter: # this uses MSC2197 (Search Filtering over Federation) path = _create_v1_path("/publicRooms") - data = {"include_all_networks": "true" if include_all_networks else "false"} + data = { + "include_all_networks": "true" if include_all_networks else "false" + } # type: Dict[str, Any] if third_party_instance_id: data["third_party_instance_id"] = third_party_instance_id if limit: @@ -347,9 +355,19 @@ class TransportLayerClient(object): data["filter"] = search_filter - response = yield self.client.post_json( - destination=remote_server, path=path, data=data, ignore_backoff=True - ) + try: + response = yield self.client.post_json( + destination=remote_server, path=path, data=data, ignore_backoff=True + ) + except HttpResponseException as e: + if e.code == 403: + raise SynapseError( + 403, + "You are not allowed to view the public rooms list of %s" + % (remote_server,), + errcode=Codes.FORBIDDEN, + ) + raise else: path = _create_v1_path("/publicRooms") @@ -363,9 +381,19 @@ class TransportLayerClient(object): if since_token: args["since"] = [since_token] - response = yield self.client.get_json( - destination=remote_server, path=path, args=args, ignore_backoff=True - ) + try: + response = yield self.client.get_json( + destination=remote_server, path=path, args=args, ignore_backoff=True + ) + except HttpResponseException as e: + if e.code == 403: + raise SynapseError( + 403, + "You are not allowed to view the public rooms list of %s" + % (remote_server,), + errcode=Codes.FORBIDDEN, + ) + raise return response diff --git a/synapse/groups/attestations.py b/synapse/groups/attestations.py index d950a8b24..27b0c0265 100644 --- a/synapse/groups/attestations.py +++ b/synapse/groups/attestations.py @@ -37,13 +37,13 @@ An attestation is a signed blob of json that looks like: import logging import random +from typing import Tuple from signedjson.sign import sign_json from twisted.internet import defer from synapse.api.errors import HttpResponseException, RequestSendFailed, SynapseError -from synapse.logging.context import run_in_background from synapse.metrics.background_process_metrics import run_as_background_process from synapse.types import get_domain_from_id @@ -162,19 +162,19 @@ class GroupAttestionRenewer(object): def _start_renew_attestations(self): return run_as_background_process("renew_attestations", self._renew_attestations) - @defer.inlineCallbacks - def _renew_attestations(self): + async def _renew_attestations(self): """Called periodically to check if we need to update any of our attestations """ now = self.clock.time_msec() - rows = yield self.store.get_attestations_need_renewals( + rows = await self.store.get_attestations_need_renewals( now + UPDATE_ATTESTATION_TIME_MS ) @defer.inlineCallbacks - def _renew_attestation(group_id, user_id): + def _renew_attestation(group_user: Tuple[str, str]): + group_id, user_id = group_user try: if not self.is_mine_id(group_id): destination = get_domain_from_id(group_id) @@ -208,7 +208,4 @@ class GroupAttestionRenewer(object): ) for row in rows: - group_id = row["group_id"] - user_id = row["user_id"] - - run_in_background(_renew_attestation, group_id, user_id) + await _renew_attestation((row["group_id"], row["user_id"])) diff --git a/synapse/groups/groups_server.py b/synapse/groups/groups_server.py index 4f0dc0a20..4acb4fa48 100644 --- a/synapse/groups/groups_server.py +++ b/synapse/groups/groups_server.py @@ -748,17 +748,18 @@ class GroupsServerHandler(GroupsServerWorkerHandler): raise NotImplementedError() - @defer.inlineCallbacks - def remove_user_from_group(self, group_id, user_id, requester_user_id, content): + async def remove_user_from_group( + self, group_id, user_id, requester_user_id, content + ): """Remove a user from the group; either a user is leaving or an admin kicked them. """ - yield self.check_group_is_ours(group_id, requester_user_id, and_exists=True) + await self.check_group_is_ours(group_id, requester_user_id, and_exists=True) is_kick = False if requester_user_id != user_id: - is_admin = yield self.store.is_user_admin_in_group( + is_admin = await self.store.is_user_admin_in_group( group_id, requester_user_id ) if not is_admin: @@ -766,30 +767,29 @@ class GroupsServerHandler(GroupsServerWorkerHandler): is_kick = True - yield self.store.remove_user_from_group(group_id, user_id) + await self.store.remove_user_from_group(group_id, user_id) if is_kick: if self.hs.is_mine_id(user_id): groups_local = self.hs.get_groups_local_handler() - yield groups_local.user_removed_from_group(group_id, user_id, {}) + await groups_local.user_removed_from_group(group_id, user_id, {}) else: - yield self.transport_client.remove_user_from_group_notification( + await self.transport_client.remove_user_from_group_notification( get_domain_from_id(user_id), group_id, user_id, {} ) if not self.hs.is_mine_id(user_id): - yield self.store.maybe_delete_remote_profile_cache(user_id) + await self.store.maybe_delete_remote_profile_cache(user_id) # Delete group if the last user has left - users = yield self.store.get_users_in_group(group_id, include_private=True) + users = await self.store.get_users_in_group(group_id, include_private=True) if not users: - yield self.store.delete_group(group_id) + await self.store.delete_group(group_id) return {} - @defer.inlineCallbacks - def create_group(self, group_id, requester_user_id, content): - group = yield self.check_group_is_ours(group_id, requester_user_id) + async def create_group(self, group_id, requester_user_id, content): + group = await self.check_group_is_ours(group_id, requester_user_id) logger.info("Attempting to create group with ID: %r", group_id) @@ -799,7 +799,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): if group: raise SynapseError(400, "Group already exists") - is_admin = yield self.auth.is_server_admin( + is_admin = await self.auth.is_server_admin( UserID.from_string(requester_user_id) ) if not is_admin: @@ -822,7 +822,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): long_description = profile.get("long_description") user_profile = content.get("user_profile", {}) - yield self.store.create_group( + await self.store.create_group( group_id, requester_user_id, name=name, @@ -834,7 +834,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): if not self.hs.is_mine_id(requester_user_id): remote_attestation = content["attestation"] - yield self.attestations.verify_attestation( + await self.attestations.verify_attestation( remote_attestation, user_id=requester_user_id, group_id=group_id ) @@ -845,7 +845,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): local_attestation = None remote_attestation = None - yield self.store.add_user_to_group( + await self.store.add_user_to_group( group_id, requester_user_id, is_admin=True, @@ -855,7 +855,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): ) if not self.hs.is_mine_id(requester_user_id): - yield self.store.add_remote_profile_cache( + await self.store.add_remote_profile_cache( requester_user_id, displayname=user_profile.get("displayname"), avatar_url=user_profile.get("avatar_url"), @@ -863,8 +863,7 @@ class GroupsServerHandler(GroupsServerWorkerHandler): return {"group_id": group_id} - @defer.inlineCallbacks - def delete_group(self, group_id, requester_user_id): + async def delete_group(self, group_id, requester_user_id): """Deletes a group, kicking out all current members. Only group admins or server admins can call this request @@ -877,14 +876,14 @@ class GroupsServerHandler(GroupsServerWorkerHandler): Deferred """ - yield self.check_group_is_ours(group_id, requester_user_id, and_exists=True) + await self.check_group_is_ours(group_id, requester_user_id, and_exists=True) # Only server admins or group admins can delete groups. - is_admin = yield self.store.is_user_admin_in_group(group_id, requester_user_id) + is_admin = await self.store.is_user_admin_in_group(group_id, requester_user_id) if not is_admin: - is_admin = yield self.auth.is_server_admin( + is_admin = await self.auth.is_server_admin( UserID.from_string(requester_user_id) ) @@ -892,18 +891,17 @@ class GroupsServerHandler(GroupsServerWorkerHandler): raise SynapseError(403, "User is not an admin") # Before deleting the group lets kick everyone out of it - users = yield self.store.get_users_in_group(group_id, include_private=True) + users = await self.store.get_users_in_group(group_id, include_private=True) - @defer.inlineCallbacks - def _kick_user_from_group(user_id): + async def _kick_user_from_group(user_id): if self.hs.is_mine_id(user_id): groups_local = self.hs.get_groups_local_handler() - yield groups_local.user_removed_from_group(group_id, user_id, {}) + await groups_local.user_removed_from_group(group_id, user_id, {}) else: - yield self.transport_client.remove_user_from_group_notification( + await self.transport_client.remove_user_from_group_notification( get_domain_from_id(user_id), group_id, user_id, {} ) - yield self.store.maybe_delete_remote_profile_cache(user_id) + await self.store.maybe_delete_remote_profile_cache(user_id) # We kick users out in the order of: # 1. Non-admins @@ -922,11 +920,11 @@ class GroupsServerHandler(GroupsServerWorkerHandler): else: non_admins.append(u["user_id"]) - yield concurrently_execute(_kick_user_from_group, non_admins, 10) - yield concurrently_execute(_kick_user_from_group, admins, 10) - yield _kick_user_from_group(requester_user_id) + await concurrently_execute(_kick_user_from_group, non_admins, 10) + await concurrently_execute(_kick_user_from_group, admins, 10) + await _kick_user_from_group(requester_user_id) - yield self.store.delete_group(group_id) + await self.store.delete_group(group_id) def _parse_join_policy_from_contents(content): diff --git a/synapse/handlers/_base.py b/synapse/handlers/_base.py index 51413d910..3b781d983 100644 --- a/synapse/handlers/_base.py +++ b/synapse/handlers/_base.py @@ -126,30 +126,28 @@ class BaseHandler(object): retry_after_ms=int(1000 * (time_allowed - time_now)) ) - @defer.inlineCallbacks - def maybe_kick_guest_users(self, event, context=None): + async def maybe_kick_guest_users(self, event, context=None): # Technically this function invalidates current_state by changing it. # Hopefully this isn't that important to the caller. if event.type == EventTypes.GuestAccess: guest_access = event.content.get("guest_access", "forbidden") if guest_access != "can_join": if context: - current_state_ids = yield context.get_current_state_ids() - current_state = yield self.store.get_events( + current_state_ids = await context.get_current_state_ids() + current_state = await self.store.get_events( list(current_state_ids.values()) ) else: - current_state = yield self.state_handler.get_current_state( + current_state = await self.state_handler.get_current_state( event.room_id ) current_state = list(current_state.values()) logger.info("maybe_kick_guest_users %r", current_state) - yield self.kick_guest_users(current_state) + await self.kick_guest_users(current_state) - @defer.inlineCallbacks - def kick_guest_users(self, current_state): + async def kick_guest_users(self, current_state): for member_event in current_state: try: if member_event.type != EventTypes.Member: @@ -180,7 +178,7 @@ class BaseHandler(object): # homeserver. requester = synapse.types.create_requester(target_user, is_guest=True) handler = self.hs.get_room_member_handler() - yield handler.update_membership( + await handler.update_membership( requester, target_user, member_event.room_id, diff --git a/synapse/handlers/auth.py b/synapse/handlers/auth.py index 7860f9625..5c20e2917 100644 --- a/synapse/handlers/auth.py +++ b/synapse/handlers/auth.py @@ -18,14 +18,12 @@ import logging import time import unicodedata import urllib.parse -from typing import Any, Dict, Iterable, List, Optional +from typing import Any, Callable, Dict, Iterable, List, Optional, Tuple, Union import attr import bcrypt # type: ignore[import] import pymacaroons -from twisted.internet import defer - import synapse.util.stringutils as stringutils from synapse.api.constants import LoginType from synapse.api.errors import ( @@ -43,10 +41,10 @@ from synapse.handlers.ui_auth.checkers import UserInteractiveAuthChecker from synapse.http.server import finish_request from synapse.http.site import SynapseRequest from synapse.logging.context import defer_to_thread +from synapse.metrics.background_process_metrics import run_as_background_process from synapse.module_api import ModuleApi from synapse.push.mailer import load_jinja2_templates from synapse.types import Requester, UserID -from synapse.util.caches.expiringcache import ExpiringCache from ._base import BaseHandler @@ -71,15 +69,6 @@ class AuthHandler(BaseHandler): self.bcrypt_rounds = hs.config.bcrypt_rounds - # This is not a cache per se, but a store of all current sessions that - # expire after N hours - self.sessions = ExpiringCache( - cache_name="register_sessions", - clock=hs.get_clock(), - expiry_ms=self.SESSION_EXPIRE_MS, - reset_expiry_on_get=True, - ) - account_handler = ModuleApi(hs, self) self.password_providers = [ module(config=config, account_handler=account_handler) @@ -91,6 +80,7 @@ class AuthHandler(BaseHandler): self.hs = hs # FIXME better possibility to access registrationHandler later? self.macaroon_gen = hs.get_macaroon_generator() self._password_enabled = hs.config.password_enabled + self._sso_enabled = hs.config.saml2_enabled or hs.config.cas_enabled # we keep this as a list despite the O(N^2) implication so that we can # keep PASSWORD first and avoid confusing clients which pick the first @@ -106,6 +96,13 @@ class AuthHandler(BaseHandler): if t not in login_types: login_types.append(t) self._supported_login_types = login_types + # Login types and UI Auth types have a heavy overlap, but are not + # necessarily identical. Login types have SSO (and other login types) + # added in the rest layer, see synapse.rest.client.v1.login.LoginRestServerlet.on_GET. + ui_auth_types = login_types.copy() + if self._sso_enabled: + ui_auth_types.append(LoginType.SSO) + self._supported_ui_auth_types = ui_auth_types # Ratelimiter for failed auth during UIA. Uses same ratelimit config # as per `rc_login.failed_attempts`. @@ -113,20 +110,52 @@ class AuthHandler(BaseHandler): self._clock = self.hs.get_clock() - # Load the SSO redirect confirmation page HTML template + # Expire old UI auth sessions after a period of time. + if hs.config.worker_app is None: + self._clock.looping_call( + run_as_background_process, + 5 * 60 * 1000, + "expire_old_sessions", + self._expire_old_sessions, + ) + + # Load the SSO HTML templates. + + # The following template is shown to the user during a client login via SSO, + # after the SSO completes and before redirecting them back to their client. + # It notifies the user they are about to give access to their matrix account + # to the client. self._sso_redirect_confirm_template = load_jinja2_templates( hs.config.sso_redirect_confirm_template_dir, ["sso_redirect_confirm.html"], )[0] + # The following template is shown during user interactive authentication + # in the fallback auth scenario. It notifies the user that they are + # authenticating for an operation to occur on their account. + self._sso_auth_confirm_template = load_jinja2_templates( + hs.config.sso_redirect_confirm_template_dir, ["sso_auth_confirm.html"], + )[0] + # The following template is shown after a successful user interactive + # authentication session. It tells the user they can close the window. + self._sso_auth_success_template = hs.config.sso_auth_success_template + # The following template is shown during the SSO authentication process if + # the account is deactivated. + self._sso_account_deactivated_template = ( + hs.config.sso_account_deactivated_template + ) self._server_name = hs.config.server_name # cast to tuple for use with str.startswith self._whitelisted_sso_clients = tuple(hs.config.sso_client_whitelist) - @defer.inlineCallbacks - def validate_user_via_ui_auth( - self, requester: Requester, request_body: Dict[str, Any], clientip: str - ): + async def validate_user_via_ui_auth( + self, + requester: Requester, + request: SynapseRequest, + request_body: Dict[str, Any], + clientip: str, + description: str, + ) -> dict: """ Checks that the user is who they claim to be, via a UI auth. @@ -137,12 +166,17 @@ class AuthHandler(BaseHandler): Args: requester: The user, as given by the access token + request: The request sent by the client. + request_body: The body of the request sent by the client clientip: The IP address of the client. + description: A human readable string to be displayed to the user that + describes the operation happening on their account. + Returns: - defer.Deferred[dict]: the parameters for this request (which may + The parameters for this request (which may have been given only in a previous call). Raises: @@ -169,10 +203,12 @@ class AuthHandler(BaseHandler): ) # build a list of supported flows - flows = [[login_type] for login_type in self._supported_login_types] + flows = [[login_type] for login_type in self._supported_ui_auth_types] try: - result, params, _ = yield self.check_auth(flows, request_body, clientip) + result, params, _ = await self.check_auth( + flows, request, request_body, clientip, description + ) except LoginError: # Update the ratelimite to say we failed (`can_do_action` doesn't raise). self._failed_uia_attempts_ratelimiter.can_do_action( @@ -185,7 +221,7 @@ class AuthHandler(BaseHandler): raise # find the completed login type - for login_type in self._supported_login_types: + for login_type in self._supported_ui_auth_types: if login_type not in result: continue @@ -209,18 +245,18 @@ class AuthHandler(BaseHandler): """ return self.checkers.keys() - @defer.inlineCallbacks - def check_auth( - self, flows: List[List[str]], clientdict: Dict[str, Any], clientip: str - ): + async def check_auth( + self, + flows: List[List[str]], + request: SynapseRequest, + clientdict: Dict[str, Any], + clientip: str, + description: str, + ) -> Tuple[dict, dict, str]: """ Takes a dictionary sent by the client in the login / registration protocol and handles the User-Interactive Auth flow. - As a side effect, this function fills in the 'creds' key on the user's - session with a map, which maps each auth-type (str) to the relevant - identity authenticated by that auth-type (mostly str, but for captcha, bool). - If no auth flows have been completed successfully, raises an InteractiveAuthIncompleteError. To handle this, you can use synapse.rest.client.v2_alpha._base.interactive_auth_handler as a @@ -231,14 +267,18 @@ class AuthHandler(BaseHandler): strings representing auth-types. At least one full flow must be completed in order for auth to be successful. + request: The request sent by the client. + clientdict: The dictionary from the client root level, not the 'auth' key: this method prompts for auth if none is sent. clientip: The IP address of the client. + description: A human readable string to be displayed to the user that + describes the operation happening on their account. + Returns: - defer.Deferred[dict, dict, str]: a deferred tuple of - (creds, params, session_id). + A tuple of (creds, params, session_id). 'creds' contains the authenticated credentials of each stage. @@ -260,9 +300,26 @@ class AuthHandler(BaseHandler): del clientdict["auth"] if "session" in authdict: sid = authdict["session"] - session = self._get_session_info(sid) - if len(clientdict) > 0: + # Convert the URI and method to strings. + uri = request.uri.decode("utf-8") + method = request.uri.decode("utf-8") + + # If there's no session ID, create a new session. + if not sid: + session = await self.store.create_ui_auth_session( + clientdict, uri, method, description + ) + + else: + try: + session = await self.store.get_ui_auth_session(sid) + except StoreError: + raise SynapseError(400, "Unknown session ID: %s" % (sid,)) + + # If the client provides parameters, update what is persisted, + # otherwise use whatever was last provided. + # # This was designed to allow the client to omit the parameters # and just supply the session in subsequent calls so it split # auth between devices by just sharing the session, (eg. so you @@ -270,31 +327,60 @@ class AuthHandler(BaseHandler): # email auth link on there). It's probably too open to abuse # because it lets unauthenticated clients store arbitrary objects # on a homeserver. - # Revisit: Assumimg the REST APIs do sensible validation, the data - # isn't arbintrary. - session["clientdict"] = clientdict - self._save_session(session) - elif "clientdict" in session: - clientdict = session["clientdict"] + # + # Revisit: Assuming the REST APIs do sensible validation, the data + # isn't arbitrary. + # + # Note that the registration endpoint explicitly removes the + # "initial_device_display_name" parameter if it is provided + # without a "password" parameter. See the changes to + # synapse.rest.client.v2_alpha.register.RegisterRestServlet.on_POST + # in commit 544722bad23fc31056b9240189c3cbbbf0ffd3f9. + if not clientdict: + clientdict = session.clientdict + + # Ensure that the queried operation does not vary between stages of + # the UI authentication session. This is done by generating a stable + # comparator and storing it during the initial query. Subsequent + # queries ensure that this comparator has not changed. + # + # The comparator is based on the requested URI and HTTP method. The + # client dict (minus the auth dict) should also be checked, but some + # clients are not spec compliant, just warn for now if the client + # dict changes. + if (session.uri, session.method) != (uri, method): + raise SynapseError( + 403, + "Requested operation has changed during the UI authentication session.", + ) + + if session.clientdict != clientdict: + logger.warning( + "Requested operation has changed during the UI " + "authentication session. A future version of Synapse " + "will remove this capability." + ) + + # For backwards compatibility, changes to the client dict are + # persisted as clients modify them throughout their user interactive + # authentication flow. + await self.store.set_ui_auth_clientdict(sid, clientdict) if not authdict: raise InteractiveAuthIncompleteError( - self._auth_dict_for_flows(flows, session) + self._auth_dict_for_flows(flows, session.session_id) ) - if "creds" not in session: - session["creds"] = {} - creds = session["creds"] - # check auth type currently being presented errordict = {} # type: Dict[str, Any] if "type" in authdict: login_type = authdict["type"] # type: str try: - result = yield self._check_auth_dict(authdict, clientip) + result = await self._check_auth_dict(authdict, clientip) if result: - creds[login_type] = result - self._save_session(session) + await self.store.mark_ui_auth_stage_complete( + session.session_id, login_type, result + ) except LoginError as e: if login_type == LoginType.EMAIL_IDENTITY: # riot used to have a bug where it would request a new @@ -310,6 +396,7 @@ class AuthHandler(BaseHandler): # so that the client can have another go. errordict = e.error_dict() + creds = await self.store.get_completed_ui_auth_stages(session.session_id) for f in flows: if len(set(f) - set(creds)) == 0: # it's very useful to know what args are stored, but this can @@ -322,15 +409,17 @@ class AuthHandler(BaseHandler): creds, list(clientdict), ) - return creds, clientdict, session["id"] - ret = self._auth_dict_for_flows(flows, session) + return creds, clientdict, session.session_id + + ret = self._auth_dict_for_flows(flows, session.session_id) ret["completed"] = list(creds) ret.update(errordict) raise InteractiveAuthIncompleteError(ret) - @defer.inlineCallbacks - def add_oob_auth(self, stagetype: str, authdict: Dict[str, Any], clientip: str): + async def add_oob_auth( + self, stagetype: str, authdict: Dict[str, Any], clientip: str + ) -> bool: """ Adds the result of out-of-band authentication into an existing auth session. Currently used for adding the result of fallback auth. @@ -340,15 +429,11 @@ class AuthHandler(BaseHandler): if "session" not in authdict: raise LoginError(400, "", Codes.MISSING_PARAM) - sess = self._get_session_info(authdict["session"]) - if "creds" not in sess: - sess["creds"] = {} - creds = sess["creds"] - - result = yield self.checkers[stagetype].check_auth(authdict, clientip) + result = await self.checkers[stagetype].check_auth(authdict, clientip) if result: - creds[stagetype] = result - self._save_session(sess) + await self.store.mark_ui_auth_stage_complete( + authdict["session"], stagetype, result + ) return True return False @@ -370,7 +455,7 @@ class AuthHandler(BaseHandler): sid = authdict["session"] return sid - def set_session_data(self, session_id: str, key: str, value: Any) -> None: + async def set_session_data(self, session_id: str, key: str, value: Any) -> None: """ Store a key-value pair into the sessions data associated with this request. This data is stored server-side and cannot be modified by @@ -381,11 +466,12 @@ class AuthHandler(BaseHandler): key: The key to store the data under value: The data to store """ - sess = self._get_session_info(session_id) - sess.setdefault("serverdict", {})[key] = value - self._save_session(sess) + try: + await self.store.set_ui_auth_session_data(session_id, key, value) + except StoreError: + raise SynapseError(400, "Unknown session ID: %s" % (session_id,)) - def get_session_data( + async def get_session_data( self, session_id: str, key: str, default: Optional[Any] = None ) -> Any: """ @@ -396,11 +482,22 @@ class AuthHandler(BaseHandler): key: The key to store the data under default: Value to return if the key has not been set """ - sess = self._get_session_info(session_id) - return sess.setdefault("serverdict", {}).get(key, default) + try: + return await self.store.get_ui_auth_session_data(session_id, key, default) + except StoreError: + raise SynapseError(400, "Unknown session ID: %s" % (session_id,)) - @defer.inlineCallbacks - def _check_auth_dict(self, authdict: Dict[str, Any], clientip: str): + async def _expire_old_sessions(self): + """ + Invalidate any user interactive authentication sessions that have expired. + """ + now = self._clock.time_msec() + expiration_time = now - self.SESSION_EXPIRE_MS + await self.store.delete_old_ui_auth_sessions(expiration_time) + + async def _check_auth_dict( + self, authdict: Dict[str, Any], clientip: str + ) -> Union[Dict[str, Any], str]: """Attempt to validate the auth dict provided by a client Args: @@ -408,7 +505,7 @@ class AuthHandler(BaseHandler): clientip: IP address of the client Returns: - Deferred: result of the stage verification. + Result of the stage verification. Raises: StoreError if there was a problem accessing the database @@ -418,7 +515,7 @@ class AuthHandler(BaseHandler): login_type = authdict["type"] checker = self.checkers.get(login_type) if checker is not None: - res = yield checker.check_auth(authdict, clientip=clientip) + res = await checker.check_auth(authdict, clientip=clientip) return res # build a v1-login-style dict out of the authdict and fall back to the @@ -428,7 +525,7 @@ class AuthHandler(BaseHandler): if user_id is None: raise SynapseError(400, "", Codes.MISSING_PARAM) - (canonical_id, callback) = yield self.validate_login(user_id, authdict) + (canonical_id, callback) = await self.validate_login(user_id, authdict) return canonical_id def _get_params_recaptcha(self) -> dict: @@ -452,7 +549,7 @@ class AuthHandler(BaseHandler): } def _auth_dict_for_flows( - self, flows: List[List[str]], session: Dict[str, Any] + self, flows: List[List[str]], session_id: str, ) -> Dict[str, Any]: public_flows = [] for f in flows: @@ -471,31 +568,12 @@ class AuthHandler(BaseHandler): params[stage] = get_params[stage]() return { - "session": session["id"], + "session": session_id, "flows": [{"stages": f} for f in public_flows], "params": params, } - def _get_session_info(self, session_id: Optional[str]) -> dict: - """ - Gets or creates a session given a session ID. - - The session can be used to track data across multiple requests, e.g. for - interactive authentication. - """ - if session_id not in self.sessions: - session_id = None - - if not session_id: - # create a new session - while session_id is None or session_id in self.sessions: - session_id = stringutils.random_string(24) - self.sessions[session_id] = {"id": session_id} - - return self.sessions[session_id] - - @defer.inlineCallbacks - def get_access_token_for_user_id( + async def get_access_token_for_user_id( self, user_id: str, device_id: Optional[str], valid_until_ms: Optional[int] ): """ @@ -525,10 +603,10 @@ class AuthHandler(BaseHandler): ) logger.info("Logging in user %s on device %s%s", user_id, device_id, fmt_expiry) - yield self.auth.check_auth_blocking(user_id) + await self.auth.check_auth_blocking(user_id) access_token = self.macaroon_gen.generate_access_token(user_id) - yield self.store.add_access_token_to_user( + await self.store.add_access_token_to_user( user_id, access_token, device_id, valid_until_ms ) @@ -538,15 +616,14 @@ class AuthHandler(BaseHandler): # device, so we double-check it here. if device_id is not None: try: - yield self.store.get_device(user_id, device_id) + await self.store.get_device(user_id, device_id) except StoreError: - yield self.store.delete_access_token(access_token) + await self.store.delete_access_token(access_token) raise StoreError(400, "Login raced against device deletion") return access_token - @defer.inlineCallbacks - def check_user_exists(self, user_id: str): + async def check_user_exists(self, user_id: str) -> Optional[str]: """ Checks to see if a user with the given id exists. Will check case insensitively, but return None if there are multiple inexact matches. @@ -555,28 +632,25 @@ class AuthHandler(BaseHandler): user_id: complete @user:id Returns: - defer.Deferred: (unicode) canonical_user_id, or None if zero or - multiple matches - - Raises: - UserDeactivatedError if a user is found but is deactivated. + The canonical_user_id, or None if zero or multiple matches """ - res = yield self._find_user_id_and_pwd_hash(user_id) + res = await self._find_user_id_and_pwd_hash(user_id) if res is not None: return res[0] return None - @defer.inlineCallbacks - def _find_user_id_and_pwd_hash(self, user_id: str): + async def _find_user_id_and_pwd_hash( + self, user_id: str + ) -> Optional[Tuple[str, str]]: """Checks to see if a user with the given id exists. Will check case insensitively, but will return None if there are multiple inexact matches. Returns: - tuple: A 2-tuple of `(canonical_user_id, password_hash)` - None: if there is not exactly one match + A 2-tuple of `(canonical_user_id, password_hash)` or `None` + if there is not exactly one match """ - user_infos = yield self.store.get_users_by_id_case_insensitive(user_id) + user_infos = await self.store.get_users_by_id_case_insensitive(user_id) result = None if not user_infos: @@ -609,8 +683,9 @@ class AuthHandler(BaseHandler): """ return self._supported_login_types - @defer.inlineCallbacks - def validate_login(self, username: str, login_submission: Dict[str, Any]): + async def validate_login( + self, username: str, login_submission: Dict[str, Any] + ) -> Tuple[str, Optional[Callable[[Dict[str, str]], None]]]: """Authenticates the user for the /login API Also used by the user-interactive auth flow to validate @@ -621,7 +696,7 @@ class AuthHandler(BaseHandler): login_submission: the whole of the login submission (including 'type' and other relevant fields) Returns: - Deferred[str, func]: canonical user id, and optional callback + A tuple of the canonical user id, and optional callback to be called once the access token and device id are issued Raises: StoreError if there was a problem accessing the database @@ -650,7 +725,7 @@ class AuthHandler(BaseHandler): for provider in self.password_providers: if hasattr(provider, "check_password") and login_type == LoginType.PASSWORD: known_login_type = True - is_valid = yield provider.check_password(qualified_user_id, password) + is_valid = await provider.check_password(qualified_user_id, password) if is_valid: return qualified_user_id, None @@ -682,7 +757,7 @@ class AuthHandler(BaseHandler): % (login_type, missing_fields), ) - result = yield provider.check_auth(username, login_type, login_dict) + result = await provider.check_auth(username, login_type, login_dict) if result: if isinstance(result, str): result = (result, None) @@ -691,8 +766,8 @@ class AuthHandler(BaseHandler): if login_type == LoginType.PASSWORD and self.hs.config.password_localdb_enabled: known_login_type = True - canonical_user_id = yield self._check_local_password( - qualified_user_id, password + canonical_user_id = await self._check_local_password( + qualified_user_id, password # type: ignore ) if canonical_user_id: @@ -705,8 +780,9 @@ class AuthHandler(BaseHandler): # login, it turns all LoginErrors into a 401 anyway. raise LoginError(403, "Invalid password", errcode=Codes.FORBIDDEN) - @defer.inlineCallbacks - def check_password_provider_3pid(self, medium: str, address: str, password: str): + async def check_password_provider_3pid( + self, medium: str, address: str, password: str + ) -> Tuple[Optional[str], Optional[Callable[[Dict[str, str]], None]]]: """Check if a password provider is able to validate a thirdparty login Args: @@ -715,9 +791,8 @@ class AuthHandler(BaseHandler): password: The password of the user. Returns: - Deferred[(str|None, func|None)]: A tuple of `(user_id, - callback)`. If authentication is successful, `user_id` is a `str` - containing the authenticated, canonical user ID. `callback` is + A tuple of `(user_id, callback)`. If authentication is successful, + `user_id`is the authenticated, canonical user ID. `callback` is then either a function to be later run after the server has completed login/registration, or `None`. If authentication was unsuccessful, `user_id` and `callback` are both `None`. @@ -729,7 +804,7 @@ class AuthHandler(BaseHandler): # success, to a str (which is the user_id) or a tuple of # (user_id, callback_func), where callback_func should be run # after we've finished everything else - result = yield provider.check_3pid_auth(medium, address, password) + result = await provider.check_3pid_auth(medium, address, password) if result: # Check if the return value is a str or a tuple if isinstance(result, str): @@ -739,8 +814,7 @@ class AuthHandler(BaseHandler): return None, None - @defer.inlineCallbacks - def _check_local_password(self, user_id: str, password: str): + async def _check_local_password(self, user_id: str, password: str) -> Optional[str]: """Authenticate a user against the local password database. user_id is checked case insensitively, but will return None if there are @@ -750,28 +824,26 @@ class AuthHandler(BaseHandler): user_id: complete @user:id password: the provided password Returns: - Deferred[unicode] the canonical_user_id, or Deferred[None] if - unknown user/bad password + The canonical_user_id, or None if unknown user/bad password """ - lookupres = yield self._find_user_id_and_pwd_hash(user_id) + lookupres = await self._find_user_id_and_pwd_hash(user_id) if not lookupres: return None (user_id, password_hash) = lookupres # If the password hash is None, the account has likely been deactivated if not password_hash: - deactivated = yield self.store.get_user_deactivated_status(user_id) + deactivated = await self.store.get_user_deactivated_status(user_id) if deactivated: raise UserDeactivatedError("This account has been deactivated") - result = yield self.validate_hash(password, password_hash) + result = await self.validate_hash(password, password_hash) if not result: logger.warning("Failed password login for user %s", user_id) return None return user_id - @defer.inlineCallbacks - def validate_short_term_login_token_and_get_user_id(self, login_token: str): + async def validate_short_term_login_token_and_get_user_id(self, login_token: str): auth_api = self.hs.get_auth() user_id = None try: @@ -781,26 +853,23 @@ class AuthHandler(BaseHandler): except Exception: raise AuthError(403, "Invalid token", errcode=Codes.FORBIDDEN) - yield self.auth.check_auth_blocking(user_id) + await self.auth.check_auth_blocking(user_id) return user_id - @defer.inlineCallbacks - def delete_access_token(self, access_token: str): + async def delete_access_token(self, access_token: str): """Invalidate a single access token Args: access_token: access token to be deleted - Returns: - Deferred """ - user_info = yield self.auth.get_user_by_access_token(access_token) - yield self.store.delete_access_token(access_token) + user_info = await self.auth.get_user_by_access_token(access_token) + await self.store.delete_access_token(access_token) # see if any of our auth providers want to know about this for provider in self.password_providers: if hasattr(provider, "on_logged_out"): - yield provider.on_logged_out( + await provider.on_logged_out( user_id=str(user_info["user"]), device_id=user_info["device_id"], access_token=access_token, @@ -808,12 +877,11 @@ class AuthHandler(BaseHandler): # delete pushers associated with this access token if user_info["token_id"] is not None: - yield self.hs.get_pusherpool().remove_pushers_by_access_token( + await self.hs.get_pusherpool().remove_pushers_by_access_token( str(user_info["user"]), (user_info["token_id"],) ) - @defer.inlineCallbacks - def delete_access_tokens_for_user( + async def delete_access_tokens_for_user( self, user_id: str, except_token_id: Optional[str] = None, @@ -827,10 +895,8 @@ class AuthHandler(BaseHandler): device_id: ID of device the tokens are associated with. If None, tokens associated with any device (or no device) will be deleted - Returns: - Deferred """ - tokens_and_devices = yield self.store.user_delete_access_tokens( + tokens_and_devices = await self.store.user_delete_access_tokens( user_id, except_token_id=except_token_id, device_id=device_id ) @@ -838,17 +904,18 @@ class AuthHandler(BaseHandler): for provider in self.password_providers: if hasattr(provider, "on_logged_out"): for token, token_id, device_id in tokens_and_devices: - yield provider.on_logged_out( + await provider.on_logged_out( user_id=user_id, device_id=device_id, access_token=token ) # delete pushers associated with the access tokens - yield self.hs.get_pusherpool().remove_pushers_by_access_token( + await self.hs.get_pusherpool().remove_pushers_by_access_token( user_id, (token_id for _, token_id, _ in tokens_and_devices) ) - @defer.inlineCallbacks - def add_threepid(self, user_id: str, medium: str, address: str, validated_at: int): + async def add_threepid( + self, user_id: str, medium: str, address: str, validated_at: int + ): # check if medium has a valid value if medium not in ["email", "msisdn"]: raise SynapseError( @@ -869,14 +936,13 @@ class AuthHandler(BaseHandler): if medium == "email": address = address.lower() - yield self.store.user_add_threepid( + await self.store.user_add_threepid( user_id, medium, address, validated_at, self.hs.get_clock().time_msec() ) - @defer.inlineCallbacks - def delete_threepid( + async def delete_threepid( self, user_id: str, medium: str, address: str, id_server: Optional[str] = None - ): + ) -> bool: """Attempts to unbind the 3pid on the identity servers and deletes it from the local database. @@ -889,7 +955,7 @@ class AuthHandler(BaseHandler): identity server specified when binding (if known). Returns: - Deferred[bool]: Returns True if successfully unbound the 3pid on + Returns True if successfully unbound the 3pid on the identity server, False if identity server doesn't support the unbind API. """ @@ -899,28 +965,21 @@ class AuthHandler(BaseHandler): address = address.lower() identity_handler = self.hs.get_handlers().identity_handler - result = yield identity_handler.try_unbind_threepid( + result = await identity_handler.try_unbind_threepid( user_id, {"medium": medium, "address": address, "id_server": id_server} ) - yield self.store.user_delete_threepid(user_id, medium, address) + await self.store.user_delete_threepid(user_id, medium, address) return result - def _save_session(self, session: Dict[str, Any]) -> None: - """Update the last used time on the session to now and add it back to the session store.""" - # TODO: Persistent storage - logger.debug("Saving session %s", session) - session["last_used"] = self.hs.get_clock().time_msec() - self.sessions[session["id"]] = session - - def hash(self, password: str): + async def hash(self, password: str) -> str: """Computes a secure hash of password. Args: password: Password to hash. Returns: - Deferred(unicode): Hashed password. + Hashed password. """ def _do_hash(): @@ -932,9 +991,11 @@ class AuthHandler(BaseHandler): bcrypt.gensalt(self.bcrypt_rounds), ).decode("ascii") - return defer_to_thread(self.hs.get_reactor(), _do_hash) + return await defer_to_thread(self.hs.get_reactor(), _do_hash) - def validate_hash(self, password: str, stored_hash: bytes): + async def validate_hash( + self, password: str, stored_hash: Union[bytes, str] + ) -> bool: """Validates that self.hash(password) == stored_hash. Args: @@ -942,7 +1003,7 @@ class AuthHandler(BaseHandler): stored_hash: Expected hash value. Returns: - Deferred(bool): Whether self.hash(password) == stored_hash. + Whether self.hash(password) == stored_hash. """ def _do_validate_hash(): @@ -958,11 +1019,57 @@ class AuthHandler(BaseHandler): if not isinstance(stored_hash, bytes): stored_hash = stored_hash.encode("ascii") - return defer_to_thread(self.hs.get_reactor(), _do_validate_hash) + return await defer_to_thread(self.hs.get_reactor(), _do_validate_hash) else: - return defer.succeed(False) + return False - def complete_sso_login( + async def start_sso_ui_auth(self, redirect_url: str, session_id: str) -> str: + """ + Get the HTML for the SSO redirect confirmation page. + + Args: + redirect_url: The URL to redirect to the SSO provider. + session_id: The user interactive authentication session ID. + + Returns: + The HTML to render. + """ + try: + session = await self.store.get_ui_auth_session(session_id) + except StoreError: + raise SynapseError(400, "Unknown session ID: %s" % (session_id,)) + return self._sso_auth_confirm_template.render( + description=session.description, redirect_url=redirect_url, + ) + + async def complete_sso_ui_auth( + self, registered_user_id: str, session_id: str, request: SynapseRequest, + ): + """Having figured out a mxid for this user, complete the HTTP request + + Args: + registered_user_id: The registered user ID to complete SSO login for. + request: The request to complete. + client_redirect_url: The URL to which to redirect the user at the end of the + process. + """ + # Mark the stage of the authentication as successful. + # Save the user who authenticated with SSO, this will be used to ensure + # that the account be modified is also the person who logged in. + await self.store.mark_ui_auth_stage_complete( + session_id, LoginType.SSO, registered_user_id + ) + + # Render the HTML and return. + html_bytes = self._sso_auth_success_template.encode("utf-8") + request.setResponseCode(200) + request.setHeader(b"Content-Type", b"text/html; charset=utf-8") + request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) + + request.write(html_bytes) + finish_request(request) + + async def complete_sso_login( self, registered_user_id: str, request: SynapseRequest, @@ -976,6 +1083,32 @@ class AuthHandler(BaseHandler): client_redirect_url: The URL to which to redirect the user at the end of the process. """ + # If the account has been deactivated, do not proceed with the login + # flow. + deactivated = await self.store.get_user_deactivated_status(registered_user_id) + if deactivated: + html_bytes = self._sso_account_deactivated_template.encode("utf-8") + + request.setResponseCode(403) + request.setHeader(b"Content-Type", b"text/html; charset=utf-8") + request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) + request.write(html_bytes) + finish_request(request) + return + + self._complete_sso_login(registered_user_id, request, client_redirect_url) + + def _complete_sso_login( + self, + registered_user_id: str, + request: SynapseRequest, + client_redirect_url: str, + ): + """ + The synchronous portion of complete_sso_login. + + This exists purely for backwards compatibility of synapse.module_api.ModuleApi. + """ # Create a login token login_token = self.macaroon_gen.generate_short_term_login_token( registered_user_id @@ -1001,7 +1134,7 @@ class AuthHandler(BaseHandler): # URL we redirect users to. redirect_url_no_params = client_redirect_url.split("?")[0] - html = self._sso_redirect_confirm_template.render( + html_bytes = self._sso_redirect_confirm_template.render( display_url=redirect_url_no_params, redirect_url=redirect_url, server_name=self._server_name, @@ -1009,8 +1142,8 @@ class AuthHandler(BaseHandler): request.setResponseCode(200) request.setHeader(b"Content-Type", b"text/html; charset=utf-8") - request.setHeader(b"Content-Length", b"%d" % (len(html),)) - request.write(html) + request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) + request.write(html_bytes) finish_request(request) @staticmethod diff --git a/synapse/handlers/cas_handler.py b/synapse/handlers/cas_handler.py new file mode 100644 index 000000000..64aaa1335 --- /dev/null +++ b/synapse/handlers/cas_handler.py @@ -0,0 +1,221 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +import logging +import xml.etree.ElementTree as ET +from typing import Dict, Optional, Tuple + +from six.moves import urllib + +from twisted.web.client import PartialDownloadError + +from synapse.api.errors import Codes, LoginError +from synapse.http.site import SynapseRequest +from synapse.types import UserID, map_username_to_mxid_localpart + +logger = logging.getLogger(__name__) + + +class CasHandler: + """ + Utility class for to handle the response from a CAS SSO service. + + Args: + hs (synapse.server.HomeServer) + """ + + def __init__(self, hs): + self._hostname = hs.hostname + self._auth_handler = hs.get_auth_handler() + self._registration_handler = hs.get_registration_handler() + + self._cas_server_url = hs.config.cas_server_url + self._cas_service_url = hs.config.cas_service_url + self._cas_displayname_attribute = hs.config.cas_displayname_attribute + self._cas_required_attributes = hs.config.cas_required_attributes + + self._http_client = hs.get_proxied_http_client() + + def _build_service_param(self, args: Dict[str, str]) -> str: + """ + Generates a value to use as the "service" parameter when redirecting or + querying the CAS service. + + Args: + args: Additional arguments to include in the final redirect URL. + + Returns: + The URL to use as a "service" parameter. + """ + return "%s%s?%s" % ( + self._cas_service_url, + "/_matrix/client/r0/login/cas/ticket", + urllib.parse.urlencode(args), + ) + + async def _validate_ticket( + self, ticket: str, service_args: Dict[str, str] + ) -> Tuple[str, Optional[str]]: + """ + Validate a CAS ticket with the server, parse the response, and return the user and display name. + + Args: + ticket: The CAS ticket from the client. + service_args: Additional arguments to include in the service URL. + Should be the same as those passed to `get_redirect_url`. + """ + uri = self._cas_server_url + "/proxyValidate" + args = { + "ticket": ticket, + "service": self._build_service_param(service_args), + } + try: + body = await self._http_client.get_raw(uri, args) + except PartialDownloadError as pde: + # Twisted raises this error if the connection is closed, + # even if that's being used old-http style to signal end-of-data + body = pde.response + + user, attributes = self._parse_cas_response(body) + displayname = attributes.pop(self._cas_displayname_attribute, None) + + for required_attribute, required_value in self._cas_required_attributes.items(): + # If required attribute was not in CAS Response - Forbidden + if required_attribute not in attributes: + raise LoginError(401, "Unauthorized", errcode=Codes.UNAUTHORIZED) + + # Also need to check value + if required_value is not None: + actual_value = attributes[required_attribute] + # If required attribute value does not match expected - Forbidden + if required_value != actual_value: + raise LoginError(401, "Unauthorized", errcode=Codes.UNAUTHORIZED) + + return user, displayname + + def _parse_cas_response( + self, cas_response_body: str + ) -> Tuple[str, Dict[str, Optional[str]]]: + """ + Retrieve the user and other parameters from the CAS response. + + Args: + cas_response_body: The response from the CAS query. + + Returns: + A tuple of the user and a mapping of other attributes. + """ + user = None + attributes = {} + try: + root = ET.fromstring(cas_response_body) + if not root.tag.endswith("serviceResponse"): + raise Exception("root of CAS response is not serviceResponse") + success = root[0].tag.endswith("authenticationSuccess") + for child in root[0]: + if child.tag.endswith("user"): + user = child.text + if child.tag.endswith("attributes"): + for attribute in child: + # ElementTree library expands the namespace in + # attribute tags to the full URL of the namespace. + # We don't care about namespace here and it will always + # be encased in curly braces, so we remove them. + tag = attribute.tag + if "}" in tag: + tag = tag.split("}")[1] + attributes[tag] = attribute.text + if user is None: + raise Exception("CAS response does not contain user") + except Exception: + logger.exception("Error parsing CAS response") + raise LoginError(401, "Invalid CAS response", errcode=Codes.UNAUTHORIZED) + if not success: + raise LoginError( + 401, "Unsuccessful CAS response", errcode=Codes.UNAUTHORIZED + ) + return user, attributes + + def get_redirect_url(self, service_args: Dict[str, str]) -> str: + """ + Generates a URL for the CAS server where the client should be redirected. + + Args: + service_args: Additional arguments to include in the final redirect URL. + + Returns: + The URL to redirect the client to. + """ + args = urllib.parse.urlencode( + {"service": self._build_service_param(service_args)} + ) + + return "%s/login?%s" % (self._cas_server_url, args) + + async def handle_ticket( + self, + request: SynapseRequest, + ticket: str, + client_redirect_url: Optional[str], + session: Optional[str], + ) -> None: + """ + Called once the user has successfully authenticated with the SSO. + Validates a CAS ticket sent by the client and completes the auth process. + + If the user interactive authentication session is provided, marks the + UI Auth session as complete, then returns an HTML page notifying the + user they are done. + + Otherwise, this registers the user if necessary, and then returns a + redirect (with a login token) to the client. + + Args: + request: the incoming request from the browser. We'll + respond to it with a redirect or an HTML page. + + ticket: The CAS ticket provided by the client. + + client_redirect_url: the redirectUrl parameter from the `/cas/ticket` HTTP request, if given. + This should be the same as the redirectUrl from the original `/login/sso/redirect` request. + + session: The session parameter from the `/cas/ticket` HTTP request, if given. + This should be the UI Auth session id. + """ + args = {} + if client_redirect_url: + args["redirectUrl"] = client_redirect_url + if session: + args["session"] = session + username, user_display_name = await self._validate_ticket(ticket, args) + + localpart = map_username_to_mxid_localpart(username) + user_id = UserID(localpart, self._hostname).to_string() + registered_user_id = await self._auth_handler.check_user_exists(user_id) + + if session: + await self._auth_handler.complete_sso_ui_auth( + registered_user_id, session, request, + ) + + else: + if not registered_user_id: + registered_user_id = await self._registration_handler.register_user( + localpart=localpart, default_display_name=user_display_name + ) + + await self._auth_handler.complete_sso_login( + registered_user_id, request, client_redirect_url + ) diff --git a/synapse/handlers/device.py b/synapse/handlers/device.py index 993499f44..9bd941b5a 100644 --- a/synapse/handlers/device.py +++ b/synapse/handlers/device.py @@ -338,8 +338,10 @@ class DeviceHandler(DeviceWorkerHandler): else: raise - yield self._auth_handler.delete_access_tokens_for_user( - user_id, device_id=device_id + yield defer.ensureDeferred( + self._auth_handler.delete_access_tokens_for_user( + user_id, device_id=device_id + ) ) yield self.store.delete_e2e_keys_by_device(user_id=user_id, device_id=device_id) @@ -391,8 +393,10 @@ class DeviceHandler(DeviceWorkerHandler): # Delete access tokens and e2e keys for each device. Not optimised as it is not # considered as part of a critical path. for device_id in device_ids: - yield self._auth_handler.delete_access_tokens_for_user( - user_id, device_id=device_id + yield defer.ensureDeferred( + self._auth_handler.delete_access_tokens_for_user( + user_id, device_id=device_id + ) ) yield self.store.delete_e2e_keys_by_device( user_id=user_id, device_id=device_id diff --git a/synapse/handlers/directory.py b/synapse/handlers/directory.py index 1d842c369..f2f16b1e4 100644 --- a/synapse/handlers/directory.py +++ b/synapse/handlers/directory.py @@ -86,8 +86,7 @@ class DirectoryHandler(BaseHandler): room_alias, room_id, servers, creator=creator ) - @defer.inlineCallbacks - def create_association( + async def create_association( self, requester: Requester, room_alias: RoomAlias, @@ -127,8 +126,12 @@ class DirectoryHandler(BaseHandler): errcode=Codes.EXCLUSIVE, ) else: - if self.require_membership and check_membership: - rooms_for_user = yield self.store.get_rooms_for_user(user_id) + # Server admins are not subject to the same constraints as normal + # users when creating an alias (e.g. being in the room). + is_admin = await self.auth.is_server_admin(requester.user) + + if (self.require_membership and check_membership) and not is_admin: + rooms_for_user = await self.store.get_rooms_for_user(user_id) if room_id not in rooms_for_user: raise AuthError( 403, "You must be in the room to create an alias for it" @@ -145,7 +148,7 @@ class DirectoryHandler(BaseHandler): # per alias creation rule? raise SynapseError(403, "Not allowed to create alias") - can_create = yield self.can_modify_alias(room_alias, user_id=user_id) + can_create = await self.can_modify_alias(room_alias, user_id=user_id) if not can_create: raise AuthError( 400, @@ -153,10 +156,9 @@ class DirectoryHandler(BaseHandler): errcode=Codes.EXCLUSIVE, ) - yield self._create_association(room_alias, room_id, servers, creator=user_id) + await self._create_association(room_alias, room_id, servers, creator=user_id) - @defer.inlineCallbacks - def delete_association(self, requester: Requester, room_alias: RoomAlias): + async def delete_association(self, requester: Requester, room_alias: RoomAlias): """Remove an alias from the directory (this is only meant for human users; AS users should call @@ -180,7 +182,7 @@ class DirectoryHandler(BaseHandler): user_id = requester.user.to_string() try: - can_delete = yield self._user_can_delete_alias(room_alias, user_id) + can_delete = await self._user_can_delete_alias(room_alias, user_id) except StoreError as e: if e.code == 404: raise NotFoundError("Unknown room alias") @@ -189,7 +191,7 @@ class DirectoryHandler(BaseHandler): if not can_delete: raise AuthError(403, "You don't have permission to delete the alias.") - can_delete = yield self.can_modify_alias(room_alias, user_id=user_id) + can_delete = await self.can_modify_alias(room_alias, user_id=user_id) if not can_delete: raise SynapseError( 400, @@ -197,10 +199,10 @@ class DirectoryHandler(BaseHandler): errcode=Codes.EXCLUSIVE, ) - room_id = yield self._delete_association(room_alias) + room_id = await self._delete_association(room_alias) try: - yield self._update_canonical_alias(requester, user_id, room_id, room_alias) + await self._update_canonical_alias(requester, user_id, room_id, room_alias) except AuthError as e: logger.info("Failed to update alias events: %s", e) @@ -292,15 +294,14 @@ class DirectoryHandler(BaseHandler): Codes.NOT_FOUND, ) - @defer.inlineCallbacks - def _update_canonical_alias( + async def _update_canonical_alias( self, requester: Requester, user_id: str, room_id: str, room_alias: RoomAlias ): """ Send an updated canonical alias event if the removed alias was set as the canonical alias or listed in the alt_aliases field. """ - alias_event = yield self.state.get_current_state( + alias_event = await self.state.get_current_state( room_id, EventTypes.CanonicalAlias, "" ) @@ -331,7 +332,7 @@ class DirectoryHandler(BaseHandler): del content["alt_aliases"] if send_update: - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.CanonicalAlias, @@ -372,8 +373,7 @@ class DirectoryHandler(BaseHandler): # either no interested services, or no service with an exclusive lock return defer.succeed(True) - @defer.inlineCallbacks - def _user_can_delete_alias(self, alias: RoomAlias, user_id: str): + async def _user_can_delete_alias(self, alias: RoomAlias, user_id: str): """Determine whether a user can delete an alias. One of the following must be true: @@ -384,24 +384,23 @@ class DirectoryHandler(BaseHandler): for the current room. """ - creator = yield self.store.get_room_alias_creator(alias.to_string()) + creator = await self.store.get_room_alias_creator(alias.to_string()) if creator is not None and creator == user_id: return True # Resolve the alias to the corresponding room. - room_mapping = yield self.get_association(alias) + room_mapping = await self.get_association(alias) room_id = room_mapping["room_id"] if not room_id: return False - res = yield self.auth.check_can_change_room_list( + res = await self.auth.check_can_change_room_list( room_id, UserID.from_string(user_id) ) return res - @defer.inlineCallbacks - def edit_published_room_list( + async def edit_published_room_list( self, requester: Requester, room_id: str, visibility: str ): """Edit the entry of the room in the published room list. @@ -429,11 +428,11 @@ class DirectoryHandler(BaseHandler): 403, "This user is not permitted to publish rooms to the room list" ) - room = yield self.store.get_room(room_id) + room = await self.store.get_room(room_id) if room is None: raise SynapseError(400, "Unknown room") - can_change_room_list = yield self.auth.check_can_change_room_list( + can_change_room_list = await self.auth.check_can_change_room_list( room_id, requester.user ) if not can_change_room_list: @@ -445,8 +444,8 @@ class DirectoryHandler(BaseHandler): making_public = visibility == "public" if making_public: - room_aliases = yield self.store.get_aliases_for_room(room_id) - canonical_alias = yield self.store.get_canonical_alias_for_room(room_id) + room_aliases = await self.store.get_aliases_for_room(room_id) + canonical_alias = await self.store.get_canonical_alias_for_room(room_id) if canonical_alias: room_aliases.append(canonical_alias) @@ -458,7 +457,7 @@ class DirectoryHandler(BaseHandler): # per alias creation rule? raise SynapseError(403, "Not allowed to publish room") - yield self.store.set_room_is_public(room_id, making_public) + await self.store.set_room_is_public(room_id, making_public) @defer.inlineCallbacks def edit_published_appservice_room_list( diff --git a/synapse/handlers/events.py b/synapse/handlers/events.py index ec18a42a6..71a89f09c 100644 --- a/synapse/handlers/events.py +++ b/synapse/handlers/events.py @@ -19,6 +19,7 @@ import random from synapse.api.constants import EventTypes, Membership from synapse.api.errors import AuthError, SynapseError from synapse.events import EventBase +from synapse.handlers.presence import format_user_presence_state from synapse.logging.utils import log_function from synapse.types import UserID from synapse.visibility import filter_events_for_client @@ -97,6 +98,8 @@ class EventStreamHandler(BaseHandler): explicit_room_id=room_id, ) + time_now = self.clock.time_msec() + # When the user joins a new room, or another user joins a currently # joined room, we need to send down presence for those users. to_add = [] @@ -112,19 +115,20 @@ class EventStreamHandler(BaseHandler): users = await self.state.get_current_users_in_room( event.room_id ) - states = await presence_handler.get_states(users, as_event=True) - to_add.extend(states) else: + users = [event.state_key] - ev = await presence_handler.get_state( - UserID.from_string(event.state_key), as_event=True - ) - to_add.append(ev) + states = await presence_handler.get_states(users) + to_add.extend( + { + "type": EventTypes.Presence, + "content": format_user_presence_state(state, time_now), + } + for state in states + ) events.extend(to_add) - time_now = self.clock.time_msec() - chunks = await self._event_serializer.serialize_events( events, time_now, diff --git a/synapse/handlers/federation.py b/synapse/handlers/federation.py index 38ab6a8fc..4e5c64552 100644 --- a/synapse/handlers/federation.py +++ b/synapse/handlers/federation.py @@ -49,6 +49,7 @@ from synapse.event_auth import auth_types_for_event from synapse.events import EventBase from synapse.events.snapshot import EventContext from synapse.events.validator import EventValidator +from synapse.handlers._base import BaseHandler from synapse.logging.context import ( make_deferred_yieldable, nested_logging_context, @@ -69,10 +70,9 @@ from synapse.types import JsonDict, StateMap, UserID, get_domain_from_id from synapse.util.async_helpers import Linearizer, concurrently_execute from synapse.util.distributor import user_joined_room from synapse.util.retryutils import NotRetryingDestination +from synapse.util.stringutils import shortstr from synapse.visibility import filter_events_for_server -from ._base import BaseHandler - logger = logging.getLogger(__name__) @@ -93,27 +93,6 @@ class _NewEventInfo: auth_events = attr.ib(type=Optional[StateMap[EventBase]], default=None) -def shortstr(iterable, maxitems=5): - """If iterable has maxitems or fewer, return the stringification of a list - containing those items. - - Otherwise, return the stringification of a a list with the first maxitems items, - followed by "...". - - Args: - iterable (Iterable): iterable to truncate - maxitems (int): number of items to return before truncating - - Returns: - unicode - """ - - items = list(itertools.islice(iterable, maxitems + 1)) - if len(items) <= maxitems: - return str(items) - return "[" + ", ".join(repr(r) for r in items[:maxitems]) + ", ...]" - - class FederationHandler(BaseHandler): """Handles events that originated from federation. Responsible for: @@ -364,7 +343,7 @@ class FederationHandler(BaseHandler): ours = await self.state_store.get_state_groups_ids(room_id, seen) # state_maps is a list of mappings from (type, state_key) to event_id - state_maps = list(ours.values()) # type: list[StateMap[str]] + state_maps = list(ours.values()) # type: List[StateMap[str]] # we don't need this any more, let's delete it. del ours @@ -1715,16 +1694,15 @@ class FederationHandler(BaseHandler): return None - @defer.inlineCallbacks - def get_state_for_pdu(self, room_id, event_id): + async def get_state_for_pdu(self, room_id: str, event_id: str) -> List[EventBase]: """Returns the state at the event. i.e. not including said event. """ - event = yield self.store.get_event( + event = await self.store.get_event( event_id, allow_none=False, check_room_id=room_id ) - state_groups = yield self.state_store.get_state_groups(room_id, [event_id]) + state_groups = await self.state_store.get_state_groups(room_id, [event_id]) if state_groups: _, state = list(iteritems(state_groups)).pop() @@ -1735,7 +1713,7 @@ class FederationHandler(BaseHandler): if "replaces_state" in event.unsigned: prev_id = event.unsigned["replaces_state"] if prev_id != event.event_id: - prev_event = yield self.store.get_event(prev_id) + prev_event = await self.store.get_event(prev_id) results[(event.type, event.state_key)] = prev_event else: del results[(event.type, event.state_key)] @@ -1745,15 +1723,14 @@ class FederationHandler(BaseHandler): else: return [] - @defer.inlineCallbacks - def get_state_ids_for_pdu(self, room_id, event_id): + 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 = yield self.store.get_event( + event = await self.store.get_event( event_id, allow_none=False, check_room_id=room_id ) - state_groups = yield self.state_store.get_state_groups_ids(room_id, [event_id]) + state_groups = await self.state_store.get_state_groups_ids(room_id, [event_id]) if state_groups: _, state = list(state_groups.items()).pop() @@ -1772,49 +1749,50 @@ class FederationHandler(BaseHandler): else: return [] - @defer.inlineCallbacks @log_function - def on_backfill_request(self, origin, room_id, pdu_list, limit): - in_room = yield self.auth.check_host_in_room(room_id, origin) + async def on_backfill_request( + self, origin: str, room_id: str, pdu_list: List[str], limit: int + ) -> List[EventBase]: + in_room = await self.auth.check_host_in_room(room_id, origin) if not in_room: raise AuthError(403, "Host not in room.") # Synapse asks for 100 events per backfill request. Do not allow more. limit = min(limit, 100) - events = yield self.store.get_backfill_events(room_id, pdu_list, limit) + events = await self.store.get_backfill_events(room_id, pdu_list, limit) - events = yield filter_events_for_server(self.storage, origin, events) + events = await filter_events_for_server(self.storage, origin, events) return events - @defer.inlineCallbacks @log_function - def get_persisted_pdu(self, origin, event_id): + async def get_persisted_pdu( + self, origin: str, event_id: str + ) -> Optional[EventBase]: """Get an event from the database for the given server. Args: - origin [str]: hostname of server which is requesting the event; we + origin: hostname of server which is requesting the event; we will check that the server is allowed to see it. - event_id [str]: id of the event being requested + event_id: id of the event being requested Returns: - Deferred[EventBase|None]: None if we know nothing about the event; - otherwise the (possibly-redacted) event. + None if we know nothing about the event; otherwise the (possibly-redacted) event. Raises: AuthError if the server is not currently in the room """ - event = yield self.store.get_event( + event = await self.store.get_event( event_id, allow_none=True, allow_rejected=True ) if event: - in_room = yield self.auth.check_host_in_room(event.room_id, origin) + in_room = await self.auth.check_host_in_room(event.room_id, origin) if not in_room: raise AuthError(403, "Host not in room.") - events = yield filter_events_for_server(self.storage, origin, [event]) + events = await filter_events_for_server(self.storage, origin, [event]) event = events[0] return event else: @@ -2418,7 +2396,7 @@ class FederationHandler(BaseHandler): """ # exclude the state key of the new event from the current_state in the context. if event.is_state(): - event_key = (event.type, event.state_key) + event_key = (event.type, event.state_key) # type: Optional[Tuple[str, str]] else: event_key = None state_updates = { @@ -2584,9 +2562,8 @@ class FederationHandler(BaseHandler): "missing": [e.event_id for e in missing_locals], } - @defer.inlineCallbacks @log_function - def exchange_third_party_invite( + async def exchange_third_party_invite( self, sender_user_id, target_user_id, room_id, signed ): third_party_invite = {"signed": signed} @@ -2602,16 +2579,16 @@ class FederationHandler(BaseHandler): "state_key": target_user_id, } - if (yield self.auth.check_host_in_room(room_id, self.hs.hostname)): - room_version = yield self.store.get_room_version_id(room_id) + if await self.auth.check_host_in_room(room_id, self.hs.hostname): + room_version = await self.store.get_room_version_id(room_id) builder = self.event_builder_factory.new(room_version, event_dict) EventValidator().validate_builder(builder) - event, context = yield self.event_creation_handler.create_new_client_event( + event, context = await self.event_creation_handler.create_new_client_event( builder=builder ) - event_allowed = yield self.third_party_event_rules.check_event_allowed( + event_allowed = await self.third_party_event_rules.check_event_allowed( event, context ) if not event_allowed: @@ -2623,7 +2600,7 @@ class FederationHandler(BaseHandler): 403, "This event is not allowed in this context", Codes.FORBIDDEN ) - event, context = yield self.add_display_name_to_third_party_invite( + event, context = await self.add_display_name_to_third_party_invite( room_version, event_dict, event, context ) @@ -2634,19 +2611,19 @@ class FederationHandler(BaseHandler): event.internal_metadata.send_on_behalf_of = self.hs.hostname try: - yield self.auth.check_from_context(room_version, event, context) + await self.auth.check_from_context(room_version, event, context) except AuthError as e: logger.warning("Denying new third party invite %r because %s", event, e) raise e - yield self._check_signature(event, context) + await self._check_signature(event, context) # We retrieve the room member handler here as to not cause a cyclic dependency member_handler = self.hs.get_room_member_handler() - yield member_handler.send_membership_event(None, event, context) + await member_handler.send_membership_event(None, event, context) else: destinations = {x.split(":", 1)[-1] for x in (sender_user_id, room_id)} - yield self.federation_client.forward_third_party_invite( + await self.federation_client.forward_third_party_invite( destinations, room_id, event_dict ) diff --git a/synapse/handlers/groups_local.py b/synapse/handlers/groups_local.py index ad2241578..ca5c83811 100644 --- a/synapse/handlers/groups_local.py +++ b/synapse/handlers/groups_local.py @@ -284,15 +284,14 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): set_group_join_policy = _create_rerouter("set_group_join_policy") - @defer.inlineCallbacks - def create_group(self, group_id, user_id, content): + async def create_group(self, group_id, user_id, content): """Create a group """ logger.info("Asking to create group with ID: %r", group_id) if self.is_mine_id(group_id): - res = yield self.groups_server_handler.create_group( + res = await self.groups_server_handler.create_group( group_id, user_id, content ) local_attestation = None @@ -301,10 +300,10 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): local_attestation = self.attestations.create_attestation(group_id, user_id) content["attestation"] = local_attestation - content["user_profile"] = yield self.profile_handler.get_profile(user_id) + content["user_profile"] = await self.profile_handler.get_profile(user_id) try: - res = yield self.transport_client.create_group( + res = await self.transport_client.create_group( get_domain_from_id(group_id), group_id, user_id, content ) except HttpResponseException as e: @@ -313,7 +312,7 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): raise SynapseError(502, "Failed to contact group server") remote_attestation = res["attestation"] - yield self.attestations.verify_attestation( + await self.attestations.verify_attestation( remote_attestation, group_id=group_id, user_id=user_id, @@ -321,7 +320,7 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): ) is_publicised = content.get("publicise", False) - token = yield self.store.register_user_group_membership( + token = await self.store.register_user_group_membership( group_id, user_id, membership="join", @@ -482,12 +481,13 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): return {"state": "invite", "user_profile": user_profile} - @defer.inlineCallbacks - def remove_user_from_group(self, group_id, user_id, requester_user_id, content): + async def remove_user_from_group( + self, group_id, user_id, requester_user_id, content + ): """Remove a user from a group """ if user_id == requester_user_id: - token = yield self.store.register_user_group_membership( + token = await self.store.register_user_group_membership( group_id, user_id, membership="leave" ) self.notifier.on_new_event("groups_key", token, users=[user_id]) @@ -496,13 +496,13 @@ class GroupsLocalHandler(GroupsLocalWorkerHandler): # retry if the group server is currently down. if self.is_mine_id(group_id): - res = yield self.groups_server_handler.remove_user_from_group( + res = await self.groups_server_handler.remove_user_from_group( group_id, user_id, requester_user_id, content ) else: content["requester_user_id"] = requester_user_id try: - res = yield self.transport_client.remove_user_from_group( + res = await self.transport_client.remove_user_from_group( get_domain_from_id(group_id), group_id, requester_user_id, diff --git a/synapse/handlers/identity.py b/synapse/handlers/identity.py index 23f07832e..0f0e632b6 100644 --- a/synapse/handlers/identity.py +++ b/synapse/handlers/identity.py @@ -18,7 +18,7 @@ """Utilities for interacting with Identity Servers""" import logging -import urllib +import urllib.parse from canonicaljson import json from signedjson.key import decode_verify_key_bytes diff --git a/synapse/handlers/initial_sync.py b/synapse/handlers/initial_sync.py index b116500c7..f88bad5f2 100644 --- a/synapse/handlers/initial_sync.py +++ b/synapse/handlers/initial_sync.py @@ -381,10 +381,16 @@ class InitialSyncHandler(BaseHandler): return [] states = await presence_handler.get_states( - [m.user_id for m in room_members], as_event=True + [m.user_id for m in room_members] ) - return states + return [ + { + "type": EventTypes.Presence, + "content": format_user_presence_state(s, time_now), + } + for s in states + ] async def get_receipts(): receipts = await self.store.get_linearized_receipts_for_room( diff --git a/synapse/handlers/message.py b/synapse/handlers/message.py index b743fc2dc..a622a600b 100644 --- a/synapse/handlers/message.py +++ b/synapse/handlers/message.py @@ -419,6 +419,8 @@ class EventCreationHandler(object): self._ephemeral_events_enabled = hs.config.enable_ephemeral_messages + self._dummy_events_threshold = hs.config.dummy_events_threshold + @defer.inlineCallbacks def create_event( self, @@ -626,8 +628,7 @@ class EventCreationHandler(object): msg = self._block_events_without_consent_error % {"consent_uri": consent_uri} raise ConsentNotGivenError(msg=msg, consent_uri=consent_uri) - @defer.inlineCallbacks - def send_nonmember_event(self, requester, event, context, ratelimit=True): + async def send_nonmember_event(self, requester, event, context, ratelimit=True): """ Persists and notifies local clients and federation of an event. @@ -647,7 +648,7 @@ class EventCreationHandler(object): assert self.hs.is_mine(user), "User must be our own: %s" % (user,) if event.is_state(): - prev_state = yield self.deduplicate_state_event(event, context) + prev_state = await self.deduplicate_state_event(event, context) if prev_state is not None: logger.info( "Not bothering to persist state event %s duplicated by %s", @@ -656,7 +657,7 @@ class EventCreationHandler(object): ) return prev_state - yield self.handle_new_client_event( + await self.handle_new_client_event( requester=requester, event=event, context=context, ratelimit=ratelimit ) @@ -683,8 +684,7 @@ class EventCreationHandler(object): return prev_event return - @defer.inlineCallbacks - def create_and_send_nonmember_event( + async def create_and_send_nonmember_event( self, requester, event_dict, ratelimit=True, txn_id=None ): """ @@ -698,8 +698,8 @@ class EventCreationHandler(object): # a situation where event persistence can't keep up, causing # extremities to pile up, which in turn leads to state resolution # taking longer. - with (yield self.limiter.queue(event_dict["room_id"])): - event, context = yield self.create_event( + with (await self.limiter.queue(event_dict["room_id"])): + event, context = await self.create_event( requester, event_dict, token_id=requester.access_token_id, txn_id=txn_id ) @@ -709,7 +709,7 @@ class EventCreationHandler(object): spam_error = "Spam is not permitted here" raise SynapseError(403, spam_error, Codes.FORBIDDEN) - yield self.send_nonmember_event( + await self.send_nonmember_event( requester, event, context, ratelimit=ratelimit ) return event @@ -770,8 +770,7 @@ class EventCreationHandler(object): return (event, context) @measure_func("handle_new_client_event") - @defer.inlineCallbacks - def handle_new_client_event( + async def handle_new_client_event( self, requester, event, context, ratelimit=True, extra_users=[] ): """Processes a new event. This includes checking auth, persisting it, @@ -794,9 +793,9 @@ class EventCreationHandler(object): ): room_version = event.content.get("room_version", RoomVersions.V1.identifier) else: - room_version = yield self.store.get_room_version_id(event.room_id) + room_version = await self.store.get_room_version_id(event.room_id) - event_allowed = yield self.third_party_event_rules.check_event_allowed( + event_allowed = await self.third_party_event_rules.check_event_allowed( event, context ) if not event_allowed: @@ -805,7 +804,7 @@ class EventCreationHandler(object): ) try: - yield self.auth.check_from_context(room_version, event, context) + await self.auth.check_from_context(room_version, event, context) except AuthError as err: logger.warning("Denying new event %r because %s", event, err) raise err @@ -818,7 +817,7 @@ class EventCreationHandler(object): logger.exception("Failed to encode content: %r", event.content) raise - yield self.action_generator.handle_push_actions_for_event(event, context) + await self.action_generator.handle_push_actions_for_event(event, context) # reraise does not allow inlineCallbacks to preserve the stacktrace, so we # hack around with a try/finally instead. @@ -826,7 +825,7 @@ class EventCreationHandler(object): try: # If we're a worker we need to hit out to the master. if self.config.worker_app: - yield self.send_event_to_master( + await self.send_event_to_master( event_id=event.event_id, store=self.store, requester=requester, @@ -838,7 +837,7 @@ class EventCreationHandler(object): success = True return - yield self.persist_and_notify_client_event( + await self.persist_and_notify_client_event( requester, event, context, ratelimit=ratelimit, extra_users=extra_users ) @@ -852,7 +851,38 @@ class EventCreationHandler(object): ) @defer.inlineCallbacks - def persist_and_notify_client_event( + def _validate_canonical_alias( + self, directory_handler, room_alias_str, expected_room_id + ): + """ + Ensure that the given room alias points to the expected room ID. + + Args: + directory_handler: The directory handler object. + room_alias_str: The room alias to check. + expected_room_id: The room ID that the alias should point to. + """ + room_alias = RoomAlias.from_string(room_alias_str) + try: + mapping = yield directory_handler.get_association(room_alias) + except SynapseError as e: + # Turn M_NOT_FOUND errors into M_BAD_ALIAS errors. + if e.errcode == Codes.NOT_FOUND: + raise SynapseError( + 400, + "Room alias %s does not point to the room" % (room_alias_str,), + Codes.BAD_ALIAS, + ) + raise + + if mapping["room_id"] != expected_room_id: + raise SynapseError( + 400, + "Room alias %s does not point to the room" % (room_alias_str,), + Codes.BAD_ALIAS, + ) + + async def persist_and_notify_client_event( self, requester, event, context, ratelimit=True, extra_users=[] ): """Called when we have fully built the event, have already @@ -869,7 +899,7 @@ class EventCreationHandler(object): # user is actually admin or not). is_admin_redaction = False if event.type == EventTypes.Redaction: - original_event = yield self.store.get_event( + original_event = await self.store.get_event( event.redacts, redact_behaviour=EventRedactBehaviour.AS_IS, get_prev_content=False, @@ -881,11 +911,11 @@ class EventCreationHandler(object): original_event and event.sender != original_event.sender ) - yield self.base_handler.ratelimit( + await self.base_handler.ratelimit( requester, is_admin_redaction=is_admin_redaction ) - yield self.base_handler.maybe_kick_guest_users(event, context) + await self.base_handler.maybe_kick_guest_users(event, context) if event.type == EventTypes.CanonicalAlias: # Validate a newly added alias or newly added alt_aliases. @@ -895,7 +925,7 @@ class EventCreationHandler(object): original_event_id = event.unsigned.get("replaces_state") if original_event_id: - original_event = yield self.store.get_event(original_event_id) + original_event = await self.store.get_event(original_event_id) if original_event: original_alias = original_event.content.get("alias", None) @@ -905,15 +935,9 @@ class EventCreationHandler(object): room_alias_str = event.content.get("alias", None) directory_handler = self.hs.get_handlers().directory_handler if room_alias_str and room_alias_str != original_alias: - room_alias = RoomAlias.from_string(room_alias_str) - mapping = yield directory_handler.get_association(room_alias) - - if mapping["room_id"] != event.room_id: - raise SynapseError( - 400, - "Room alias %s does not point to the room" % (room_alias_str,), - Codes.BAD_ALIAS, - ) + await self._validate_canonical_alias( + directory_handler, room_alias_str, event.room_id + ) # Check that alt_aliases is the proper form. alt_aliases = event.content.get("alt_aliases", []) @@ -931,16 +955,9 @@ class EventCreationHandler(object): new_alt_aliases = set(alt_aliases) - set(original_alt_aliases) if new_alt_aliases: for alias_str in new_alt_aliases: - room_alias = RoomAlias.from_string(alias_str) - mapping = yield directory_handler.get_association(room_alias) - - if mapping["room_id"] != event.room_id: - raise SynapseError( - 400, - "Room alias %s does not point to the room" - % (room_alias_str,), - Codes.BAD_ALIAS, - ) + await self._validate_canonical_alias( + directory_handler, alias_str, event.room_id + ) federation_handler = self.hs.get_handlers().federation_handler @@ -950,7 +967,7 @@ class EventCreationHandler(object): def is_inviter_member_event(e): return e.type == EventTypes.Member and e.sender == event.sender - current_state_ids = yield context.get_current_state_ids() + current_state_ids = await context.get_current_state_ids() state_to_include_ids = [ e_id @@ -959,7 +976,7 @@ class EventCreationHandler(object): or k == (EventTypes.Member, event.sender) ] - state_to_include = yield self.store.get_events(state_to_include_ids) + state_to_include = await self.store.get_events(state_to_include_ids) event.unsigned["invite_room_state"] = [ { @@ -977,8 +994,8 @@ class EventCreationHandler(object): # way? If we have been invited by a remote server, we need # to get them to sign the event. - returned_invite = yield defer.ensureDeferred( - federation_handler.send_invite(invitee.domain, event) + returned_invite = await federation_handler.send_invite( + invitee.domain, event ) event.unsigned.pop("room_state", None) @@ -986,7 +1003,7 @@ class EventCreationHandler(object): event.signatures.update(returned_invite.signatures) if event.type == EventTypes.Redaction: - original_event = yield self.store.get_event( + original_event = await self.store.get_event( event.redacts, redact_behaviour=EventRedactBehaviour.AS_IS, get_prev_content=False, @@ -1002,14 +1019,14 @@ class EventCreationHandler(object): if original_event.room_id != event.room_id: raise SynapseError(400, "Cannot redact event from a different room") - prev_state_ids = yield context.get_prev_state_ids() - auth_events_ids = yield self.auth.compute_auth_events( + prev_state_ids = await context.get_prev_state_ids() + auth_events_ids = await self.auth.compute_auth_events( event, prev_state_ids, for_verification=True ) - auth_events = yield self.store.get_events(auth_events_ids) + auth_events = await self.store.get_events(auth_events_ids) auth_events = {(e.type, e.state_key): e for e in auth_events.values()} - room_version = yield self.store.get_room_version_id(event.room_id) + room_version = await self.store.get_room_version_id(event.room_id) room_version_obj = KNOWN_ROOM_VERSIONS[room_version] if event_auth.check_redaction( @@ -1028,11 +1045,11 @@ class EventCreationHandler(object): event.internal_metadata.recheck_redaction = False if event.type == EventTypes.Create: - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = await context.get_prev_state_ids() if prev_state_ids: raise AuthError(403, "Changing the room create event is forbidden") - event_stream_id, max_stream_id = yield self.storage.persistence.persist_event( + event_stream_id, max_stream_id = await self.storage.persistence.persist_event( event, context=context ) @@ -1040,7 +1057,7 @@ class EventCreationHandler(object): # If there's an expiry timestamp on the event, schedule its expiry. self._message_handler.maybe_schedule_expiry(event) - yield self.pusher_pool.on_new_notifications(event_stream_id, max_stream_id) + await self.pusher_pool.on_new_notifications(event_stream_id, max_stream_id) def _notify(): try: @@ -1064,14 +1081,13 @@ class EventCreationHandler(object): except Exception: logger.exception("Error bumping presence active time") - @defer.inlineCallbacks - def _send_dummy_events_to_fill_extremities(self): + async def _send_dummy_events_to_fill_extremities(self): """Background task to send dummy events into rooms that have a large number of extremities """ self._expire_rooms_to_exclude_from_dummy_event_insertion() - room_ids = yield self.store.get_rooms_with_many_extremities( - min_count=10, + room_ids = await self.store.get_rooms_with_many_extremities( + min_count=self._dummy_events_threshold, limit=5, room_id_filter=self._rooms_to_exclude_from_dummy_event_insertion.keys(), ) @@ -1080,9 +1096,9 @@ class EventCreationHandler(object): # For each room we need to find a joined member we can use to send # the dummy event with. - latest_event_ids = yield self.store.get_prev_events_for_room(room_id) + latest_event_ids = await self.store.get_prev_events_for_room(room_id) - members = yield self.state.get_current_users_in_room( + members = await self.state.get_current_users_in_room( room_id, latest_event_ids=latest_event_ids ) dummy_event_sent = False @@ -1091,7 +1107,7 @@ class EventCreationHandler(object): continue requester = create_requester(user_id) try: - event, context = yield self.create_event( + event, context = await self.create_event( requester, { "type": "org.matrix.dummy_event", @@ -1104,7 +1120,7 @@ class EventCreationHandler(object): event.internal_metadata.proactively_send = False - yield self.send_nonmember_event( + await self.send_nonmember_event( requester, event, context, ratelimit=False ) dummy_event_sent = True diff --git a/synapse/handlers/password_policy.py b/synapse/handlers/password_policy.py new file mode 100644 index 000000000..d06b11026 --- /dev/null +++ b/synapse/handlers/password_policy.py @@ -0,0 +1,93 @@ +# -*- coding: utf-8 -*- +# Copyright 2019 New Vector Ltd +# Copyright 2019 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. + +import logging +import re + +from synapse.api.errors import Codes, PasswordRefusedError + +logger = logging.getLogger(__name__) + + +class PasswordPolicyHandler(object): + def __init__(self, hs): + self.policy = hs.config.password_policy + self.enabled = hs.config.password_policy_enabled + + # Regexps for the spec'd policy parameters. + self.regexp_digit = re.compile("[0-9]") + self.regexp_symbol = re.compile("[^a-zA-Z0-9]") + self.regexp_uppercase = re.compile("[A-Z]") + self.regexp_lowercase = re.compile("[a-z]") + + def validate_password(self, password): + """Checks whether a given password complies with the server's policy. + + Args: + password (str): The password to check against the server's policy. + + Raises: + PasswordRefusedError: The password doesn't comply with the server's policy. + """ + + if not self.enabled: + return + + minimum_accepted_length = self.policy.get("minimum_length", 0) + if len(password) < minimum_accepted_length: + raise PasswordRefusedError( + msg=( + "The password must be at least %d characters long" + % minimum_accepted_length + ), + errcode=Codes.PASSWORD_TOO_SHORT, + ) + + if ( + self.policy.get("require_digit", False) + and self.regexp_digit.search(password) is None + ): + raise PasswordRefusedError( + msg="The password must include at least one digit", + errcode=Codes.PASSWORD_NO_DIGIT, + ) + + if ( + self.policy.get("require_symbol", False) + and self.regexp_symbol.search(password) is None + ): + raise PasswordRefusedError( + msg="The password must include at least one symbol", + errcode=Codes.PASSWORD_NO_SYMBOL, + ) + + if ( + self.policy.get("require_uppercase", False) + and self.regexp_uppercase.search(password) is None + ): + raise PasswordRefusedError( + msg="The password must include at least one uppercase letter", + errcode=Codes.PASSWORD_NO_UPPERCASE, + ) + + if ( + self.policy.get("require_lowercase", False) + and self.regexp_lowercase.search(password) is None + ): + raise PasswordRefusedError( + msg="The password must include at least one lowercase letter", + errcode=Codes.PASSWORD_NO_LOWERCASE, + ) diff --git a/synapse/handlers/presence.py b/synapse/handlers/presence.py index 5526015dd..5cbefae17 100644 --- a/synapse/handlers/presence.py +++ b/synapse/handlers/presence.py @@ -1,5 +1,6 @@ # -*- coding: utf-8 -*- # Copyright 2014-2016 OpenMarket Ltd +# Copyright 2020 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. @@ -21,10 +22,10 @@ The methods that define policy are: - PresenceHandler._handle_timeouts - should_notify """ - +import abc import logging from contextlib import contextmanager -from typing import Dict, List, Set +from typing import Dict, Iterable, List, Set from six import iteritems, itervalues @@ -41,7 +42,7 @@ from synapse.logging.utils import log_function from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process from synapse.storage.presence import UserPresenceState -from synapse.types import UserID, get_domain_from_id +from synapse.types import JsonDict, UserID, get_domain_from_id from synapse.util.async_helpers import Linearizer from synapse.util.caches.descriptors import cached from synapse.util.metrics import Measure @@ -99,13 +100,106 @@ EXTERNAL_PROCESS_EXPIRY = 5 * 60 * 1000 assert LAST_ACTIVE_GRANULARITY < IDLE_TIMER -class PresenceHandler(object): +class BasePresenceHandler(abc.ABC): + """Parts of the PresenceHandler that are shared between workers and master""" + def __init__(self, hs: "synapse.server.HomeServer"): + self.clock = hs.get_clock() + self.store = hs.get_datastore() + + active_presence = self.store.take_presence_startup_info() + self.user_to_current_state = {state.user_id: state for state in active_presence} + + @abc.abstractmethod + async def user_syncing( + self, user_id: str, affect_presence: bool + ) -> ContextManager[None]: + """Returns a context manager that should surround any stream requests + from the user. + + This allows us to keep track of who is currently streaming and who isn't + without having to have timers outside of this module to avoid flickering + when users disconnect/reconnect. + + Args: + user_id: the user that is starting a sync + affect_presence: If false this function will be a no-op. + Useful for streams that are not associated with an actual + client that is being used by a user. + """ + + @abc.abstractmethod + def get_currently_syncing_users_for_replication(self) -> Iterable[str]: + """Get an iterable of syncing users on this worker, to send to the presence handler + + This is called when a replication connection is established. It should return + a list of user ids, which are then sent as USER_SYNC commands to inform the + process handling presence about those users. + + Returns: + An iterable of user_id strings. + """ + + async def get_state(self, target_user: UserID) -> UserPresenceState: + results = await self.get_states([target_user.to_string()]) + return results[0] + + async def get_states( + self, target_user_ids: Iterable[str] + ) -> List[UserPresenceState]: + """Get the presence state for users.""" + + updates_d = await self.current_state_for_users(target_user_ids) + updates = list(updates_d.values()) + + for user_id in set(target_user_ids) - {u.user_id for u in updates}: + updates.append(UserPresenceState.default(user_id)) + + return updates + + async def current_state_for_users( + self, user_ids: Iterable[str] + ) -> Dict[str, UserPresenceState]: + """Get the current presence state for multiple users. + + Returns: + dict: `user_id` -> `UserPresenceState` + """ + states = { + user_id: self.user_to_current_state.get(user_id, None) + for user_id in user_ids + } + + missing = [user_id for user_id, state in iteritems(states) if not state] + if missing: + # There are things not in our in memory cache. Lets pull them out of + # the database. + res = await self.store.get_presence_for_users(missing) + states.update(res) + + missing = [user_id for user_id, state in iteritems(states) if not state] + if missing: + new = { + user_id: UserPresenceState.default(user_id) for user_id in missing + } + states.update(new) + self.user_to_current_state.update(new) + + return states + + @abc.abstractmethod + async def set_state( + self, target_user: UserID, state: JsonDict, ignore_status_msg: bool = False + ) -> None: + """Set the presence state of the user. """ + + +class PresenceHandler(BasePresenceHandler): + def __init__(self, hs: "synapse.server.HomeServer"): + super().__init__(hs) self.hs = hs self.is_mine_id = hs.is_mine_id self.server_name = hs.hostname - self.clock = hs.get_clock() - self.store = hs.get_datastore() self.wheel_timer = WheelTimer() self.notifier = hs.get_notifier() self.federation = hs.get_federation_sender() @@ -115,13 +209,6 @@ class PresenceHandler(object): federation_registry.register_edu_handler("m.presence", self.incoming_presence) - active_presence = self.store.take_presence_startup_info() - - # A dictionary of the current state of users. This is prefilled with - # non-offline presence from the DB. We should fetch from the DB if - # we can't find a users presence in here. - self.user_to_current_state = {state.user_id: state for state in active_presence} - LaterGauge( "synapse_handlers_presence_user_to_current_state_size", "", @@ -130,7 +217,7 @@ class PresenceHandler(object): ) now = self.clock.time_msec() - for state in active_presence: + for state in self.user_to_current_state.values(): self.wheel_timer.insert( now=now, obj=state.user_id, then=state.last_active_ts + IDLE_TIMER ) @@ -361,10 +448,18 @@ class PresenceHandler(object): timers_fired_counter.inc(len(states)) + syncing_user_ids = { + user_id + for user_id, count in self.user_to_num_current_syncs.items() + if count + } + for user_ids in self.external_process_to_current_syncs.values(): + syncing_user_ids.update(user_ids) + changes = handle_timeouts( states, is_mine_fn=self.is_mine_id, - syncing_user_ids=self.get_currently_syncing_users(), + syncing_user_ids=syncing_user_ids, now=now, ) @@ -462,22 +557,9 @@ class PresenceHandler(object): return _user_syncing() - def get_currently_syncing_users(self): - """Get the set of user ids that are currently syncing on this HS. - Returns: - set(str): A set of user_id strings. - """ - if self.hs.config.use_presence: - syncing_user_ids = { - user_id - for user_id, count in self.user_to_num_current_syncs.items() - if count - } - for user_ids in self.external_process_to_current_syncs.values(): - syncing_user_ids.update(user_ids) - return syncing_user_ids - else: - return set() + def get_currently_syncing_users_for_replication(self) -> Iterable[str]: + # since we are the process handling presence, there is nothing to do here. + return [] async def update_external_syncs_row( self, process_id, user_id, is_syncing, sync_time_msec @@ -554,34 +636,6 @@ class PresenceHandler(object): res = await self.current_state_for_users([user_id]) return res[user_id] - async def current_state_for_users(self, user_ids): - """Get the current presence state for multiple users. - - Returns: - dict: `user_id` -> `UserPresenceState` - """ - states = { - user_id: self.user_to_current_state.get(user_id, None) - for user_id in user_ids - } - - missing = [user_id for user_id, state in iteritems(states) if not state] - if missing: - # There are things not in our in memory cache. Lets pull them out of - # the database. - res = await self.store.get_presence_for_users(missing) - states.update(res) - - missing = [user_id for user_id, state in iteritems(states) if not state] - if missing: - new = { - user_id: UserPresenceState.default(user_id) for user_id in missing - } - states.update(new) - self.user_to_current_state.update(new) - - return states - async def _persist_and_notify(self, states): """Persist states in the database, poke the notifier and send to interested remote servers @@ -669,40 +723,6 @@ class PresenceHandler(object): federation_presence_counter.inc(len(updates)) await self._update_states(updates) - async def get_state(self, target_user, as_event=False): - results = await self.get_states([target_user.to_string()], as_event=as_event) - - return results[0] - - async def get_states(self, target_user_ids, as_event=False): - """Get the presence state for users. - - Args: - target_user_ids (list) - as_event (bool): Whether to format it as a client event or not. - - Returns: - list - """ - - updates = await self.current_state_for_users(target_user_ids) - updates = list(updates.values()) - - for user_id in set(target_user_ids) - {u.user_id for u in updates}: - updates.append(UserPresenceState.default(user_id)) - - now = self.clock.time_msec() - if as_event: - return [ - { - "type": "m.presence", - "content": format_user_presence_state(state, now), - } - for state in updates - ] - else: - return updates - async def set_state(self, target_user, state, ignore_status_msg=False): """Set the presence state of the user. """ @@ -747,7 +767,7 @@ class PresenceHandler(object): return False - async def get_all_presence_updates(self, last_id, current_id): + async def get_all_presence_updates(self, last_id, current_id, limit): """ Gets a list of presence update rows from between the given stream ids. Each row has: @@ -762,7 +782,7 @@ class PresenceHandler(object): """ # TODO(markjh): replicate the unpersisted changes. # This could use the in-memory stores for recent changes. - rows = await self.store.get_all_presence_updates(last_id, current_id) + rows = await self.store.get_all_presence_updates(last_id, current_id, limit) return rows def notify_new_event(self): @@ -889,7 +909,7 @@ class PresenceHandler(object): user_ids = await self.state.get_current_users_in_room(room_id) user_ids = list(filter(self.is_mine_id, user_ids)) - states = await self.current_state_for_users(user_ids) + states_d = await self.current_state_for_users(user_ids) # Filter out old presence, i.e. offline presence states where # the user hasn't been active for a week. We can change this @@ -899,7 +919,7 @@ class PresenceHandler(object): now = self.clock.time_msec() states = [ state - for state in states.values() + for state in states_d.values() if state.state != PresenceState.OFFLINE or now - state.last_active_ts < 7 * 24 * 60 * 60 * 1000 or state.status_msg is not None diff --git a/synapse/handlers/profile.py b/synapse/handlers/profile.py index 50ce0c585..302efc1b9 100644 --- a/synapse/handlers/profile.py +++ b/synapse/handlers/profile.py @@ -141,8 +141,9 @@ class BaseProfileHandler(BaseHandler): return result["displayname"] - @defer.inlineCallbacks - def set_displayname(self, target_user, requester, new_displayname, by_admin=False): + async def set_displayname( + self, target_user, requester, new_displayname, by_admin=False + ): """Set the displayname of a user Args: @@ -157,6 +158,15 @@ class BaseProfileHandler(BaseHandler): if not by_admin and target_user != requester.user: raise AuthError(400, "Cannot set another user's displayname") + if not by_admin and not self.hs.config.enable_set_displayname: + profile = await self.store.get_profileinfo(target_user.localpart) + if profile.display_name: + raise SynapseError( + 400, + "Changing display name is disabled on this server", + Codes.FORBIDDEN, + ) + if len(new_displayname) > MAX_DISPLAYNAME_LEN: raise SynapseError( 400, "Displayname is too long (max %i)" % (MAX_DISPLAYNAME_LEN,) @@ -171,15 +181,15 @@ class BaseProfileHandler(BaseHandler): if by_admin: requester = create_requester(target_user) - yield self.store.set_profile_displayname(target_user.localpart, new_displayname) + await self.store.set_profile_displayname(target_user.localpart, new_displayname) if self.hs.config.user_directory_search_all_users: - profile = yield self.store.get_profileinfo(target_user.localpart) - yield self.user_directory_handler.handle_local_profile_change( + profile = await self.store.get_profileinfo(target_user.localpart) + await self.user_directory_handler.handle_local_profile_change( target_user.to_string(), profile ) - yield self._update_join_states(requester, target_user) + await self._update_join_states(requester, target_user) @defer.inlineCallbacks def get_avatar_url(self, target_user): @@ -208,8 +218,9 @@ class BaseProfileHandler(BaseHandler): return result["avatar_url"] - @defer.inlineCallbacks - def set_avatar_url(self, target_user, requester, new_avatar_url, by_admin=False): + async def set_avatar_url( + self, target_user, requester, new_avatar_url, by_admin=False + ): """target_user is the user whose avatar_url is to be changed; auth_user is the user attempting to make this change.""" if not self.hs.is_mine(target_user): @@ -218,6 +229,13 @@ class BaseProfileHandler(BaseHandler): if not by_admin and target_user != requester.user: raise AuthError(400, "Cannot set another user's avatar_url") + if not by_admin and not self.hs.config.enable_set_avatar_url: + profile = await self.store.get_profileinfo(target_user.localpart) + if profile.avatar_url: + raise SynapseError( + 400, "Changing avatar is disabled on this server", Codes.FORBIDDEN + ) + if len(new_avatar_url) > MAX_AVATAR_URL_LEN: raise SynapseError( 400, "Avatar URL is too long (max %i)" % (MAX_AVATAR_URL_LEN,) @@ -227,15 +245,15 @@ class BaseProfileHandler(BaseHandler): if by_admin: requester = create_requester(target_user) - yield self.store.set_profile_avatar_url(target_user.localpart, new_avatar_url) + await self.store.set_profile_avatar_url(target_user.localpart, new_avatar_url) if self.hs.config.user_directory_search_all_users: - profile = yield self.store.get_profileinfo(target_user.localpart) - yield self.user_directory_handler.handle_local_profile_change( + profile = await self.store.get_profileinfo(target_user.localpart) + await self.user_directory_handler.handle_local_profile_change( target_user.to_string(), profile ) - yield self._update_join_states(requester, target_user) + await self._update_join_states(requester, target_user) @defer.inlineCallbacks def on_profile_query(self, args): @@ -263,21 +281,20 @@ class BaseProfileHandler(BaseHandler): return response - @defer.inlineCallbacks - def _update_join_states(self, requester, target_user): + async def _update_join_states(self, requester, target_user): if not self.hs.is_mine(target_user): return - yield self.ratelimit(requester) + await self.ratelimit(requester) - room_ids = yield self.store.get_rooms_for_user(target_user.to_string()) + room_ids = await self.store.get_rooms_for_user(target_user.to_string()) for room_id in room_ids: handler = self.hs.get_room_member_handler() try: # Assume the target_user isn't a guest, # because we don't let guests set profile or avatar data. - yield handler.update_membership( + await handler.update_membership( requester, target_user, room_id, diff --git a/synapse/handlers/register.py b/synapse/handlers/register.py index 7ffc194f0..a6178e74a 100644 --- a/synapse/handlers/register.py +++ b/synapse/handlers/register.py @@ -132,7 +132,7 @@ class RegistrationHandler(BaseHandler): def register_user( self, localpart=None, - password=None, + password_hash=None, guest_access_token=None, make_guest=False, admin=False, @@ -145,9 +145,9 @@ class RegistrationHandler(BaseHandler): """Registers a new client on the server. Args: - localpart : The local part of the user ID to register. If None, + localpart: The local part of the user ID to register. If None, one will be generated. - password (unicode) : The password to assign to this user so they can + password_hash (str|None): The hashed password to assign to this user so they can login again. This can be None which means they cannot login again via a password (e.g. the user is an application service user). user_type (str|None): type of user. One of the values from @@ -164,9 +164,6 @@ class RegistrationHandler(BaseHandler): yield self.check_registration_ratelimit(address) yield self.auth.check_auth_blocking(threepid=threepid) - password_hash = None - if password: - password_hash = yield self._auth_handler.hash(password) if localpart is not None: yield self.check_username(localpart, guest_access_token=guest_access_token) @@ -242,7 +239,7 @@ class RegistrationHandler(BaseHandler): fail_count += 1 if not self.hs.config.user_consent_at_registration: - yield self._auto_join_rooms(user_id) + yield defer.ensureDeferred(self._auto_join_rooms(user_id)) else: logger.info( "Skipping auto-join for %s because consent is required at registration", @@ -264,8 +261,7 @@ class RegistrationHandler(BaseHandler): return user_id - @defer.inlineCallbacks - def _auto_join_rooms(self, user_id): + async def _auto_join_rooms(self, user_id): """Automatically joins users to auto join rooms - creating the room in the first place if the user is the first to be created. @@ -279,9 +275,9 @@ class RegistrationHandler(BaseHandler): # that an auto-generated support or bot user is not a real user and will never be # the user to create the room should_auto_create_rooms = False - is_real_user = yield self.store.is_real_user(user_id) + is_real_user = await self.store.is_real_user(user_id) if self.hs.config.autocreate_auto_join_rooms and is_real_user: - count = yield self.store.count_real_users() + count = await self.store.count_real_users() should_auto_create_rooms = count == 1 for r in self.hs.config.auto_join_rooms: logger.info("Auto-joining %s to %s", user_id, r) @@ -300,7 +296,7 @@ class RegistrationHandler(BaseHandler): # getting the RoomCreationHandler during init gives a dependency # loop - yield self.hs.get_room_creation_handler().create_room( + await self.hs.get_room_creation_handler().create_room( fake_requester, config={ "preset": "public_chat", @@ -309,7 +305,7 @@ class RegistrationHandler(BaseHandler): ratelimit=False, ) else: - yield self._join_user_to_room(fake_requester, r) + await self._join_user_to_room(fake_requester, r) except ConsentNotGivenError as e: # Technically not necessary to pull out this error though # moving away from bare excepts is a good thing to do. @@ -317,15 +313,14 @@ class RegistrationHandler(BaseHandler): except Exception as e: logger.error("Failed to join new user to %r: %r", r, e) - @defer.inlineCallbacks - def post_consent_actions(self, user_id): + async def post_consent_actions(self, user_id): """A series of registration actions that can only be carried out once consent has been granted Args: user_id (str): The user to join """ - yield self._auto_join_rooms(user_id) + await self._auto_join_rooms(user_id) @defer.inlineCallbacks def appservice_register(self, user_localpart, as_token): @@ -392,14 +387,13 @@ class RegistrationHandler(BaseHandler): self._next_generated_user_id += 1 return str(id) - @defer.inlineCallbacks - def _join_user_to_room(self, requester, room_identifier): + async def _join_user_to_room(self, requester, room_identifier): room_member_handler = self.hs.get_room_member_handler() if RoomID.is_valid(room_identifier): room_id = room_identifier elif RoomAlias.is_valid(room_identifier): room_alias = RoomAlias.from_string(room_identifier) - room_id, remote_room_hosts = yield room_member_handler.lookup_room_alias( + room_id, remote_room_hosts = await room_member_handler.lookup_room_alias( room_alias ) room_id = room_id.to_string() @@ -408,7 +402,7 @@ class RegistrationHandler(BaseHandler): 400, "%s was not legal room ID or room alias" % (room_identifier,) ) - yield room_member_handler.update_membership( + await room_member_handler.update_membership( requester=requester, target=requester.user, room_id=room_id, @@ -540,14 +534,15 @@ class RegistrationHandler(BaseHandler): user_id, ["guest = true"] ) else: - access_token = yield self._auth_handler.get_access_token_for_user_id( - user_id, device_id=device_id, valid_until_ms=valid_until_ms + access_token = yield defer.ensureDeferred( + self._auth_handler.get_access_token_for_user_id( + user_id, device_id=device_id, valid_until_ms=valid_until_ms + ) ) return (device_id, access_token) - @defer.inlineCallbacks - def post_registration_actions(self, user_id, auth_result, access_token): + async def post_registration_actions(self, user_id, auth_result, access_token): """A user has completed registration Args: @@ -558,7 +553,7 @@ class RegistrationHandler(BaseHandler): device, or None if `inhibit_login` enabled. """ if self.hs.config.worker_app: - yield self._post_registration_client( + await self._post_registration_client( user_id=user_id, auth_result=auth_result, access_token=access_token ) return @@ -570,19 +565,18 @@ class RegistrationHandler(BaseHandler): if is_threepid_reserved( self.hs.config.mau_limits_reserved_threepids, threepid ): - yield self.store.upsert_monthly_active_user(user_id) + await self.store.upsert_monthly_active_user(user_id) - yield self._register_email_threepid(user_id, threepid, access_token) + await self._register_email_threepid(user_id, threepid, access_token) if auth_result and LoginType.MSISDN in auth_result: threepid = auth_result[LoginType.MSISDN] - yield self._register_msisdn_threepid(user_id, threepid) + await self._register_msisdn_threepid(user_id, threepid) if auth_result and LoginType.TERMS in auth_result: - yield self._on_user_consented(user_id, self.hs.config.user_consent_version) + await self._on_user_consented(user_id, self.hs.config.user_consent_version) - @defer.inlineCallbacks - def _on_user_consented(self, user_id, consent_version): + async def _on_user_consented(self, user_id, consent_version): """A user consented to the terms on registration Args: @@ -591,8 +585,8 @@ class RegistrationHandler(BaseHandler): consented to. """ logger.info("%s has consented to the privacy policy", user_id) - yield self.store.user_set_consent_version(user_id, consent_version) - yield self.post_consent_actions(user_id) + await self.store.user_set_consent_version(user_id, consent_version) + await self.post_consent_actions(user_id) @defer.inlineCallbacks def _register_email_threepid(self, user_id, threepid, token): @@ -617,8 +611,13 @@ class RegistrationHandler(BaseHandler): logger.info("Can't add incomplete 3pid") return - yield self._auth_handler.add_threepid( - user_id, threepid["medium"], threepid["address"], threepid["validated_at"] + yield defer.ensureDeferred( + self._auth_handler.add_threepid( + user_id, + threepid["medium"], + threepid["address"], + threepid["validated_at"], + ) ) # And we add an email pusher for them by default, but only @@ -670,6 +669,11 @@ class RegistrationHandler(BaseHandler): return None raise - yield self._auth_handler.add_threepid( - user_id, threepid["medium"], threepid["address"], threepid["validated_at"] + yield defer.ensureDeferred( + self._auth_handler.add_threepid( + user_id, + threepid["medium"], + threepid["address"], + threepid["validated_at"], + ) ) diff --git a/synapse/handlers/room.py b/synapse/handlers/room.py index f580ab2e9..da12df7f5 100644 --- a/synapse/handlers/room.py +++ b/synapse/handlers/room.py @@ -148,17 +148,16 @@ class RoomCreationHandler(BaseHandler): return ret - @defer.inlineCallbacks - def _upgrade_room( + async def _upgrade_room( self, requester: Requester, old_room_id: str, new_version: RoomVersion ): user_id = requester.user.to_string() # start by allocating a new room id - r = yield self.store.get_room(old_room_id) + r = await self.store.get_room(old_room_id) if r is None: raise NotFoundError("Unknown room id %s" % (old_room_id,)) - new_room_id = yield self._generate_room_id( + new_room_id = await self._generate_room_id( creator_id=user_id, is_public=r["is_public"], room_version=new_version, ) @@ -169,7 +168,7 @@ class RoomCreationHandler(BaseHandler): ( tombstone_event, tombstone_context, - ) = yield self.event_creation_handler.create_event( + ) = await self.event_creation_handler.create_event( requester, { "type": EventTypes.Tombstone, @@ -183,12 +182,12 @@ class RoomCreationHandler(BaseHandler): }, token_id=requester.access_token_id, ) - old_room_version = yield self.store.get_room_version_id(old_room_id) - yield self.auth.check_from_context( + old_room_version = await self.store.get_room_version_id(old_room_id) + await self.auth.check_from_context( old_room_version, tombstone_event, tombstone_context ) - yield self.clone_existing_room( + await self.clone_existing_room( requester, old_room_id=old_room_id, new_room_id=new_room_id, @@ -197,32 +196,31 @@ class RoomCreationHandler(BaseHandler): ) # now send the tombstone - yield self.event_creation_handler.send_nonmember_event( + await self.event_creation_handler.send_nonmember_event( requester, tombstone_event, tombstone_context ) - old_room_state = yield tombstone_context.get_current_state_ids() + old_room_state = await tombstone_context.get_current_state_ids() # update any aliases - yield self._move_aliases_to_new_room( + await self._move_aliases_to_new_room( requester, old_room_id, new_room_id, old_room_state ) # Copy over user push rules, tags and migrate room directory state - yield self.room_member_handler.transfer_room_state_on_room_upgrade( + await self.room_member_handler.transfer_room_state_on_room_upgrade( old_room_id, new_room_id ) # finally, shut down the PLs in the old room, and update them in the new # room. - yield self._update_upgraded_room_pls( + await self._update_upgraded_room_pls( requester, old_room_id, new_room_id, old_room_state, ) return new_room_id - @defer.inlineCallbacks - def _update_upgraded_room_pls( + async def _update_upgraded_room_pls( self, requester: Requester, old_room_id: str, @@ -249,7 +247,7 @@ class RoomCreationHandler(BaseHandler): ) return - old_room_pl_state = yield self.store.get_event(old_room_pl_event_id) + old_room_pl_state = await self.store.get_event(old_room_pl_event_id) # we try to stop regular users from speaking by setting the PL required # to send regular events and invites to 'Moderator' level. That's normally @@ -278,7 +276,7 @@ class RoomCreationHandler(BaseHandler): if updated: try: - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.PowerLevels, @@ -292,7 +290,7 @@ class RoomCreationHandler(BaseHandler): except AuthError as e: logger.warning("Unable to update PLs in old room: %s", e) - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.PowerLevels, @@ -304,8 +302,7 @@ class RoomCreationHandler(BaseHandler): ratelimit=False, ) - @defer.inlineCallbacks - def clone_existing_room( + async def clone_existing_room( self, requester: Requester, old_room_id: str, @@ -338,7 +335,7 @@ class RoomCreationHandler(BaseHandler): # Check if old room was non-federatable # Get old room's create event - old_room_create_event = yield self.store.get_create_event_for_room(old_room_id) + old_room_create_event = await self.store.get_create_event_for_room(old_room_id) # Check if the create event specified a non-federatable room if not old_room_create_event.content.get("m.federate", True): @@ -361,11 +358,11 @@ class RoomCreationHandler(BaseHandler): (EventTypes.PowerLevels, ""), ) - old_room_state_ids = yield self.store.get_filtered_current_state_ids( + old_room_state_ids = await self.store.get_filtered_current_state_ids( old_room_id, StateFilter.from_types(types_to_copy) ) # map from event_id to BaseEvent - old_room_state_events = yield self.store.get_events(old_room_state_ids.values()) + old_room_state_events = await self.store.get_events(old_room_state_ids.values()) for k, old_event_id in iteritems(old_room_state_ids): old_event = old_room_state_events.get(old_event_id) @@ -400,7 +397,7 @@ class RoomCreationHandler(BaseHandler): if current_power_level < needed_power_level: power_levels["users"][user_id] = needed_power_level - yield self._send_events_for_new_room( + await self._send_events_for_new_room( requester, new_room_id, # we expect to override all the presets with initial_state, so this is @@ -412,12 +409,12 @@ class RoomCreationHandler(BaseHandler): ) # Transfer membership events - old_room_member_state_ids = yield self.store.get_filtered_current_state_ids( + old_room_member_state_ids = await self.store.get_filtered_current_state_ids( old_room_id, StateFilter.from_types([(EventTypes.Member, None)]) ) # map from event_id to BaseEvent - old_room_member_state_events = yield self.store.get_events( + old_room_member_state_events = await self.store.get_events( old_room_member_state_ids.values() ) for k, old_event in iteritems(old_room_member_state_events): @@ -426,7 +423,7 @@ class RoomCreationHandler(BaseHandler): "membership" in old_event.content and old_event.content["membership"] == "ban" ): - yield self.room_member_handler.update_membership( + await self.room_member_handler.update_membership( requester, UserID.from_string(old_event["state_key"]), new_room_id, @@ -438,8 +435,7 @@ class RoomCreationHandler(BaseHandler): # XXX invites/joins # XXX 3pid invites - @defer.inlineCallbacks - def _move_aliases_to_new_room( + async def _move_aliases_to_new_room( self, requester: Requester, old_room_id: str, @@ -448,13 +444,13 @@ class RoomCreationHandler(BaseHandler): ): directory_handler = self.hs.get_handlers().directory_handler - aliases = yield self.store.get_aliases_for_room(old_room_id) + aliases = await self.store.get_aliases_for_room(old_room_id) # check to see if we have a canonical alias. canonical_alias_event = None canonical_alias_event_id = old_room_state.get((EventTypes.CanonicalAlias, "")) if canonical_alias_event_id: - canonical_alias_event = yield self.store.get_event(canonical_alias_event_id) + canonical_alias_event = await self.store.get_event(canonical_alias_event_id) # first we try to remove the aliases from the old room (we suppress sending # the room_aliases event until the end). @@ -472,7 +468,7 @@ class RoomCreationHandler(BaseHandler): for alias_str in aliases: alias = RoomAlias.from_string(alias_str) try: - yield directory_handler.delete_association(requester, alias) + await directory_handler.delete_association(requester, alias) removed_aliases.append(alias_str) except SynapseError as e: logger.warning("Unable to remove alias %s from old room: %s", alias, e) @@ -485,7 +481,7 @@ class RoomCreationHandler(BaseHandler): # we can now add any aliases we successfully removed to the new room. for alias in removed_aliases: try: - yield directory_handler.create_association( + await directory_handler.create_association( requester, RoomAlias.from_string(alias), new_room_id, @@ -502,7 +498,7 @@ class RoomCreationHandler(BaseHandler): # alias event for the new room with a copy of the information. try: if canonical_alias_event: - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.CanonicalAlias, @@ -518,8 +514,9 @@ class RoomCreationHandler(BaseHandler): # we returned the new room to the client at this point. logger.error("Unable to send updated alias events in new room: %s", e) - @defer.inlineCallbacks - def create_room(self, requester, config, ratelimit=True, creator_join_profile=None): + async def create_room( + self, requester, config, ratelimit=True, creator_join_profile=None + ): """ Creates a new room. Args: @@ -547,7 +544,7 @@ class RoomCreationHandler(BaseHandler): """ user_id = requester.user.to_string() - yield self.auth.check_auth_blocking(user_id) + await self.auth.check_auth_blocking(user_id) if ( self._server_notices_mxid is not None @@ -556,11 +553,11 @@ class RoomCreationHandler(BaseHandler): # allow the server notices mxid to create rooms is_requester_admin = True else: - is_requester_admin = yield self.auth.is_server_admin(requester.user) + is_requester_admin = await self.auth.is_server_admin(requester.user) # Check whether the third party rules allows/changes the room create # request. - event_allowed = yield self.third_party_event_rules.on_create_room( + event_allowed = await self.third_party_event_rules.on_create_room( requester, config, is_requester_admin=is_requester_admin ) if not event_allowed: @@ -574,7 +571,7 @@ class RoomCreationHandler(BaseHandler): raise SynapseError(403, "You are not permitted to create rooms") if ratelimit: - yield self.ratelimit(requester) + await self.ratelimit(requester) room_version_id = config.get( "room_version", self.config.default_room_version.identifier @@ -597,7 +594,7 @@ class RoomCreationHandler(BaseHandler): raise SynapseError(400, "Invalid characters in room alias") room_alias = RoomAlias(config["room_alias_name"], self.hs.hostname) - mapping = yield self.store.get_association_from_room_alias(room_alias) + mapping = await self.store.get_association_from_room_alias(room_alias) if mapping: raise SynapseError(400, "Room alias already taken", Codes.ROOM_IN_USE) @@ -612,7 +609,7 @@ class RoomCreationHandler(BaseHandler): except Exception: raise SynapseError(400, "Invalid user_id: %s" % (i,)) - yield self.event_creation_handler.assert_accepted_privacy_policy(requester) + await self.event_creation_handler.assert_accepted_privacy_policy(requester) power_level_content_override = config.get("power_level_content_override") if ( @@ -631,13 +628,13 @@ class RoomCreationHandler(BaseHandler): visibility = config.get("visibility", None) is_public = visibility == "public" - room_id = yield self._generate_room_id( + room_id = await self._generate_room_id( creator_id=user_id, is_public=is_public, room_version=room_version, ) directory_handler = self.hs.get_handlers().directory_handler if room_alias: - yield directory_handler.create_association( + await directory_handler.create_association( requester=requester, room_id=room_id, room_alias=room_alias, @@ -645,6 +642,13 @@ class RoomCreationHandler(BaseHandler): check_membership=False, ) + if is_public: + if not self.config.is_publishing_room_allowed(user_id, room_id, room_alias): + # Lets just return a generic message, as there may be all sorts of + # reasons why we said no. TODO: Allow configurable error messages + # per alias creation rule? + raise SynapseError(403, "Not allowed to publish room") + preset_config = config.get( "preset", RoomCreationPreset.PRIVATE_CHAT @@ -663,7 +667,7 @@ class RoomCreationHandler(BaseHandler): # override any attempt to set room versions via the creation_content creation_content["room_version"] = room_version.identifier - yield self._send_events_for_new_room( + await self._send_events_for_new_room( requester, room_id, preset_config=preset_config, @@ -677,7 +681,7 @@ class RoomCreationHandler(BaseHandler): if "name" in config: name = config["name"] - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.Name, @@ -691,7 +695,7 @@ class RoomCreationHandler(BaseHandler): if "topic" in config: topic = config["topic"] - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.Topic, @@ -709,7 +713,7 @@ class RoomCreationHandler(BaseHandler): if is_direct: content["is_direct"] = is_direct - yield self.room_member_handler.update_membership( + await self.room_member_handler.update_membership( requester, UserID.from_string(invitee), room_id, @@ -723,7 +727,7 @@ class RoomCreationHandler(BaseHandler): id_access_token = invite_3pid.get("id_access_token") # optional address = invite_3pid["address"] medium = invite_3pid["medium"] - yield self.hs.get_room_member_handler().do_3pid_invite( + await self.hs.get_room_member_handler().do_3pid_invite( room_id, requester.user, medium, @@ -741,8 +745,7 @@ class RoomCreationHandler(BaseHandler): return result - @defer.inlineCallbacks - def _send_events_for_new_room( + async def _send_events_for_new_room( self, creator, # A Requester object. room_id, @@ -762,11 +765,10 @@ class RoomCreationHandler(BaseHandler): return e - @defer.inlineCallbacks - def send(etype, content, **kwargs): + async def send(etype, content, **kwargs): event = create(etype, content, **kwargs) logger.debug("Sending %s in new room", etype) - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( creator, event, ratelimit=False ) @@ -777,10 +779,10 @@ class RoomCreationHandler(BaseHandler): event_keys = {"room_id": room_id, "sender": creator_id, "state_key": ""} creation_content.update({"creator": creator_id}) - yield send(etype=EventTypes.Create, content=creation_content) + await send(etype=EventTypes.Create, content=creation_content) logger.debug("Sending %s in new room", EventTypes.Member) - yield self.room_member_handler.update_membership( + await self.room_member_handler.update_membership( creator, creator.user, room_id, @@ -793,7 +795,7 @@ class RoomCreationHandler(BaseHandler): # of the first events that get sent into a room. pl_content = initial_state.pop((EventTypes.PowerLevels, ""), None) if pl_content is not None: - yield send(etype=EventTypes.PowerLevels, content=pl_content) + await send(etype=EventTypes.PowerLevels, content=pl_content) else: power_level_content = { "users": {creator_id: 100}, @@ -806,6 +808,7 @@ class RoomCreationHandler(BaseHandler): EventTypes.RoomAvatar: 50, EventTypes.Tombstone: 100, EventTypes.ServerACL: 100, + EventTypes.RoomEncryption: 100, }, "events_default": 0, "state_default": 50, @@ -825,33 +828,33 @@ class RoomCreationHandler(BaseHandler): if power_level_content_override: power_level_content.update(power_level_content_override) - yield send(etype=EventTypes.PowerLevels, content=power_level_content) + await send(etype=EventTypes.PowerLevels, content=power_level_content) if room_alias and (EventTypes.CanonicalAlias, "") not in initial_state: - yield send( + await send( etype=EventTypes.CanonicalAlias, content={"alias": room_alias.to_string()}, ) if (EventTypes.JoinRules, "") not in initial_state: - yield send( + await send( etype=EventTypes.JoinRules, content={"join_rule": config["join_rules"]} ) if (EventTypes.RoomHistoryVisibility, "") not in initial_state: - yield send( + await send( etype=EventTypes.RoomHistoryVisibility, content={"history_visibility": config["history_visibility"]}, ) if config["guest_can_join"]: if (EventTypes.GuestAccess, "") not in initial_state: - yield send( + await send( etype=EventTypes.GuestAccess, content={"guest_access": "can_join"} ) for (etype, state_key), content in initial_state.items(): - yield send(etype=etype, state_key=state_key, content=content) + await send(etype=etype, state_key=state_key, content=content) @defer.inlineCallbacks def _generate_room_id( diff --git a/synapse/handlers/room_list.py b/synapse/handlers/room_list.py index 0b7d3da68..e75dabcd7 100644 --- a/synapse/handlers/room_list.py +++ b/synapse/handlers/room_list.py @@ -15,6 +15,7 @@ import logging from collections import namedtuple +from typing import Any, Dict, Optional from six import iteritems @@ -89,7 +90,11 @@ class RoomListHandler(BaseHandler): logger.info("Bypassing cache as search request.") return self._get_public_room_list( - limit, since_token, search_filter, network_tuple=network_tuple + limit, + since_token, + search_filter, + network_tuple=network_tuple, + from_federation=from_federation, ) key = (limit, since_token, network_tuple) @@ -105,22 +110,22 @@ class RoomListHandler(BaseHandler): @defer.inlineCallbacks def _get_public_room_list( self, - limit=None, - since_token=None, - search_filter=None, - network_tuple=EMPTY_THIRD_PARTY_ID, - from_federation=False, - ): + limit: Optional[int] = None, + since_token: Optional[str] = None, + search_filter: Optional[Dict] = None, + network_tuple: ThirdPartyInstanceID = EMPTY_THIRD_PARTY_ID, + from_federation: bool = False, + ) -> Dict[str, Any]: """Generate a public room list. Args: - limit (int|None): Maximum amount of rooms to return. - since_token (str|None) - search_filter (dict|None): Dictionary to filter rooms by. - network_tuple (ThirdPartyInstanceID): Which public list to use. + limit: Maximum amount of rooms to return. + since_token: + search_filter: Dictionary to filter rooms by. + network_tuple: Which public list to use. This can be (None, None) to indicate the main list, or a particular appservice and network id to use an appservice specific one. Setting to None returns all public rooms across all lists. - from_federation (bool): Whether this request originated from a + from_federation: Whether this request originated from a federating server or a client. Used for room filtering. """ diff --git a/synapse/handlers/room_member.py b/synapse/handlers/room_member.py index 426042636..53b49bc15 100644 --- a/synapse/handlers/room_member.py +++ b/synapse/handlers/room_member.py @@ -142,8 +142,7 @@ class RoomMemberHandler(object): """ raise NotImplementedError() - @defer.inlineCallbacks - def _local_membership_update( + async def _local_membership_update( self, requester, target, @@ -164,7 +163,7 @@ class RoomMemberHandler(object): if requester.is_guest: content["kind"] = "guest" - event, context = yield self.event_creation_handler.create_event( + event, context = await self.event_creation_handler.create_event( requester, { "type": EventTypes.Member, @@ -182,18 +181,18 @@ class RoomMemberHandler(object): ) # Check if this event matches the previous membership event for the user. - duplicate = yield self.event_creation_handler.deduplicate_state_event( + duplicate = await self.event_creation_handler.deduplicate_state_event( event, context ) if duplicate is not None: # Discard the new event since this membership change is a no-op. return duplicate - yield self.event_creation_handler.handle_new_client_event( + await self.event_creation_handler.handle_new_client_event( requester, event, context, extra_users=[target], ratelimit=ratelimit ) - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = await context.get_prev_state_ids() prev_member_event_id = prev_state_ids.get((EventTypes.Member, user_id), None) @@ -203,15 +202,15 @@ class RoomMemberHandler(object): # info. newly_joined = True if prev_member_event_id: - prev_member_event = yield self.store.get_event(prev_member_event_id) + prev_member_event = await self.store.get_event(prev_member_event_id) newly_joined = prev_member_event.membership != Membership.JOIN if newly_joined: - yield self._user_joined_room(target, room_id) + await self._user_joined_room(target, room_id) elif event.membership == Membership.LEAVE: if prev_member_event_id: - prev_member_event = yield self.store.get_event(prev_member_event_id) + prev_member_event = await self.store.get_event(prev_member_event_id) if prev_member_event.membership == Membership.JOIN: - yield self._user_left_room(target, room_id) + await self._user_left_room(target, room_id) return event @@ -253,8 +252,7 @@ class RoomMemberHandler(object): for tag, tag_content in room_tags.items(): yield self.store.add_tag_to_room(user_id, new_room_id, tag, tag_content) - @defer.inlineCallbacks - def update_membership( + async def update_membership( self, requester, target, @@ -269,8 +267,8 @@ class RoomMemberHandler(object): ): key = (room_id,) - with (yield self.member_linearizer.queue(key)): - result = yield self._update_membership( + with (await self.member_linearizer.queue(key)): + result = await self._update_membership( requester, target, room_id, @@ -285,8 +283,7 @@ class RoomMemberHandler(object): return result - @defer.inlineCallbacks - def _update_membership( + async def _update_membership( self, requester, target, @@ -321,7 +318,7 @@ class RoomMemberHandler(object): # if this is a join with a 3pid signature, we may need to turn a 3pid # invite into a normal invite before we can handle the join. if third_party_signed is not None: - yield self.federation_handler.exchange_third_party_invite( + await self.federation_handler.exchange_third_party_invite( third_party_signed["sender"], target.to_string(), room_id, @@ -332,7 +329,7 @@ class RoomMemberHandler(object): remote_room_hosts = [] if effective_membership_state not in ("leave", "ban"): - is_blocked = yield self.store.is_room_blocked(room_id) + is_blocked = await self.store.is_room_blocked(room_id) if is_blocked: raise SynapseError(403, "This room has been blocked on this server") @@ -351,7 +348,7 @@ class RoomMemberHandler(object): is_requester_admin = True else: - is_requester_admin = yield self.auth.is_server_admin(requester.user) + is_requester_admin = await self.auth.is_server_admin(requester.user) if not is_requester_admin: if self.config.block_non_admin_invites: @@ -370,9 +367,9 @@ class RoomMemberHandler(object): if block_invite: raise SynapseError(403, "Invites have been disabled on this server") - latest_event_ids = yield self.store.get_prev_events_for_room(room_id) + latest_event_ids = await self.store.get_prev_events_for_room(room_id) - current_state_ids = yield self.state_handler.get_current_state_ids( + current_state_ids = await self.state_handler.get_current_state_ids( room_id, latest_event_ids=latest_event_ids ) @@ -381,7 +378,7 @@ class RoomMemberHandler(object): # transitions and generic otherwise old_state_id = current_state_ids.get((EventTypes.Member, target.to_string())) if old_state_id: - old_state = yield self.store.get_event(old_state_id, allow_none=True) + old_state = await self.store.get_event(old_state_id, allow_none=True) old_membership = old_state.content.get("membership") if old_state else None if action == "unban" and old_membership != "ban": raise SynapseError( @@ -413,7 +410,7 @@ class RoomMemberHandler(object): old_membership == Membership.INVITE and effective_membership_state == Membership.LEAVE ): - is_blocked = yield self._is_server_notice_room(room_id) + is_blocked = await self._is_server_notice_room(room_id) if is_blocked: raise SynapseError( http_client.FORBIDDEN, @@ -424,18 +421,18 @@ class RoomMemberHandler(object): if action == "kick": raise AuthError(403, "The target user is not in the room") - is_host_in_room = yield self._is_host_in_room(current_state_ids) + is_host_in_room = await self._is_host_in_room(current_state_ids) if effective_membership_state == Membership.JOIN: if requester.is_guest: - guest_can_join = yield self._can_guest_join(current_state_ids) + guest_can_join = await self._can_guest_join(current_state_ids) if not guest_can_join: # This should be an auth check, but guests are a local concept, # so don't really fit into the general auth process. raise AuthError(403, "Guest access not allowed") if not is_host_in_room: - inviter = yield self._get_inviter(target.to_string(), room_id) + inviter = await self._get_inviter(target.to_string(), room_id) if inviter and not self.hs.is_mine(inviter): remote_room_hosts.append(inviter.domain) @@ -443,13 +440,13 @@ class RoomMemberHandler(object): profile = self.profile_handler if not content_specified: - content["displayname"] = yield profile.get_displayname(target) - content["avatar_url"] = yield profile.get_avatar_url(target) + content["displayname"] = await profile.get_displayname(target) + content["avatar_url"] = await profile.get_avatar_url(target) if requester.is_guest: content["kind"] = "guest" - remote_join_response = yield self._remote_join( + remote_join_response = await self._remote_join( requester, remote_room_hosts, room_id, target, content ) @@ -458,7 +455,7 @@ class RoomMemberHandler(object): elif effective_membership_state == Membership.LEAVE: if not is_host_in_room: # perhaps we've been invited - inviter = yield self._get_inviter(target.to_string(), room_id) + inviter = await self._get_inviter(target.to_string(), room_id) if not inviter: raise SynapseError(404, "Not a known room") @@ -472,12 +469,12 @@ class RoomMemberHandler(object): else: # send the rejection to the inviter's HS. remote_room_hosts = remote_room_hosts + [inviter.domain] - res = yield self._remote_reject_invite( + res = await self._remote_reject_invite( requester, remote_room_hosts, room_id, target, content, ) return res - res = yield self._local_membership_update( + res = await self._local_membership_update( requester=requester, target=target, room_id=room_id, @@ -519,6 +516,9 @@ class RoomMemberHandler(object): yield self.store.set_room_is_public(old_room_id, False) yield self.store.set_room_is_public(room_id, True) + # Transfer alias mappings in the room directory + yield self.store.update_aliases_for_room(old_room_id, room_id) + # Check if any groups we own contain the predecessor room local_group_ids = yield self.store.get_local_groups_for_room(old_room_id) for group_id in local_group_ids: @@ -569,8 +569,7 @@ class RoomMemberHandler(object): ) continue - @defer.inlineCallbacks - def send_membership_event(self, requester, event, context, ratelimit=True): + async def send_membership_event(self, requester, event, context, ratelimit=True): """ Change the membership status of a user in a room. @@ -596,27 +595,27 @@ class RoomMemberHandler(object): else: requester = types.create_requester(target_user) - prev_event = yield self.event_creation_handler.deduplicate_state_event( + prev_event = await self.event_creation_handler.deduplicate_state_event( event, context ) if prev_event is not None: return - prev_state_ids = yield context.get_prev_state_ids() + prev_state_ids = await context.get_prev_state_ids() if event.membership == Membership.JOIN: if requester.is_guest: - guest_can_join = yield self._can_guest_join(prev_state_ids) + guest_can_join = await self._can_guest_join(prev_state_ids) if not guest_can_join: # This should be an auth check, but guests are a local concept, # so don't really fit into the general auth process. raise AuthError(403, "Guest access not allowed") if event.membership not in (Membership.LEAVE, Membership.BAN): - is_blocked = yield self.store.is_room_blocked(room_id) + is_blocked = await self.store.is_room_blocked(room_id) if is_blocked: raise SynapseError(403, "This room has been blocked on this server") - yield self.event_creation_handler.handle_new_client_event( + await self.event_creation_handler.handle_new_client_event( requester, event, context, extra_users=[target_user], ratelimit=ratelimit ) @@ -630,15 +629,15 @@ class RoomMemberHandler(object): # info. newly_joined = True if prev_member_event_id: - prev_member_event = yield self.store.get_event(prev_member_event_id) + prev_member_event = await self.store.get_event(prev_member_event_id) newly_joined = prev_member_event.membership != Membership.JOIN if newly_joined: - yield self._user_joined_room(target_user, room_id) + await self._user_joined_room(target_user, room_id) elif event.membership == Membership.LEAVE: if prev_member_event_id: - prev_member_event = yield self.store.get_event(prev_member_event_id) + prev_member_event = await self.store.get_event(prev_member_event_id) if prev_member_event.membership == Membership.JOIN: - yield self._user_left_room(target_user, room_id) + await self._user_left_room(target_user, room_id) @defer.inlineCallbacks def _can_guest_join(self, current_state_ids): @@ -696,8 +695,7 @@ class RoomMemberHandler(object): if invite: return UserID.from_string(invite.sender) - @defer.inlineCallbacks - def do_3pid_invite( + async def do_3pid_invite( self, room_id, inviter, @@ -709,7 +707,7 @@ class RoomMemberHandler(object): id_access_token=None, ): if self.config.block_non_admin_invites: - is_requester_admin = yield self.auth.is_server_admin(requester.user) + is_requester_admin = await self.auth.is_server_admin(requester.user) if not is_requester_admin: raise SynapseError( 403, "Invites have been disabled on this server", Codes.FORBIDDEN @@ -717,9 +715,9 @@ class RoomMemberHandler(object): # We need to rate limit *before* we send out any 3PID invites, so we # can't just rely on the standard ratelimiting of events. - yield self.base_handler.ratelimit(requester) + await self.base_handler.ratelimit(requester) - can_invite = yield self.third_party_event_rules.check_threepid_can_be_invited( + can_invite = await self.third_party_event_rules.check_threepid_can_be_invited( medium, address, room_id ) if not can_invite: @@ -734,16 +732,16 @@ class RoomMemberHandler(object): 403, "Looking up third-party identifiers is denied from this server" ) - invitee = yield self.identity_handler.lookup_3pid( + invitee = await self.identity_handler.lookup_3pid( id_server, medium, address, id_access_token ) if invitee: - yield self.update_membership( + await self.update_membership( requester, UserID.from_string(invitee), room_id, "invite", txn_id=txn_id ) else: - yield self._make_and_store_3pid_invite( + await self._make_and_store_3pid_invite( requester, id_server, medium, @@ -754,8 +752,7 @@ class RoomMemberHandler(object): id_access_token=id_access_token, ) - @defer.inlineCallbacks - def _make_and_store_3pid_invite( + async def _make_and_store_3pid_invite( self, requester, id_server, @@ -766,7 +763,7 @@ class RoomMemberHandler(object): txn_id, id_access_token=None, ): - room_state = yield self.state_handler.get_current_state(room_id) + room_state = await self.state_handler.get_current_state(room_id) inviter_display_name = "" inviter_avatar_url = "" @@ -804,7 +801,7 @@ class RoomMemberHandler(object): public_keys, fallback_public_key, display_name, - ) = yield self.identity_handler.ask_id_server_for_third_party_invite( + ) = await self.identity_handler.ask_id_server_for_third_party_invite( requester=requester, id_server=id_server, medium=medium, @@ -820,7 +817,7 @@ class RoomMemberHandler(object): id_access_token=id_access_token, ) - yield self.event_creation_handler.create_and_send_nonmember_event( + await self.event_creation_handler.create_and_send_nonmember_event( requester, { "type": EventTypes.ThirdPartyInvite, @@ -914,8 +911,7 @@ class RoomMemberMasterHandler(RoomMemberHandler): return complexity["v1"] > max_complexity - @defer.inlineCallbacks - def _remote_join(self, requester, remote_room_hosts, room_id, user, content): + async def _remote_join(self, requester, remote_room_hosts, room_id, user, content): """Implements RoomMemberHandler._remote_join """ # filter ourselves out of remote_room_hosts: do_invite_join ignores it @@ -930,7 +926,7 @@ class RoomMemberMasterHandler(RoomMemberHandler): if self.hs.config.limit_remote_rooms.enabled: # Fetch the room complexity - too_complex = yield self._is_remote_room_too_complex( + too_complex = await self._is_remote_room_too_complex( room_id, remote_room_hosts ) if too_complex is True: @@ -944,12 +940,10 @@ class RoomMemberMasterHandler(RoomMemberHandler): # join dance for now, since we're kinda implicitly checking # that we are allowed to join when we decide whether or not we # need to do the invite/join dance. - yield defer.ensureDeferred( - self.federation_handler.do_invite_join( - remote_room_hosts, room_id, user.to_string(), content - ) + await self.federation_handler.do_invite_join( + remote_room_hosts, room_id, user.to_string(), content ) - yield self._user_joined_room(user, room_id) + await self._user_joined_room(user, room_id) # Check the room we just joined wasn't too large, if we didn't fetch the # complexity of it before. @@ -959,7 +953,7 @@ class RoomMemberMasterHandler(RoomMemberHandler): return # Check again, but with the local state events - too_complex = yield self._is_local_room_too_complex(room_id) + too_complex = await self._is_local_room_too_complex(room_id) if too_complex is False: # We're under the limit. @@ -967,7 +961,7 @@ class RoomMemberMasterHandler(RoomMemberHandler): # The room is too large. Leave. requester = types.create_requester(user, None, False, None) - yield self.update_membership( + await self.update_membership( requester=requester, target=user, room_id=room_id, action="leave" ) raise SynapseError( @@ -1005,12 +999,12 @@ class RoomMemberMasterHandler(RoomMemberHandler): def _user_joined_room(self, target, room_id): """Implements RoomMemberHandler._user_joined_room """ - return user_joined_room(self.distributor, target, room_id) + return defer.succeed(user_joined_room(self.distributor, target, room_id)) def _user_left_room(self, target, room_id): """Implements RoomMemberHandler._user_left_room """ - return user_left_room(self.distributor, target, room_id) + return defer.succeed(user_left_room(self.distributor, target, room_id)) @defer.inlineCallbacks def forget(self, user, room_id): diff --git a/synapse/handlers/saml_handler.py b/synapse/handlers/saml_handler.py index 72c109981..96f2dd36a 100644 --- a/synapse/handlers/saml_handler.py +++ b/synapse/handlers/saml_handler.py @@ -14,7 +14,7 @@ # limitations under the License. import logging import re -from typing import Tuple +from typing import Optional, Tuple import attr import saml2 @@ -26,6 +26,7 @@ from synapse.config import ConfigError from synapse.http.server import finish_request from synapse.http.servlet import parse_string from synapse.module_api import ModuleApi +from synapse.module_api.errors import RedirectException from synapse.types import ( UserID, map_username_to_mxid_localpart, @@ -43,11 +44,15 @@ class Saml2SessionData: # time the session was created, in milliseconds creation_time = attr.ib() + # The user interactive authentication session ID associated with this SAML + # session (or None if this SAML session is for an initial login). + ui_auth_session_id = attr.ib(type=Optional[str], default=None) class SamlHandler: def __init__(self, hs): self._saml_client = Saml2Client(hs.config.saml2_sp_config) + self._auth = hs.get_auth() self._auth_handler = hs.get_auth_handler() self._registration_handler = hs.get_registration_handler() @@ -76,12 +81,14 @@ class SamlHandler: self._error_html_content = hs.config.saml2_error_html_content - def handle_redirect_request(self, client_redirect_url): + def handle_redirect_request(self, client_redirect_url, ui_auth_session_id=None): """Handle an incoming request to /login/sso/redirect Args: client_redirect_url (bytes): the URL that we should redirect the client to when everything is done + ui_auth_session_id (Optional[str]): The session ID of the ongoing UI Auth (or + None if this is a login). Returns: bytes: URL to redirect to @@ -91,7 +98,9 @@ class SamlHandler: ) now = self._clock.time_msec() - self._outstanding_requests_dict[reqid] = Saml2SessionData(creation_time=now) + self._outstanding_requests_dict[reqid] = Saml2SessionData( + creation_time=now, ui_auth_session_id=ui_auth_session_id, + ) for key, value in info["headers"]: if key == "Location": @@ -118,7 +127,12 @@ class SamlHandler: self.expire_sessions() try: - user_id = await self._map_saml_response_to_user(resp_bytes, relay_state) + user_id, current_session = await self._map_saml_response_to_user( + resp_bytes, relay_state + ) + except RedirectException: + # Raise the exception as per the wishes of the SAML module response + raise except Exception as e: # If decoding the response or mapping it to a user failed, then log the # error and tell the user that something went wrong. @@ -133,9 +147,28 @@ class SamlHandler: finish_request(request) return - self._auth_handler.complete_sso_login(user_id, request, relay_state) + # Complete the interactive auth session or the login. + if current_session and current_session.ui_auth_session_id: + await self._auth_handler.complete_sso_ui_auth( + user_id, current_session.ui_auth_session_id, request + ) - async def _map_saml_response_to_user(self, resp_bytes, client_redirect_url): + else: + await self._auth_handler.complete_sso_login(user_id, request, relay_state) + + async def _map_saml_response_to_user( + self, resp_bytes: str, client_redirect_url: str + ) -> Tuple[str, Optional[Saml2SessionData]]: + """ + Given a sample response, retrieve the cached session and user for it. + + Args: + resp_bytes: The SAML response. + client_redirect_url: The redirect URL passed in by the client. + + Returns: + Tuple of the user ID and SAML session associated with this response. + """ try: saml2_auth = self._saml_client.parse_authn_request_response( resp_bytes, @@ -163,7 +196,9 @@ class SamlHandler: logger.info("SAML2 mapped attributes: %s", saml2_auth.ava) - self._outstanding_requests_dict.pop(saml2_auth.in_response_to, None) + current_session = self._outstanding_requests_dict.pop( + saml2_auth.in_response_to, None + ) remote_user_id = self._user_mapping_provider.get_remote_user_id( saml2_auth, client_redirect_url @@ -184,7 +219,7 @@ class SamlHandler: ) if registered_user_id is not None: logger.info("Found existing mapping %s", registered_user_id) - return registered_user_id + return registered_user_id, current_session # backwards-compatibility hack: see if there is an existing user with a # suitable mapping from the uid @@ -209,7 +244,7 @@ class SamlHandler: await self._datastore.record_user_external_id( self._auth_provider_id, remote_user_id, registered_user_id ) - return registered_user_id + return registered_user_id, current_session # Map saml response to user attributes using the configured mapping provider for i in range(1000): @@ -256,7 +291,7 @@ class SamlHandler: await self._datastore.record_user_external_id( self._auth_provider_id, remote_user_id, registered_user_id ) - return registered_user_id + return registered_user_id, current_session def expire_sessions(self): expire_before = self._clock.time_msec() - self._saml2_session_lifetime diff --git a/synapse/handlers/set_password.py b/synapse/handlers/set_password.py index 12657ca69..63d8f9aa0 100644 --- a/synapse/handlers/set_password.py +++ b/synapse/handlers/set_password.py @@ -15,8 +15,6 @@ import logging from typing import Optional -from twisted.internet import defer - from synapse.api.errors import Codes, StoreError, SynapseError from synapse.types import Requester @@ -32,9 +30,9 @@ class SetPasswordHandler(BaseHandler): super(SetPasswordHandler, self).__init__(hs) self._auth_handler = hs.get_auth_handler() self._device_handler = hs.get_device_handler() + self._password_policy_handler = hs.get_password_policy_handler() - @defer.inlineCallbacks - def set_password( + async def set_password( self, user_id: str, new_password: str, @@ -44,10 +42,11 @@ class SetPasswordHandler(BaseHandler): if not self.hs.config.password_localdb_enabled: raise SynapseError(403, "Password change disabled", errcode=Codes.FORBIDDEN) - password_hash = yield self._auth_handler.hash(new_password) + self._password_policy_handler.validate_password(new_password) + password_hash = await self._auth_handler.hash(new_password) try: - yield self.store.user_set_password_hash(user_id, password_hash) + await self.store.user_set_password_hash(user_id, password_hash) except StoreError as e: if e.code == 404: raise SynapseError(404, "Unknown user", Codes.NOT_FOUND) @@ -59,12 +58,12 @@ class SetPasswordHandler(BaseHandler): except_access_token_id = requester.access_token_id if requester else None # First delete all of their other devices. - yield self._device_handler.delete_all_devices_for_user( + await self._device_handler.delete_all_devices_for_user( user_id, except_device_id=except_device_id ) # and now delete any access tokens which weren't associated with # devices (or were associated with this device). - yield self._auth_handler.delete_access_tokens_for_user( + await self._auth_handler.delete_access_tokens_for_user( user_id, except_token_id=except_access_token_id ) diff --git a/synapse/handlers/sync.py b/synapse/handlers/sync.py index cfd5dfc9e..00718d7f2 100644 --- a/synapse/handlers/sync.py +++ b/synapse/handlers/sync.py @@ -26,7 +26,7 @@ from prometheus_client import Counter from synapse.api.constants import EventTypes, Membership from synapse.api.filtering import FilterCollection from synapse.events import EventBase -from synapse.logging.context import LoggingContext +from synapse.logging.context import current_context from synapse.push.clientformat import format_push_rules_for_user from synapse.storage.roommember import MemberSummary from synapse.storage.state import StateFilter @@ -301,7 +301,7 @@ class SyncHandler(object): else: sync_type = "incremental_sync" - context = LoggingContext.current_context() + context = current_context() if context: context.tag = sync_type @@ -1143,10 +1143,14 @@ class SyncHandler(object): user_id ) - tracked_users = set(users_who_share_room) + # Always tell the user about their own devices. We check as the user + # ID is almost certainly already included (unless they're not in any + # rooms) and taking a copy of the set is relatively expensive. + if user_id not in users_who_share_room: + users_who_share_room = set(users_who_share_room) + users_who_share_room.add(user_id) - # Always tell the user about their own devices - tracked_users.add(user_id) + tracked_users = users_who_share_room # Step 1a, check for changes in devices of users we share a room with users_that_have_changed = await self.store.get_users_whose_devices_changed( @@ -1639,7 +1643,7 @@ class SyncHandler(object): ) # We loop through all room ids, even if there are no new events, in case - # there are non room events taht we need to notify about. + # there are non room events that we need to notify about. for room_id in sync_result_builder.joined_room_ids: room_entry = room_to_events.get(room_id, None) diff --git a/synapse/handlers/typing.py b/synapse/handlers/typing.py index 391bceb0c..c7bc14c62 100644 --- a/synapse/handlers/typing.py +++ b/synapse/handlers/typing.py @@ -15,6 +15,7 @@ import logging from collections import namedtuple +from typing import List from twisted.internet import defer @@ -257,7 +258,13 @@ class TypingHandler(object): "typing_key", self._latest_room_serial, rooms=[member.room_id] ) - async def get_all_typing_updates(self, last_id, current_id): + async def get_all_typing_updates( + self, last_id: int, current_id: int, limit: int + ) -> List[dict]: + """Get up to `limit` typing updates between the given tokens, earliest + updates first. + """ + if last_id == current_id: return [] @@ -275,7 +282,7 @@ class TypingHandler(object): typing = self._room_typing[room_id] rows.append((serial, room_id, list(typing))) rows.sort() - return rows + return rows[:limit] def get_current_token(self): return self._latest_room_serial diff --git a/synapse/http/matrixfederationclient.py b/synapse/http/matrixfederationclient.py index 6f1bb04d8..6b0a532c0 100644 --- a/synapse/http/matrixfederationclient.py +++ b/synapse/http/matrixfederationclient.py @@ -434,21 +434,27 @@ class MatrixFederationHttpClient(object): logger.info("Failed to send request: %s", e) raise_from(RequestSendFailed(e, can_retry=True), e) - logger.info( - "{%s} [%s] Got response headers: %d %s", - request.txn_id, - request.destination, - response.code, - response.phrase.decode("ascii", errors="replace"), - ) - incoming_responses_counter.labels(method_bytes, response.code).inc() set_tag(tags.HTTP_STATUS_CODE, response.code) if 200 <= response.code < 300: + logger.debug( + "{%s} [%s] Got response headers: %d %s", + request.txn_id, + request.destination, + response.code, + response.phrase.decode("ascii", errors="replace"), + ) pass else: + logger.info( + "{%s} [%s] Got response headers: %d %s", + request.txn_id, + request.destination, + response.code, + response.phrase.decode("ascii", errors="replace"), + ) # :'( # Update transactions table? d = treq.content(response) diff --git a/synapse/http/request_metrics.py b/synapse/http/request_metrics.py index 58f9cc61c..b58ae3d9d 100644 --- a/synapse/http/request_metrics.py +++ b/synapse/http/request_metrics.py @@ -19,7 +19,7 @@ import threading from prometheus_client.core import Counter, Histogram -from synapse.logging.context import LoggingContext +from synapse.logging.context import current_context from synapse.metrics import LaterGauge logger = logging.getLogger(__name__) @@ -148,7 +148,7 @@ LaterGauge( class RequestMetrics(object): def start(self, time_sec, name, method): self.start = time_sec - self.start_context = LoggingContext.current_context() + self.start_context = current_context() self.name = name self.method = method @@ -163,7 +163,7 @@ class RequestMetrics(object): with _in_flight_requests_lock: _in_flight_requests.discard(self) - context = LoggingContext.current_context() + context = current_context() tag = "" if context: diff --git a/synapse/http/site.py b/synapse/http/site.py index e092193c9..32feb0d96 100644 --- a/synapse/http/site.py +++ b/synapse/http/site.py @@ -193,6 +193,12 @@ class SynapseRequest(Request): self.finish_time = time.time() Request.connectionLost(self, reason) + if self.logcontext is None: + logger.info( + "Connection from %s lost before request headers were read", self.client + ) + return + # we only get here if the connection to the client drops before we send # the response. # @@ -236,13 +242,6 @@ class SynapseRequest(Request): def _finished_processing(self): """Log the completion of this request and update the metrics """ - - if self.logcontext is None: - # this can happen if the connection closed before we read the - # headers (so render was never called). In that case we'll already - # have logged a warning, so just bail out. - return - usage = self.logcontext.get_resource_usage() if self._processing_finished_time is None: diff --git a/synapse/logging/_structured.py b/synapse/logging/_structured.py index ffa7b20ca..7372450b4 100644 --- a/synapse/logging/_structured.py +++ b/synapse/logging/_structured.py @@ -42,7 +42,7 @@ from synapse.logging._terse_json import ( TerseJSONToConsoleLogObserver, TerseJSONToTCPLogObserver, ) -from synapse.logging.context import LoggingContext +from synapse.logging.context import current_context def stdlib_log_level_to_twisted(level: str) -> LogLevel: @@ -86,7 +86,7 @@ class LogContextObserver(object): ].startswith("Timing out client"): return - context = LoggingContext.current_context() + context = current_context() # Copy the context information to the log event. if context is not None: diff --git a/synapse/logging/context.py b/synapse/logging/context.py index 860b99a4c..8b9c4e38b 100644 --- a/synapse/logging/context.py +++ b/synapse/logging/context.py @@ -27,6 +27,7 @@ import inspect import logging import threading import types +import warnings from typing import TYPE_CHECKING, Optional, Tuple, TypeVar, Union from typing_extensions import Literal @@ -51,7 +52,7 @@ try: is_thread_resource_usage_supported = True - def get_thread_resource_usage(): + def get_thread_resource_usage() -> "Optional[resource._RUsage]": return resource.getrusage(RUSAGE_THREAD) @@ -60,7 +61,7 @@ except Exception: # won't track resource usage. is_thread_resource_usage_supported = False - def get_thread_resource_usage(): + def get_thread_resource_usage() -> "Optional[resource._RUsage]": return None @@ -175,7 +176,54 @@ class ContextResourceUsage(object): return res -LoggingContextOrSentinel = Union["LoggingContext", "LoggingContext.Sentinel"] +LoggingContextOrSentinel = Union["LoggingContext", "_Sentinel"] + + +class _Sentinel(object): + """Sentinel to represent the root context""" + + __slots__ = ["previous_context", "finished", "request", "scope", "tag"] + + def __init__(self) -> None: + # Minimal set for compatibility with LoggingContext + self.previous_context = None + self.finished = False + self.request = None + self.scope = None + self.tag = None + + def __str__(self): + return "sentinel" + + def copy_to(self, record): + pass + + def copy_to_twisted_log_entry(self, record): + record["request"] = None + record["scope"] = None + + def start(self, rusage: "Optional[resource._RUsage]"): + pass + + def stop(self, rusage: "Optional[resource._RUsage]"): + pass + + def add_database_transaction(self, duration_sec): + pass + + def add_database_scheduled(self, sched_sec): + pass + + def record_event_fetch(self, event_count): + pass + + def __nonzero__(self): + return False + + __bool__ = __nonzero__ # python3 + + +SENTINEL_CONTEXT = _Sentinel() class LoggingContext(object): @@ -199,76 +247,33 @@ class LoggingContext(object): "_resource_usage", "usage_start", "main_thread", - "alive", + "finished", "request", "tag", "scope", ] - thread_local = threading.local() - - class Sentinel(object): - """Sentinel to represent the root context""" - - __slots__ = ["previous_context", "alive", "request", "scope", "tag"] - - def __init__(self) -> None: - # Minimal set for compatibility with LoggingContext - self.previous_context = None - self.alive = None - self.request = None - self.scope = None - self.tag = None - - def __str__(self): - return "sentinel" - - def copy_to(self, record): - pass - - def copy_to_twisted_log_entry(self, record): - record["request"] = None - record["scope"] = None - - def start(self): - pass - - def stop(self): - pass - - def add_database_transaction(self, duration_sec): - pass - - def add_database_scheduled(self, sched_sec): - pass - - def record_event_fetch(self, event_count): - pass - - def __nonzero__(self): - return False - - __bool__ = __nonzero__ # python3 - - sentinel = Sentinel() - def __init__(self, name=None, parent_context=None, request=None) -> None: - self.previous_context = LoggingContext.current_context() + self.previous_context = current_context() self.name = name # track the resources used by this context so far self._resource_usage = ContextResourceUsage() - # If alive has the thread resource usage when the logcontext last - # became active. - self.usage_start = None + # The thread resource usage when the logcontext became active. None + # if the context is not currently active. + self.usage_start = None # type: Optional[resource._RUsage] self.main_thread = get_thread_id() self.request = None self.tag = "" - self.alive = True self.scope = None # type: Optional[_LogContextScope] + # keep track of whether we have hit the __exit__ block for this context + # (suggesting that the the thing that created the context thinks it should + # be finished, and that re-activating it would suggest an error). + self.finished = False + self.parent_context = parent_context if self.parent_context is not None: @@ -287,40 +292,51 @@ class LoggingContext(object): def current_context(cls) -> LoggingContextOrSentinel: """Get the current logging context from thread local storage + This exists for backwards compatibility. ``current_context()`` should be + called directly. + Returns: LoggingContext: the current logging context """ - return getattr(cls.thread_local, "current_context", cls.sentinel) + warnings.warn( + "synapse.logging.context.LoggingContext.current_context() is deprecated " + "in favor of synapse.logging.context.current_context().", + DeprecationWarning, + stacklevel=2, + ) + return current_context() @classmethod def set_current_context( cls, context: LoggingContextOrSentinel ) -> LoggingContextOrSentinel: """Set the current logging context in thread local storage + + This exists for backwards compatibility. ``set_current_context()`` should be + called directly. + Args: context(LoggingContext): The context to activate. Returns: The context that was previously active """ - current = cls.current_context() - - if current is not context: - current.stop() - cls.thread_local.current_context = context - context.start() - return current + warnings.warn( + "synapse.logging.context.LoggingContext.set_current_context() is deprecated " + "in favor of synapse.logging.context.set_current_context().", + DeprecationWarning, + stacklevel=2, + ) + return set_current_context(context) def __enter__(self) -> "LoggingContext": """Enters this logging context into thread local storage""" - old_context = self.set_current_context(self) + old_context = set_current_context(self) if self.previous_context != old_context: logger.warning( "Expected previous context %r, found %r", self.previous_context, old_context, ) - self.alive = True - return self def __exit__(self, type, value, traceback) -> None: @@ -329,24 +345,19 @@ class LoggingContext(object): Returns: None to avoid suppressing any exceptions that were thrown. """ - current = self.set_current_context(self.previous_context) + current = set_current_context(self.previous_context) if current is not self: - if current is self.sentinel: + if current is SENTINEL_CONTEXT: logger.warning("Expected logging context %s was lost", self) else: logger.warning( "Expected logging context %s but found %s", self, current ) - self.alive = False - # if we have a parent, pass our CPU usage stats on - if self.parent_context is not None and hasattr( - self.parent_context, "_resource_usage" - ): - self.parent_context._resource_usage += self._resource_usage - - # reset them in case we get entered again - self._resource_usage.reset() + # the fact that we are here suggests that the caller thinks that everything + # is done and dusted for this logcontext, and further activity will not get + # recorded against the correct metrics. + self.finished = True def copy_to(self, record) -> None: """Copy logging fields from this context to a log record or @@ -366,35 +377,63 @@ class LoggingContext(object): record["request"] = self.request record["scope"] = self.scope - def start(self) -> None: + def start(self, rusage: "Optional[resource._RUsage]") -> None: + """ + Record that this logcontext is currently running. + + This should not be called directly: use set_current_context + + Args: + rusage: the resources used by the current thread, at the point of + switching to this logcontext. May be None if this platform doesn't + support getrusuage. + """ if get_thread_id() != self.main_thread: logger.warning("Started logcontext %s on different thread", self) return + if self.finished: + logger.warning("Re-starting finished log context %s", self) + # If we haven't already started record the thread resource usage so # far - if not self.usage_start: - self.usage_start = get_thread_resource_usage() + if self.usage_start: + logger.warning("Re-starting already-active log context %s", self) + else: + self.usage_start = rusage - def stop(self) -> None: - if get_thread_id() != self.main_thread: - logger.warning("Stopped logcontext %s on different thread", self) - return + def stop(self, rusage: "Optional[resource._RUsage]") -> None: + """ + Record that this logcontext is no longer running. - # When we stop, let's record the cpu used since we started - if not self.usage_start: - # Log a warning on platforms that support thread usage tracking - if is_thread_resource_usage_supported: + This should not be called directly: use set_current_context + + Args: + rusage: the resources used by the current thread, at the point of + switching away from this logcontext. May be None if this platform + doesn't support getrusuage. + """ + + try: + if get_thread_id() != self.main_thread: + logger.warning("Stopped logcontext %s on different thread", self) + return + + if not rusage: + return + + # Record the cpu used since we started + if not self.usage_start: logger.warning( - "Called stop on logcontext %s without calling start", self + "Called stop on logcontext %s without recording a start rusage", + self, ) - return + return - utime_delta, stime_delta = self._get_cputime() - self._resource_usage.ru_utime += utime_delta - self._resource_usage.ru_stime += stime_delta - - self.usage_start = None + utime_delta, stime_delta = self._get_cputime(rusage) + self.add_cputime(utime_delta, stime_delta) + finally: + self.usage_start = None def get_resource_usage(self) -> ContextResourceUsage: """Get resources used by this logcontext so far. @@ -409,25 +448,25 @@ class LoggingContext(object): # If we are on the correct thread and we're currently running then we # can include resource usage so far. is_main_thread = get_thread_id() == self.main_thread - if self.alive and self.usage_start and is_main_thread: - utime_delta, stime_delta = self._get_cputime() + if self.usage_start and is_main_thread: + rusage = get_thread_resource_usage() + assert rusage is not None + utime_delta, stime_delta = self._get_cputime(rusage) res.ru_utime += utime_delta res.ru_stime += stime_delta return res - def _get_cputime(self) -> Tuple[float, float]: - """Get the cpu usage time so far + def _get_cputime(self, current: "resource._RUsage") -> Tuple[float, float]: + """Get the cpu usage time between start() and the given rusage + + Args: + rusage: the current resource usage Returns: Tuple[float, float]: seconds in user mode, seconds in system mode """ assert self.usage_start is not None - current = get_thread_resource_usage() - - # Indicate to mypy that we know that self.usage_start is None. - assert self.usage_start is not None - utime_delta = current.ru_utime - self.usage_start.ru_utime stime_delta = current.ru_stime - self.usage_start.ru_stime @@ -450,30 +489,52 @@ class LoggingContext(object): return utime_delta, stime_delta + def add_cputime(self, utime_delta: float, stime_delta: float) -> None: + """Update the CPU time usage of this context (and any parents, recursively). + + Args: + utime_delta: additional user time, in seconds, spent in this context. + stime_delta: additional system time, in seconds, spent in this context. + """ + self._resource_usage.ru_utime += utime_delta + self._resource_usage.ru_stime += stime_delta + if self.parent_context: + self.parent_context.add_cputime(utime_delta, stime_delta) + def add_database_transaction(self, duration_sec: float) -> None: + """Record the use of a database transaction and the length of time it took. + + Args: + duration_sec: The number of seconds the database transaction took. + """ if duration_sec < 0: raise ValueError("DB txn time can only be non-negative") self._resource_usage.db_txn_count += 1 self._resource_usage.db_txn_duration_sec += duration_sec + if self.parent_context: + self.parent_context.add_database_transaction(duration_sec) def add_database_scheduled(self, sched_sec: float) -> None: """Record a use of the database pool Args: - sched_sec (float): number of seconds it took us to get a - connection + sched_sec: number of seconds it took us to get a connection """ if sched_sec < 0: raise ValueError("DB scheduling time can only be non-negative") self._resource_usage.db_sched_duration_sec += sched_sec + if self.parent_context: + self.parent_context.add_database_scheduled(sched_sec) def record_event_fetch(self, event_count: int) -> None: """Record a number of events being fetched from the db Args: - event_count (int): number of events being fetched + event_count: number of events being fetched """ self._resource_usage.evt_db_fetch_count += event_count + if self.parent_context: + self.parent_context.record_event_fetch(event_count) class LoggingContextFilter(logging.Filter): @@ -492,7 +553,7 @@ class LoggingContextFilter(logging.Filter): Returns: True to include the record in the log output. """ - context = LoggingContext.current_context() + context = current_context() for key, value in self.defaults.items(): setattr(record, key, value) @@ -512,27 +573,24 @@ class PreserveLoggingContext(object): __slots__ = ["current_context", "new_context", "has_parent"] - def __init__(self, new_context: Optional[LoggingContextOrSentinel] = None) -> None: - if new_context is None: - self.new_context = LoggingContext.sentinel # type: LoggingContextOrSentinel - else: - self.new_context = new_context + def __init__( + self, new_context: LoggingContextOrSentinel = SENTINEL_CONTEXT + ) -> None: + self.new_context = new_context def __enter__(self) -> None: """Captures the current logging context""" - self.current_context = LoggingContext.set_current_context(self.new_context) + self.current_context = set_current_context(self.new_context) if self.current_context: self.has_parent = self.current_context.previous_context is not None - if not self.current_context.alive: - logger.debug("Entering dead context: %s", self.current_context) def __exit__(self, type, value, traceback) -> None: """Restores the current logging context""" - context = LoggingContext.set_current_context(self.current_context) + context = set_current_context(self.current_context) if context != self.new_context: - if context is LoggingContext.sentinel: + if not context: logger.warning("Expected logging context %s was lost", self.new_context) else: logger.warning( @@ -541,9 +599,37 @@ class PreserveLoggingContext(object): context, ) - if self.current_context is not LoggingContext.sentinel: - if not self.current_context.alive: - logger.debug("Restoring dead context: %s", self.current_context) + +_thread_local = threading.local() +_thread_local.current_context = SENTINEL_CONTEXT + + +def current_context() -> LoggingContextOrSentinel: + """Get the current logging context from thread local storage""" + return getattr(_thread_local, "current_context", SENTINEL_CONTEXT) + + +def set_current_context(context: LoggingContextOrSentinel) -> LoggingContextOrSentinel: + """Set the current logging context in thread local storage + Args: + context(LoggingContext): The context to activate. + Returns: + The context that was previously active + """ + # everything blows up if we allow current_context to be set to None, so sanity-check + # that now. + if context is None: + raise TypeError("'context' argument may not be None") + + current = current_context() + + if current is not context: + rusage = get_thread_resource_usage() + current.stop(rusage) + _thread_local.current_context = context + context.start(rusage) + + return current def nested_logging_context( @@ -572,7 +658,7 @@ def nested_logging_context( if parent_context is not None: context = parent_context # type: LoggingContextOrSentinel else: - context = LoggingContext.current_context() + context = current_context() return LoggingContext( parent_context=context, request=str(context.request) + "-" + suffix ) @@ -604,7 +690,7 @@ def run_in_background(f, *args, **kwargs): CRITICAL error about an unhandled error will be logged without much indication about where it came from. """ - current = LoggingContext.current_context() + current = current_context() try: res = f(*args, **kwargs) except: # noqa: E722 @@ -625,7 +711,7 @@ def run_in_background(f, *args, **kwargs): # The function may have reset the context before returning, so # we need to restore it now. - ctx = LoggingContext.set_current_context(current) + ctx = set_current_context(current) # The original context will be restored when the deferred # completes, but there is nothing waiting for it, so it will @@ -674,7 +760,7 @@ def make_deferred_yieldable(deferred): # ok, we can't be sure that a yield won't block, so let's reset the # logcontext, and add a callback to the deferred to restore it. - prev_context = LoggingContext.set_current_context(LoggingContext.sentinel) + prev_context = set_current_context(SENTINEL_CONTEXT) deferred.addBoth(_set_context_cb, prev_context) return deferred @@ -684,7 +770,7 @@ ResultT = TypeVar("ResultT") def _set_context_cb(result: ResultT, context: LoggingContext) -> ResultT: """A callback function which just sets the logging context""" - LoggingContext.set_current_context(context) + set_current_context(context) return result @@ -752,7 +838,7 @@ def defer_to_threadpool(reactor, threadpool, f, *args, **kwargs): Deferred: A Deferred which fires a callback with the result of `f`, or an errback if `f` throws an exception. """ - logcontext = LoggingContext.current_context() + logcontext = current_context() def g(): with LoggingContext(parent_context=logcontext): diff --git a/synapse/logging/opentracing.py b/synapse/logging/opentracing.py index 0638cec42..5dddf5700 100644 --- a/synapse/logging/opentracing.py +++ b/synapse/logging/opentracing.py @@ -171,7 +171,7 @@ import logging import re import types from functools import wraps -from typing import Dict +from typing import TYPE_CHECKING, Dict from canonicaljson import json @@ -179,6 +179,9 @@ from twisted.internet import defer from synapse.config import ConfigError +if TYPE_CHECKING: + from synapse.server import HomeServer + # Helper class @@ -297,14 +300,11 @@ def _noop_context_manager(*args, **kwargs): # Setup -def init_tracer(config): +def init_tracer(hs: "HomeServer"): """Set the whitelists and initialise the JaegerClient tracer - - Args: - config (HomeserverConfig): The config used by the homeserver """ global opentracing - if not config.opentracer_enabled: + if not hs.config.opentracer_enabled: # We don't have a tracer opentracing = None return @@ -315,18 +315,15 @@ def init_tracer(config): "installed." ) - # Include the worker name - name = config.worker_name if config.worker_name else "master" - # Pull out the jaeger config if it was given. Otherwise set it to something sensible. # See https://github.com/jaegertracing/jaeger-client-python/blob/master/jaeger_client/config.py - set_homeserver_whitelist(config.opentracer_whitelist) + set_homeserver_whitelist(hs.config.opentracer_whitelist) JaegerConfig( - config=config.jaeger_config, - service_name="{} {}".format(config.server_name, name), - scope_manager=LogContextScopeManager(config), + config=hs.config.jaeger_config, + service_name="{} {}".format(hs.config.server_name, hs.get_instance_name()), + scope_manager=LogContextScopeManager(hs.config), ).initialize_tracer() diff --git a/synapse/logging/scopecontextmanager.py b/synapse/logging/scopecontextmanager.py index 4eed4f233..dc3ab00cb 100644 --- a/synapse/logging/scopecontextmanager.py +++ b/synapse/logging/scopecontextmanager.py @@ -19,7 +19,7 @@ from opentracing import Scope, ScopeManager import twisted -from synapse.logging.context import LoggingContext, nested_logging_context +from synapse.logging.context import current_context, nested_logging_context logger = logging.getLogger(__name__) @@ -49,11 +49,8 @@ class LogContextScopeManager(ScopeManager): (Scope) : the Scope that is active, or None if not available. """ - ctx = LoggingContext.current_context() - if ctx is LoggingContext.sentinel: - return None - else: - return ctx.scope + ctx = current_context() + return ctx.scope def activate(self, span, finish_on_close): """ @@ -70,9 +67,9 @@ class LogContextScopeManager(ScopeManager): """ enter_logcontext = False - ctx = LoggingContext.current_context() + ctx = current_context() - if ctx is LoggingContext.sentinel: + if not ctx: # We don't want this scope to affect. logger.error("Tried to activate scope outside of loggingcontext") return Scope(None, span) diff --git a/synapse/module_api/__init__.py b/synapse/module_api/__init__.py index c7fffd72f..d678c0eb9 100644 --- a/synapse/module_api/__init__.py +++ b/synapse/module_api/__init__.py @@ -86,7 +86,7 @@ class ModuleApi(object): Deferred[str|None]: Canonical (case-corrected) user_id, or None if the user is not registered. """ - return self._auth_handler.check_user_exists(user_id) + return defer.ensureDeferred(self._auth_handler.check_user_exists(user_id)) @defer.inlineCallbacks def register(self, localpart, displayname=None, emails=[]): @@ -196,7 +196,9 @@ class ModuleApi(object): yield self._hs.get_device_handler().delete_device(user_id, device_id) else: # no associated device. Just delete the access token. - yield self._auth_handler.delete_access_token(access_token) + yield defer.ensureDeferred( + self._auth_handler.delete_access_token(access_token) + ) def run_db_interaction(self, desc, func, *args, **kwargs): """Run a function with a database connection @@ -220,6 +222,8 @@ class ModuleApi(object): want their access token sent to `client_redirect_url`, or redirect them to that URL with a token directly if the URL matches with one of the whitelisted clients. + This is deprecated in favor of complete_sso_login_async. + Args: registered_user_id: The MXID that has been registered as a previous step of of this SSO login. @@ -227,6 +231,24 @@ class ModuleApi(object): client_redirect_url: The URL to which to offer to redirect the user (or to redirect them directly if whitelisted). """ - self._auth_handler.complete_sso_login( + self._auth_handler._complete_sso_login( + registered_user_id, request, client_redirect_url, + ) + + async def complete_sso_login_async( + self, registered_user_id: str, request: SynapseRequest, client_redirect_url: str + ): + """Complete a SSO login by redirecting the user to a page to confirm whether they + want their access token sent to `client_redirect_url`, or redirect them to that + URL with a token directly if the URL matches with one of the whitelisted clients. + + Args: + registered_user_id: The MXID that has been registered as a previous step of + of this SSO login. + request: The request to respond to. + client_redirect_url: The URL to which to offer to redirect the user (or to + redirect them directly if whitelisted). + """ + await self._auth_handler.complete_sso_login( registered_user_id, request, client_redirect_url, ) diff --git a/synapse/notifier.py b/synapse/notifier.py index 6132727cb..87c120a59 100644 --- a/synapse/notifier.py +++ b/synapse/notifier.py @@ -15,7 +15,7 @@ import logging from collections import namedtuple -from typing import Callable, List +from typing import Callable, Iterable, List, TypeVar from prometheus_client import Counter @@ -42,12 +42,14 @@ users_woken_by_stream_counter = Counter( "synapse_notifier_users_woken_by_stream", "", ["stream"] ) +T = TypeVar("T") + # TODO(paul): Should be shared somewhere -def count(func, l): - """Return the number of items in l for which func returns true.""" +def count(func: Callable[[T], bool], it: Iterable[T]) -> int: + """Return the number of items in it for which func returns true.""" n = 0 - for x in l: + for x in it: if func(x): n += 1 return n @@ -220,12 +222,6 @@ class Notifier(object): """ self.replication_callbacks.append(cb) - def add_remote_server_up_callback(self, cb: Callable[[str], None]): - """Add a callback that will be called when synapse detects a server - has been - """ - self.remote_server_up_callbacks.append(cb) - def on_new_room_event( self, event, room_stream_id, max_room_stream_id, extra_users=[] ): @@ -279,10 +275,9 @@ class Notifier(object): "room_key", room_stream_id, users=extra_users, rooms=[event.room_id] ) - @defer.inlineCallbacks - def _notify_app_services(self, room_stream_id): + async def _notify_app_services(self, room_stream_id): try: - yield self.appservice_handler.notify_interested_services(room_stream_id) + await self.appservice_handler.notify_interested_services(room_stream_id) except Exception: logger.exception("Error notifying application services of event") @@ -481,20 +476,18 @@ class Notifier(object): return result - @defer.inlineCallbacks - def _get_room_ids(self, user, explicit_room_id): - joined_room_ids = yield self.store.get_rooms_for_user(user.to_string()) + async def _get_room_ids(self, user, explicit_room_id): + joined_room_ids = await self.store.get_rooms_for_user(user.to_string()) if explicit_room_id: if explicit_room_id in joined_room_ids: return [explicit_room_id], True - if (yield self._is_world_readable(explicit_room_id)): + if await self._is_world_readable(explicit_room_id): return [explicit_room_id], False raise AuthError(403, "Non-joined access not allowed") return joined_room_ids, True - @defer.inlineCallbacks - def _is_world_readable(self, room_id): - state = yield self.state_handler.get_current_state( + async def _is_world_readable(self, room_id): + state = await self.state_handler.get_current_state( room_id, EventTypes.RoomHistoryVisibility, "" ) if state and "history_visibility" in state.content: @@ -544,6 +537,3 @@ class Notifier(object): # circular dependencies. if self.federation_sender: self.federation_sender.wake_destination(server) - - for cb in self.remote_server_up_callbacks: - cb(server) diff --git a/synapse/push/mailer.py b/synapse/push/mailer.py index 73580c1c6..ab33abbee 100644 --- a/synapse/push/mailer.py +++ b/synapse/push/mailer.py @@ -19,6 +19,7 @@ import logging import time from email.mime.multipart import MIMEMultipart from email.mime.text import MIMEText +from typing import Iterable, List, TypeVar from six.moves import urllib @@ -41,6 +42,8 @@ from synapse.visibility import filter_events_for_client logger = logging.getLogger(__name__) +T = TypeVar("T") + MESSAGE_FROM_PERSON_IN_ROOM = ( "You have a message on %(app)s from %(person)s in the %(room)s room..." @@ -638,10 +641,10 @@ def safe_text(raw_text): ) -def deduped_ordered_list(l): +def deduped_ordered_list(it: Iterable[T]) -> List[T]: seen = set() ret = [] - for item in l: + for item in it: if item not in seen: seen.add(item) ret.append(item) diff --git a/synapse/python_dependencies.py b/synapse/python_dependencies.py index 8de8cb2c1..39c99a280 100644 --- a/synapse/python_dependencies.py +++ b/synapse/python_dependencies.py @@ -98,6 +98,9 @@ CONDITIONAL_REQUIREMENTS = { "sentry": ["sentry-sdk>=0.7.2"], "opentracing": ["jaeger-client>=4.0.0", "opentracing>=2.2.0"], "jwt": ["pyjwt>=1.6.4"], + # hiredis is not a *strict* dependency, but it makes things much faster. + # (if it is not installed, we fall back to slow code.) + "redis": ["txredisapi>=1.4.7", "hiredis"], } ALL_OPTIONAL_REQUIREMENTS = set() # type: Set[str] diff --git a/synapse/replication/http/__init__.py b/synapse/replication/http/__init__.py index 28dbc6fcb..4613b2538 100644 --- a/synapse/replication/http/__init__.py +++ b/synapse/replication/http/__init__.py @@ -21,6 +21,7 @@ from synapse.replication.http import ( membership, register, send_event, + streams, ) REPLICATION_PREFIX = "/_synapse/replication" @@ -38,3 +39,4 @@ class ReplicationRestResource(JsonResource): login.register_servlets(hs, self) register.register_servlets(hs, self) devices.register_servlets(hs, self) + streams.register_servlets(hs, self) diff --git a/synapse/replication/http/_base.py b/synapse/replication/http/_base.py index 1be1ccbdf..f88c80ae8 100644 --- a/synapse/replication/http/_base.py +++ b/synapse/replication/http/_base.py @@ -16,6 +16,7 @@ import abc import logging import re +from inspect import signature from typing import Dict, List, Tuple from six import raise_from @@ -60,6 +61,8 @@ class ReplicationEndpoint(object): must call `register` to register the path with the HTTP server. Requests can be sent by calling the client returned by `make_client`. + Requests are sent to master process by default, but can be sent to other + named processes by specifying an `instance_name` keyword argument. Attributes: NAME (str): A name for the endpoint, added to the path as well as used @@ -91,6 +94,16 @@ class ReplicationEndpoint(object): hs, "repl." + self.NAME, timeout_ms=30 * 60 * 1000 ) + # We reserve `instance_name` as a parameter to sending requests, so we + # assert here that sub classes don't try and use the name. + assert ( + "instance_name" not in self.PATH_ARGS + ), "`instance_name` is a reserved paramater name" + assert ( + "instance_name" + not in signature(self.__class__._serialize_payload).parameters + ), "`instance_name` is a reserved paramater name" + assert self.METHOD in ("PUT", "POST", "GET") @abc.abstractmethod @@ -135,7 +148,11 @@ class ReplicationEndpoint(object): @trace(opname="outgoing_replication_request") @defer.inlineCallbacks - def send_request(**kwargs): + def send_request(instance_name="master", **kwargs): + # Currently we only support sending requests to master process. + if instance_name != "master": + raise Exception("Unknown instance") + data = yield cls._serialize_payload(**kwargs) url_args = [ diff --git a/synapse/replication/http/streams.py b/synapse/replication/http/streams.py new file mode 100644 index 000000000..0459f582b --- /dev/null +++ b/synapse/replication/http/streams.py @@ -0,0 +1,82 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +import logging + +from synapse.api.errors import SynapseError +from synapse.http.servlet import parse_integer +from synapse.replication.http._base import ReplicationEndpoint + +logger = logging.getLogger(__name__) + + +class ReplicationGetStreamUpdates(ReplicationEndpoint): + """Fetches stream updates from a server. Used for streams not persisted to + the database, e.g. typing notifications. + + The API looks like: + + GET /_synapse/replication/get_repl_stream_updates/?from_token=0&to_token=10 + + 200 OK + + { + updates: [ ... ], + upto_token: 10, + limited: False, + } + + If there are more rows than can sensibly be returned in one lump, `limited` will be + set to true, and the caller should call again with a new `from_token`. + + """ + + NAME = "get_repl_stream_updates" + PATH_ARGS = ("stream_name",) + METHOD = "GET" + + def __init__(self, hs): + super().__init__(hs) + + self._instance_name = hs.get_instance_name() + + # We pull the streams from the replication steamer (if we try and make + # them ourselves we end up in an import loop). + self.streams = hs.get_replication_streamer().get_streams() + + @staticmethod + def _serialize_payload(stream_name, from_token, upto_token): + return {"from_token": from_token, "upto_token": upto_token} + + async def _handle_request(self, request, stream_name): + stream = self.streams.get(stream_name) + if stream is None: + raise SynapseError(400, "Unknown stream") + + from_token = parse_integer(request, "from_token", required=True) + upto_token = parse_integer(request, "upto_token", required=True) + + updates, upto_token, limited = await stream.get_updates_since( + self._instance_name, from_token, upto_token + ) + + return ( + 200, + {"updates": updates, "upto_token": upto_token, "limited": limited}, + ) + + +def register_servlets(hs, http_server): + ReplicationGetStreamUpdates(hs).register(http_server) diff --git a/synapse/replication/slave/storage/_base.py b/synapse/replication/slave/storage/_base.py index f45cbd37a..5d7c8871a 100644 --- a/synapse/replication/slave/storage/_base.py +++ b/synapse/replication/slave/storage/_base.py @@ -14,12 +14,14 @@ # limitations under the License. import logging -from typing import Dict, Optional +from typing import Optional import six -from synapse.storage._base import SQLBaseStore -from synapse.storage.data_stores.main.cache import CURRENT_STATE_CACHE_NAME +from synapse.storage.data_stores.main.cache import ( + CURRENT_STATE_CACHE_NAME, + CacheInvalidationWorkerStore, +) from synapse.storage.database import Database from synapse.storage.engines import PostgresEngine @@ -35,7 +37,7 @@ def __func__(inp): return inp.__func__ -class BaseSlavedStore(SQLBaseStore): +class BaseSlavedStore(CacheInvalidationWorkerStore): def __init__(self, database: Database, db_conn, hs): super(BaseSlavedStore, self).__init__(database, db_conn, hs) if isinstance(self.database_engine, PostgresEngine): @@ -47,18 +49,11 @@ class BaseSlavedStore(SQLBaseStore): self.hs = hs - def stream_positions(self) -> Dict[str, int]: - """ - Get the current positions of all the streams this store wants to subscribe to - - Returns: - map from stream name to the most recent update we have for - that stream (ie, the point we want to start replicating from) - """ - pos = {} + def get_cache_stream_token(self): if self._cache_id_gen: - pos["caches"] = self._cache_id_gen.get_current_token() - return pos + return self._cache_id_gen.get_current_token() + else: + return 0 def process_replication_rows(self, stream_name, token, rows): if stream_name == "caches": diff --git a/synapse/replication/slave/storage/account_data.py b/synapse/replication/slave/storage/account_data.py index ebe94909c..65e54b1c7 100644 --- a/synapse/replication/slave/storage/account_data.py +++ b/synapse/replication/slave/storage/account_data.py @@ -32,14 +32,6 @@ class SlavedAccountDataStore(TagsWorkerStore, AccountDataWorkerStore, BaseSlaved def get_max_account_data_stream_id(self): return self._account_data_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedAccountDataStore, self).stream_positions() - position = self._account_data_id_gen.get_current_token() - result["user_account_data"] = position - result["room_account_data"] = position - result["tag_account_data"] = position - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "tag_account_data": self._account_data_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/deviceinbox.py b/synapse/replication/slave/storage/deviceinbox.py index 0c237c6e0..c923751e5 100644 --- a/synapse/replication/slave/storage/deviceinbox.py +++ b/synapse/replication/slave/storage/deviceinbox.py @@ -43,11 +43,6 @@ class SlavedDeviceInboxStore(DeviceInboxWorkerStore, BaseSlavedStore): expiry_ms=30 * 60 * 1000, ) - def stream_positions(self): - result = super(SlavedDeviceInboxStore, self).stream_positions() - result["to_device"] = self._device_inbox_id_gen.get_current_token() - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "to_device": self._device_inbox_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/devices.py b/synapse/replication/slave/storage/devices.py index 1c77687ee..58fb0eaae 100644 --- a/synapse/replication/slave/storage/devices.py +++ b/synapse/replication/slave/storage/devices.py @@ -29,7 +29,13 @@ class SlavedDeviceStore(EndToEndKeyWorkerStore, DeviceWorkerStore, BaseSlavedSto self.hs = hs self._device_list_id_gen = SlavedIdTracker( - db_conn, "device_lists_stream", "stream_id" + db_conn, + "device_lists_stream", + "stream_id", + extra_tables=[ + ("user_signature_stream", "stream_id"), + ("device_lists_outbound_pokes", "stream_id"), + ], ) device_list_max = self._device_list_id_gen.get_current_token() self._device_list_stream_cache = StreamChangeCache( @@ -42,36 +48,30 @@ class SlavedDeviceStore(EndToEndKeyWorkerStore, DeviceWorkerStore, BaseSlavedSto "DeviceListFederationStreamChangeCache", device_list_max ) - def stream_positions(self): - result = super(SlavedDeviceStore, self).stream_positions() - # The user signature stream uses the same stream ID generator as the - # device list stream, so set them both to the device list ID - # generator's current token. - current_token = self._device_list_id_gen.get_current_token() - result[DeviceListsStream.NAME] = current_token - result[UserSignatureStream.NAME] = current_token - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == DeviceListsStream.NAME: self._device_list_id_gen.advance(token) - for row in rows: - self._invalidate_caches_for_devices(token, row.user_id, row.destination) + self._invalidate_caches_for_devices(token, rows) elif stream_name == UserSignatureStream.NAME: + self._device_list_id_gen.advance(token) for row in rows: self._user_signature_stream_cache.entity_has_changed(row.user_id, token) return super(SlavedDeviceStore, self).process_replication_rows( stream_name, token, rows ) - def _invalidate_caches_for_devices(self, token, user_id, destination): - self._device_list_stream_cache.entity_has_changed(user_id, token) + def _invalidate_caches_for_devices(self, token, rows): + for row in rows: + # The entities are either user IDs (starting with '@') whose devices + # have changed, or remote servers that we need to tell about + # changes. + if row.entity.startswith("@"): + self._device_list_stream_cache.entity_has_changed(row.entity, token) + self.get_cached_devices_for_user.invalidate((row.entity,)) + self._get_cached_user_device.invalidate_many((row.entity,)) + self.get_device_list_last_stream_id_for_remote.invalidate((row.entity,)) - if destination: - self._device_list_federation_stream_cache.entity_has_changed( - destination, token - ) - - self.get_cached_devices_for_user.invalidate((user_id,)) - self._get_cached_user_device.invalidate_many((user_id,)) - self.get_device_list_last_stream_id_for_remote.invalidate((user_id,)) + else: + self._device_list_federation_stream_cache.entity_has_changed( + row.entity, token + ) diff --git a/synapse/replication/slave/storage/events.py b/synapse/replication/slave/storage/events.py index e73342c65..15011259d 100644 --- a/synapse/replication/slave/storage/events.py +++ b/synapse/replication/slave/storage/events.py @@ -93,12 +93,6 @@ class SlavedEventStore( def get_room_min_stream_ordering(self): return self._backfill_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedEventStore, self).stream_positions() - result["events"] = self._stream_id_gen.get_current_token() - result["backfill"] = -self._backfill_id_gen.get_current_token() - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "events": self._stream_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/groups.py b/synapse/replication/slave/storage/groups.py index 2d4fd08cf..01bcf0e88 100644 --- a/synapse/replication/slave/storage/groups.py +++ b/synapse/replication/slave/storage/groups.py @@ -37,11 +37,6 @@ class SlavedGroupServerStore(GroupServerWorkerStore, BaseSlavedStore): def get_group_stream_token(self): return self._group_updates_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedGroupServerStore, self).stream_positions() - result["groups"] = self._group_updates_id_gen.get_current_token() - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "groups": self._group_updates_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/presence.py b/synapse/replication/slave/storage/presence.py index ad8f0c15a..fae312507 100644 --- a/synapse/replication/slave/storage/presence.py +++ b/synapse/replication/slave/storage/presence.py @@ -41,15 +41,6 @@ class SlavedPresenceStore(BaseSlavedStore): def get_current_presence_token(self): return self._presence_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedPresenceStore, self).stream_positions() - - if self.hs.config.use_presence: - position = self._presence_id_gen.get_current_token() - result["presence"] = position - - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "presence": self._presence_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/push_rule.py b/synapse/replication/slave/storage/push_rule.py index eebd5a1fb..6138796da 100644 --- a/synapse/replication/slave/storage/push_rule.py +++ b/synapse/replication/slave/storage/push_rule.py @@ -37,11 +37,6 @@ class SlavedPushRuleStore(SlavedEventStore, PushRulesWorkerStore): def get_max_push_rules_stream_id(self): return self._push_rules_stream_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedPushRuleStore, self).stream_positions() - result["push_rules"] = self._push_rules_stream_id_gen.get_current_token() - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "push_rules": self._push_rules_stream_id_gen.advance(token) diff --git a/synapse/replication/slave/storage/pushers.py b/synapse/replication/slave/storage/pushers.py index f22c2d44a..67be33794 100644 --- a/synapse/replication/slave/storage/pushers.py +++ b/synapse/replication/slave/storage/pushers.py @@ -28,10 +28,8 @@ class SlavedPusherStore(PusherWorkerStore, BaseSlavedStore): db_conn, "pushers", "id", extra_tables=[("deleted_pushers", "stream_id")] ) - def stream_positions(self): - result = super(SlavedPusherStore, self).stream_positions() - result["pushers"] = self._pushers_id_gen.get_current_token() - return result + def get_pushers_stream_token(self): + return self._pushers_id_gen.get_current_token() def process_replication_rows(self, stream_name, token, rows): if stream_name == "pushers": diff --git a/synapse/replication/slave/storage/receipts.py b/synapse/replication/slave/storage/receipts.py index d40dc6e1f..993432edc 100644 --- a/synapse/replication/slave/storage/receipts.py +++ b/synapse/replication/slave/storage/receipts.py @@ -42,11 +42,6 @@ class SlavedReceiptsStore(ReceiptsWorkerStore, BaseSlavedStore): def get_max_receipt_stream_id(self): return self._receipts_id_gen.get_current_token() - def stream_positions(self): - result = super(SlavedReceiptsStore, self).stream_positions() - result["receipts"] = self._receipts_id_gen.get_current_token() - return result - def invalidate_caches_for_receipt(self, room_id, receipt_type, user_id): self.get_receipts_for_user.invalidate((user_id, receipt_type)) self._get_linearized_receipts_for_room.invalidate_many((room_id,)) diff --git a/synapse/replication/slave/storage/room.py b/synapse/replication/slave/storage/room.py index 3a20f4531..10dda8708 100644 --- a/synapse/replication/slave/storage/room.py +++ b/synapse/replication/slave/storage/room.py @@ -30,11 +30,6 @@ class RoomStore(RoomWorkerStore, BaseSlavedStore): def get_current_public_room_stream_id(self): return self._public_room_id_gen.get_current_token() - def stream_positions(self): - result = super(RoomStore, self).stream_positions() - result["public_rooms"] = self._public_room_id_gen.get_current_token() - return result - def process_replication_rows(self, stream_name, token, rows): if stream_name == "public_rooms": self._public_room_id_gen.advance(token) diff --git a/synapse/replication/tcp/__init__.py b/synapse/replication/tcp/__init__.py index 81c2ea7ee..523a1358d 100644 --- a/synapse/replication/tcp/__init__.py +++ b/synapse/replication/tcp/__init__.py @@ -20,11 +20,31 @@ Further details can be found in docs/tcp_replication.rst Structure of the module: - * client.py - the client classes used for workers to connect to master + * handler.py - the classes used to handle sending/receiving commands to + replication * command.py - the definitions of all the valid commands - * protocol.py - contains bot the client and server protocol implementations, - these should not be used directly - * resource.py - the server classes that accepts and handle client connections - * streams.py - the definitons of all the valid streams + * protocol.py - the TCP protocol classes + * resource.py - handles streaming stream updates to replications + * streams/ - the definitons of all the valid streams + +The general interaction of the classes are: + + +---------------------+ + | ReplicationStreamer | + +---------------------+ + | + v + +---------------------------+ +----------------------+ + | ReplicationCommandHandler |---->|ReplicationDataHandler| + +---------------------------+ +----------------------+ + | ^ + v | + +-------------+ + | Protocols | + | (TCP/redis) | + +-------------+ + +Where the ReplicationDataHandler (or subclasses) handles incoming stream +updates. """ diff --git a/synapse/replication/tcp/client.py b/synapse/replication/tcp/client.py index 02ab5b66e..3bbf3c356 100644 --- a/synapse/replication/tcp/client.py +++ b/synapse/replication/tcp/client.py @@ -16,45 +16,40 @@ """ import logging -from typing import Dict, List, Optional +from typing import TYPE_CHECKING -from twisted.internet import defer from twisted.internet.protocol import ReconnectingClientFactory from synapse.replication.slave.storage._base import BaseSlavedStore -from synapse.replication.tcp.protocol import ( - AbstractReplicationClientHandler, - ClientReplicationStreamProtocol, -) +from synapse.replication.tcp.protocol import ClientReplicationStreamProtocol -from .commands import ( - Command, - FederationAckCommand, - InvalidateCacheCommand, - RemoteServerUpCommand, - RemovePusherCommand, - UserIpCommand, - UserSyncCommand, -) +if TYPE_CHECKING: + from synapse.server import HomeServer + from synapse.replication.tcp.handler import ReplicationCommandHandler logger = logging.getLogger(__name__) -class ReplicationClientFactory(ReconnectingClientFactory): +class DirectTcpReplicationClientFactory(ReconnectingClientFactory): """Factory for building connections to the master. Will reconnect if the connection is lost. - Accepts a handler that will be called when new data is available or data - is required. + Accepts a handler that is passed to `ClientReplicationStreamProtocol`. """ initialDelay = 0.1 maxDelay = 1 # Try at least once every N seconds - def __init__(self, hs, client_name, handler: AbstractReplicationClientHandler): + def __init__( + self, + hs: "HomeServer", + client_name: str, + command_handler: "ReplicationCommandHandler", + ): self.client_name = client_name - self.handler = handler + self.command_handler = command_handler self.server_name = hs.config.server_name + self.hs = hs self._clock = hs.get_clock() # As self.clock is defined in super class hs.get_reactor().addSystemEventTrigger("before", "shutdown", self.stopTrying) @@ -65,7 +60,11 @@ class ReplicationClientFactory(ReconnectingClientFactory): def buildProtocol(self, addr): logger.info("Connected to replication: %r", addr) return ClientReplicationStreamProtocol( - self.client_name, self.server_name, self._clock, self.handler + self.hs, + self.client_name, + self.server_name, + self._clock, + self.command_handler, ) def clientConnectionLost(self, connector, reason): @@ -77,168 +76,34 @@ class ReplicationClientFactory(ReconnectingClientFactory): ReconnectingClientFactory.clientConnectionFailed(self, connector, reason) -class ReplicationClientHandler(AbstractReplicationClientHandler): - """A base handler that can be passed to the ReplicationClientFactory. +class ReplicationDataHandler: + """Handles incoming stream updates from replication. - By default proxies incoming replication data to the SlaveStore. + This instance notifies the slave data store about updates. Can be subclassed + to handle updates in additional ways. """ def __init__(self, store: BaseSlavedStore): self.store = store - # The current connection. None if we are currently (re)connecting - self.connection = None - - # Any pending commands to be sent once a new connection has been - # established - self.pending_commands = [] # type: List[Command] - - # Map from string -> deferred, to wake up when receiveing a SYNC with - # the given string. - # Used for tests. - self.awaiting_syncs = {} # type: Dict[str, defer.Deferred] - - # The factory used to create connections. - self.factory = None # type: Optional[ReplicationClientFactory] - - def start_replication(self, hs): - """Helper method to start a replication connection to the remote server - using TCP. - """ - client_name = hs.config.worker_name - self.factory = ReplicationClientFactory(hs, client_name, self) - host = hs.config.worker_replication_host - port = hs.config.worker_replication_port - hs.get_reactor().connectTCP(host, port, self.factory) - - async def on_rdata(self, stream_name, token, rows): + async def on_rdata( + self, stream_name: str, instance_name: str, token: int, rows: list + ): """Called to handle a batch of replication data with a given stream token. By default this just pokes the slave store. Can be overridden in subclasses to handle more. Args: - stream_name (str): name of the replication stream for this batch of rows - token (int): stream token for this batch of rows - rows (list): a list of Stream.ROW_TYPE objects as returned by - Stream.parse_row. + stream_name: name of the replication stream for this batch of rows + instance_name: the instance that wrote the rows. + token: stream token for this batch of rows + rows: a list of Stream.ROW_TYPE objects as returned by Stream.parse_row. """ - logger.debug("Received rdata %s -> %s", stream_name, token) self.store.process_replication_rows(stream_name, token, rows) - async def on_position(self, stream_name, token): - """Called when we get new position data. By default this just pokes - the slave store. - - Can be overriden in subclasses to handle more. - """ + async def on_position(self, stream_name: str, token: int): self.store.process_replication_rows(stream_name, token, []) - def on_sync(self, data): - """When we received a SYNC we wake up any deferreds that were waiting - for the sync with the given data. - - Used by tests. - """ - d = self.awaiting_syncs.pop(data, None) - if d: - d.callback(data) - def on_remote_server_up(self, server: str): """Called when get a new REMOTE_SERVER_UP command.""" - - def get_streams_to_replicate(self) -> Dict[str, int]: - """Called when a new connection has been established and we need to - subscribe to streams. - - Returns: - map from stream name to the most recent update we have for - that stream (ie, the point we want to start replicating from) - """ - args = self.store.stream_positions() - user_account_data = args.pop("user_account_data", None) - room_account_data = args.pop("room_account_data", None) - if user_account_data: - args["account_data"] = user_account_data - elif room_account_data: - args["account_data"] = room_account_data - - return args - - def get_currently_syncing_users(self): - """Get the list of currently syncing users (if any). This is called - when a connection has been established and we need to send the - currently syncing users. (Overriden by the synchrotron's only) - """ - return [] - - def send_command(self, cmd): - """Send a command to master (when we get establish a connection if we - don't have one already.) - """ - if self.connection: - self.connection.send_command(cmd) - else: - logger.warning("Queuing command as not connected: %r", cmd.NAME) - self.pending_commands.append(cmd) - - def send_federation_ack(self, token): - """Ack data for the federation stream. This allows the master to drop - data stored purely in memory. - """ - self.send_command(FederationAckCommand(token)) - - def send_user_sync(self, user_id, is_syncing, last_sync_ms): - """Poke the master that a user has started/stopped syncing. - """ - self.send_command(UserSyncCommand(user_id, is_syncing, last_sync_ms)) - - def send_remove_pusher(self, app_id, push_key, user_id): - """Poke the master to remove a pusher for a user - """ - cmd = RemovePusherCommand(app_id, push_key, user_id) - self.send_command(cmd) - - def send_invalidate_cache(self, cache_func, keys): - """Poke the master to invalidate a cache. - """ - cmd = InvalidateCacheCommand(cache_func.__name__, keys) - self.send_command(cmd) - - def send_user_ip(self, user_id, access_token, ip, user_agent, device_id, last_seen): - """Tell the master that the user made a request. - """ - cmd = UserIpCommand(user_id, access_token, ip, user_agent, device_id, last_seen) - self.send_command(cmd) - - def send_remote_server_up(self, server: str): - self.send_command(RemoteServerUpCommand(server)) - - def await_sync(self, data): - """Returns a deferred that is resolved when we receive a SYNC command - with given data. - - [Not currently] used by tests. - """ - return self.awaiting_syncs.setdefault(data, defer.Deferred()) - - def update_connection(self, connection): - """Called when a connection has been established (or lost with None). - """ - self.connection = connection - if connection: - for cmd in self.pending_commands: - connection.send_command(cmd) - self.pending_commands = [] - - def finished_connecting(self): - """Called when we have successfully subscribed and caught up to all - streams we're interested in. - """ - logger.info("Finished connecting to server") - - # We don't reset the delay any earlier as otherwise if there is a - # problem during start up we'll end up tight looping connecting to the - # server. - if self.factory: - self.factory.resetDelay() diff --git a/synapse/replication/tcp/commands.py b/synapse/replication/tcp/commands.py index 451671412..f58e384d1 100644 --- a/synapse/replication/tcp/commands.py +++ b/synapse/replication/tcp/commands.py @@ -17,7 +17,7 @@ The VALID_SERVER_COMMANDS and VALID_CLIENT_COMMANDS define which commands are allowed to be sent by which side. """ - +import abc import logging import platform from typing import Tuple, Type @@ -34,34 +34,29 @@ else: logger = logging.getLogger(__name__) -class Command(object): +class Command(metaclass=abc.ABCMeta): """The base command class. All subclasses must set the NAME variable which equates to the name of the command on the wire. A full command line on the wire is constructed from `NAME + " " + to_line()` - - The default implementation creates a command of form ` ` """ NAME = None # type: str - def __init__(self, data): - self.data = data - @classmethod + @abc.abstractmethod def from_line(cls, line): """Deserialises a line from the wire into this command. `line` does not include the command. """ - return cls(line) - def to_line(self): + @abc.abstractmethod + def to_line(self) -> str: """Serialises the comamnd for the wire. Does not include the command prefix. """ - return self.data def get_logcontext_id(self): """Get a suitable string for the logcontext when processing this command""" @@ -70,7 +65,21 @@ class Command(object): return self.NAME -class ServerCommand(Command): +class _SimpleCommand(Command): + """An implementation of Command whose argument is just a 'data' string.""" + + def __init__(self, data): + self.data = data + + @classmethod + def from_line(cls, line): + return cls(line) + + def to_line(self) -> str: + return self.data + + +class ServerCommand(_SimpleCommand): """Sent by the server on new connection and includes the server_name. Format:: @@ -86,7 +95,7 @@ class RdataCommand(Command): Format:: - RDATA + RDATA The `` may either be a numeric stream id OR "batch". The latter case is used to support sending multiple updates with the same stream ID. This @@ -96,33 +105,40 @@ class RdataCommand(Command): The client should batch all incoming RDATA with a token of "batch" (per stream_name) until it sees an RDATA with a numeric stream ID. + The `` is the source of the new data (usually "master"). + `` of "batch" maps to the instance variable `token` being None. An example of a batched series of RDATA:: - RDATA presence batch ["@foo:example.com", "online", ...] - RDATA presence batch ["@bar:example.com", "online", ...] - RDATA presence 59 ["@baz:example.com", "online", ...] + RDATA presence master batch ["@foo:example.com", "online", ...] + RDATA presence master batch ["@bar:example.com", "online", ...] + RDATA presence master 59 ["@baz:example.com", "online", ...] """ NAME = "RDATA" - def __init__(self, stream_name, token, row): + def __init__(self, stream_name, instance_name, token, row): self.stream_name = stream_name + self.instance_name = instance_name self.token = token self.row = row @classmethod def from_line(cls, line): - stream_name, token, row_json = line.split(" ", 2) + stream_name, instance_name, token, row_json = line.split(" ", 3) return cls( - stream_name, None if token == "batch" else int(token), json.loads(row_json) + stream_name, + instance_name, + None if token == "batch" else int(token), + json.loads(row_json), ) def to_line(self): return " ".join( ( self.stream_name, + self.instance_name, str(self.token) if self.token is not None else "batch", _json_encoder.encode(self.row), ) @@ -136,26 +152,34 @@ class PositionCommand(Command): """Sent by the server to tell the client the stream postition without needing to send an RDATA. - Sent to the client after all missing updates for a stream have been sent - to the client and they're now up to date. + Format:: + + POSITION + + On receipt of a POSITION command clients should check if they have missed + any updates, and if so then fetch them out of band. + + The `` is the process that sent the command and is the source + of the stream. """ NAME = "POSITION" - def __init__(self, stream_name, token): + def __init__(self, stream_name, instance_name, token): self.stream_name = stream_name + self.instance_name = instance_name self.token = token @classmethod def from_line(cls, line): - stream_name, token = line.split(" ", 1) - return cls(stream_name, int(token)) + stream_name, instance_name, token = line.split(" ", 2) + return cls(stream_name, instance_name, int(token)) def to_line(self): - return " ".join((self.stream_name, str(self.token))) + return " ".join((self.stream_name, self.instance_name, str(self.token))) -class ErrorCommand(Command): +class ErrorCommand(_SimpleCommand): """Sent by either side if there was an ERROR. The data is a string describing the error. """ @@ -163,14 +187,14 @@ class ErrorCommand(Command): NAME = "ERROR" -class PingCommand(Command): +class PingCommand(_SimpleCommand): """Sent by either side as a keep alive. The data is arbitary (often timestamp) """ NAME = "PING" -class NameCommand(Command): +class NameCommand(_SimpleCommand): """Sent by client to inform the server of the client's identity. The data is the name """ @@ -179,76 +203,63 @@ class NameCommand(Command): class ReplicateCommand(Command): - """Sent by the client to subscribe to the stream. + """Sent by the client to subscribe to streams. Format:: - REPLICATE - - Where may be either: - * a numeric stream_id to stream updates from - * "NOW" to stream all subsequent updates. - - The can be "ALL" to subscribe to all known streams, in which - case the must be set to "NOW", i.e.:: - - REPLICATE ALL NOW + REPLICATE """ NAME = "REPLICATE" - def __init__(self, stream_name, token): - self.stream_name = stream_name - self.token = token + def __init__(self): + pass @classmethod def from_line(cls, line): - stream_name, token = line.split(" ", 1) - if token in ("NOW", "now"): - token = "NOW" - else: - token = int(token) - return cls(stream_name, token) + return cls() def to_line(self): - return " ".join((self.stream_name, str(self.token))) - - def get_logcontext_id(self): - return "REPLICATE-" + self.stream_name + return "" class UserSyncCommand(Command): """Sent by the client to inform the server that a user has started or - stopped syncing. Used to calculate presence on the master. + stopped syncing on this process. + + This is used by the process handling presence (typically the master) to + calculate who is online and who is not. Includes a timestamp of when the last user sync was. Format:: - USER_SYNC + USER_SYNC - Where is either "start" or "stop" + Where is either "start" or "end" """ NAME = "USER_SYNC" - def __init__(self, user_id, is_syncing, last_sync_ms): + def __init__(self, instance_id, user_id, is_syncing, last_sync_ms): + self.instance_id = instance_id self.user_id = user_id self.is_syncing = is_syncing self.last_sync_ms = last_sync_ms @classmethod def from_line(cls, line): - user_id, state, last_sync_ms = line.split(" ", 2) + instance_id, user_id, state, last_sync_ms = line.split(" ", 3) if state not in ("start", "end"): raise Exception("Invalid USER_SYNC state %r" % (state,)) - return cls(user_id, state == "start", int(last_sync_ms)) + return cls(instance_id, user_id, state == "start", int(last_sync_ms)) def to_line(self): return " ".join( ( + self.instance_id, self.user_id, "start" if self.is_syncing else "end", str(self.last_sync_ms), @@ -256,6 +267,30 @@ class UserSyncCommand(Command): ) +class ClearUserSyncsCommand(Command): + """Sent by the client to inform the server that it should drop all + information about syncing users sent by the client. + + Mainly used when client is about to shut down. + + Format:: + + CLEAR_USER_SYNC + """ + + NAME = "CLEAR_USER_SYNC" + + def __init__(self, instance_id): + self.instance_id = instance_id + + @classmethod + def from_line(cls, line): + return cls(line) + + def to_line(self): + return self.instance_id + + class FederationAckCommand(Command): """Sent by the client when it has processed up to a given point in the federation stream. This allows the master to drop in-memory caches of the @@ -281,14 +316,6 @@ class FederationAckCommand(Command): return str(self.token) -class SyncCommand(Command): - """Used for testing. The client protocol implementation allows waiting - on a SYNC command with a specified data. - """ - - NAME = "SYNC" - - class RemovePusherCommand(Command): """Sent by the client to request the master remove the given pusher. @@ -387,7 +414,7 @@ class UserIpCommand(Command): ) -class RemoteServerUpCommand(Command): +class RemoteServerUpCommand(_SimpleCommand): """Sent when a worker has detected that a remote server is no longer "down" and retry timings should be reset. @@ -411,11 +438,11 @@ _COMMANDS = ( ReplicateCommand, UserSyncCommand, FederationAckCommand, - SyncCommand, RemovePusherCommand, InvalidateCacheCommand, UserIpCommand, RemoteServerUpCommand, + ClearUserSyncsCommand, ) # type: Tuple[Type[Command], ...] # Map of command name to command type. @@ -428,7 +455,6 @@ VALID_SERVER_COMMANDS = ( PositionCommand.NAME, ErrorCommand.NAME, PingCommand.NAME, - SyncCommand.NAME, RemoteServerUpCommand.NAME, ) @@ -438,6 +464,7 @@ VALID_CLIENT_COMMANDS = ( ReplicateCommand.NAME, PingCommand.NAME, UserSyncCommand.NAME, + ClearUserSyncsCommand.NAME, FederationAckCommand.NAME, RemovePusherCommand.NAME, InvalidateCacheCommand.NAME, @@ -445,3 +472,21 @@ VALID_CLIENT_COMMANDS = ( ErrorCommand.NAME, RemoteServerUpCommand.NAME, ) + + +def parse_command_from_line(line: str) -> Command: + """Parses a command from a received line. + + Line should already be stripped of whitespace and be checked if blank. + """ + + idx = line.find(" ") + if idx >= 0: + cmd_name = line[:idx] + rest_of_line = line[idx + 1 :] + else: + cmd_name = line + rest_of_line = "" + + cmd_cls = COMMAND_MAP[cmd_name] + return cmd_cls.from_line(rest_of_line) diff --git a/synapse/replication/tcp/handler.py b/synapse/replication/tcp/handler.py new file mode 100644 index 000000000..4328b38e9 --- /dev/null +++ b/synapse/replication/tcp/handler.py @@ -0,0 +1,569 @@ +# -*- coding: utf-8 -*- +# Copyright 2017 Vector Creations Ltd +# Copyright 2020 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. + +import logging +from typing import ( + Any, + Callable, + Dict, + Iterable, + Iterator, + List, + Optional, + Set, + Tuple, + TypeVar, +) + +from prometheus_client import Counter + +from twisted.internet.protocol import ReconnectingClientFactory + +from synapse.metrics import LaterGauge +from synapse.replication.tcp.client import DirectTcpReplicationClientFactory +from synapse.replication.tcp.commands import ( + ClearUserSyncsCommand, + Command, + FederationAckCommand, + InvalidateCacheCommand, + PositionCommand, + RdataCommand, + RemoteServerUpCommand, + RemovePusherCommand, + ReplicateCommand, + UserIpCommand, + UserSyncCommand, +) +from synapse.replication.tcp.protocol import AbstractConnection +from synapse.replication.tcp.streams import STREAMS_MAP, Stream +from synapse.util.async_helpers import Linearizer + +logger = logging.getLogger(__name__) + + +# number of updates received for each RDATA stream +inbound_rdata_count = Counter( + "synapse_replication_tcp_protocol_inbound_rdata_count", "", ["stream_name"] +) +user_sync_counter = Counter("synapse_replication_tcp_resource_user_sync", "") +federation_ack_counter = Counter("synapse_replication_tcp_resource_federation_ack", "") +remove_pusher_counter = Counter("synapse_replication_tcp_resource_remove_pusher", "") +invalidate_cache_counter = Counter( + "synapse_replication_tcp_resource_invalidate_cache", "" +) +user_ip_cache_counter = Counter("synapse_replication_tcp_resource_user_ip_cache", "") + + +class ReplicationCommandHandler: + """Handles incoming commands from replication as well as sending commands + back out to connections. + """ + + def __init__(self, hs): + self._replication_data_handler = hs.get_replication_data_handler() + self._presence_handler = hs.get_presence_handler() + self._store = hs.get_datastore() + self._notifier = hs.get_notifier() + self._clock = hs.get_clock() + self._instance_id = hs.get_instance_id() + self._instance_name = hs.get_instance_name() + + self._streams = { + stream.NAME: stream(hs) for stream in STREAMS_MAP.values() + } # type: Dict[str, Stream] + + self._position_linearizer = Linearizer( + "replication_position", clock=self._clock + ) + + # Map of stream to batched updates. See RdataCommand for info on how + # batching works. + self._pending_batches = {} # type: Dict[str, List[Any]] + + # The factory used to create connections. + self._factory = None # type: Optional[ReconnectingClientFactory] + + # The currently connected connections. (The list of places we need to send + # outgoing replication commands to.) + self._connections = [] # type: List[AbstractConnection] + + # For each connection, the incoming streams that are coming from that connection + self._streams_by_connection = {} # type: Dict[AbstractConnection, Set[str]] + + LaterGauge( + "synapse_replication_tcp_resource_total_connections", + "", + [], + lambda: len(self._connections), + ) + + self._is_master = hs.config.worker_app is None + + self._federation_sender = None + if self._is_master and not hs.config.send_federation: + self._federation_sender = hs.get_federation_sender() + + self._server_notices_sender = None + if self._is_master: + self._server_notices_sender = hs.get_server_notices_sender() + + def start_replication(self, hs): + """Helper method to start a replication connection to the remote server + using TCP. + """ + if hs.config.redis.redis_enabled: + from synapse.replication.tcp.redis import ( + RedisDirectTcpReplicationClientFactory, + ) + import txredisapi + + logger.info( + "Connecting to redis (host=%r port=%r)", + hs.config.redis_host, + hs.config.redis_port, + ) + + # We need two connections to redis, one for the subscription stream and + # one to send commands to (as you can't send further redis commands to a + # connection after SUBSCRIBE is called). + + # First create the connection for sending commands. + outbound_redis_connection = txredisapi.lazyConnection( + host=hs.config.redis_host, + port=hs.config.redis_port, + password=hs.config.redis.redis_password, + reconnect=True, + ) + + # Now create the factory/connection for the subscription stream. + self._factory = RedisDirectTcpReplicationClientFactory( + hs, outbound_redis_connection + ) + hs.get_reactor().connectTCP( + hs.config.redis.redis_host, hs.config.redis.redis_port, self._factory, + ) + else: + client_name = hs.get_instance_name() + self._factory = DirectTcpReplicationClientFactory(hs, client_name, self) + host = hs.config.worker_replication_host + port = hs.config.worker_replication_port + hs.get_reactor().connectTCP(host, port, self._factory) + + async def on_REPLICATE(self, conn: AbstractConnection, cmd: ReplicateCommand): + # We only want to announce positions by the writer of the streams. + # Currently this is just the master process. + if not self._is_master: + return + + for stream_name, stream in self._streams.items(): + current_token = stream.current_token() + self.send_command( + PositionCommand(stream_name, self._instance_name, current_token) + ) + + async def on_USER_SYNC(self, conn: AbstractConnection, cmd: UserSyncCommand): + user_sync_counter.inc() + + if self._is_master: + await self._presence_handler.update_external_syncs_row( + cmd.instance_id, cmd.user_id, cmd.is_syncing, cmd.last_sync_ms + ) + + async def on_CLEAR_USER_SYNC( + self, conn: AbstractConnection, cmd: ClearUserSyncsCommand + ): + if self._is_master: + await self._presence_handler.update_external_syncs_clear(cmd.instance_id) + + async def on_FEDERATION_ACK( + self, conn: AbstractConnection, cmd: FederationAckCommand + ): + federation_ack_counter.inc() + + if self._federation_sender: + self._federation_sender.federation_ack(cmd.token) + + async def on_REMOVE_PUSHER( + self, conn: AbstractConnection, cmd: RemovePusherCommand + ): + remove_pusher_counter.inc() + + if self._is_master: + await self._store.delete_pusher_by_app_id_pushkey_user_id( + app_id=cmd.app_id, pushkey=cmd.push_key, user_id=cmd.user_id + ) + + self._notifier.on_new_replication_data() + + async def on_INVALIDATE_CACHE( + self, conn: AbstractConnection, cmd: InvalidateCacheCommand + ): + invalidate_cache_counter.inc() + + if self._is_master: + # We invalidate the cache locally, but then also stream that to other + # workers. + await self._store.invalidate_cache_and_stream( + cmd.cache_func, tuple(cmd.keys) + ) + + async def on_USER_IP(self, conn: AbstractConnection, cmd: UserIpCommand): + user_ip_cache_counter.inc() + + if self._is_master: + await self._store.insert_client_ip( + cmd.user_id, + cmd.access_token, + cmd.ip, + cmd.user_agent, + cmd.device_id, + cmd.last_seen, + ) + + if self._server_notices_sender: + await self._server_notices_sender.on_user_ip(cmd.user_id) + + async def on_RDATA(self, conn: AbstractConnection, cmd: RdataCommand): + if cmd.instance_name == self._instance_name: + # Ignore RDATA that are just our own echoes + return + + stream_name = cmd.stream_name + inbound_rdata_count.labels(stream_name).inc() + + try: + row = STREAMS_MAP[stream_name].parse_row(cmd.row) + except Exception: + logger.exception("Failed to parse RDATA: %r %r", stream_name, cmd.row) + raise + + # We linearize here for two reasons: + # 1. so we don't try and concurrently handle multiple rows for the + # same stream, and + # 2. so we don't race with getting a POSITION command and fetching + # missing RDATA. + with await self._position_linearizer.queue(cmd.stream_name): + # make sure that we've processed a POSITION for this stream *on this + # connection*. (A POSITION on another connection is no good, as there + # is no guarantee that we have seen all the intermediate updates.) + sbc = self._streams_by_connection.get(conn) + if not sbc or stream_name not in sbc: + # Let's drop the row for now, on the assumption we'll receive a + # `POSITION` soon and we'll catch up correctly then. + logger.debug( + "Discarding RDATA for unconnected stream %s -> %s", + stream_name, + cmd.token, + ) + return + + if cmd.token is None: + # I.e. this is part of a batch of updates for this stream (in + # which case batch until we get an update for the stream with a non + # None token). + self._pending_batches.setdefault(stream_name, []).append(row) + else: + # Check if this is the last of a batch of updates + rows = self._pending_batches.pop(stream_name, []) + rows.append(row) + await self.on_rdata(stream_name, cmd.instance_name, cmd.token, rows) + + async def on_rdata( + self, stream_name: str, instance_name: str, token: int, rows: list + ): + """Called to handle a batch of replication data with a given stream token. + + Args: + stream_name: name of the replication stream for this batch of rows + instance_name: the instance that wrote the rows. + token: stream token for this batch of rows + rows: a list of Stream.ROW_TYPE objects as returned by + Stream.parse_row. + """ + logger.debug("Received rdata %s -> %s", stream_name, token) + await self._replication_data_handler.on_rdata( + stream_name, instance_name, token, rows + ) + + async def on_POSITION(self, conn: AbstractConnection, cmd: PositionCommand): + if cmd.instance_name == self._instance_name: + # Ignore POSITION that are just our own echoes + return + + logger.info("Handling '%s %s'", cmd.NAME, cmd.to_line()) + + stream_name = cmd.stream_name + stream = self._streams.get(stream_name) + if not stream: + logger.error("Got POSITION for unknown stream: %s", stream_name) + return + + # We protect catching up with a linearizer in case the replication + # connection reconnects under us. + with await self._position_linearizer.queue(stream_name): + # We're about to go and catch up with the stream, so remove from set + # of connected streams. + for streams in self._streams_by_connection.values(): + streams.discard(stream_name) + + # We clear the pending batches for the stream as the fetching of the + # missing updates below will fetch all rows in the batch. + self._pending_batches.pop(stream_name, []) + + # Find where we previously streamed up to. + current_token = stream.current_token() + + # If the position token matches our current token then we're up to + # date and there's nothing to do. Otherwise, fetch all updates + # between then and now. + missing_updates = cmd.token != current_token + while missing_updates: + logger.info( + "Fetching replication rows for '%s' between %i and %i", + stream_name, + current_token, + cmd.token, + ) + ( + updates, + current_token, + missing_updates, + ) = await stream.get_updates_since( + cmd.instance_name, current_token, cmd.token + ) + + # TODO: add some tests for this + + # Some streams return multiple rows with the same stream IDs, + # which need to be processed in batches. + + for token, rows in _batch_updates(updates): + await self.on_rdata( + stream_name, + cmd.instance_name, + token, + [stream.parse_row(row) for row in rows], + ) + + logger.info("Caught up with stream '%s' to %i", stream_name, cmd.token) + + # We've now caught up to position sent to us, notify handler. + await self._replication_data_handler.on_position(stream_name, cmd.token) + + self._streams_by_connection.setdefault(conn, set()).add(stream_name) + + async def on_REMOTE_SERVER_UP( + self, conn: AbstractConnection, cmd: RemoteServerUpCommand + ): + """"Called when get a new REMOTE_SERVER_UP command.""" + self._replication_data_handler.on_remote_server_up(cmd.data) + + self._notifier.notify_remote_server_up(cmd.data) + + # We relay to all other connections to ensure every instance gets the + # notification. + # + # When configured to use redis we'll always only have one connection and + # so this is a no-op (all instances will have already received the same + # REMOTE_SERVER_UP command). + # + # For direct TCP connections this will relay to all other connections + # connected to us. When on master this will correctly fan out to all + # other direct TCP clients and on workers there'll only be the one + # connection to master. + # + # (The logic here should also be sound if we have a mix of Redis and + # direct TCP connections so long as there is only one traffic route + # between two instances, but that is not currently supported). + self.send_command(cmd, ignore_conn=conn) + + def new_connection(self, connection: AbstractConnection): + """Called when we have a new connection. + """ + self._connections.append(connection) + + # If we are connected to replication as a client (rather than a server) + # we need to reset the reconnection delay on the client factory (which + # is used to do exponential back off when the connection drops). + # + # Ideally we would reset the delay when we've "fully established" the + # connection (for some definition thereof) to stop us from tightlooping + # on reconnection if something fails after this point and we drop the + # connection. Unfortunately, we don't really have a better definition of + # "fully established" than the connection being established. + if self._factory: + self._factory.resetDelay() + + # Tell the other end if we have any users currently syncing. + currently_syncing = ( + self._presence_handler.get_currently_syncing_users_for_replication() + ) + + now = self._clock.time_msec() + for user_id in currently_syncing: + connection.send_command( + UserSyncCommand(self._instance_id, user_id, True, now) + ) + + def lost_connection(self, connection: AbstractConnection): + """Called when a connection is closed/lost. + """ + # we no longer need _streams_by_connection for this connection. + streams = self._streams_by_connection.pop(connection, None) + if streams: + logger.info( + "Lost replication connection; streams now disconnected: %s", streams + ) + try: + self._connections.remove(connection) + except ValueError: + pass + + def connected(self) -> bool: + """Do we have any replication connections open? + + Is used by e.g. `ReplicationStreamer` to no-op if nothing is connected. + """ + return bool(self._connections) + + def send_command( + self, cmd: Command, ignore_conn: Optional[AbstractConnection] = None + ): + """Send a command to all connected connections. + + Args: + cmd + ignore_conn: If set don't send command to the given connection. + Used when relaying commands from one connection to all others. + """ + if self._connections: + for connection in self._connections: + if connection == ignore_conn: + continue + + try: + connection.send_command(cmd) + except Exception: + # We probably want to catch some types of exceptions here + # and log them as warnings (e.g. connection gone), but I + # can't find what those exception types they would be. + logger.exception( + "Failed to write command %s to connection %s", + cmd.NAME, + connection, + ) + else: + logger.warning("Dropping command as not connected: %r", cmd.NAME) + + def send_federation_ack(self, token: int): + """Ack data for the federation stream. This allows the master to drop + data stored purely in memory. + """ + self.send_command(FederationAckCommand(token)) + + def send_user_sync( + self, instance_id: str, user_id: str, is_syncing: bool, last_sync_ms: int + ): + """Poke the master that a user has started/stopped syncing. + """ + self.send_command( + UserSyncCommand(instance_id, user_id, is_syncing, last_sync_ms) + ) + + def send_remove_pusher(self, app_id: str, push_key: str, user_id: str): + """Poke the master to remove a pusher for a user + """ + cmd = RemovePusherCommand(app_id, push_key, user_id) + self.send_command(cmd) + + def send_invalidate_cache(self, cache_func: Callable, keys: tuple): + """Poke the master to invalidate a cache. + """ + cmd = InvalidateCacheCommand(cache_func.__name__, keys) + self.send_command(cmd) + + def send_user_ip( + self, + user_id: str, + access_token: str, + ip: str, + user_agent: str, + device_id: str, + last_seen: int, + ): + """Tell the master that the user made a request. + """ + cmd = UserIpCommand(user_id, access_token, ip, user_agent, device_id, last_seen) + self.send_command(cmd) + + def send_remote_server_up(self, server: str): + self.send_command(RemoteServerUpCommand(server)) + + def stream_update(self, stream_name: str, token: str, data: Any): + """Called when a new update is available to stream to clients. + + We need to check if the client is interested in the stream or not + """ + self.send_command(RdataCommand(stream_name, self._instance_name, token, data)) + + +UpdateToken = TypeVar("UpdateToken") +UpdateRow = TypeVar("UpdateRow") + + +def _batch_updates( + updates: Iterable[Tuple[UpdateToken, UpdateRow]] +) -> Iterator[Tuple[UpdateToken, List[UpdateRow]]]: + """Collect stream updates with the same token together + + Given a series of updates returned by Stream.get_updates_since(), collects + the updates which share the same stream_id together. + + For example: + + [(1, a), (1, b), (2, c), (3, d), (3, e)] + + becomes: + + [ + (1, [a, b]), + (2, [c]), + (3, [d, e]), + ] + """ + + update_iter = iter(updates) + + first_update = next(update_iter, None) + if first_update is None: + # empty input + return + + current_batch_token = first_update[0] + current_batch = [first_update[1]] + + for token, row in update_iter: + if token != current_batch_token: + # different token to the previous row: flush the previous + # batch and start anew + yield current_batch_token, current_batch + current_batch_token = token + current_batch = [] + + current_batch.append(row) + + # flush the final batch + yield current_batch_token, current_batch diff --git a/synapse/replication/tcp/protocol.py b/synapse/replication/tcp/protocol.py index bc1482a9b..4198eece7 100644 --- a/synapse/replication/tcp/protocol.py +++ b/synapse/replication/tcp/protocol.py @@ -35,9 +35,7 @@ indicate which side is sending, these are *not* included on the wire:: > PING 1490197665618 < NAME synapse.app.appservice < PING 1490197665618 - < REPLICATE events 1 - < REPLICATE backfill 1 - < REPLICATE caches 1 + < REPLICATE > POSITION events 1 > POSITION backfill 1 > POSITION caches 1 @@ -52,45 +50,51 @@ import abc import fcntl import logging import struct -from collections import defaultdict -from typing import Any, DefaultDict, Dict, List, Set, Tuple - -from six import iteritems, iterkeys +from typing import TYPE_CHECKING, List from prometheus_client import Counter -from twisted.internet import defer from twisted.protocols.basic import LineOnlyReceiver from twisted.python.failure import Failure -from synapse.logging.context import make_deferred_yieldable, run_in_background from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process from synapse.replication.tcp.commands import ( - COMMAND_MAP, VALID_CLIENT_COMMANDS, VALID_SERVER_COMMANDS, Command, ErrorCommand, NameCommand, PingCommand, - PositionCommand, - RdataCommand, - RemoteServerUpCommand, ReplicateCommand, ServerCommand, - SyncCommand, - UserSyncCommand, + parse_command_from_line, ) -from synapse.replication.tcp.streams import STREAMS_MAP from synapse.types import Collection from synapse.util import Clock from synapse.util.stringutils import random_string +if TYPE_CHECKING: + from synapse.replication.tcp.handler import ReplicationCommandHandler + from synapse.server import HomeServer + + connection_close_counter = Counter( "synapse_replication_tcp_protocol_close_reason", "", ["reason_type"] ) +tcp_inbound_commands_counter = Counter( + "synapse_replication_tcp_protocol_inbound_commands", + "Number of commands received from replication, by command and name of process connected to", + ["command", "name"], +) + +tcp_outbound_commands_counter = Counter( + "synapse_replication_tcp_protocol_outbound_commands", + "Number of commands sent to replication, by command and name of process connected to", + ["command", "name"], +) + # A list of all connected protocols. This allows us to send metrics about the # connections. connected_connections = [] @@ -119,7 +123,7 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): are only sent by the server. On receiving a new command it calls `on_` with the parsed - command. + command before delegating to `ReplicationCommandHandler.on_`. It also sends `PING` periodically, and correctly times out remote connections (if they send a `PING` command) @@ -135,8 +139,9 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): max_line_buffer = 10000 - def __init__(self, clock): + def __init__(self, clock: Clock, handler: "ReplicationCommandHandler"): self.clock = clock + self.command_handler = handler self.last_received_command = self.clock.time_msec() self.last_sent_command = 0 @@ -155,9 +160,6 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): # The LoopingCall for sending pings. self._send_ping_loop = None - self.inbound_commands_counter = defaultdict(int) # type: DefaultDict[str, int] - self.outbound_commands_counter = defaultdict(int) # type: DefaultDict[str, int] - def connectionMade(self): logger.info("[%s] Connection established", self.id()) @@ -176,6 +178,8 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): # can time us out. self.send_command(PingCommand(self.clock.time_msec())) + self.command_handler.new_connection(self) + def send_ping(self): """Periodically sends a ping and checks if we should close the connection due to the other side timing out. @@ -203,38 +207,30 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): ) self.send_error("ping timeout") - def lineReceived(self, line): + def lineReceived(self, line: bytes): """Called when we've received a line """ if line.strip() == "": # Ignore blank lines return - line = line.decode("utf-8") - cmd_name, rest_of_line = line.split(" ", 1) + linestr = line.decode("utf-8") - if cmd_name not in self.VALID_INBOUND_COMMANDS: - logger.error("[%s] invalid command %s", self.id(), cmd_name) - self.send_error("invalid command: %s", cmd_name) + try: + cmd = parse_command_from_line(linestr) + except Exception as e: + logger.exception("[%s] failed to parse line: %r", self.id(), linestr) + self.send_error("failed to parse line: %r (%r):" % (e, linestr)) + return + + if cmd.NAME not in self.VALID_INBOUND_COMMANDS: + logger.error("[%s] invalid command %s", self.id(), cmd.NAME) + self.send_error("invalid command: %s", cmd.NAME) return self.last_received_command = self.clock.time_msec() - self.inbound_commands_counter[cmd_name] = ( - self.inbound_commands_counter[cmd_name] + 1 - ) - - cmd_cls = COMMAND_MAP[cmd_name] - try: - cmd = cmd_cls.from_line(rest_of_line) - except Exception as e: - logger.exception( - "[%s] failed to parse line %r: %r", self.id(), cmd_name, rest_of_line - ) - self.send_error( - "failed to parse line for %r: %r (%r):" % (cmd_name, e, rest_of_line) - ) - return + tcp_inbound_commands_counter.labels(cmd.NAME, self.name).inc() # Now lets try and call on_ function run_as_background_process( @@ -244,13 +240,31 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): async def handle_command(self, cmd: Command): """Handle a command we have received over the replication stream. - By default delegates to on_, which should return an awaitable. + First calls `self.on_` if it exists, then calls + `self.command_handler.on_` if it exists. This allows for + protocol level handling of commands (e.g. PINGs), before delegating to + the handler. Args: cmd: received command """ - handler = getattr(self, "on_%s" % (cmd.NAME,)) - await handler(cmd) + handled = False + + # First call any command handlers on this instance. These are for TCP + # specific handling. + cmd_func = getattr(self, "on_%s" % (cmd.NAME,), None) + if cmd_func: + await cmd_func(cmd) + handled = True + + # Then call out to the handler. + cmd_func = getattr(self.command_handler, "on_%s" % (cmd.NAME,), None) + if cmd_func: + await cmd_func(self, cmd) + handled = True + + if not handled: + logger.warning("Unhandled command: %r", cmd) def close(self): logger.warning("[%s] Closing connection", self.id()) @@ -282,9 +296,8 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): self._queue_command(cmd) return - self.outbound_commands_counter[cmd.NAME] = ( - self.outbound_commands_counter[cmd.NAME] + 1 - ) + tcp_outbound_commands_counter.labels(cmd.NAME, self.name).inc() + string = "%s %s" % (cmd.NAME, cmd.to_line()) if "\n" in string: raise Exception("Unexpected newline in command: %r", string) @@ -379,6 +392,8 @@ class BaseReplicationStreamProtocol(LineOnlyReceiver): self.state = ConnectionStates.CLOSED self.pending_commands = [] + self.command_handler.lost_connection(self) + if self.transport: self.transport.unregisterProducer() @@ -405,232 +420,21 @@ class ServerReplicationStreamProtocol(BaseReplicationStreamProtocol): VALID_INBOUND_COMMANDS = VALID_CLIENT_COMMANDS VALID_OUTBOUND_COMMANDS = VALID_SERVER_COMMANDS - def __init__(self, server_name, clock, streamer): - BaseReplicationStreamProtocol.__init__(self, clock) # Old style class + def __init__( + self, server_name: str, clock: Clock, handler: "ReplicationCommandHandler" + ): + super().__init__(clock, handler) self.server_name = server_name - self.streamer = streamer - - # The streams the client has subscribed to and is up to date with - self.replication_streams = set() # type: Set[str] - - # The streams the client is currently subscribing to. - self.connecting_streams = set() # type: Set[str] - - # Map from stream name to list of updates to send once we've finished - # subscribing the client to the stream. - self.pending_rdata = {} # type: Dict[str, List[Tuple[int, Any]]] def connectionMade(self): self.send_command(ServerCommand(self.server_name)) - BaseReplicationStreamProtocol.connectionMade(self) - self.streamer.new_connection(self) + super().connectionMade() async def on_NAME(self, cmd): logger.info("[%s] Renamed to %r", self.id(), cmd.data) self.name = cmd.data - async def on_USER_SYNC(self, cmd): - await self.streamer.on_user_sync( - self.conn_id, cmd.user_id, cmd.is_syncing, cmd.last_sync_ms - ) - - async def on_REPLICATE(self, cmd): - stream_name = cmd.stream_name - token = cmd.token - - if stream_name == "ALL": - # Subscribe to all streams we're publishing to. - deferreds = [ - run_in_background(self.subscribe_to_stream, stream, token) - for stream in iterkeys(self.streamer.streams_by_name) - ] - - await make_deferred_yieldable( - defer.gatherResults(deferreds, consumeErrors=True) - ) - else: - await self.subscribe_to_stream(stream_name, token) - - async def on_FEDERATION_ACK(self, cmd): - self.streamer.federation_ack(cmd.token) - - async def on_REMOVE_PUSHER(self, cmd): - await self.streamer.on_remove_pusher(cmd.app_id, cmd.push_key, cmd.user_id) - - async def on_INVALIDATE_CACHE(self, cmd): - await self.streamer.on_invalidate_cache(cmd.cache_func, cmd.keys) - - async def on_REMOTE_SERVER_UP(self, cmd: RemoteServerUpCommand): - self.streamer.on_remote_server_up(cmd.data) - - async def on_USER_IP(self, cmd): - self.streamer.on_user_ip( - cmd.user_id, - cmd.access_token, - cmd.ip, - cmd.user_agent, - cmd.device_id, - cmd.last_seen, - ) - - async def subscribe_to_stream(self, stream_name, token): - """Subscribe the remote to a stream. - - This invloves checking if they've missed anything and sending those - updates down if they have. During that time new updates for the stream - are queued and sent once we've sent down any missed updates. - """ - self.replication_streams.discard(stream_name) - self.connecting_streams.add(stream_name) - - try: - # Get missing updates - updates, current_token = await self.streamer.get_stream_updates( - stream_name, token - ) - - # Send all the missing updates - for update in updates: - token, row = update[0], update[1] - self.send_command(RdataCommand(stream_name, token, row)) - - # We send a POSITION command to ensure that they have an up to - # date token (especially useful if we didn't send any updates - # above) - self.send_command(PositionCommand(stream_name, current_token)) - - # Now we can send any updates that came in while we were subscribing - pending_rdata = self.pending_rdata.pop(stream_name, []) - updates = [] - for token, update in pending_rdata: - # If the token is null, it is part of a batch update. Batches - # are multiple updates that share a single token. To denote - # this, the token is set to None for all tokens in the batch - # except for the last. If we find a None token, we keep looking - # through tokens until we find one that is not None and then - # process all previous updates in the batch as if they had the - # final token. - if token is None: - # Store this update as part of a batch - updates.append(update) - continue - - if token <= current_token: - # This update or batch of updates is older than - # current_token, dismiss it - updates = [] - continue - - updates.append(update) - - # Send all updates that are part of this batch with the - # found token - for update in updates: - self.send_command(RdataCommand(stream_name, token, update)) - - # Clear stored updates - updates = [] - - # They're now fully subscribed - self.replication_streams.add(stream_name) - except Exception as e: - logger.exception("[%s] Failed to handle REPLICATE command", self.id()) - self.send_error("failed to handle replicate: %r", e) - finally: - self.connecting_streams.discard(stream_name) - - def stream_update(self, stream_name, token, data): - """Called when a new update is available to stream to clients. - - We need to check if the client is interested in the stream or not - """ - if stream_name in self.replication_streams: - # The client is subscribed to the stream - self.send_command(RdataCommand(stream_name, token, data)) - elif stream_name in self.connecting_streams: - # The client is being subscribed to the stream - logger.debug("[%s] Queuing RDATA %r %r", self.id(), stream_name, token) - self.pending_rdata.setdefault(stream_name, []).append((token, data)) - else: - # The client isn't subscribed - logger.debug("[%s] Dropping RDATA %r %r", self.id(), stream_name, token) - - def send_sync(self, data): - self.send_command(SyncCommand(data)) - - def send_remote_server_up(self, server: str): - self.send_command(RemoteServerUpCommand(server)) - - def on_connection_closed(self): - BaseReplicationStreamProtocol.on_connection_closed(self) - self.streamer.lost_connection(self) - - -class AbstractReplicationClientHandler(metaclass=abc.ABCMeta): - """ - The interface for the handler that should be passed to - ClientReplicationStreamProtocol - """ - - @abc.abstractmethod - async def on_rdata(self, stream_name, token, rows): - """Called to handle a batch of replication data with a given stream token. - - Args: - stream_name (str): name of the replication stream for this batch of rows - token (int): stream token for this batch of rows - rows (list): a list of Stream.ROW_TYPE objects as returned by - Stream.parse_row. - """ - raise NotImplementedError() - - @abc.abstractmethod - async def on_position(self, stream_name, token): - """Called when we get new position data.""" - raise NotImplementedError() - - @abc.abstractmethod - def on_sync(self, data): - """Called when get a new SYNC command.""" - raise NotImplementedError() - - @abc.abstractmethod - async def on_remote_server_up(self, server: str): - """Called when get a new REMOTE_SERVER_UP command.""" - raise NotImplementedError() - - @abc.abstractmethod - def get_streams_to_replicate(self): - """Called when a new connection has been established and we need to - subscribe to streams. - - Returns: - map from stream name to the most recent update we have for - that stream (ie, the point we want to start replicating from) - """ - raise NotImplementedError() - - @abc.abstractmethod - def get_currently_syncing_users(self): - """Get the list of currently syncing users (if any). This is called - when a connection has been established and we need to send the - currently syncing users.""" - raise NotImplementedError() - - @abc.abstractmethod - def update_connection(self, connection): - """Called when a connection has been established (or lost with None). - """ - raise NotImplementedError() - - @abc.abstractmethod - def finished_connecting(self): - """Called when we have successfully subscribed and caught up to all - streams we're interested in. - """ - raise NotImplementedError() - class ClientReplicationStreamProtocol(BaseReplicationStreamProtocol): VALID_INBOUND_COMMANDS = VALID_SERVER_COMMANDS @@ -638,110 +442,51 @@ class ClientReplicationStreamProtocol(BaseReplicationStreamProtocol): def __init__( self, + hs: "HomeServer", client_name: str, server_name: str, clock: Clock, - handler: AbstractReplicationClientHandler, + command_handler: "ReplicationCommandHandler", ): - BaseReplicationStreamProtocol.__init__(self, clock) + super().__init__(clock, command_handler) self.client_name = client_name self.server_name = server_name - self.handler = handler - - # Set of stream names that have been subscribe to, but haven't yet - # caught up with. This is used to track when the client has been fully - # connected to the remote. - self.streams_connecting = set() # type: Set[str] - - # Map of stream to batched updates. See RdataCommand for info on how - # batching works. - self.pending_batches = {} # type: Dict[str, Any] def connectionMade(self): self.send_command(NameCommand(self.client_name)) - BaseReplicationStreamProtocol.connectionMade(self) + super().connectionMade() # Once we've connected subscribe to the necessary streams - for stream_name, token in iteritems(self.handler.get_streams_to_replicate()): - self.replicate(stream_name, token) - - # Tell the server if we have any users currently syncing (should only - # happen on synchrotrons) - currently_syncing = self.handler.get_currently_syncing_users() - now = self.clock.time_msec() - for user_id in currently_syncing: - self.send_command(UserSyncCommand(user_id, True, now)) - - # We've now finished connecting to so inform the client handler - self.handler.update_connection(self) - - # This will happen if we don't actually subscribe to any streams - if not self.streams_connecting: - self.handler.finished_connecting() + self.replicate() async def on_SERVER(self, cmd): if cmd.data != self.server_name: logger.error("[%s] Connected to wrong remote: %r", self.id(), cmd.data) self.send_error("Wrong remote") - async def on_RDATA(self, cmd): - stream_name = cmd.stream_name - inbound_rdata_count.labels(stream_name).inc() - - try: - row = STREAMS_MAP[stream_name].parse_row(cmd.row) - except Exception: - logger.exception( - "[%s] Failed to parse RDATA: %r %r", self.id(), stream_name, cmd.row - ) - raise - - if cmd.token is None: - # I.e. this is part of a batch of updates for this stream. Batch - # until we get an update for the stream with a non None token - self.pending_batches.setdefault(stream_name, []).append(row) - else: - # Check if this is the last of a batch of updates - rows = self.pending_batches.pop(stream_name, []) - rows.append(row) - await self.handler.on_rdata(stream_name, cmd.token, rows) - - async def on_POSITION(self, cmd): - # When we get a `POSITION` command it means we've finished getting - # missing updates for the given stream, and are now up to date. - self.streams_connecting.discard(cmd.stream_name) - if not self.streams_connecting: - self.handler.finished_connecting() - - await self.handler.on_position(cmd.stream_name, cmd.token) - - async def on_SYNC(self, cmd): - self.handler.on_sync(cmd.data) - - async def on_REMOTE_SERVER_UP(self, cmd: RemoteServerUpCommand): - self.handler.on_remote_server_up(cmd.data) - - def replicate(self, stream_name, token): + def replicate(self): """Send the subscription request to the server """ - if stream_name not in STREAMS_MAP: - raise Exception("Invalid stream name %r" % (stream_name,)) + logger.info("[%s] Subscribing to replication streams", self.id()) - logger.info( - "[%s] Subscribing to replication stream: %r from %r", - self.id(), - stream_name, - token, - ) + self.send_command(ReplicateCommand()) - self.streams_connecting.add(stream_name) - self.send_command(ReplicateCommand(stream_name, token)) +class AbstractConnection(abc.ABC): + """An interface for replication connections. + """ - def on_connection_closed(self): - BaseReplicationStreamProtocol.on_connection_closed(self) - self.handler.update_connection(None) + @abc.abstractmethod + def send_command(self, cmd: Command): + """Send the command down the connection + """ + pass + + +# This tells python that `BaseReplicationStreamProtocol` implements the +# interface. +AbstractConnection.register(BaseReplicationStreamProtocol) # The following simply registers metrics for the replication connections @@ -804,31 +549,3 @@ tcp_transport_kernel_read_buffer = LaterGauge( for p in connected_connections }, ) - - -tcp_inbound_commands = LaterGauge( - "synapse_replication_tcp_protocol_inbound_commands", - "", - ["command", "name"], - lambda: { - (k, p.name): count - for p in connected_connections - for k, count in iteritems(p.inbound_commands_counter) - }, -) - -tcp_outbound_commands = LaterGauge( - "synapse_replication_tcp_protocol_outbound_commands", - "", - ["command", "name"], - lambda: { - (k, p.name): count - for p in connected_connections - for k, count in iteritems(p.outbound_commands_counter) - }, -) - -# number of updates received for each RDATA stream -inbound_rdata_count = Counter( - "synapse_replication_tcp_protocol_inbound_rdata_count", "", ["stream_name"] -) diff --git a/synapse/replication/tcp/redis.py b/synapse/replication/tcp/redis.py new file mode 100644 index 000000000..55bfa71df --- /dev/null +++ b/synapse/replication/tcp/redis.py @@ -0,0 +1,210 @@ +# -*- coding: utf-8 -*- +# Copyright 2020 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. + +import logging +from typing import TYPE_CHECKING + +import txredisapi + +from synapse.logging.context import make_deferred_yieldable +from synapse.metrics.background_process_metrics import run_as_background_process +from synapse.replication.tcp.commands import ( + Command, + ReplicateCommand, + parse_command_from_line, +) +from synapse.replication.tcp.protocol import ( + AbstractConnection, + tcp_inbound_commands_counter, + tcp_outbound_commands_counter, +) + +if TYPE_CHECKING: + from synapse.replication.tcp.handler import ReplicationCommandHandler + from synapse.server import HomeServer + +logger = logging.getLogger(__name__) + + +class RedisSubscriber(txredisapi.SubscriberProtocol, AbstractConnection): + """Connection to redis subscribed to replication stream. + + This class fulfils two functions: + + (a) it implements the twisted Protocol API, where it handles the SUBSCRIBEd redis + connection, parsing *incoming* messages into replication commands, and passing them + to `ReplicationCommandHandler` + + (b) it implements the AbstractConnection API, where it sends *outgoing* commands + onto outbound_redis_connection. + + Due to the vagaries of `txredisapi` we don't want to have a custom + constructor, so instead we expect the defined attributes below to be set + immediately after initialisation. + + Attributes: + handler: The command handler to handle incoming commands. + stream_name: The *redis* stream name to subscribe to and publish from + (not anything to do with Synapse replication streams). + outbound_redis_connection: The connection to redis to use to send + commands. + """ + + handler = None # type: ReplicationCommandHandler + stream_name = None # type: str + outbound_redis_connection = None # type: txredisapi.RedisProtocol + + def connectionMade(self): + logger.info("Connected to redis") + super().connectionMade() + run_as_background_process("subscribe-replication", self._send_subscribe) + self.handler.new_connection(self) + + async def _send_subscribe(self): + # it's important to make sure that we only send the REPLICATE command once we + # have successfully subscribed to the stream - otherwise we might miss the + # POSITION response sent back by the other end. + logger.info("Sending redis SUBSCRIBE for %s", self.stream_name) + await make_deferred_yieldable(self.subscribe(self.stream_name)) + logger.info( + "Successfully subscribed to redis stream, sending REPLICATE command" + ) + await self._async_send_command(ReplicateCommand()) + logger.info("REPLICATE successfully sent") + + def messageReceived(self, pattern: str, channel: str, message: str): + """Received a message from redis. + """ + + if message.strip() == "": + # Ignore blank lines + return + + try: + cmd = parse_command_from_line(message) + except Exception: + logger.exception( + "Failed to parse replication line: %r", message, + ) + return + + # We use "redis" as the name here as we don't have 1:1 connections to + # remote instances. + tcp_inbound_commands_counter.labels(cmd.NAME, "redis").inc() + + # Now lets try and call on_ function + run_as_background_process( + "replication-" + cmd.get_logcontext_id(), self.handle_command, cmd + ) + + async def handle_command(self, cmd: Command): + """Handle a command we have received over the replication stream. + + By default delegates to on_, which should return an awaitable. + + Args: + cmd: received command + """ + handled = False + + # First call any command handlers on this instance. These are for redis + # specific handling. + cmd_func = getattr(self, "on_%s" % (cmd.NAME,), None) + if cmd_func: + await cmd_func(cmd) + handled = True + + # Then call out to the handler. + cmd_func = getattr(self.handler, "on_%s" % (cmd.NAME,), None) + if cmd_func: + await cmd_func(self, cmd) + handled = True + + if not handled: + logger.warning("Unhandled command: %r", cmd) + + def connectionLost(self, reason): + logger.info("Lost connection to redis") + super().connectionLost(reason) + self.handler.lost_connection(self) + + def send_command(self, cmd: Command): + """Send a command if connection has been established. + + Args: + cmd (Command) + """ + run_as_background_process("send-cmd", self._async_send_command, cmd) + + async def _async_send_command(self, cmd: Command): + """Encode a replication command and send it over our outbound connection""" + string = "%s %s" % (cmd.NAME, cmd.to_line()) + if "\n" in string: + raise Exception("Unexpected newline in command: %r", string) + + encoded_string = string.encode("utf-8") + + # We use "redis" as the name here as we don't have 1:1 connections to + # remote instances. + tcp_outbound_commands_counter.labels(cmd.NAME, "redis").inc() + + await make_deferred_yieldable( + self.outbound_redis_connection.publish(self.stream_name, encoded_string) + ) + + +class RedisDirectTcpReplicationClientFactory(txredisapi.SubscriberFactory): + """This is a reconnecting factory that connects to redis and immediately + subscribes to a stream. + + Args: + hs + outbound_redis_connection: A connection to redis that will be used to + send outbound commands (this is seperate to the redis connection + used to subscribe). + """ + + maxDelay = 5 + continueTrying = True + protocol = RedisSubscriber + + def __init__( + self, hs: "HomeServer", outbound_redis_connection: txredisapi.RedisProtocol + ): + + super().__init__() + + # This sets the password on the RedisFactory base class (as + # SubscriberFactory constructor doesn't pass it through). + self.password = hs.config.redis.redis_password + + self.handler = hs.get_tcp_replication() + self.stream_name = hs.hostname + + self.outbound_redis_connection = outbound_redis_connection + + def buildProtocol(self, addr): + p = super().buildProtocol(addr) # type: RedisSubscriber + + # We do this here rather than add to the constructor of `RedisSubcriber` + # as to do so would involve overriding `buildProtocol` entirely, however + # the base method does some other things than just instantiating the + # protocol. + p.handler = self.handler + p.outbound_redis_connection = self.outbound_redis_connection + p.stream_name = self.stream_name + p.password = self.password + + return p diff --git a/synapse/replication/tcp/resource.py b/synapse/replication/tcp/resource.py index ce9d1fae1..33d2f589a 100644 --- a/synapse/replication/tcp/resource.py +++ b/synapse/replication/tcp/resource.py @@ -17,32 +17,20 @@ import logging import random -from typing import Any, List - -from six import itervalues +from typing import Dict, List from prometheus_client import Counter from twisted.internet.protocol import Factory -from synapse.metrics import LaterGauge from synapse.metrics.background_process_metrics import run_as_background_process -from synapse.util.metrics import Measure, measure_func - -from .protocol import ServerReplicationStreamProtocol -from .streams import STREAMS_MAP -from .streams.federation import FederationStream +from synapse.replication.tcp.protocol import ServerReplicationStreamProtocol +from synapse.replication.tcp.streams import STREAMS_MAP, FederationStream, Stream +from synapse.util.metrics import Measure stream_updates_counter = Counter( "synapse_replication_tcp_resource_stream_updates", "", ["stream_name"] ) -user_sync_counter = Counter("synapse_replication_tcp_resource_user_sync", "") -federation_ack_counter = Counter("synapse_replication_tcp_resource_federation_ack", "") -remove_pusher_counter = Counter("synapse_replication_tcp_resource_remove_pusher", "") -invalidate_cache_counter = Counter( - "synapse_replication_tcp_resource_invalidate_cache", "" -) -user_ip_cache_counter = Counter("synapse_replication_tcp_resource_user_ip_cache", "") logger = logging.getLogger(__name__) @@ -52,13 +40,23 @@ class ReplicationStreamProtocolFactory(Factory): """ def __init__(self, hs): - self.streamer = ReplicationStreamer(hs) + self.command_handler = hs.get_tcp_replication() self.clock = hs.get_clock() self.server_name = hs.config.server_name + # If we've created a `ReplicationStreamProtocolFactory` then we're + # almost certainly registering a replication listener, so let's ensure + # that we've started a `ReplicationStreamer` instance to actually push + # data. + # + # (This is a bit of a weird place to do this, but the alternatives such + # as putting this in `HomeServer.setup()`, requires either passing the + # listener config again or always starting a `ReplicationStreamer`.) + hs.get_replication_streamer() + def buildProtocol(self, addr): return ServerReplicationStreamProtocol( - self.server_name, self.clock, self.streamer + self.server_name, self.clock, self.command_handler ) @@ -71,67 +69,39 @@ class ReplicationStreamer(object): def __init__(self, hs): self.store = hs.get_datastore() - self.presence_handler = hs.get_presence_handler() self.clock = hs.get_clock() self.notifier = hs.get_notifier() - self._server_notices_sender = hs.get_server_notices_sender() self._replication_torture_level = hs.config.replication_torture_level - # Current connections. - self.connections = [] # type: List[ServerReplicationStreamProtocol] + # Work out list of streams that this instance is the source of. + self.streams = [] # type: List[Stream] + if hs.config.worker_app is None: + for stream in STREAMS_MAP.values(): + if stream == FederationStream and hs.config.send_federation: + # We only support federation stream if federation sending + # hase been disabled on the master. + continue - LaterGauge( - "synapse_replication_tcp_resource_total_connections", - "", - [], - lambda: len(self.connections), - ) - - # List of streams that clients can subscribe to. - # We only support federation stream if federation sending hase been - # disabled on the master. - self.streams = [ - stream(hs) - for stream in itervalues(STREAMS_MAP) - if stream != FederationStream or not hs.config.send_federation - ] + self.streams.append(stream(hs)) self.streams_by_name = {stream.NAME: stream for stream in self.streams} - LaterGauge( - "synapse_replication_tcp_resource_connections_per_stream", - "", - ["stream_name"], - lambda: { - (stream_name,): len( - [ - conn - for conn in self.connections - if stream_name in conn.replication_streams - ] - ) - for stream_name in self.streams_by_name - }, - ) - - self.federation_sender = None - if not hs.config.send_federation: - self.federation_sender = hs.get_federation_sender() - - self.notifier.add_replication_callback(self.on_notifier_poke) - self.notifier.add_remote_server_up_callback(self.send_remote_server_up) + # Only bother registering the notifier callback if we have streams to + # publish. + if self.streams: + self.notifier.add_replication_callback(self.on_notifier_poke) # Keeps track of whether we are currently checking for updates self.is_looping = False self.pending_updates = False - hs.get_reactor().addSystemEventTrigger("before", "shutdown", self.on_shutdown) + self.command_handler = hs.get_tcp_replication() - def on_shutdown(self): - # close all connections on shutdown - for conn in self.connections: - conn.send_error("server shutting down") + def get_streams(self) -> Dict[str, Stream]: + """Get a mapp from stream name to stream instance. + """ + return self.streams_by_name def on_notifier_poke(self): """Checks if there is actually any new data and sends it to the @@ -140,7 +110,7 @@ class ReplicationStreamer(object): This should get called each time new data is available, even if it is currently being executed, so that nothing gets missed """ - if not self.connections: + if not self.command_handler.connected(): # Don't bother if nothing is listening. We still need to advance # the stream tokens otherwise they'll fall beihind forever for stream in self.streams: @@ -166,11 +136,6 @@ class ReplicationStreamer(object): self.pending_updates = False with Measure(self.clock, "repl.stream.get_updates"): - # First we tell the streams that they should update their - # current tokens. - for stream in self.streams: - stream.advance_current_token() - all_streams = self.streams if self._replication_torture_level is not None: @@ -180,7 +145,7 @@ class ReplicationStreamer(object): random.shuffle(all_streams) for stream in all_streams: - if stream.last_token == stream.upto_token: + if stream.last_token == stream.current_token(): continue if self._replication_torture_level: @@ -192,18 +157,17 @@ class ReplicationStreamer(object): "Getting stream: %s: %s -> %s", stream.NAME, stream.last_token, - stream.upto_token, + stream.current_token(), ) try: - updates, current_token = await stream.get_updates() + updates, current_token, limited = await stream.get_updates() + self.pending_updates |= limited except Exception: logger.info("Failed to handle stream %s", stream.NAME) raise logger.debug( - "Sending %d updates to %d connections", - len(updates), - len(self.connections), + "Sending %d updates", len(updates), ) if updates: @@ -219,116 +183,19 @@ class ReplicationStreamer(object): # token. See RdataCommand for more details. batched_updates = _batch_updates(updates) - for conn in self.connections: - for token, row in batched_updates: - try: - conn.stream_update(stream.NAME, token, row) - except Exception: - logger.exception("Failed to replicate") + for token, row in batched_updates: + try: + self.command_handler.stream_update( + stream.NAME, token, row + ) + except Exception: + logger.exception("Failed to replicate") logger.debug("No more pending updates, breaking poke loop") finally: self.pending_updates = False self.is_looping = False - @measure_func("repl.get_stream_updates") - async def get_stream_updates(self, stream_name, token): - """For a given stream get all updates since token. This is called when - a client first subscribes to a stream. - """ - stream = self.streams_by_name.get(stream_name, None) - if not stream: - raise Exception("unknown stream %s", stream_name) - - return await stream.get_updates_since(token) - - @measure_func("repl.federation_ack") - def federation_ack(self, token): - """We've received an ack for federation stream from a client. - """ - federation_ack_counter.inc() - if self.federation_sender: - self.federation_sender.federation_ack(token) - - @measure_func("repl.on_user_sync") - async def on_user_sync(self, conn_id, user_id, is_syncing, last_sync_ms): - """A client has started/stopped syncing on a worker. - """ - user_sync_counter.inc() - await self.presence_handler.update_external_syncs_row( - conn_id, user_id, is_syncing, last_sync_ms - ) - - @measure_func("repl.on_remove_pusher") - async def on_remove_pusher(self, app_id, push_key, user_id): - """A client has asked us to remove a pusher - """ - remove_pusher_counter.inc() - await self.store.delete_pusher_by_app_id_pushkey_user_id( - app_id=app_id, pushkey=push_key, user_id=user_id - ) - - self.notifier.on_new_replication_data() - - @measure_func("repl.on_invalidate_cache") - async def on_invalidate_cache(self, cache_func: str, keys: List[Any]): - """The client has asked us to invalidate a cache - """ - invalidate_cache_counter.inc() - - # We invalidate the cache locally, but then also stream that to other - # workers. - await self.store.invalidate_cache_and_stream(cache_func, tuple(keys)) - - @measure_func("repl.on_user_ip") - async def on_user_ip( - self, user_id, access_token, ip, user_agent, device_id, last_seen - ): - """The client saw a user request - """ - user_ip_cache_counter.inc() - await self.store.insert_client_ip( - user_id, access_token, ip, user_agent, device_id, last_seen - ) - await self._server_notices_sender.on_user_ip(user_id) - - @measure_func("repl.on_remote_server_up") - def on_remote_server_up(self, server: str): - self.notifier.notify_remote_server_up(server) - - def send_remote_server_up(self, server: str): - for conn in self.connections: - conn.send_remote_server_up(server) - - def send_sync_to_all_connections(self, data): - """Sends a SYNC command to all clients. - - Used in tests. - """ - for conn in self.connections: - conn.send_sync(data) - - def new_connection(self, connection): - """A new client connection has been established - """ - self.connections.append(connection) - - def lost_connection(self, connection): - """A client connection has been lost - """ - try: - self.connections.remove(connection) - except ValueError: - pass - - # We need to tell the presence handler that the connection has been - # lost so that it can handle any ongoing syncs on that connection. - run_as_background_process( - "update_external_syncs_clear", - self.presence_handler.update_external_syncs_clear, - connection.conn_id, - ) - def _batch_updates(updates): """Takes a list of updates of form [(token, row)] and sets the token to diff --git a/synapse/replication/tcp/streams/__init__.py b/synapse/replication/tcp/streams/__init__.py index 5f52264e8..d1a61c331 100644 --- a/synapse/replication/tcp/streams/__init__.py +++ b/synapse/replication/tcp/streams/__init__.py @@ -25,26 +25,63 @@ Each stream is defined by the following information: update_function: The function that returns a list of updates between two tokens """ -from . import _base, events, federation +from synapse.replication.tcp.streams._base import ( + AccountDataStream, + BackfillStream, + CachesStream, + DeviceListsStream, + GroupServerStream, + PresenceStream, + PublicRoomsStream, + PushersStream, + PushRulesStream, + ReceiptsStream, + Stream, + TagAccountDataStream, + ToDeviceStream, + TypingStream, + UserSignatureStream, +) +from synapse.replication.tcp.streams.events import EventsStream +from synapse.replication.tcp.streams.federation import FederationStream STREAMS_MAP = { stream.NAME: stream for stream in ( - events.EventsStream, - _base.BackfillStream, - _base.PresenceStream, - _base.TypingStream, - _base.ReceiptsStream, - _base.PushRulesStream, - _base.PushersStream, - _base.CachesStream, - _base.PublicRoomsStream, - _base.DeviceListsStream, - _base.ToDeviceStream, - federation.FederationStream, - _base.TagAccountDataStream, - _base.AccountDataStream, - _base.GroupServerStream, - _base.UserSignatureStream, + EventsStream, + BackfillStream, + PresenceStream, + TypingStream, + ReceiptsStream, + PushRulesStream, + PushersStream, + CachesStream, + PublicRoomsStream, + DeviceListsStream, + ToDeviceStream, + FederationStream, + TagAccountDataStream, + AccountDataStream, + GroupServerStream, + UserSignatureStream, ) } + +__all__ = [ + "STREAMS_MAP", + "Stream", + "BackfillStream", + "PresenceStream", + "TypingStream", + "ReceiptsStream", + "PushRulesStream", + "PushersStream", + "CachesStream", + "PublicRoomsStream", + "DeviceListsStream", + "ToDeviceStream", + "TagAccountDataStream", + "AccountDataStream", + "GroupServerStream", + "UserSignatureStream", +] diff --git a/synapse/replication/tcp/streams/_base.py b/synapse/replication/tcp/streams/_base.py index 208e8a667..b0f87c365 100644 --- a/synapse/replication/tcp/streams/_base.py +++ b/synapse/replication/tcp/streams/_base.py @@ -14,117 +14,71 @@ # See the License for the specific language governing permissions and # limitations under the License. -import itertools import logging from collections import namedtuple -from typing import Any, List, Optional +from typing import Any, Awaitable, Callable, List, Optional, Tuple import attr +from synapse.replication.http.streams import ReplicationGetStreamUpdates + logger = logging.getLogger(__name__) - -MAX_EVENTS_BEHIND = 500000 - -BackfillStreamRow = namedtuple( - "BackfillStreamRow", - ( - "event_id", # str - "room_id", # str - "type", # str - "state_key", # str, optional - "redacts", # str, optional - "relates_to", # str, optional - ), -) -PresenceStreamRow = namedtuple( - "PresenceStreamRow", - ( - "user_id", # str - "state", # str - "last_active_ts", # int - "last_federation_update_ts", # int - "last_user_sync_ts", # int - "status_msg", # str - "currently_active", # bool - ), -) -TypingStreamRow = namedtuple( - "TypingStreamRow", ("room_id", "user_ids") # str # list(str) -) -ReceiptsStreamRow = namedtuple( - "ReceiptsStreamRow", - ( - "room_id", # str - "receipt_type", # str - "user_id", # str - "event_id", # str - "data", # dict - ), -) -PushRulesStreamRow = namedtuple("PushRulesStreamRow", ("user_id",)) # str -PushersStreamRow = namedtuple( - "PushersStreamRow", - ("user_id", "app_id", "pushkey", "deleted"), # str # str # str # bool -) +# the number of rows to request from an update_function. +_STREAM_UPDATE_TARGET_ROW_COUNT = 100 -@attr.s -class CachesStreamRow: - """Stream to inform workers they should invalidate their cache. +# Some type aliases to make things a bit easier. - Attributes: - cache_func: Name of the cached function. - keys: The entry in the cache to invalidate. If None then will - invalidate all. - invalidation_ts: Timestamp of when the invalidation took place. - """ +# A stream position token +Token = int - cache_func = attr.ib(type=str) - keys = attr.ib(type=Optional[List[Any]]) - invalidation_ts = attr.ib(type=int) +# The type of a stream update row, after JSON deserialisation, but before +# parsing with Stream.parse_row (which turns it into a `ROW_TYPE`). Normally it's +# just a row from a database query, though this is dependent on the stream in question. +# +StreamRow = Tuple +# The type returned by the update_function of a stream, as well as get_updates(), +# get_updates_since, etc. +# +# It consists of a triplet `(updates, new_last_token, limited)`, where: +# * `updates` is a list of `(token, row)` entries. +# * `new_last_token` is the new position in stream. +# * `limited` is whether there are more updates to fetch. +# +StreamUpdateResult = Tuple[List[Tuple[Token, StreamRow]], Token, bool] -PublicRoomsStreamRow = namedtuple( - "PublicRoomsStreamRow", - ( - "room_id", # str - "visibility", # str - "appservice_id", # str, optional - "network_id", # str, optional - ), -) -DeviceListsStreamRow = namedtuple( - "DeviceListsStreamRow", ("user_id", "destination") # str # str -) -ToDeviceStreamRow = namedtuple("ToDeviceStreamRow", ("entity",)) # str -TagAccountDataStreamRow = namedtuple( - "TagAccountDataStreamRow", ("user_id", "room_id", "data") # str # str # dict -) -AccountDataStreamRow = namedtuple( - "AccountDataStream", ("user_id", "room_id", "data_type") # str # str # str -) -GroupsStreamRow = namedtuple( - "GroupsStreamRow", - ("group_id", "user_id", "type", "content"), # str # str # str # dict -) -UserSignatureStreamRow = namedtuple("UserSignatureStreamRow", ("user_id")) # str +# The type of an update_function for a stream +# +# The arguments are: +# +# * instance_name: the writer of the stream +# * from_token: the previous stream token: the starting point for fetching the +# updates +# * to_token: the new stream token: the point to get updates up to +# * target_row_count: a target for the number of rows to be returned. +# +# The update_function is expected to return up to _approximately_ target_row_count rows. +# If there are more updates available, it should set `limited` in the result, and +# it will be called again to get the next batch. +# +UpdateFunction = Callable[[str, Token, Token, int], Awaitable[StreamUpdateResult]] class Stream(object): """Base class for the streams. Provides a `get_updates()` function that returns new updates since the last - time it was called up until the point `advance_current_token` was called. + time it was called. """ NAME = None # type: str # The name of the stream # The type of the row. Used by the default impl of parse_row. ROW_TYPE = None # type: Any - _LIMITED = True # Whether the update function takes a limit @classmethod - def parse_row(cls, row): + def parse_row(cls, row: StreamRow): """Parse a row received over replication By default, assumes that the row data is an array object and passes its contents @@ -138,101 +92,122 @@ class Stream(object): """ return cls.ROW_TYPE(*row) - def __init__(self, hs): + def __init__( + self, + local_instance_name: str, + current_token_function: Callable[[], Token], + update_function: UpdateFunction, + ): + """Instantiate a Stream + + current_token_function and update_function are callbacks which should be + implemented by subclasses. + + current_token_function is called to get the current token of the underlying + stream. + + update_function is called to get updates for this stream between a pair of + stream tokens. See the UpdateFunction type definition for more info. + + Args: + local_instance_name: The instance name of the current process + current_token_function: callback to get the current token, as above + update_function: callback go get stream updates, as above + """ + self.local_instance_name = local_instance_name + self.current_token = current_token_function + self.update_function = update_function + # The token from which we last asked for updates self.last_token = self.current_token() - # The token that we will get updates up to - self.upto_token = self.current_token() - - def advance_current_token(self): - """Updates `upto_token` to "now", which updates up until which point - get_updates[_since] will fetch rows till. - """ - self.upto_token = self.current_token() - def discard_updates_and_advance(self): """Called when the stream should advance but the updates would be discarded, e.g. when there are no currently connected workers. """ - self.upto_token = self.current_token() - self.last_token = self.upto_token + self.last_token = self.current_token() - async def get_updates(self): + async def get_updates(self) -> StreamUpdateResult: """Gets all updates since the last time this function was called (or - since the stream was constructed if it hadn't been called before), - until the `upto_token` + since the stream was constructed if it hadn't been called before). Returns: - Deferred[Tuple[List[Tuple[int, Any]], int]: - Resolves to a pair ``(updates, current_token)``, where ``updates`` is a - list of ``(token, row)`` entries. ``row`` will be json-serialised and - sent over the replication steam. + A triplet `(updates, new_last_token, limited)`, where `updates` is + a list of `(token, row)` entries, `new_last_token` is the new + position in stream, and `limited` is whether there are more updates + to fetch. """ - updates, current_token = await self.get_updates_since(self.last_token) + current_token = self.current_token() + updates, current_token, limited = await self.get_updates_since( + self.local_instance_name, self.last_token, current_token + ) self.last_token = current_token - return updates, current_token + return updates, current_token, limited - async def get_updates_since(self, from_token): + async def get_updates_since( + self, instance_name: str, from_token: Token, upto_token: Token + ) -> StreamUpdateResult: """Like get_updates except allows specifying from when we should stream updates Returns: - Deferred[Tuple[List[Tuple[int, Any]], int]: - Resolves to a pair ``(updates, current_token)``, where ``updates`` is a - list of ``(token, row)`` entries. ``row`` will be json-serialised and - sent over the replication steam. + A triplet `(updates, new_last_token, limited)`, where `updates` is + a list of `(token, row)` entries, `new_last_token` is the new + position in stream, and `limited` is whether there are more updates + to fetch. """ - if from_token in ("NOW", "now"): - return [], self.upto_token - - current_token = self.upto_token from_token = int(from_token) - if from_token == current_token: - return [], current_token + if from_token == upto_token: + return [], upto_token, False - logger.info("get_updates_since: %s", self.__class__) - if self._LIMITED: - rows = await self.update_function( - from_token, current_token, limit=MAX_EVENTS_BEHIND + 1 - ) + updates, upto_token, limited = await self.update_function( + instance_name, from_token, upto_token, _STREAM_UPDATE_TARGET_ROW_COUNT, + ) + return updates, upto_token, limited - # never turn more than MAX_EVENTS_BEHIND + 1 into updates. - rows = itertools.islice(rows, MAX_EVENTS_BEHIND + 1) - else: - rows = await self.update_function(from_token, current_token) +def db_query_to_update_function( + query_function: Callable[[Token, Token, int], Awaitable[List[tuple]]] +) -> UpdateFunction: + """Wraps a db query function which returns a list of rows to make it + suitable for use as an `update_function` for the Stream class + """ + + async def update_function(instance_name, from_token, upto_token, limit): + rows = await query_function(from_token, upto_token, limit) updates = [(row[0], row[1:]) for row in rows] + limited = False + if len(updates) >= limit: + upto_token = updates[-1][0] + limited = True - # check we didn't get more rows than the limit. - # doing it like this allows the update_function to be a generator. - if self._LIMITED and len(updates) >= MAX_EVENTS_BEHIND: - raise Exception("stream %s has fallen behind" % (self.NAME)) + return updates, upto_token, limited - return updates, current_token + return update_function - def current_token(self): - """Gets the current token of the underlying streams. Should be provided - by the sub classes - Returns: - int - """ - raise NotImplementedError() +def make_http_update_function(hs, stream_name: str) -> UpdateFunction: + """Makes a suitable function for use as an `update_function` that queries + the master process for updates. + """ - def update_function(self, from_token, current_token, limit=None): - """Get updates between from_token and to_token. If Stream._LIMITED is - True then limit is provided, otherwise it's not. + client = ReplicationGetStreamUpdates.make_client(hs) - Returns: - Deferred(list(tuple)): the first entry in the tuple is the token for - that update, and the rest of the tuple gets used to construct - a ``ROW_TYPE`` instance - """ - raise NotImplementedError() + async def update_function( + instance_name: str, from_token: int, upto_token: int, limit: int + ) -> StreamUpdateResult: + result = await client( + instance_name=instance_name, + stream_name=stream_name, + from_token=from_token, + upto_token=upto_token, + ) + return result["updates"], result["upto_token"], result["limited"] + + return update_function class BackfillStream(Stream): @@ -240,93 +215,166 @@ class BackfillStream(Stream): or it went from being an outlier to not. """ + BackfillStreamRow = namedtuple( + "BackfillStreamRow", + ( + "event_id", # str + "room_id", # str + "type", # str + "state_key", # str, optional + "redacts", # str, optional + "relates_to", # str, optional + ), + ) + NAME = "backfill" ROW_TYPE = BackfillStreamRow def __init__(self, hs): store = hs.get_datastore() - self.current_token = store.get_current_backfill_token # type: ignore - self.update_function = store.get_all_new_backfill_event_rows # type: ignore - - super(BackfillStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_current_backfill_token, + db_query_to_update_function(store.get_all_new_backfill_event_rows), + ) class PresenceStream(Stream): + PresenceStreamRow = namedtuple( + "PresenceStreamRow", + ( + "user_id", # str + "state", # str + "last_active_ts", # int + "last_federation_update_ts", # int + "last_user_sync_ts", # int + "status_msg", # str + "currently_active", # bool + ), + ) + NAME = "presence" - _LIMITED = False ROW_TYPE = PresenceStreamRow def __init__(self, hs): store = hs.get_datastore() - presence_handler = hs.get_presence_handler() - self.current_token = store.get_current_presence_token # type: ignore - self.update_function = presence_handler.get_all_presence_updates # type: ignore + if hs.config.worker_app is None: + # on the master, query the presence handler + presence_handler = hs.get_presence_handler() + update_function = db_query_to_update_function( + presence_handler.get_all_presence_updates + ) + else: + # Query master process + update_function = make_http_update_function(hs, self.NAME) - super(PresenceStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), store.get_current_presence_token, update_function + ) class TypingStream(Stream): + TypingStreamRow = namedtuple( + "TypingStreamRow", ("room_id", "user_ids") # str # list(str) + ) + NAME = "typing" - _LIMITED = False ROW_TYPE = TypingStreamRow def __init__(self, hs): typing_handler = hs.get_typing_handler() - self.current_token = typing_handler.get_current_token # type: ignore - self.update_function = typing_handler.get_all_typing_updates # type: ignore + if hs.config.worker_app is None: + # on the master, query the typing handler + update_function = db_query_to_update_function( + typing_handler.get_all_typing_updates + ) + else: + # Query master process + update_function = make_http_update_function(hs, self.NAME) - super(TypingStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), typing_handler.get_current_token, update_function + ) class ReceiptsStream(Stream): + ReceiptsStreamRow = namedtuple( + "ReceiptsStreamRow", + ( + "room_id", # str + "receipt_type", # str + "user_id", # str + "event_id", # str + "data", # dict + ), + ) + NAME = "receipts" ROW_TYPE = ReceiptsStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_max_receipt_stream_id # type: ignore - self.update_function = store.get_all_updated_receipts # type: ignore - - super(ReceiptsStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_max_receipt_stream_id, + db_query_to_update_function(store.get_all_updated_receipts), + ) class PushRulesStream(Stream): """A user has changed their push rules """ + PushRulesStreamRow = namedtuple("PushRulesStreamRow", ("user_id",)) # str + NAME = "push_rules" ROW_TYPE = PushRulesStreamRow def __init__(self, hs): self.store = hs.get_datastore() - super(PushRulesStream, self).__init__(hs) + super(PushRulesStream, self).__init__( + hs.get_instance_name(), self._current_token, self._update_function + ) - def current_token(self): + def _current_token(self) -> int: push_rules_token, _ = self.store.get_push_rules_stream_token() return push_rules_token - async def update_function(self, from_token, to_token, limit): + async def _update_function( + self, instance_name: str, from_token: Token, to_token: Token, limit: int + ): rows = await self.store.get_all_push_rule_updates(from_token, to_token, limit) - return [(row[0], row[2]) for row in rows] + + limited = False + if len(rows) == limit: + to_token = rows[-1][0] + limited = True + + return [(row[0], (row[2],)) for row in rows], to_token, limited class PushersStream(Stream): """A user has added/changed/removed a pusher """ + PushersStreamRow = namedtuple( + "PushersStreamRow", + ("user_id", "app_id", "pushkey", "deleted"), # str # str # str # bool + ) + NAME = "pushers" ROW_TYPE = PushersStreamRow def __init__(self, hs): store = hs.get_datastore() - self.current_token = store.get_pushers_stream_token # type: ignore - self.update_function = store.get_all_updated_pushers_rows # type: ignore - - super(PushersStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_pushers_stream_token, + db_query_to_update_function(store.get_all_updated_pushers_rows), + ) class CachesStream(Stream): @@ -334,98 +382,138 @@ class CachesStream(Stream): the cache on the workers """ + @attr.s + class CachesStreamRow: + """Stream to inform workers they should invalidate their cache. + + Attributes: + cache_func: Name of the cached function. + keys: The entry in the cache to invalidate. If None then will + invalidate all. + invalidation_ts: Timestamp of when the invalidation took place. + """ + + cache_func = attr.ib(type=str) + keys = attr.ib(type=Optional[List[Any]]) + invalidation_ts = attr.ib(type=int) + NAME = "caches" ROW_TYPE = CachesStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_cache_stream_token # type: ignore - self.update_function = store.get_all_updated_caches # type: ignore - - super(CachesStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_cache_stream_token, + db_query_to_update_function(store.get_all_updated_caches), + ) class PublicRoomsStream(Stream): """The public rooms list changed """ + PublicRoomsStreamRow = namedtuple( + "PublicRoomsStreamRow", + ( + "room_id", # str + "visibility", # str + "appservice_id", # str, optional + "network_id", # str, optional + ), + ) + NAME = "public_rooms" ROW_TYPE = PublicRoomsStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_current_public_room_stream_id # type: ignore - self.update_function = store.get_all_new_public_rooms # type: ignore - - super(PublicRoomsStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_current_public_room_stream_id, + db_query_to_update_function(store.get_all_new_public_rooms), + ) class DeviceListsStream(Stream): - """Someone added/changed/removed a device + """Either a user has updated their devices or a remote server needs to be + told about a device update. """ + @attr.s + class DeviceListsStreamRow: + entity = attr.ib(type=str) + NAME = "device_lists" - _LIMITED = False ROW_TYPE = DeviceListsStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_device_stream_token # type: ignore - self.update_function = store.get_all_device_list_changes_for_remotes # type: ignore - - super(DeviceListsStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_device_stream_token, + db_query_to_update_function(store.get_all_device_list_changes_for_remotes), + ) class ToDeviceStream(Stream): """New to_device messages for a client """ + ToDeviceStreamRow = namedtuple("ToDeviceStreamRow", ("entity",)) # str + NAME = "to_device" ROW_TYPE = ToDeviceStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_to_device_stream_token # type: ignore - self.update_function = store.get_all_new_device_messages # type: ignore - - super(ToDeviceStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_to_device_stream_token, + db_query_to_update_function(store.get_all_new_device_messages), + ) class TagAccountDataStream(Stream): """Someone added/removed a tag for a room """ + TagAccountDataStreamRow = namedtuple( + "TagAccountDataStreamRow", ("user_id", "room_id", "data") # str # str # dict + ) + NAME = "tag_account_data" ROW_TYPE = TagAccountDataStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_max_account_data_stream_id # type: ignore - self.update_function = store.get_all_updated_tags # type: ignore - - super(TagAccountDataStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_max_account_data_stream_id, + db_query_to_update_function(store.get_all_updated_tags), + ) class AccountDataStream(Stream): """Global or per room account data was changed """ + AccountDataStreamRow = namedtuple( + "AccountDataStream", ("user_id", "room_id", "data_type") # str # str # str + ) + NAME = "account_data" ROW_TYPE = AccountDataStreamRow def __init__(self, hs): self.store = hs.get_datastore() + super().__init__( + hs.get_instance_name(), + self.store.get_max_account_data_stream_id, + db_query_to_update_function(self._update_function), + ) - self.current_token = self.store.get_max_account_data_stream_id # type: ignore - - super(AccountDataStream, self).__init__(hs) - - async def update_function(self, from_token, to_token, limit): + async def _update_function(self, from_token, to_token, limit): global_results, room_results = await self.store.get_all_updated_account_data( from_token, from_token, to_token, limit ) @@ -440,30 +528,38 @@ class AccountDataStream(Stream): class GroupServerStream(Stream): + GroupsStreamRow = namedtuple( + "GroupsStreamRow", + ("group_id", "user_id", "type", "content"), # str # str # str # dict + ) + NAME = "groups" ROW_TYPE = GroupsStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_group_stream_token # type: ignore - self.update_function = store.get_all_groups_changes # type: ignore - - super(GroupServerStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_group_stream_token, + db_query_to_update_function(store.get_all_groups_changes), + ) class UserSignatureStream(Stream): """A user has signed their own device with their user-signing key """ + UserSignatureStreamRow = namedtuple("UserSignatureStreamRow", ("user_id")) # str + NAME = "user_signature" - _LIMITED = False ROW_TYPE = UserSignatureStreamRow def __init__(self, hs): store = hs.get_datastore() - - self.current_token = store.get_device_stream_token # type: ignore - self.update_function = store.get_all_user_signature_changes_for_remotes # type: ignore - - super(UserSignatureStream, self).__init__(hs) + super().__init__( + hs.get_instance_name(), + store.get_device_stream_token, + db_query_to_update_function( + store.get_all_user_signature_changes_for_remotes + ), + ) diff --git a/synapse/replication/tcp/streams/events.py b/synapse/replication/tcp/streams/events.py index b3afabb8c..890e75d82 100644 --- a/synapse/replication/tcp/streams/events.py +++ b/synapse/replication/tcp/streams/events.py @@ -15,11 +15,12 @@ # limitations under the License. import heapq -from typing import Tuple, Type +from collections import Iterable +from typing import List, Tuple, Type import attr -from ._base import Stream +from ._base import Stream, StreamUpdateResult, Token """Handling of the 'events' replication stream @@ -116,28 +117,107 @@ class EventsStream(Stream): def __init__(self, hs): self._store = hs.get_datastore() - self.current_token = self._store.get_current_events_token # type: ignore + super().__init__( + hs.get_instance_name(), + self._store.get_current_events_token, + self._update_function, + ) - super(EventsStream, self).__init__(hs) + async def _update_function( + self, + instance_name: str, + from_token: Token, + current_token: Token, + target_row_count: int, + ) -> StreamUpdateResult: + + # the events stream merges together three separate sources: + # * new events + # * current_state changes + # * events which were previously outliers, but have now been de-outliered. + # + # The merge operation is complicated by the fact that we only have a single + # "stream token" which is supposed to indicate how far we have got through + # all three streams. It's therefore no good to return rows 1-1000 from the + # "new events" table if the state_deltas are limited to rows 1-100 by the + # target_row_count. + # + # In other words: we must pick a new upper limit, and must return *all* rows + # up to that point for each of the three sources. + # + # Start by trying to split the target_row_count up. We expect to have a + # negligible number of ex-outliers, and a rough approximation based on recent + # traffic on sw1v.org shows that there are approximately the same number of + # event rows between a given pair of stream ids as there are state + # updates, so let's split our target_row_count among those two types. The target + # is only an approximation - it doesn't matter if we end up going a bit over it. + + target_row_count //= 2 + + # now we fetch up to that many rows from the events table - async def update_function(self, from_token, current_token, limit=None): event_rows = await self._store.get_all_new_forward_event_rows( - from_token, current_token, limit + from_token, current_token, target_row_count + ) # type: List[Tuple] + + # we rely on get_all_new_forward_event_rows strictly honouring the limit, so + # that we know it is safe to just take upper_limit = event_rows[-1][0]. + assert ( + len(event_rows) <= target_row_count + ), "get_all_new_forward_event_rows did not honour row limit" + + # if we hit the limit on event_updates, there's no point in going beyond the + # last stream_id in the batch for the other sources. + + if len(event_rows) == target_row_count: + limited = True + upper_limit = event_rows[-1][0] # type: int + else: + limited = False + upper_limit = current_token + + # next up is the state delta table. + ( + state_rows, + upper_limit, + state_rows_limited, + ) = await self._store.get_all_updated_current_state_deltas( + from_token, upper_limit, target_row_count ) + + limited = limited or state_rows_limited + + # finally, fetch the ex-outliers rows. We assume there are few enough of these + # not to bother with the limit. + + ex_outliers_rows = await self._store.get_ex_outlier_stream_rows( + from_token, upper_limit + ) # type: List[Tuple] + + # we now need to turn the raw database rows returned into tuples suitable + # for the replication protocol (basically, we add an identifier to + # distinguish the row type). At the same time, we can limit the event_rows + # to the max stream_id from state_rows. + event_updates = ( - (row[0], EventsStreamEventRow.TypeId, row[1:]) for row in event_rows - ) + (stream_id, (EventsStreamEventRow.TypeId, rest)) + for (stream_id, *rest) in event_rows + if stream_id <= upper_limit + ) # type: Iterable[Tuple[int, Tuple]] - state_rows = await self._store.get_all_updated_current_state_deltas( - from_token, current_token, limit - ) state_updates = ( - (row[0], EventsStreamCurrentStateRow.TypeId, row[1:]) for row in state_rows - ) + (stream_id, (EventsStreamCurrentStateRow.TypeId, rest)) + for (stream_id, *rest) in state_rows + ) # type: Iterable[Tuple[int, Tuple]] - all_updates = heapq.merge(event_updates, state_updates) + ex_outliers_updates = ( + (stream_id, (EventsStreamEventRow.TypeId, rest)) + for (stream_id, *rest) in ex_outliers_rows + ) # type: Iterable[Tuple[int, Tuple]] - return all_updates + # we need to return a sorted list, so merge them together. + updates = list(heapq.merge(event_updates, state_updates, ex_outliers_updates)) + return updates, upper_limit, limited @classmethod def parse_row(cls, row): diff --git a/synapse/replication/tcp/streams/federation.py b/synapse/replication/tcp/streams/federation.py index 615f3dc9a..e8bd52e38 100644 --- a/synapse/replication/tcp/streams/federation.py +++ b/synapse/replication/tcp/streams/federation.py @@ -15,15 +15,7 @@ # limitations under the License. from collections import namedtuple -from ._base import Stream - -FederationStreamRow = namedtuple( - "FederationStreamRow", - ( - "type", # str, the type of data as defined in the BaseFederationRows - "data", # dict, serialization of a federation.send_queue.BaseFederationRow - ), -) +from synapse.replication.tcp.streams._base import Stream, db_query_to_update_function class FederationStream(Stream): @@ -31,13 +23,33 @@ class FederationStream(Stream): sending disabled. """ + FederationStreamRow = namedtuple( + "FederationStreamRow", + ( + "type", # str, the type of data as defined in the BaseFederationRows + "data", # dict, serialization of a federation.send_queue.BaseFederationRow + ), + ) + NAME = "federation" ROW_TYPE = FederationStreamRow def __init__(self, hs): - federation_sender = hs.get_federation_sender() + # Not all synapse instances will have a federation sender instance, + # whether that's a `FederationSender` or a `FederationRemoteSendQueue`, + # so we stub the stream out when that is the case. + if hs.config.worker_app is None or hs.should_send_federation(): + federation_sender = hs.get_federation_sender() + current_token = federation_sender.get_current_token + update_function = db_query_to_update_function( + federation_sender.get_replication_rows + ) + else: + current_token = lambda: 0 + update_function = self._stub_update_function - self.current_token = federation_sender.get_current_token # type: ignore - self.update_function = federation_sender.get_replication_rows # type: ignore + super().__init__(hs.get_instance_name(), current_token, update_function) - super(FederationStream, self).__init__(hs) + @staticmethod + async def _stub_update_function(instance_name, from_token, upto_token, limit): + return [], upto_token, False diff --git a/synapse/res/templates/sso_account_deactivated.html b/synapse/res/templates/sso_account_deactivated.html new file mode 100644 index 000000000..4eb8db9fb --- /dev/null +++ b/synapse/res/templates/sso_account_deactivated.html @@ -0,0 +1,10 @@ + + + + + SSO account deactivated + + +

This account has been deactivated.

+ + diff --git a/synapse/res/templates/sso_auth_confirm.html b/synapse/res/templates/sso_auth_confirm.html new file mode 100644 index 000000000..0d9de9d46 --- /dev/null +++ b/synapse/res/templates/sso_auth_confirm.html @@ -0,0 +1,14 @@ + + + Authentication + + +
+

+ A client is trying to {{ description | e }}. To confirm this action, + re-authenticate with single sign-on. + If you did not expect this, your account may be compromised! +

+
+ + diff --git a/synapse/res/templates/sso_auth_success.html b/synapse/res/templates/sso_auth_success.html new file mode 100644 index 000000000..03f141946 --- /dev/null +++ b/synapse/res/templates/sso_auth_success.html @@ -0,0 +1,18 @@ + + + Authentication Successful + + + +
+

Thank you

+

You may now close this window and return to the application

+
+ + diff --git a/synapse/rest/__init__.py b/synapse/rest/__init__.py index 4a1fc2ec2..46e458e95 100644 --- a/synapse/rest/__init__.py +++ b/synapse/rest/__init__.py @@ -41,6 +41,7 @@ from synapse.rest.client.v2_alpha import ( keys, notifications, openid, + password_policy, read_marker, receipts, register, @@ -118,6 +119,7 @@ class ClientRestResource(JsonResource): capabilities.register_servlets(hs, client_resource) account_validity.register_servlets(hs, client_resource) relations.register_servlets(hs, client_resource) + password_policy.register_servlets(hs, client_resource) # moving to /_synapse/admin synapse.rest.admin.register_servlets_for_client_rest_resource( diff --git a/synapse/rest/admin/__init__.py b/synapse/rest/admin/__init__.py index 42cc2b062..ed70d448a 100644 --- a/synapse/rest/admin/__init__.py +++ b/synapse/rest/admin/__init__.py @@ -29,7 +29,11 @@ from synapse.rest.admin._base import ( from synapse.rest.admin.groups import DeleteGroupAdminRestServlet from synapse.rest.admin.media import ListMediaInRoom, register_servlets_for_media_repo from synapse.rest.admin.purge_room_servlet import PurgeRoomServlet -from synapse.rest.admin.rooms import ListRoomRestServlet, ShutdownRoomRestServlet +from synapse.rest.admin.rooms import ( + JoinRoomAliasServlet, + ListRoomRestServlet, + ShutdownRoomRestServlet, +) from synapse.rest.admin.server_notice_servlet import SendServerNoticeServlet from synapse.rest.admin.users import ( AccountValidityRenewServlet, @@ -189,6 +193,7 @@ def register_servlets(hs, http_server): """ register_servlets_for_client_rest_resource(hs, http_server) ListRoomRestServlet(hs).register(http_server) + JoinRoomAliasServlet(hs).register(http_server) PurgeRoomServlet(hs).register(http_server) SendServerNoticeServlet(hs).register(http_server) VersionServlet(hs).register(http_server) diff --git a/synapse/rest/admin/rooms.py b/synapse/rest/admin/rooms.py index f9b8c0a4f..d1bdb6411 100644 --- a/synapse/rest/admin/rooms.py +++ b/synapse/rest/admin/rooms.py @@ -13,9 +13,10 @@ # See the License for the specific language governing permissions and # limitations under the License. import logging +from typing import List, Optional -from synapse.api.constants import Membership -from synapse.api.errors import Codes, SynapseError +from synapse.api.constants import EventTypes, JoinRules, Membership +from synapse.api.errors import Codes, NotFoundError, SynapseError from synapse.http.servlet import ( RestServlet, assert_params_in_dict, @@ -29,7 +30,7 @@ from synapse.rest.admin._base import ( historical_admin_path_patterns, ) from synapse.storage.data_stores.main.room import RoomSortOrder -from synapse.types import create_requester +from synapse.types import RoomAlias, RoomID, UserID, create_requester from synapse.util.async_helpers import maybe_awaitable logger = logging.getLogger(__name__) @@ -182,10 +183,23 @@ class ListRoomRestServlet(RestServlet): # Extract query parameters start = parse_integer(request, "from", default=0) limit = parse_integer(request, "limit", default=100) - order_by = parse_string(request, "order_by", default="alphabetical") + order_by = parse_string(request, "order_by", default=RoomSortOrder.NAME.value) if order_by not in ( RoomSortOrder.ALPHABETICAL.value, RoomSortOrder.SIZE.value, + RoomSortOrder.NAME.value, + RoomSortOrder.CANONICAL_ALIAS.value, + RoomSortOrder.JOINED_MEMBERS.value, + RoomSortOrder.JOINED_LOCAL_MEMBERS.value, + RoomSortOrder.VERSION.value, + RoomSortOrder.CREATOR.value, + RoomSortOrder.ENCRYPTION.value, + RoomSortOrder.FEDERATABLE.value, + RoomSortOrder.PUBLIC.value, + RoomSortOrder.JOIN_RULES.value, + RoomSortOrder.GUEST_ACCESS.value, + RoomSortOrder.HISTORY_VISIBILITY.value, + RoomSortOrder.STATE_EVENTS.value, ): raise SynapseError( 400, @@ -237,3 +251,75 @@ class ListRoomRestServlet(RestServlet): response["prev_batch"] = 0 return 200, response + + +class JoinRoomAliasServlet(RestServlet): + + PATTERNS = admin_patterns("/join/(?P[^/]*)") + + def __init__(self, hs): + self.hs = hs + self.auth = hs.get_auth() + self.room_member_handler = hs.get_room_member_handler() + self.admin_handler = hs.get_handlers().admin_handler + self.state_handler = hs.get_state_handler() + + async def on_POST(self, request, room_identifier): + requester = await self.auth.get_user_by_req(request) + await assert_user_is_admin(self.auth, requester.user) + + content = parse_json_object_from_request(request) + + assert_params_in_dict(content, ["user_id"]) + target_user = UserID.from_string(content["user_id"]) + + if not self.hs.is_mine(target_user): + raise SynapseError(400, "This endpoint can only be used with local users") + + if not await self.admin_handler.get_user(target_user): + raise NotFoundError("User not found") + + if RoomID.is_valid(room_identifier): + room_id = room_identifier + try: + remote_room_hosts = [ + x.decode("ascii") for x in request.args[b"server_name"] + ] # type: Optional[List[str]] + except Exception: + remote_room_hosts = None + elif RoomAlias.is_valid(room_identifier): + handler = self.room_member_handler + room_alias = RoomAlias.from_string(room_identifier) + room_id, remote_room_hosts = await handler.lookup_room_alias(room_alias) + room_id = room_id.to_string() + else: + raise SynapseError( + 400, "%s was not legal room ID or room alias" % (room_identifier,) + ) + + fake_requester = create_requester(target_user) + + # send invite if room has "JoinRules.INVITE" + room_state = await self.state_handler.get_current_state(room_id) + join_rules_event = room_state.get((EventTypes.JoinRules, "")) + if join_rules_event: + if not (join_rules_event.content.get("join_rule") == JoinRules.PUBLIC): + await self.room_member_handler.update_membership( + requester=requester, + target=fake_requester.user, + room_id=room_id, + action="invite", + remote_room_hosts=remote_room_hosts, + ratelimit=False, + ) + + await self.room_member_handler.update_membership( + requester=fake_requester, + target=fake_requester.user, + room_id=room_id, + action="join", + remote_room_hosts=remote_room_hosts, + ratelimit=False, + ) + + return 200, {"room_id": room_id} diff --git a/synapse/rest/admin/users.py b/synapse/rest/admin/users.py index 8551ac19b..326682fbd 100644 --- a/synapse/rest/admin/users.py +++ b/synapse/rest/admin/users.py @@ -94,10 +94,10 @@ class UsersRestServletV2(RestServlet): guests = parse_boolean(request, "guests", default=True) deactivated = parse_boolean(request, "deactivated", default=False) - users = await self.store.get_users_paginate( + users, total = await self.store.get_users_paginate( start, limit, user_id, guests, deactivated ) - ret = {"users": users} + ret = {"users": users, "total": total} if len(users) >= limit: ret["next_token"] = str(start + len(users)) @@ -199,7 +199,7 @@ class UserRestServletV2(RestServlet): user_id, threepid["medium"], threepid["address"], current_time ) - if "avatar_url" in body: + if "avatar_url" in body and type(body["avatar_url"]) == str: await self.profile_handler.set_avatar_url( target_user, requester, body["avatar_url"], True ) @@ -243,11 +243,11 @@ class UserRestServletV2(RestServlet): else: # create user password = body.get("password") - if password is not None and ( - not isinstance(body["password"], text_type) - or len(body["password"]) > 512 - ): - raise SynapseError(400, "Invalid password") + password_hash = None + if password is not None: + if not isinstance(password, text_type) or len(password) > 512: + raise SynapseError(400, "Invalid password") + password_hash = await self.auth_handler.hash(password) admin = body.get("admin", None) user_type = body.get("user_type", None) @@ -259,7 +259,7 @@ class UserRestServletV2(RestServlet): user_id = await self.registration_handler.register_user( localpart=target_user.localpart, - password=password, + password_hash=password_hash, admin=bool(admin), default_display_name=displayname, user_type=user_type, @@ -276,7 +276,7 @@ class UserRestServletV2(RestServlet): user_id, threepid["medium"], threepid["address"], current_time ) - if "avatar_url" in body: + if "avatar_url" in body and type(body["avatar_url"]) == str: await self.profile_handler.set_avatar_url( user_id, requester, body["avatar_url"], True ) @@ -298,7 +298,7 @@ class UserRegisterServlet(RestServlet): NONCE_TIMEOUT = 60 def __init__(self, hs): - self.handlers = hs.get_handlers() + self.auth_handler = hs.get_auth_handler() self.reactor = hs.get_reactor() self.nonces = {} self.hs = hs @@ -362,16 +362,16 @@ class UserRegisterServlet(RestServlet): 400, "password must be specified", errcode=Codes.BAD_JSON ) else: - if ( - not isinstance(body["password"], text_type) - or len(body["password"]) > 512 - ): + password = body["password"] + if not isinstance(password, text_type) or len(password) > 512: raise SynapseError(400, "Invalid password") - password = body["password"].encode("utf-8") - if b"\x00" in password: + password_bytes = password.encode("utf-8") + if b"\x00" in password_bytes: raise SynapseError(400, "Invalid password") + password_hash = await self.auth_handler.hash(password) + admin = body.get("admin", None) user_type = body.get("user_type", None) @@ -388,7 +388,7 @@ class UserRegisterServlet(RestServlet): want_mac_builder.update(b"\x00") want_mac_builder.update(username) want_mac_builder.update(b"\x00") - want_mac_builder.update(password) + want_mac_builder.update(password_bytes) want_mac_builder.update(b"\x00") want_mac_builder.update(b"admin" if admin else b"notadmin") if user_type: @@ -407,7 +407,7 @@ class UserRegisterServlet(RestServlet): user_id = await register.registration_handler.register_user( localpart=body["username"].lower(), - password=body["password"], + password_hash=password_hash, admin=bool(admin), user_type=user_type, ) diff --git a/synapse/rest/client/v1/login.py b/synapse/rest/client/v1/login.py index d0d499979..4de2f97d0 100644 --- a/synapse/rest/client/v1/login.py +++ b/synapse/rest/client/v1/login.py @@ -14,11 +14,6 @@ # limitations under the License. import logging -import xml.etree.ElementTree as ET - -from six.moves import urllib - -from twisted.web.client import PartialDownloadError from synapse.api.errors import Codes, LoginError, SynapseError from synapse.api.ratelimiting import Ratelimiter @@ -28,10 +23,10 @@ from synapse.http.servlet import ( parse_json_object_from_request, parse_string, ) -from synapse.push.mailer import load_jinja2_templates +from synapse.http.site import SynapseRequest from synapse.rest.client.v2_alpha._base import client_patterns from synapse.rest.well_known import WellKnownBuilder -from synapse.types import UserID, map_username_to_mxid_localpart +from synapse.types import UserID from synapse.util.msisdn import phone_number_to_msisdn logger = logging.getLogger(__name__) @@ -402,7 +397,7 @@ class BaseSSORedirectServlet(RestServlet): PATTERNS = client_patterns("/login/(cas|sso)/redirect", v1=True) - def on_GET(self, request): + def on_GET(self, request: SynapseRequest): args = request.args if b"redirectUrl" not in args: return 400, "Redirect URL not specified for SSO auth" @@ -411,15 +406,15 @@ class BaseSSORedirectServlet(RestServlet): request.redirect(sso_url) finish_request(request) - def get_sso_url(self, client_redirect_url): + def get_sso_url(self, client_redirect_url: bytes) -> bytes: """Get the URL to redirect to, to perform SSO auth Args: - client_redirect_url (bytes): the URL that we should redirect the + client_redirect_url: the URL that we should redirect the client to when everything is done Returns: - bytes: URL to redirect to + URL to redirect to """ # to be implemented by subclasses raise NotImplementedError() @@ -427,19 +422,12 @@ class BaseSSORedirectServlet(RestServlet): class CasRedirectServlet(BaseSSORedirectServlet): def __init__(self, hs): - super(CasRedirectServlet, self).__init__() - self.cas_server_url = hs.config.cas_server_url.encode("ascii") - self.cas_service_url = hs.config.cas_service_url.encode("ascii") + self._cas_handler = hs.get_cas_handler() - def get_sso_url(self, client_redirect_url): - client_redirect_url_param = urllib.parse.urlencode( - {b"redirectUrl": client_redirect_url} + def get_sso_url(self, client_redirect_url: bytes) -> bytes: + return self._cas_handler.get_redirect_url( + {"redirectUrl": client_redirect_url} ).encode("ascii") - hs_redirect_url = self.cas_service_url + b"/_matrix/client/r0/login/cas/ticket" - service_param = urllib.parse.urlencode( - {b"service": b"%s?%s" % (hs_redirect_url, client_redirect_url_param)} - ).encode("ascii") - return b"%s/login?%s" % (self.cas_server_url, service_param) class CasTicketServlet(RestServlet): @@ -447,81 +435,25 @@ class CasTicketServlet(RestServlet): def __init__(self, hs): super(CasTicketServlet, self).__init__() - self.cas_server_url = hs.config.cas_server_url - self.cas_service_url = hs.config.cas_service_url - self.cas_displayname_attribute = hs.config.cas_displayname_attribute - self.cas_required_attributes = hs.config.cas_required_attributes - self._sso_auth_handler = SSOAuthHandler(hs) - self._http_client = hs.get_proxied_http_client() + self._cas_handler = hs.get_cas_handler() - async def on_GET(self, request): - client_redirect_url = parse_string(request, "redirectUrl", required=True) - uri = self.cas_server_url + "/proxyValidate" - args = { - "ticket": parse_string(request, "ticket", required=True), - "service": self.cas_service_url, - } - try: - body = await self._http_client.get_raw(uri, args) - except PartialDownloadError as pde: - # Twisted raises this error if the connection is closed, - # even if that's being used old-http style to signal end-of-data - body = pde.response - result = await self.handle_cas_response(request, body, client_redirect_url) - return result + async def on_GET(self, request: SynapseRequest) -> None: + client_redirect_url = parse_string(request, "redirectUrl") + ticket = parse_string(request, "ticket", required=True) - def handle_cas_response(self, request, cas_response_body, client_redirect_url): - user, attributes = self.parse_cas_response(cas_response_body) - displayname = attributes.pop(self.cas_displayname_attribute, None) + # Maybe get a session ID (if this ticket is from user interactive + # authentication). + session = parse_string(request, "session") - for required_attribute, required_value in self.cas_required_attributes.items(): - # If required attribute was not in CAS Response - Forbidden - if required_attribute not in attributes: - raise LoginError(401, "Unauthorized", errcode=Codes.UNAUTHORIZED) + # Either client_redirect_url or session must be provided. + if not client_redirect_url and not session: + message = "Missing string query parameter redirectUrl or session" + raise SynapseError(400, message, errcode=Codes.MISSING_PARAM) - # Also need to check value - if required_value is not None: - actual_value = attributes[required_attribute] - # If required attribute value does not match expected - Forbidden - if required_value != actual_value: - raise LoginError(401, "Unauthorized", errcode=Codes.UNAUTHORIZED) - - return self._sso_auth_handler.on_successful_auth( - user, request, client_redirect_url, displayname + await self._cas_handler.handle_ticket( + request, ticket, client_redirect_url, session ) - def parse_cas_response(self, cas_response_body): - user = None - attributes = {} - try: - root = ET.fromstring(cas_response_body) - if not root.tag.endswith("serviceResponse"): - raise Exception("root of CAS response is not serviceResponse") - success = root[0].tag.endswith("authenticationSuccess") - for child in root[0]: - if child.tag.endswith("user"): - user = child.text - if child.tag.endswith("attributes"): - for attribute in child: - # ElementTree library expands the namespace in - # attribute tags to the full URL of the namespace. - # We don't care about namespace here and it will always - # be encased in curly braces, so we remove them. - tag = attribute.tag - if "}" in tag: - tag = tag.split("}")[1] - attributes[tag] = attribute.text - if user is None: - raise Exception("CAS response does not contain user") - except Exception: - logger.exception("Error parsing CAS response") - raise LoginError(401, "Invalid CAS response", errcode=Codes.UNAUTHORIZED) - if not success: - raise LoginError( - 401, "Unsuccessful CAS response", errcode=Codes.UNAUTHORIZED - ) - return user, attributes - class SAMLRedirectServlet(BaseSSORedirectServlet): PATTERNS = client_patterns("/login/sso/redirect", v1=True) @@ -529,72 +461,10 @@ class SAMLRedirectServlet(BaseSSORedirectServlet): def __init__(self, hs): self._saml_handler = hs.get_saml_handler() - def get_sso_url(self, client_redirect_url): + def get_sso_url(self, client_redirect_url: bytes) -> bytes: return self._saml_handler.handle_redirect_request(client_redirect_url) -class SSOAuthHandler(object): - """ - Utility class for Resources and Servlets which handle the response from a SSO - service - - Args: - hs (synapse.server.HomeServer) - """ - - def __init__(self, hs): - self._hostname = hs.hostname - self._auth_handler = hs.get_auth_handler() - self._registration_handler = hs.get_registration_handler() - self._macaroon_gen = hs.get_macaroon_generator() - - # Load the redirect page HTML template - self._template = load_jinja2_templates( - hs.config.sso_redirect_confirm_template_dir, ["sso_redirect_confirm.html"], - )[0] - - self._server_name = hs.config.server_name - - # cast to tuple for use with str.startswith - self._whitelisted_sso_clients = tuple(hs.config.sso_client_whitelist) - - async def on_successful_auth( - self, username, request, client_redirect_url, user_display_name=None - ): - """Called once the user has successfully authenticated with the SSO. - - Registers the user if necessary, and then returns a redirect (with - a login token) to the client. - - Args: - username (unicode|bytes): the remote user id. We'll map this onto - something sane for a MXID localpath. - - request (SynapseRequest): the incoming request from the browser. We'll - respond to it with a redirect. - - client_redirect_url (unicode): the redirect_url the client gave us when - it first started the process. - - user_display_name (unicode|None): if set, and we have to register a new user, - we will set their displayname to this. - - Returns: - Deferred[none]: Completes once we have handled the request. - """ - localpart = map_username_to_mxid_localpart(username) - user_id = UserID(localpart, self._hostname).to_string() - registered_user_id = await self._auth_handler.check_user_exists(user_id) - if not registered_user_id: - registered_user_id = await self._registration_handler.register_user( - localpart=localpart, default_display_name=user_display_name - ) - - self._auth_handler.complete_sso_login( - registered_user_id, request, client_redirect_url - ) - - def register_servlets(hs, http_server): LoginRestServlet(hs).register(http_server) if hs.config.cas_enabled: diff --git a/synapse/rest/client/v1/room.py b/synapse/rest/client/v1/room.py index bffd43de5..6b5830cc3 100644 --- a/synapse/rest/client/v1/room.py +++ b/synapse/rest/client/v1/room.py @@ -27,6 +27,7 @@ from synapse.api.constants import EventTypes, Membership from synapse.api.errors import ( AuthError, Codes, + HttpResponseException, InvalidClientCredentialsError, SynapseError, ) @@ -364,10 +365,13 @@ class PublicRoomListRestServlet(TransactionRestServlet): limit = None handler = self.hs.get_room_list_handler() - if server: - data = await handler.get_remote_public_room_list( - server, limit=limit, since_token=since_token - ) + if server and server != self.hs.config.server_name: + try: + data = await handler.get_remote_public_room_list( + server, limit=limit, since_token=since_token + ) + except HttpResponseException as e: + raise e.to_synapse_error() else: data = await handler.get_local_public_room_list( limit=limit, since_token=since_token @@ -404,15 +408,18 @@ class PublicRoomListRestServlet(TransactionRestServlet): limit = None handler = self.hs.get_room_list_handler() - if server: - data = await handler.get_remote_public_room_list( - server, - limit=limit, - since_token=since_token, - search_filter=search_filter, - include_all_networks=include_all_networks, - third_party_instance_id=third_party_instance_id, - ) + if server and server != self.hs.config.server_name: + try: + data = await handler.get_remote_public_room_list( + server, + limit=limit, + since_token=since_token, + search_filter=search_filter, + include_all_networks=include_all_networks, + third_party_instance_id=third_party_instance_id, + ) + except HttpResponseException as e: + raise e.to_synapse_error() else: data = await handler.get_local_public_room_list( limit=limit, diff --git a/synapse/rest/client/v2_alpha/account.py b/synapse/rest/client/v2_alpha/account.py index 631cc74cb..1bd023477 100644 --- a/synapse/rest/client/v2_alpha/account.py +++ b/synapse/rest/client/v2_alpha/account.py @@ -30,7 +30,7 @@ from synapse.http.servlet import ( ) from synapse.push.mailer import Mailer, load_jinja2_templates from synapse.util.msisdn import phone_number_to_msisdn -from synapse.util.stringutils import assert_valid_client_secret +from synapse.util.stringutils import assert_valid_client_secret, random_string from synapse.util.threepids import check_3pid_allowed from ._base import client_patterns, interactive_auth_handler @@ -100,6 +100,11 @@ class EmailPasswordRequestTokenRestServlet(RestServlet): ) if existing_user_id is None: + if self.config.request_token_inhibit_3pid_errors: + # Make the client think the operation succeeded. See the rationale in the + # comments for request_token_inhibit_3pid_errors. + return 200, {"sid": random_string(16)} + raise SynapseError(400, "Email not found", Codes.THREEPID_NOT_FOUND) if self.config.threepid_behaviour_email == ThreepidBehaviour.REMOTE: @@ -234,13 +239,21 @@ class PasswordRestServlet(RestServlet): if self.auth.has_access_token(request): requester = await self.auth.get_user_by_req(request) params = await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "modify your account password", ) user_id = requester.user.to_string() else: requester = None result, params, _ = await self.auth_handler.check_auth( - [[LoginType.EMAIL_IDENTITY]], body, self.hs.get_ip_from_request(request) + [[LoginType.EMAIL_IDENTITY]], + request, + body, + self.hs.get_ip_from_request(request), + "modify your account password", ) if LoginType.EMAIL_IDENTITY in result: @@ -308,7 +321,11 @@ class DeactivateAccountRestServlet(RestServlet): return 200, {} await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "deactivate your account", ) result = await self._deactivate_account_handler.deactivate_account( requester.user.to_string(), erase, id_server=body.get("id_server") @@ -378,6 +395,11 @@ class EmailThreepidRequestTokenRestServlet(RestServlet): ) if existing_user_id is not None: + if self.config.request_token_inhibit_3pid_errors: + # Make the client think the operation succeeded. See the rationale in the + # comments for request_token_inhibit_3pid_errors. + return 200, {"sid": random_string(16)} + raise SynapseError(400, "Email is already in use", Codes.THREEPID_IN_USE) if self.config.threepid_behaviour_email == ThreepidBehaviour.REMOTE: @@ -441,6 +463,11 @@ class MsisdnThreepidRequestTokenRestServlet(RestServlet): existing_user_id = await self.store.get_user_id_by_threepid("msisdn", msisdn) if existing_user_id is not None: + if self.hs.config.request_token_inhibit_3pid_errors: + # Make the client think the operation succeeded. See the rationale in the + # comments for request_token_inhibit_3pid_errors. + return 200, {"sid": random_string(16)} + raise SynapseError(400, "MSISDN is already in use", Codes.THREEPID_IN_USE) if not self.hs.config.account_threepid_delegate_msisdn: @@ -602,6 +629,11 @@ class ThreepidRestServlet(RestServlet): return 200, {"threepids": threepids} async def on_POST(self, request): + if not self.hs.config.enable_3pid_changes: + raise SynapseError( + 400, "3PID changes are disabled on this server", Codes.FORBIDDEN + ) + requester = await self.auth.get_user_by_req(request) user_id = requester.user.to_string() body = parse_json_object_from_request(request) @@ -646,6 +678,11 @@ class ThreepidAddRestServlet(RestServlet): @interactive_auth_handler async def on_POST(self, request): + if not self.hs.config.enable_3pid_changes: + raise SynapseError( + 400, "3PID changes are disabled on this server", Codes.FORBIDDEN + ) + requester = await self.auth.get_user_by_req(request) user_id = requester.user.to_string() body = parse_json_object_from_request(request) @@ -656,7 +693,11 @@ class ThreepidAddRestServlet(RestServlet): assert_valid_client_secret(client_secret) await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "add a third-party identifier to your account", ) validation_session = await self.identity_handler.validate_threepid_session( @@ -741,10 +782,16 @@ class ThreepidDeleteRestServlet(RestServlet): def __init__(self, hs): super(ThreepidDeleteRestServlet, self).__init__() + self.hs = hs self.auth = hs.get_auth() self.auth_handler = hs.get_auth_handler() async def on_POST(self, request): + if not self.hs.config.enable_3pid_changes: + raise SynapseError( + 400, "3PID changes are disabled on this server", Codes.FORBIDDEN + ) + body = parse_json_object_from_request(request) assert_params_in_dict(body, ["medium", "address"]) diff --git a/synapse/rest/client/v2_alpha/auth.py b/synapse/rest/client/v2_alpha/auth.py index 50e080673..24dd3d3e9 100644 --- a/synapse/rest/client/v2_alpha/auth.py +++ b/synapse/rest/client/v2_alpha/auth.py @@ -130,7 +130,17 @@ class AuthRestServlet(RestServlet): self.auth_handler = hs.get_auth_handler() self.registration_handler = hs.get_registration_handler() - def on_GET(self, request, stagetype): + # SSO configuration. + self._saml_enabled = hs.config.saml2_enabled + if self._saml_enabled: + self._saml_handler = hs.get_saml_handler() + self._cas_enabled = hs.config.cas_enabled + if self._cas_enabled: + self._cas_handler = hs.get_cas_handler() + self._cas_server_url = hs.config.cas_server_url + self._cas_service_url = hs.config.cas_service_url + + async def on_GET(self, request, stagetype): session = parse_string(request, "session") if not session: raise SynapseError(400, "No session supplied") @@ -142,14 +152,6 @@ class AuthRestServlet(RestServlet): % (CLIENT_API_PREFIX, LoginType.RECAPTCHA), "sitekey": self.hs.config.recaptcha_public_key, } - html_bytes = html.encode("utf8") - request.setResponseCode(200) - request.setHeader(b"Content-Type", b"text/html; charset=utf-8") - request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) - - request.write(html_bytes) - finish_request(request) - return None elif stagetype == LoginType.TERMS: html = TERMS_TEMPLATE % { "session": session, @@ -158,17 +160,41 @@ class AuthRestServlet(RestServlet): "myurl": "%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.TERMS), } - html_bytes = html.encode("utf8") - request.setResponseCode(200) - request.setHeader(b"Content-Type", b"text/html; charset=utf-8") - request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) - request.write(html_bytes) - finish_request(request) - return None + elif stagetype == LoginType.SSO: + # Display a confirmation page which prompts the user to + # re-authenticate with their SSO provider. + if self._cas_enabled: + # Generate a request to CAS that redirects back to an endpoint + # to verify the successful authentication. + sso_redirect_url = self._cas_handler.get_redirect_url( + {"session": session}, + ) + + elif self._saml_enabled: + client_redirect_url = "" + sso_redirect_url = self._saml_handler.handle_redirect_request( + client_redirect_url, session + ) + + else: + raise SynapseError(400, "Homeserver not configured for SSO.") + + html = await self.auth_handler.start_sso_ui_auth(sso_redirect_url, session) + else: raise SynapseError(404, "Unknown auth stage type") + # Render the HTML and return. + html_bytes = html.encode("utf8") + request.setResponseCode(200) + request.setHeader(b"Content-Type", b"text/html; charset=utf-8") + request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) + + request.write(html_bytes) + finish_request(request) + return None + async def on_POST(self, request, stagetype): session = parse_string(request, "session") @@ -196,15 +222,6 @@ class AuthRestServlet(RestServlet): % (CLIENT_API_PREFIX, LoginType.RECAPTCHA), "sitekey": self.hs.config.recaptcha_public_key, } - html_bytes = html.encode("utf8") - request.setResponseCode(200) - request.setHeader(b"Content-Type", b"text/html; charset=utf-8") - request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) - - request.write(html_bytes) - finish_request(request) - - return None elif stagetype == LoginType.TERMS: authdict = {"session": session} @@ -225,17 +242,22 @@ class AuthRestServlet(RestServlet): "myurl": "%s/r0/auth/%s/fallback/web" % (CLIENT_API_PREFIX, LoginType.TERMS), } - html_bytes = html.encode("utf8") - request.setResponseCode(200) - request.setHeader(b"Content-Type", b"text/html; charset=utf-8") - request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) - - request.write(html_bytes) - finish_request(request) - return None + elif stagetype == LoginType.SSO: + # The SSO fallback workflow should not post here, + raise SynapseError(404, "Fallback SSO auth does not support POST requests.") else: raise SynapseError(404, "Unknown auth stage type") + # Render the HTML and return. + html_bytes = html.encode("utf8") + request.setResponseCode(200) + request.setHeader(b"Content-Type", b"text/html; charset=utf-8") + request.setHeader(b"Content-Length", b"%d" % (len(html_bytes),)) + + request.write(html_bytes) + finish_request(request) + return None + def on_OPTIONS(self, _): return 200, {} diff --git a/synapse/rest/client/v2_alpha/devices.py b/synapse/rest/client/v2_alpha/devices.py index 94ff73f38..c0714fcfb 100644 --- a/synapse/rest/client/v2_alpha/devices.py +++ b/synapse/rest/client/v2_alpha/devices.py @@ -81,7 +81,11 @@ class DeleteDevicesRestServlet(RestServlet): assert_params_in_dict(body, ["devices"]) await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "remove device(s) from your account", ) await self.device_handler.delete_devices( @@ -127,7 +131,11 @@ class DeviceRestServlet(RestServlet): raise await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "remove a device from your account", ) await self.device_handler.delete_device(requester.user.to_string(), device_id) diff --git a/synapse/rest/client/v2_alpha/keys.py b/synapse/rest/client/v2_alpha/keys.py index f7ed4daf9..8f41a3edb 100644 --- a/synapse/rest/client/v2_alpha/keys.py +++ b/synapse/rest/client/v2_alpha/keys.py @@ -263,7 +263,11 @@ class SigningKeyUploadServlet(RestServlet): body = parse_json_object_from_request(request) await self.auth_handler.validate_user_via_ui_auth( - requester, body, self.hs.get_ip_from_request(request) + requester, + request, + body, + self.hs.get_ip_from_request(request), + "add a device signing key to your account", ) result = await self.e2e_keys_handler.upload_signing_keys_for_user(user_id, body) diff --git a/synapse/rest/client/v2_alpha/password_policy.py b/synapse/rest/client/v2_alpha/password_policy.py new file mode 100644 index 000000000..968403cca --- /dev/null +++ b/synapse/rest/client/v2_alpha/password_policy.py @@ -0,0 +1,58 @@ +# -*- coding: utf-8 -*- +# Copyright 2019 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. + +import logging + +from synapse.http.servlet import RestServlet + +from ._base import client_patterns + +logger = logging.getLogger(__name__) + + +class PasswordPolicyServlet(RestServlet): + PATTERNS = client_patterns("/password_policy$") + + def __init__(self, hs): + """ + Args: + hs (synapse.server.HomeServer): server + """ + super(PasswordPolicyServlet, self).__init__() + + self.policy = hs.config.password_policy + self.enabled = hs.config.password_policy_enabled + + def on_GET(self, request): + if not self.enabled or not self.policy: + return (200, {}) + + policy = {} + + for param in [ + "minimum_length", + "require_digit", + "require_symbol", + "require_lowercase", + "require_uppercase", + ]: + if param in self.policy: + policy["m.%s" % param] = self.policy[param] + + return (200, policy) + + +def register_servlets(hs, http_server): + PasswordPolicyServlet(hs).register(http_server) diff --git a/synapse/rest/client/v2_alpha/register.py b/synapse/rest/client/v2_alpha/register.py index a09189b1b..c26927f27 100644 --- a/synapse/rest/client/v2_alpha/register.py +++ b/synapse/rest/client/v2_alpha/register.py @@ -49,7 +49,7 @@ from synapse.http.servlet import ( from synapse.push.mailer import load_jinja2_templates from synapse.util.msisdn import phone_number_to_msisdn from synapse.util.ratelimitutils import FederationRateLimiter -from synapse.util.stringutils import assert_valid_client_secret +from synapse.util.stringutils import assert_valid_client_secret, random_string from synapse.util.threepids import check_3pid_allowed from ._base import client_patterns, interactive_auth_handler @@ -135,6 +135,11 @@ class EmailRegisterRequestTokenRestServlet(RestServlet): ) if existing_user_id is not None: + if self.hs.config.request_token_inhibit_3pid_errors: + # Make the client think the operation succeeded. See the rationale in the + # comments for request_token_inhibit_3pid_errors. + return 200, {"sid": random_string(16)} + raise SynapseError(400, "Email is already in use", Codes.THREEPID_IN_USE) if self.config.threepid_behaviour_email == ThreepidBehaviour.REMOTE: @@ -202,6 +207,11 @@ class MsisdnRegisterRequestTokenRestServlet(RestServlet): ) if existing_user_id is not None: + if self.hs.config.request_token_inhibit_3pid_errors: + # Make the client think the operation succeeded. See the rationale in the + # comments for request_token_inhibit_3pid_errors. + return 200, {"sid": random_string(16)} + raise SynapseError( 400, "Phone number is already in use", Codes.THREEPID_IN_USE ) @@ -373,6 +383,7 @@ class RegisterRestServlet(RestServlet): self.room_member_handler = hs.get_room_member_handler() self.macaroon_gen = hs.get_macaroon_generator() self.ratelimiter = hs.get_registration_ratelimiter() + self.password_policy_handler = hs.get_password_policy_handler() self.clock = hs.get_clock() self._registration_flows = _calculate_registration_flows( @@ -415,11 +426,16 @@ class RegisterRestServlet(RestServlet): # we do basic sanity checks here because the auth layer will store these # in sessions. Pull out the username/password provided to us. if "password" in body: - if ( - not isinstance(body["password"], string_types) - or len(body["password"]) > 512 - ): + password = body.pop("password") + if not isinstance(password, string_types) or len(password) > 512: raise SynapseError(400, "Invalid password") + self.password_policy_handler.validate_password(password) + + # If the password is valid, hash it and store it back on the request. + # This ensures the hashed password is handled everywhere. + if "password_hash" in body: + raise SynapseError(400, "Unexpected property: password_hash") + body["password_hash"] = await self.auth_handler.hash(password) desired_username = None if "username" in body: @@ -472,7 +488,7 @@ class RegisterRestServlet(RestServlet): guest_access_token = body.get("guest_access_token", None) - if "initial_device_display_name" in body and "password" not in body: + if "initial_device_display_name" in body and "password_hash" not in body: # ignore 'initial_device_display_name' if sent without # a password to work around a client bug where it sent # the 'initial_device_display_name' param alone, wiping out @@ -487,7 +503,7 @@ class RegisterRestServlet(RestServlet): # registered a user for this session, so we could just return the # user here. We carry on and go through the auth checks though, # for paranoia. - registered_user_id = self.auth_handler.get_session_data( + registered_user_id = await self.auth_handler.get_session_data( session_id, "registered_user_id", None ) @@ -499,7 +515,11 @@ class RegisterRestServlet(RestServlet): ) auth_result, params, session_id = await self.auth_handler.check_auth( - self._registration_flows, body, self.hs.get_ip_from_request(request) + self._registration_flows, + request, + body, + self.hs.get_ip_from_request(request), + "register a new account", ) # Check that we're not trying to register a denied 3pid. @@ -530,11 +550,11 @@ class RegisterRestServlet(RestServlet): registered = False else: # NB: This may be from the auth handler and NOT from the POST - assert_params_in_dict(params, ["password"]) + assert_params_in_dict(params, ["password_hash"]) desired_username = params.get("username", None) guest_access_token = params.get("guest_access_token", None) - new_password = params.get("password", None) + new_password_hash = params.get("password_hash", None) if desired_username is not None: desired_username = desired_username.lower() @@ -567,7 +587,7 @@ class RegisterRestServlet(RestServlet): registered_user_id = await self.registration_handler.register_user( localpart=desired_username, - password=new_password, + password_hash=new_password_hash, guest_access_token=guest_access_token, threepid=threepid, address=client_addr, @@ -582,7 +602,7 @@ class RegisterRestServlet(RestServlet): # remember that we've now registered that user account, and with # what user ID (since the user may not have specified) - self.auth_handler.set_session_data( + await self.auth_handler.set_session_data( session_id, "registered_user_id", registered_user_id ) diff --git a/synapse/rest/client/v2_alpha/room_keys.py b/synapse/rest/client/v2_alpha/room_keys.py index 38952a1d2..59529707d 100644 --- a/synapse/rest/client/v2_alpha/room_keys.py +++ b/synapse/rest/client/v2_alpha/room_keys.py @@ -188,7 +188,7 @@ class RoomKeysServlet(RestServlet): """ requester = await self.auth.get_user_by_req(request, allow_guest=False) user_id = requester.user.to_string() - version = parse_string(request, "version") + version = parse_string(request, "version", required=True) room_keys = await self.e2e_room_keys_handler.get_room_keys( user_id, version, room_id, session_id diff --git a/synapse/rest/media/v1/download_resource.py b/synapse/rest/media/v1/download_resource.py index 66a01559e..24d3ae5bb 100644 --- a/synapse/rest/media/v1/download_resource.py +++ b/synapse/rest/media/v1/download_resource.py @@ -50,6 +50,9 @@ class DownloadResource(DirectServeResource): b" media-src 'self';" b" object-src 'self';", ) + request.setHeader( + b"Referrer-Policy", b"no-referrer", + ) server_name, media_id, name = parse_media_id(request) if server_name == self.server_name: await self.media_repo.get_local_media(request, media_id, name) diff --git a/synapse/rest/media/v1/media_repository.py b/synapse/rest/media/v1/media_repository.py index 490b1b45a..fd10d42f2 100644 --- a/synapse/rest/media/v1/media_repository.py +++ b/synapse/rest/media/v1/media_repository.py @@ -24,7 +24,6 @@ from six import iteritems import twisted.internet.error import twisted.web.http -from twisted.internet import defer from twisted.web.resource import Resource from synapse.api.errors import ( @@ -114,15 +113,14 @@ class MediaRepository(object): "update_recently_accessed_media", self._update_recently_accessed ) - @defer.inlineCallbacks - def _update_recently_accessed(self): + async def _update_recently_accessed(self): remote_media = self.recently_accessed_remotes self.recently_accessed_remotes = set() local_media = self.recently_accessed_locals self.recently_accessed_locals = set() - yield self.store.update_cached_last_access_time( + await self.store.update_cached_last_access_time( local_media, remote_media, self.clock.time_msec() ) @@ -138,8 +136,7 @@ class MediaRepository(object): else: self.recently_accessed_locals.add(media_id) - @defer.inlineCallbacks - def create_content( + async def create_content( self, media_type, upload_name, content, content_length, auth_user ): """Store uploaded content for a local user and return the mxc URL @@ -158,11 +155,11 @@ class MediaRepository(object): file_info = FileInfo(server_name=None, file_id=media_id) - fname = yield self.media_storage.store_file(content, file_info) + fname = await self.media_storage.store_file(content, file_info) logger.info("Stored local media in file %r", fname) - yield self.store.store_local_media( + await self.store.store_local_media( media_id=media_id, media_type=media_type, time_now_ms=self.clock.time_msec(), @@ -171,12 +168,11 @@ class MediaRepository(object): user_id=auth_user, ) - yield self._generate_thumbnails(None, media_id, media_id, media_type) + await self._generate_thumbnails(None, media_id, media_id, media_type) return "mxc://%s/%s" % (self.server_name, media_id) - @defer.inlineCallbacks - def get_local_media(self, request, media_id, name): + async def get_local_media(self, request, media_id, name): """Responds to reqests for local media, if exists, or returns 404. Args: @@ -190,7 +186,7 @@ class MediaRepository(object): Deferred: Resolves once a response has successfully been written to request """ - media_info = yield self.store.get_local_media(media_id) + media_info = await self.store.get_local_media(media_id) if not media_info or media_info["quarantined_by"]: respond_404(request) return @@ -204,13 +200,12 @@ class MediaRepository(object): file_info = FileInfo(None, media_id, url_cache=url_cache) - responder = yield self.media_storage.fetch_media(file_info) - yield respond_with_responder( + responder = await self.media_storage.fetch_media(file_info) + await respond_with_responder( request, responder, media_type, media_length, upload_name ) - @defer.inlineCallbacks - def get_remote_media(self, request, server_name, media_id, name): + async def get_remote_media(self, request, server_name, media_id, name): """Respond to requests for remote media. Args: @@ -236,8 +231,8 @@ class MediaRepository(object): # We linearize here to ensure that we don't try and download remote # media multiple times concurrently key = (server_name, media_id) - with (yield self.remote_media_linearizer.queue(key)): - responder, media_info = yield self._get_remote_media_impl( + with (await self.remote_media_linearizer.queue(key)): + responder, media_info = await self._get_remote_media_impl( server_name, media_id ) @@ -246,14 +241,13 @@ class MediaRepository(object): media_type = media_info["media_type"] media_length = media_info["media_length"] upload_name = name if name else media_info["upload_name"] - yield respond_with_responder( + await respond_with_responder( request, responder, media_type, media_length, upload_name ) else: respond_404(request) - @defer.inlineCallbacks - def get_remote_media_info(self, server_name, media_id): + async def get_remote_media_info(self, server_name, media_id): """Gets the media info associated with the remote file, downloading if necessary. @@ -274,8 +268,8 @@ class MediaRepository(object): # We linearize here to ensure that we don't try and download remote # media multiple times concurrently key = (server_name, media_id) - with (yield self.remote_media_linearizer.queue(key)): - responder, media_info = yield self._get_remote_media_impl( + with (await self.remote_media_linearizer.queue(key)): + responder, media_info = await self._get_remote_media_impl( server_name, media_id ) @@ -286,8 +280,7 @@ class MediaRepository(object): return media_info - @defer.inlineCallbacks - def _get_remote_media_impl(self, server_name, media_id): + async def _get_remote_media_impl(self, server_name, media_id): """Looks for media in local cache, if not there then attempt to download from remote server. @@ -299,7 +292,7 @@ class MediaRepository(object): Returns: Deferred[(Responder, media_info)] """ - media_info = yield self.store.get_cached_remote_media(server_name, media_id) + media_info = await self.store.get_cached_remote_media(server_name, media_id) # file_id is the ID we use to track the file locally. If we've already # seen the file then reuse the existing ID, otherwise genereate a new @@ -317,19 +310,18 @@ class MediaRepository(object): logger.info("Media is quarantined") raise NotFoundError() - responder = yield self.media_storage.fetch_media(file_info) + responder = await self.media_storage.fetch_media(file_info) if responder: return responder, media_info # Failed to find the file anywhere, lets download it. - media_info = yield self._download_remote_file(server_name, media_id, file_id) + media_info = await self._download_remote_file(server_name, media_id, file_id) - responder = yield self.media_storage.fetch_media(file_info) + responder = await self.media_storage.fetch_media(file_info) return responder, media_info - @defer.inlineCallbacks - def _download_remote_file(self, server_name, media_id, file_id): + async def _download_remote_file(self, server_name, media_id, file_id): """Attempt to download the remote file from the given server name, using the given file_id as the local id. @@ -351,7 +343,7 @@ class MediaRepository(object): ("/_matrix/media/v1/download", server_name, media_id) ) try: - length, headers = yield self.client.get_file( + length, headers = await self.client.get_file( server_name, request_path, output_stream=f, @@ -397,7 +389,7 @@ class MediaRepository(object): ) raise SynapseError(502, "Failed to fetch remote media") - yield finish() + await finish() media_type = headers[b"Content-Type"][0].decode("ascii") upload_name = get_filename_from_headers(headers) @@ -405,7 +397,7 @@ class MediaRepository(object): logger.info("Stored remote media in file %r", fname) - yield self.store.store_cached_remote_media( + await self.store.store_cached_remote_media( origin=server_name, media_id=media_id, media_type=media_type, @@ -423,7 +415,7 @@ class MediaRepository(object): "filesystem_id": file_id, } - yield self._generate_thumbnails(server_name, media_id, file_id, media_type) + await self._generate_thumbnails(server_name, media_id, file_id, media_type) return media_info @@ -458,16 +450,15 @@ class MediaRepository(object): return t_byte_source - @defer.inlineCallbacks - def generate_local_exact_thumbnail( + async def generate_local_exact_thumbnail( self, media_id, t_width, t_height, t_method, t_type, url_cache ): - input_path = yield self.media_storage.ensure_media_is_in_local_cache( + input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(None, media_id, url_cache=url_cache) ) thumbnailer = Thumbnailer(input_path) - t_byte_source = yield defer_to_thread( + t_byte_source = await defer_to_thread( self.hs.get_reactor(), self._generate_thumbnail, thumbnailer, @@ -490,7 +481,7 @@ class MediaRepository(object): thumbnail_type=t_type, ) - output_path = yield self.media_storage.store_file( + output_path = await self.media_storage.store_file( t_byte_source, file_info ) finally: @@ -500,22 +491,21 @@ class MediaRepository(object): t_len = os.path.getsize(output_path) - yield self.store.store_local_thumbnail( + await self.store.store_local_thumbnail( media_id, t_width, t_height, t_type, t_method, t_len ) return output_path - @defer.inlineCallbacks - def generate_remote_exact_thumbnail( + async def generate_remote_exact_thumbnail( self, server_name, file_id, media_id, t_width, t_height, t_method, t_type ): - input_path = yield self.media_storage.ensure_media_is_in_local_cache( + input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(server_name, file_id, url_cache=False) ) thumbnailer = Thumbnailer(input_path) - t_byte_source = yield defer_to_thread( + t_byte_source = await defer_to_thread( self.hs.get_reactor(), self._generate_thumbnail, thumbnailer, @@ -537,7 +527,7 @@ class MediaRepository(object): thumbnail_type=t_type, ) - output_path = yield self.media_storage.store_file( + output_path = await self.media_storage.store_file( t_byte_source, file_info ) finally: @@ -547,7 +537,7 @@ class MediaRepository(object): t_len = os.path.getsize(output_path) - yield self.store.store_remote_media_thumbnail( + await self.store.store_remote_media_thumbnail( server_name, media_id, file_id, @@ -560,8 +550,7 @@ class MediaRepository(object): return output_path - @defer.inlineCallbacks - def _generate_thumbnails( + async def _generate_thumbnails( self, server_name, media_id, file_id, media_type, url_cache=False ): """Generate and store thumbnails for an image. @@ -582,7 +571,7 @@ class MediaRepository(object): if not requirements: return - input_path = yield self.media_storage.ensure_media_is_in_local_cache( + input_path = await self.media_storage.ensure_media_is_in_local_cache( FileInfo(server_name, file_id, url_cache=url_cache) ) @@ -600,7 +589,7 @@ class MediaRepository(object): return if thumbnailer.transpose_method is not None: - m_width, m_height = yield defer_to_thread( + m_width, m_height = await defer_to_thread( self.hs.get_reactor(), thumbnailer.transpose ) @@ -620,11 +609,11 @@ class MediaRepository(object): for (t_width, t_height, t_type), t_method in iteritems(thumbnails): # Generate the thumbnail if t_method == "crop": - t_byte_source = yield defer_to_thread( + t_byte_source = await defer_to_thread( self.hs.get_reactor(), thumbnailer.crop, t_width, t_height, t_type ) elif t_method == "scale": - t_byte_source = yield defer_to_thread( + t_byte_source = await defer_to_thread( self.hs.get_reactor(), thumbnailer.scale, t_width, t_height, t_type ) else: @@ -646,7 +635,7 @@ class MediaRepository(object): url_cache=url_cache, ) - output_path = yield self.media_storage.store_file( + output_path = await self.media_storage.store_file( t_byte_source, file_info ) finally: @@ -656,7 +645,7 @@ class MediaRepository(object): # Write to database if server_name: - yield self.store.store_remote_media_thumbnail( + await self.store.store_remote_media_thumbnail( server_name, media_id, file_id, @@ -667,15 +656,14 @@ class MediaRepository(object): t_len, ) else: - yield self.store.store_local_thumbnail( + await self.store.store_local_thumbnail( media_id, t_width, t_height, t_type, t_method, t_len ) return {"width": m_width, "height": m_height} - @defer.inlineCallbacks - def delete_old_remote_media(self, before_ts): - old_media = yield self.store.get_remote_media_before(before_ts) + async def delete_old_remote_media(self, before_ts): + old_media = await self.store.get_remote_media_before(before_ts) deleted = 0 @@ -689,7 +677,7 @@ class MediaRepository(object): # TODO: Should we delete from the backup store - with (yield self.remote_media_linearizer.queue(key)): + with (await self.remote_media_linearizer.queue(key)): full_path = self.filepaths.remote_media_filepath(origin, file_id) try: os.remove(full_path) @@ -705,7 +693,7 @@ class MediaRepository(object): ) shutil.rmtree(thumbnail_dir, ignore_errors=True) - yield self.store.delete_remote_media(origin, media_id) + await self.store.delete_remote_media(origin, media_id) deleted += 1 return {"deleted": deleted} diff --git a/synapse/rest/media/v1/preview_url_resource.py b/synapse/rest/media/v1/preview_url_resource.py index 07e395cfd..f20660572 100644 --- a/synapse/rest/media/v1/preview_url_resource.py +++ b/synapse/rest/media/v1/preview_url_resource.py @@ -86,6 +86,7 @@ class PreviewUrlResource(DirectServeResource): self.media_storage = media_storage self.url_preview_url_blacklist = hs.config.url_preview_url_blacklist + self.url_preview_accept_language = hs.config.url_preview_accept_language # memory cache mapping urls to an ObservableDeferred returning # JSON-encoded OG metadata @@ -165,8 +166,7 @@ class PreviewUrlResource(DirectServeResource): og = await make_deferred_yieldable(defer.maybeDeferred(observable.observe)) respond_with_json_bytes(request, 200, og, send_cors=True) - @defer.inlineCallbacks - def _do_preview(self, url, user, ts): + async def _do_preview(self, url, user, ts): """Check the db, and download the URL and build a preview Args: @@ -179,7 +179,7 @@ class PreviewUrlResource(DirectServeResource): """ # check the URL cache in the DB (which will also provide us with # historical previews, if we have any) - cache_result = yield self.store.get_url_cache(url, ts) + cache_result = await self.store.get_url_cache(url, ts) if ( cache_result and cache_result["expires_ts"] > ts @@ -192,13 +192,13 @@ class PreviewUrlResource(DirectServeResource): og = og.encode("utf8") return og - media_info = yield self._download_url(url, user) + media_info = await self._download_url(url, user) logger.debug("got media_info of '%s'", media_info) if _is_media(media_info["media_type"]): file_id = media_info["filesystem_id"] - dims = yield self.media_repo._generate_thumbnails( + dims = await self.media_repo._generate_thumbnails( None, file_id, file_id, media_info["media_type"], url_cache=True ) @@ -248,14 +248,14 @@ class PreviewUrlResource(DirectServeResource): # request itself and benefit from the same caching etc. But for now we # just rely on the caching on the master request to speed things up. if "og:image" in og and og["og:image"]: - image_info = yield self._download_url( + image_info = await self._download_url( _rebase_url(og["og:image"], media_info["uri"]), user ) if _is_media(image_info["media_type"]): # TODO: make sure we don't choke on white-on-transparent images file_id = image_info["filesystem_id"] - dims = yield self.media_repo._generate_thumbnails( + dims = await self.media_repo._generate_thumbnails( None, file_id, file_id, image_info["media_type"], url_cache=True ) if dims: @@ -293,7 +293,7 @@ class PreviewUrlResource(DirectServeResource): jsonog = json.dumps(og) # store OG in history-aware DB cache - yield self.store.store_url_cache( + await self.store.store_url_cache( url, media_info["response_code"], media_info["etag"], @@ -305,8 +305,7 @@ class PreviewUrlResource(DirectServeResource): return jsonog.encode("utf8") - @defer.inlineCallbacks - def _download_url(self, url, user): + async def _download_url(self, url, user): # TODO: we should probably honour robots.txt... except in practice # we're most likely being explicitly triggered by a human rather than a # bot, so are we really a robot? @@ -317,9 +316,12 @@ class PreviewUrlResource(DirectServeResource): with self.media_storage.store_into_file(file_info) as (f, fname, finish): try: - logger.debug("Trying to get url '%s'", url) - length, headers, uri, code = yield self.client.get_file( - url, output_stream=f, max_size=self.max_spider_size + logger.debug("Trying to get preview for url '%s'", url) + length, headers, uri, code = await self.client.get_file( + url, + output_stream=f, + max_size=self.max_spider_size, + headers={"Accept-Language": self.url_preview_accept_language}, ) except SynapseError: # Pass SynapseErrors through directly, so that the servlet @@ -345,7 +347,7 @@ class PreviewUrlResource(DirectServeResource): % (traceback.format_exception_only(sys.exc_info()[0], e),), Codes.UNKNOWN, ) - yield finish() + await finish() try: if b"Content-Type" in headers: @@ -356,7 +358,7 @@ class PreviewUrlResource(DirectServeResource): download_name = get_filename_from_headers(headers) - yield self.store.store_local_media( + await self.store.store_local_media( media_id=file_id, media_type=media_type, time_now_ms=self.clock.time_msec(), @@ -393,22 +395,21 @@ class PreviewUrlResource(DirectServeResource): "expire_url_cache_data", self._expire_url_cache_data ) - @defer.inlineCallbacks - def _expire_url_cache_data(self): + async def _expire_url_cache_data(self): """Clean up expired url cache content, media and thumbnails. """ # TODO: Delete from backup media store now = self.clock.time_msec() - logger.info("Running url preview cache expiry") + logger.debug("Running url preview cache expiry") - if not (yield self.store.db.updates.has_completed_background_updates()): + if not (await self.store.db.updates.has_completed_background_updates()): logger.info("Still running DB updates; skipping expiry") return # First we delete expired url cache entries - media_ids = yield self.store.get_expired_url_cache(now) + media_ids = await self.store.get_expired_url_cache(now) removed_media = [] for media_id in media_ids: @@ -430,17 +431,19 @@ class PreviewUrlResource(DirectServeResource): except Exception: pass - yield self.store.delete_url_cache(removed_media) + await self.store.delete_url_cache(removed_media) if removed_media: logger.info("Deleted %d entries from url cache", len(removed_media)) + else: + logger.debug("No entries removed from url cache") # Now we delete old images associated with the url cache. # These may be cached for a bit on the client (i.e., they # may have a room open with a preview url thing open). # So we wait a couple of days before deleting, just in case. expire_before = now - 2 * 24 * 60 * 60 * 1000 - media_ids = yield self.store.get_url_cache_media_before(expire_before) + media_ids = await self.store.get_url_cache_media_before(expire_before) removed_media = [] for media_id in media_ids: @@ -478,9 +481,12 @@ class PreviewUrlResource(DirectServeResource): except Exception: pass - yield self.store.delete_url_cache_media(removed_media) + await self.store.delete_url_cache_media(removed_media) - logger.info("Deleted %d media from url cache", len(removed_media)) + if removed_media: + logger.info("Deleted %d media from url cache", len(removed_media)) + else: + logger.debug("No media removed from url cache") def decode_and_calc_og(body, media_uri, request_encoding=None): diff --git a/synapse/rest/media/v1/thumbnail_resource.py b/synapse/rest/media/v1/thumbnail_resource.py index d57480f76..0b8722023 100644 --- a/synapse/rest/media/v1/thumbnail_resource.py +++ b/synapse/rest/media/v1/thumbnail_resource.py @@ -16,8 +16,6 @@ import logging -from twisted.internet import defer - from synapse.http.server import ( DirectServeResource, set_cors_headers, @@ -79,11 +77,10 @@ class ThumbnailResource(DirectServeResource): ) self.media_repo.mark_recently_accessed(server_name, media_id) - @defer.inlineCallbacks - def _respond_local_thumbnail( + async def _respond_local_thumbnail( self, request, media_id, width, height, method, m_type ): - media_info = yield self.store.get_local_media(media_id) + media_info = await self.store.get_local_media(media_id) if not media_info: respond_404(request) @@ -93,7 +90,7 @@ class ThumbnailResource(DirectServeResource): respond_404(request) return - thumbnail_infos = yield self.store.get_local_media_thumbnails(media_id) + thumbnail_infos = await self.store.get_local_media_thumbnails(media_id) if thumbnail_infos: thumbnail_info = self._select_thumbnail( @@ -114,14 +111,13 @@ class ThumbnailResource(DirectServeResource): t_type = file_info.thumbnail_type t_length = thumbnail_info["thumbnail_length"] - responder = yield self.media_storage.fetch_media(file_info) - yield respond_with_responder(request, responder, t_type, t_length) + responder = await self.media_storage.fetch_media(file_info) + await respond_with_responder(request, responder, t_type, t_length) else: logger.info("Couldn't find any generated thumbnails") respond_404(request) - @defer.inlineCallbacks - def _select_or_generate_local_thumbnail( + async def _select_or_generate_local_thumbnail( self, request, media_id, @@ -130,7 +126,7 @@ class ThumbnailResource(DirectServeResource): desired_method, desired_type, ): - media_info = yield self.store.get_local_media(media_id) + media_info = await self.store.get_local_media(media_id) if not media_info: respond_404(request) @@ -140,7 +136,7 @@ class ThumbnailResource(DirectServeResource): respond_404(request) return - thumbnail_infos = yield self.store.get_local_media_thumbnails(media_id) + thumbnail_infos = await self.store.get_local_media_thumbnails(media_id) for info in thumbnail_infos: t_w = info["thumbnail_width"] == desired_width t_h = info["thumbnail_height"] == desired_height @@ -162,15 +158,15 @@ class ThumbnailResource(DirectServeResource): t_type = file_info.thumbnail_type t_length = info["thumbnail_length"] - responder = yield self.media_storage.fetch_media(file_info) + responder = await self.media_storage.fetch_media(file_info) if responder: - yield respond_with_responder(request, responder, t_type, t_length) + await respond_with_responder(request, responder, t_type, t_length) return logger.debug("We don't have a thumbnail of that size. Generating") # Okay, so we generate one. - file_path = yield self.media_repo.generate_local_exact_thumbnail( + file_path = await self.media_repo.generate_local_exact_thumbnail( media_id, desired_width, desired_height, @@ -180,13 +176,12 @@ class ThumbnailResource(DirectServeResource): ) if file_path: - yield respond_with_file(request, desired_type, file_path) + await respond_with_file(request, desired_type, file_path) else: logger.warning("Failed to generate thumbnail") respond_404(request) - @defer.inlineCallbacks - def _select_or_generate_remote_thumbnail( + async def _select_or_generate_remote_thumbnail( self, request, server_name, @@ -196,9 +191,9 @@ class ThumbnailResource(DirectServeResource): desired_method, desired_type, ): - media_info = yield self.media_repo.get_remote_media_info(server_name, media_id) + media_info = await self.media_repo.get_remote_media_info(server_name, media_id) - thumbnail_infos = yield self.store.get_remote_media_thumbnails( + thumbnail_infos = await self.store.get_remote_media_thumbnails( server_name, media_id ) @@ -224,15 +219,15 @@ class ThumbnailResource(DirectServeResource): t_type = file_info.thumbnail_type t_length = info["thumbnail_length"] - responder = yield self.media_storage.fetch_media(file_info) + responder = await self.media_storage.fetch_media(file_info) if responder: - yield respond_with_responder(request, responder, t_type, t_length) + await respond_with_responder(request, responder, t_type, t_length) return logger.debug("We don't have a thumbnail of that size. Generating") # Okay, so we generate one. - file_path = yield self.media_repo.generate_remote_exact_thumbnail( + file_path = await self.media_repo.generate_remote_exact_thumbnail( server_name, file_id, media_id, @@ -243,21 +238,20 @@ class ThumbnailResource(DirectServeResource): ) if file_path: - yield respond_with_file(request, desired_type, file_path) + await respond_with_file(request, desired_type, file_path) else: logger.warning("Failed to generate thumbnail") respond_404(request) - @defer.inlineCallbacks - def _respond_remote_thumbnail( + async def _respond_remote_thumbnail( self, request, server_name, media_id, width, height, method, m_type ): # TODO: Don't download the whole remote file # We should proxy the thumbnail from the remote server instead of # downloading the remote file and generating our own thumbnails. - media_info = yield self.media_repo.get_remote_media_info(server_name, media_id) + media_info = await self.media_repo.get_remote_media_info(server_name, media_id) - thumbnail_infos = yield self.store.get_remote_media_thumbnails( + thumbnail_infos = await self.store.get_remote_media_thumbnails( server_name, media_id ) @@ -278,8 +272,8 @@ class ThumbnailResource(DirectServeResource): t_type = file_info.thumbnail_type t_length = thumbnail_info["thumbnail_length"] - responder = yield self.media_storage.fetch_media(file_info) - yield respond_with_responder(request, responder, t_type, t_length) + responder = await self.media_storage.fetch_media(file_info) + await respond_with_responder(request, responder, t_type, t_length) else: logger.info("Failed to find any generated thumbnails") respond_404(request) diff --git a/synapse/server.py b/synapse/server.py index 1b980371d..bf97a16c0 100644 --- a/synapse/server.py +++ b/synapse/server.py @@ -56,6 +56,7 @@ from synapse.handlers.account_validity import AccountValidityHandler from synapse.handlers.acme import AcmeHandler from synapse.handlers.appservice import ApplicationServicesHandler from synapse.handlers.auth import AuthHandler, MacaroonGenerator +from synapse.handlers.cas_handler import CasHandler from synapse.handlers.deactivate_account import DeactivateAccountHandler from synapse.handlers.device import DeviceHandler, DeviceWorkerHandler from synapse.handlers.devicemessage import DeviceMessageHandler @@ -66,6 +67,7 @@ from synapse.handlers.groups_local import GroupsLocalHandler, GroupsLocalWorkerH from synapse.handlers.initial_sync import InitialSyncHandler from synapse.handlers.message import EventCreationHandler, MessageHandler from synapse.handlers.pagination import PaginationHandler +from synapse.handlers.password_policy import PasswordPolicyHandler from synapse.handlers.presence import PresenceHandler from synapse.handlers.profile import BaseProfileHandler, MasterProfileHandler from synapse.handlers.read_marker import ReadMarkerHandler @@ -85,6 +87,9 @@ from synapse.http.matrixfederationclient import MatrixFederationHttpClient from synapse.notifier import Notifier from synapse.push.action_generator import ActionGenerator from synapse.push.pusherpool import PusherPool +from synapse.replication.tcp.client import ReplicationDataHandler +from synapse.replication.tcp.handler import ReplicationCommandHandler +from synapse.replication.tcp.resource import ReplicationStreamer from synapse.rest.media.v1.media_repository import ( MediaRepository, MediaRepositoryResource, @@ -100,6 +105,7 @@ from synapse.storage import DataStores, Storage from synapse.streams.events import EventSources from synapse.util import Clock from synapse.util.distributor import Distributor +from synapse.util.stringutils import random_string logger = logging.getLogger(__name__) @@ -196,9 +202,13 @@ class HomeServer(object): "sendmail", "registration_handler", "account_validity_handler", + "cas_handler", "saml_handler", "event_client_serializer", + "password_policy_handler", "storage", + "replication_streamer", + "replication_data_handler", ] REQUIRED_ON_MASTER_STARTUP = ["user_directory_handler", "stats_handler"] @@ -224,6 +234,9 @@ class HomeServer(object): self._listening_services = [] self.start_time = None + self._instance_id = random_string(5) + self._instance_name = config.worker_name or "master" + self.clock = Clock(reactor) self.distributor = Distributor() self.ratelimiter = Ratelimiter() @@ -236,6 +249,22 @@ class HomeServer(object): for depname in kwargs: setattr(self, depname, kwargs[depname]) + def get_instance_id(self): + """A unique ID for this synapse process instance. + + This is used to distinguish running instances in worker-based + deployments. + """ + return self._instance_id + + def get_instance_name(self) -> str: + """A unique name for this synapse process. + + Used to identify the process over replication and in config. Does not + change over restarts. + """ + return self._instance_name + def setup(self): logger.info("Setting up.") self.start_time = int(self.get_clock().time()) @@ -451,7 +480,7 @@ class HomeServer(object): return ReadMarkerHandler(self) def build_tcp_replication(self): - raise NotImplementedError() + return ReplicationCommandHandler(self) def build_action_generator(self): return ActionGenerator(self) @@ -525,6 +554,9 @@ class HomeServer(object): def build_account_validity_handler(self): return AccountValidityHandler(self) + def build_cas_handler(self): + return CasHandler(self) + def build_saml_handler(self): from synapse.handlers.saml_handler import SamlHandler @@ -533,9 +565,18 @@ class HomeServer(object): def build_event_client_serializer(self): return EventClientSerializer(self) + def build_password_policy_handler(self): + return PasswordPolicyHandler(self) + def build_storage(self) -> Storage: return Storage(self, self.datastores) + def build_replication_streamer(self) -> ReplicationStreamer: + return ReplicationStreamer(self) + + def build_replication_data_handler(self): + return ReplicationDataHandler(self.get_datastore()) + def remove_pusher(self, app_id, push_key, user_id): return self.get_pusherpool().remove_pusher(app_id, push_key, user_id) @@ -557,24 +598,22 @@ def _make_dependency_method(depname): try: builder = getattr(hs, "build_%s" % (depname)) except AttributeError: - builder = None + raise NotImplementedError( + "%s has no %s nor a builder for it" % (type(hs).__name__, depname) + ) - if builder: - # Prevent cyclic dependencies from deadlocking - if depname in hs._building: - raise ValueError("Cyclic dependency while building %s" % (depname,)) - hs._building[depname] = 1 + # Prevent cyclic dependencies from deadlocking + if depname in hs._building: + raise ValueError("Cyclic dependency while building %s" % (depname,)) + hs._building[depname] = 1 + try: dep = builder() setattr(hs, depname, dep) - + finally: del hs._building[depname] - return dep - - raise NotImplementedError( - "%s has no %s nor a builder for it" % (type(hs).__name__, depname) - ) + return dep setattr(HomeServer, "get_%s" % (depname), _get) diff --git a/synapse/server.pyi b/synapse/server.pyi index 3844f0e12..18043a259 100644 --- a/synapse/server.pyi +++ b/synapse/server.pyi @@ -19,11 +19,13 @@ import synapse.handlers.set_password import synapse.http.client import synapse.notifier import synapse.replication.tcp.client +import synapse.replication.tcp.handler import synapse.rest.media.v1.media_repository import synapse.server_notices.server_notices_manager import synapse.server_notices.server_notices_sender import synapse.state import synapse.storage +from synapse.events.builder import EventBuilderFactory class HomeServer(object): @property @@ -96,7 +98,7 @@ class HomeServer(object): pass def get_notifier(self) -> synapse.notifier.Notifier: pass - def get_presence_handler(self) -> synapse.handlers.presence.PresenceHandler: + def get_presence_handler(self) -> synapse.handlers.presence.BasePresenceHandler: pass def get_clock(self) -> synapse.util.Clock: pass @@ -106,7 +108,11 @@ class HomeServer(object): pass def get_tcp_replication( self, - ) -> synapse.replication.tcp.client.ReplicationClientHandler: + ) -> synapse.replication.tcp.handler.ReplicationCommandHandler: + pass + def get_replication_data_handler( + self, + ) -> synapse.replication.tcp.client.ReplicationDataHandler: pass def get_federation_registry( self, @@ -114,3 +120,11 @@ class HomeServer(object): pass def is_mine_id(self, domain_id: str) -> bool: pass + def get_instance_id(self) -> str: + pass + def get_instance_name(self) -> str: + pass + def get_event_builder_factory(self) -> EventBuilderFactory: + pass + def get_storage(self) -> synapse.storage.Storage: + pass diff --git a/synapse/server_notices/consent_server_notices.py b/synapse/server_notices/consent_server_notices.py index 5736c5603..3bf330da4 100644 --- a/synapse/server_notices/consent_server_notices.py +++ b/synapse/server_notices/consent_server_notices.py @@ -16,8 +16,6 @@ import logging from six import iteritems, string_types -from twisted.internet import defer - from synapse.api.errors import SynapseError from synapse.api.urls import ConsentURIBuilder from synapse.config import ConfigError @@ -59,8 +57,7 @@ class ConsentServerNotices(object): self._consent_uri_builder = ConsentURIBuilder(hs.config) - @defer.inlineCallbacks - def maybe_send_server_notice_to_user(self, user_id): + async def maybe_send_server_notice_to_user(self, user_id): """Check if we need to send a notice to this user, and does so if so Args: @@ -78,7 +75,7 @@ class ConsentServerNotices(object): return self._users_in_progress.add(user_id) try: - u = yield self._store.get_user_by_id(user_id) + u = await self._store.get_user_by_id(user_id) if u["is_guest"] and not self._send_to_guests: # don't send to guests @@ -100,8 +97,8 @@ class ConsentServerNotices(object): content = copy_with_str_subst( self._server_notice_content, {"consent_uri": consent_uri} ) - yield self._server_notices_manager.send_notice(user_id, content) - yield self._store.user_set_consent_server_notice_sent( + await self._server_notices_manager.send_notice(user_id, content) + await self._store.user_set_consent_server_notice_sent( user_id, self._current_consent_version ) except SynapseError as e: diff --git a/synapse/server_notices/resource_limits_server_notices.py b/synapse/server_notices/resource_limits_server_notices.py index 9fae2e0af..d97166351 100644 --- a/synapse/server_notices/resource_limits_server_notices.py +++ b/synapse/server_notices/resource_limits_server_notices.py @@ -16,8 +16,6 @@ import logging from six import iteritems -from twisted.internet import defer - from synapse.api.constants import ( EventTypes, LimitBlockingTypes, @@ -50,8 +48,7 @@ class ResourceLimitsServerNotices(object): self._notifier = hs.get_notifier() - @defer.inlineCallbacks - def maybe_send_server_notice_to_user(self, user_id): + async def maybe_send_server_notice_to_user(self, user_id): """Check if we need to send a notice to this user, this will be true in two cases. 1. The server has reached its limit does not reflect this @@ -74,22 +71,24 @@ class ResourceLimitsServerNotices(object): # Don't try and send server notices unless they've been enabled return - timestamp = yield self._store.user_last_seen_monthly_active(user_id) + timestamp = await self._store.user_last_seen_monthly_active(user_id) if timestamp is None: # This user will be blocked from receiving the notice anyway. # In practice, not sure we can ever get here return - room_id = yield self._server_notices_manager.get_notice_room_for_user(user_id) + room_id = await self._server_notices_manager.get_or_create_notice_room_for_user( + user_id + ) if not room_id: logger.warning("Failed to get server notices room") return - yield self._check_and_set_tags(user_id, room_id) + await self._check_and_set_tags(user_id, room_id) # Determine current state of room - currently_blocked, ref_events = yield self._is_room_currently_blocked(room_id) + currently_blocked, ref_events = await self._is_room_currently_blocked(room_id) limit_msg = None limit_type = None @@ -97,7 +96,7 @@ class ResourceLimitsServerNotices(object): # Normally should always pass in user_id to check_auth_blocking # if you have it, but in this case are checking what would happen # to other users if they were to arrive. - yield self._auth.check_auth_blocking() + await self._auth.check_auth_blocking() except ResourceLimitError as e: limit_msg = e.msg limit_type = e.limit_type @@ -110,22 +109,21 @@ class ResourceLimitsServerNotices(object): # We have hit the MAU limit, but MAU alerting is disabled: # reset room if necessary and return if currently_blocked: - self._remove_limit_block_notification(user_id, ref_events) + await self._remove_limit_block_notification(user_id, ref_events) return if currently_blocked and not limit_msg: # Room is notifying of a block, when it ought not to be. - yield self._remove_limit_block_notification(user_id, ref_events) + await self._remove_limit_block_notification(user_id, ref_events) elif not currently_blocked and limit_msg: # Room is not notifying of a block, when it ought to be. - yield self._apply_limit_block_notification( + await self._apply_limit_block_notification( user_id, limit_msg, limit_type ) except SynapseError as e: logger.error("Error sending resource limits server notice: %s", e) - @defer.inlineCallbacks - def _remove_limit_block_notification(self, user_id, ref_events): + async def _remove_limit_block_notification(self, user_id, ref_events): """Utility method to remove limit block notifications from the server notices room. @@ -135,12 +133,13 @@ class ResourceLimitsServerNotices(object): limit blocking and need to be preserved. """ content = {"pinned": ref_events} - yield self._server_notices_manager.send_notice( + await self._server_notices_manager.send_notice( user_id, content, EventTypes.Pinned, "" ) - @defer.inlineCallbacks - def _apply_limit_block_notification(self, user_id, event_body, event_limit_type): + async def _apply_limit_block_notification( + self, user_id, event_body, event_limit_type + ): """Utility method to apply limit block notifications in the server notices room. @@ -157,17 +156,16 @@ class ResourceLimitsServerNotices(object): "admin_contact": self._config.admin_contact, "limit_type": event_limit_type, } - event = yield self._server_notices_manager.send_notice( + event = await self._server_notices_manager.send_notice( user_id, content, EventTypes.Message ) content = {"pinned": [event.event_id]} - yield self._server_notices_manager.send_notice( + await self._server_notices_manager.send_notice( user_id, content, EventTypes.Pinned, "" ) - @defer.inlineCallbacks - def _check_and_set_tags(self, user_id, room_id): + async def _check_and_set_tags(self, user_id, room_id): """ Since server notices rooms were originally not with tags, important to check that tags have been set correctly @@ -175,20 +173,19 @@ class ResourceLimitsServerNotices(object): user_id(str): the user in question room_id(str): the server notices room for that user """ - tags = yield self._store.get_tags_for_room(user_id, room_id) + tags = await self._store.get_tags_for_room(user_id, room_id) need_to_set_tag = True if tags: if SERVER_NOTICE_ROOM_TAG in tags: # tag already present, nothing to do here need_to_set_tag = False if need_to_set_tag: - max_id = yield self._store.add_tag_to_room( + max_id = await self._store.add_tag_to_room( user_id, room_id, SERVER_NOTICE_ROOM_TAG, {} ) self._notifier.on_new_event("account_data_key", max_id, users=[user_id]) - @defer.inlineCallbacks - def _is_room_currently_blocked(self, room_id): + async def _is_room_currently_blocked(self, room_id): """ Determines if the room is currently blocked @@ -196,7 +193,7 @@ class ResourceLimitsServerNotices(object): room_id(str): The room id of the server notices room Returns: - + Deferred[Tuple[bool, List]]: bool: Is the room currently blocked list: The list of pinned events that are unrelated to limit blocking This list can be used as a convenience in the case where the block @@ -206,7 +203,7 @@ class ResourceLimitsServerNotices(object): currently_blocked = False pinned_state_event = None try: - pinned_state_event = yield self._state.get_current_state( + pinned_state_event = await self._state.get_current_state( room_id, event_type=EventTypes.Pinned ) except AuthError: @@ -217,7 +214,7 @@ class ResourceLimitsServerNotices(object): if pinned_state_event is not None: referenced_events = list(pinned_state_event.content.get("pinned", [])) - events = yield self._store.get_events(referenced_events) + events = await self._store.get_events(referenced_events) for event_id, event in iteritems(events): if event.type != EventTypes.Message: continue diff --git a/synapse/server_notices/server_notices_manager.py b/synapse/server_notices/server_notices_manager.py index f7432c8d2..999c621b9 100644 --- a/synapse/server_notices/server_notices_manager.py +++ b/synapse/server_notices/server_notices_manager.py @@ -14,11 +14,9 @@ # limitations under the License. import logging -from twisted.internet import defer - from synapse.api.constants import EventTypes, Membership, RoomCreationPreset -from synapse.types import create_requester -from synapse.util.caches.descriptors import cachedInlineCallbacks +from synapse.types import UserID, create_requester +from synapse.util.caches.descriptors import cached logger = logging.getLogger(__name__) @@ -36,10 +34,12 @@ class ServerNoticesManager(object): self._store = hs.get_datastore() self._config = hs.config self._room_creation_handler = hs.get_room_creation_handler() + self._room_member_handler = hs.get_room_member_handler() self._event_creation_handler = hs.get_event_creation_handler() self._is_mine_id = hs.is_mine_id self._notifier = hs.get_notifier() + self.server_notices_mxid = self._config.server_notices_mxid def is_enabled(self): """Checks if server notices are enabled on this server. @@ -49,8 +49,7 @@ class ServerNoticesManager(object): """ return self._config.server_notices_mxid is not None - @defer.inlineCallbacks - def send_notice( + async def send_notice( self, user_id, event_content, type=EventTypes.Message, state_key=None ): """Send a notice to the given user @@ -66,7 +65,8 @@ class ServerNoticesManager(object): Returns: Deferred[FrozenEvent] """ - room_id = yield self.get_notice_room_for_user(user_id) + room_id = await self.get_or_create_notice_room_for_user(user_id) + await self.maybe_invite_user_to_room(user_id, room_id) system_mxid = self._config.server_notices_mxid requester = create_requester(system_mxid) @@ -83,16 +83,17 @@ class ServerNoticesManager(object): if state_key is not None: event_dict["state_key"] = state_key - res = yield self._event_creation_handler.create_and_send_nonmember_event( + res = await self._event_creation_handler.create_and_send_nonmember_event( requester, event_dict, ratelimit=False ) return res - @cachedInlineCallbacks() - def get_notice_room_for_user(self, user_id): + @cached() + async def get_or_create_notice_room_for_user(self, user_id): """Get the room for notices for a given user - If we have not yet created a notice room for this user, create it + If we have not yet created a notice room for this user, create it, but don't + invite the user to it. Args: user_id (str): complete user id for the user we want a room for @@ -105,21 +106,24 @@ class ServerNoticesManager(object): assert self._is_mine_id(user_id), "Cannot send server notices to remote users" - rooms = yield self._store.get_rooms_for_local_user_where_membership_is( + rooms = await self._store.get_rooms_for_local_user_where_membership_is( user_id, [Membership.INVITE, Membership.JOIN] ) - system_mxid = self._config.server_notices_mxid for room in rooms: # it's worth noting that there is an asymmetry here in that we # expect the user to be invited or joined, but the system user must # be joined. This is kinda deliberate, in that if somebody somehow # manages to invite the system user to a room, that doesn't make it # the server notices room. - user_ids = yield self._store.get_users_in_room(room.room_id) - if system_mxid in user_ids: + user_ids = await self._store.get_users_in_room(room.room_id) + if self.server_notices_mxid in user_ids: # we found a room which our user shares with the system notice # user - logger.info("Using room %s", room.room_id) + logger.info( + "Using existing server notices room %s for user %s", + room.room_id, + user_id, + ) return room.room_id # apparently no existing notice room: create a new one @@ -138,24 +142,49 @@ class ServerNoticesManager(object): "avatar_url": self._config.server_notices_mxid_avatar_url, } - requester = create_requester(system_mxid) - info = yield self._room_creation_handler.create_room( + requester = create_requester(self.server_notices_mxid) + info = await self._room_creation_handler.create_room( requester, config={ "preset": RoomCreationPreset.PRIVATE_CHAT, "name": self._config.server_notices_room_name, "power_level_content_override": {"users_default": -10}, - "invite": (user_id,), }, ratelimit=False, creator_join_profile=join_profile, ) room_id = info["room_id"] - max_id = yield self._store.add_tag_to_room( + max_id = await self._store.add_tag_to_room( user_id, room_id, SERVER_NOTICE_ROOM_TAG, {} ) self._notifier.on_new_event("account_data_key", max_id, users=[user_id]) logger.info("Created server notices room %s for %s", room_id, user_id) return room_id + + async def maybe_invite_user_to_room(self, user_id: str, room_id: str): + """Invite the given user to the given server room, unless the user has already + joined or been invited to it. + + Args: + user_id: The ID of the user to invite. + room_id: The ID of the room to invite the user to. + """ + requester = create_requester(self.server_notices_mxid) + + # Check whether the user has already joined or been invited to this room. If + # that's the case, there is no need to re-invite them. + joined_rooms = await self._store.get_rooms_for_local_user_where_membership_is( + user_id, [Membership.INVITE, Membership.JOIN] + ) + for room in joined_rooms: + if room.room_id == room_id: + return + + await self._room_member_handler.update_membership( + requester=requester, + target=UserID.from_string(user_id), + room_id=room_id, + action="invite", + ) diff --git a/synapse/server_notices/server_notices_sender.py b/synapse/server_notices/server_notices_sender.py index 652bab58e..be74e8664 100644 --- a/synapse/server_notices/server_notices_sender.py +++ b/synapse/server_notices/server_notices_sender.py @@ -12,8 +12,6 @@ # 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 twisted.internet import defer - from synapse.server_notices.consent_server_notices import ConsentServerNotices from synapse.server_notices.resource_limits_server_notices import ( ResourceLimitsServerNotices, @@ -36,18 +34,16 @@ class ServerNoticesSender(object): ResourceLimitsServerNotices(hs), ) - @defer.inlineCallbacks - def on_user_syncing(self, user_id): + async def on_user_syncing(self, user_id): """Called when the user performs a sync operation. Args: user_id (str): mxid of user who synced """ for sn in self._server_notices: - yield sn.maybe_send_server_notice_to_user(user_id) + await sn.maybe_send_server_notice_to_user(user_id) - @defer.inlineCallbacks - def on_user_ip(self, user_id): + async def on_user_ip(self, user_id): """Called on the master when a worker process saw a client request. Args: @@ -57,4 +53,4 @@ class ServerNoticesSender(object): # we check for notices to send to the user in on_user_ip as well as # in on_user_syncing for sn in self._server_notices: - yield sn.maybe_send_server_notice_to_user(user_id) + await sn.maybe_send_server_notice_to_user(user_id) diff --git a/synapse/static/client/login/index.html b/synapse/static/client/login/index.html index bcb6bc6bb..6fefdaaff 100644 --- a/synapse/static/client/login/index.html +++ b/synapse/static/client/login/index.html @@ -1,15 +1,16 @@ + Login - +

-

Log in with one of the following methods

+

diff --git a/synapse/static/client/login/js/jquery-2.1.3.min.js b/synapse/static/client/login/js/jquery-2.1.3.min.js deleted file mode 100644 index 25714ed29..000000000 --- a/synapse/static/client/login/js/jquery-2.1.3.min.js +++ /dev/null @@ -1,4 +0,0 @@ -/*! jQuery v2.1.3 | (c) 2005, 2014 jQuery Foundation, Inc. | jquery.org/license */ -!function(a,b){"object"==typeof module&&"object"==typeof module.exports?module.exports=a.document?b(a,!0):function(a){if(!a.document)throw new Error("jQuery requires a window with a document");return b(a)}:b(a)}("undefined"!=typeof window?window:this,function(a,b){var c=[],d=c.slice,e=c.concat,f=c.push,g=c.indexOf,h={},i=h.toString,j=h.hasOwnProperty,k={},l=a.document,m="2.1.3",n=function(a,b){return new n.fn.init(a,b)},o=/^[\s\uFEFF\xA0]+|[\s\uFEFF\xA0]+$/g,p=/^-ms-/,q=/-([\da-z])/gi,r=function(a,b){return b.toUpperCase()};n.fn=n.prototype={jquery:m,constructor:n,selector:"",length:0,toArray:function(){return d.call(this)},get:function(a){return null!=a?0>a?this[a+this.length]:this[a]:d.call(this)},pushStack:function(a){var b=n.merge(this.constructor(),a);return b.prevObject=this,b.context=this.context,b},each:function(a,b){return n.each(this,a,b)},map:function(a){return this.pushStack(n.map(this,function(b,c){return a.call(b,c,b)}))},slice:function(){return this.pushStack(d.apply(this,arguments))},first:function(){return this.eq(0)},last:function(){return this.eq(-1)},eq:function(a){var b=this.length,c=+a+(0>a?b:0);return this.pushStack(c>=0&&b>c?[this[c]]:[])},end:function(){return this.prevObject||this.constructor(null)},push:f,sort:c.sort,splice:c.splice},n.extend=n.fn.extend=function(){var a,b,c,d,e,f,g=arguments[0]||{},h=1,i=arguments.length,j=!1;for("boolean"==typeof g&&(j=g,g=arguments[h]||{},h++),"object"==typeof g||n.isFunction(g)||(g={}),h===i&&(g=this,h--);i>h;h++)if(null!=(a=arguments[h]))for(b in a)c=g[b],d=a[b],g!==d&&(j&&d&&(n.isPlainObject(d)||(e=n.isArray(d)))?(e?(e=!1,f=c&&n.isArray(c)?c:[]):f=c&&n.isPlainObject(c)?c:{},g[b]=n.extend(j,f,d)):void 0!==d&&(g[b]=d));return g},n.extend({expando:"jQuery"+(m+Math.random()).replace(/\D/g,""),isReady:!0,error:function(a){throw new Error(a)},noop:function(){},isFunction:function(a){return"function"===n.type(a)},isArray:Array.isArray,isWindow:function(a){return null!=a&&a===a.window},isNumeric:function(a){return!n.isArray(a)&&a-parseFloat(a)+1>=0},isPlainObject:function(a){return"object"!==n.type(a)||a.nodeType||n.isWindow(a)?!1:a.constructor&&!j.call(a.constructor.prototype,"isPrototypeOf")?!1:!0},isEmptyObject:function(a){var b;for(b in a)return!1;return!0},type:function(a){return null==a?a+"":"object"==typeof a||"function"==typeof a?h[i.call(a)]||"object":typeof a},globalEval:function(a){var b,c=eval;a=n.trim(a),a&&(1===a.indexOf("use strict")?(b=l.createElement("script"),b.text=a,l.head.appendChild(b).parentNode.removeChild(b)):c(a))},camelCase:function(a){return a.replace(p,"ms-").replace(q,r)},nodeName:function(a,b){return a.nodeName&&a.nodeName.toLowerCase()===b.toLowerCase()},each:function(a,b,c){var d,e=0,f=a.length,g=s(a);if(c){if(g){for(;f>e;e++)if(d=b.apply(a[e],c),d===!1)break}else for(e in a)if(d=b.apply(a[e],c),d===!1)break}else if(g){for(;f>e;e++)if(d=b.call(a[e],e,a[e]),d===!1)break}else for(e in a)if(d=b.call(a[e],e,a[e]),d===!1)break;return a},trim:function(a){return null==a?"":(a+"").replace(o,"")},makeArray:function(a,b){var c=b||[];return null!=a&&(s(Object(a))?n.merge(c,"string"==typeof a?[a]:a):f.call(c,a)),c},inArray:function(a,b,c){return null==b?-1:g.call(b,a,c)},merge:function(a,b){for(var c=+b.length,d=0,e=a.length;c>d;d++)a[e++]=b[d];return a.length=e,a},grep:function(a,b,c){for(var d,e=[],f=0,g=a.length,h=!c;g>f;f++)d=!b(a[f],f),d!==h&&e.push(a[f]);return e},map:function(a,b,c){var d,f=0,g=a.length,h=s(a),i=[];if(h)for(;g>f;f++)d=b(a[f],f,c),null!=d&&i.push(d);else for(f in a)d=b(a[f],f,c),null!=d&&i.push(d);return e.apply([],i)},guid:1,proxy:function(a,b){var c,e,f;return"string"==typeof b&&(c=a[b],b=a,a=c),n.isFunction(a)?(e=d.call(arguments,2),f=function(){return a.apply(b||this,e.concat(d.call(arguments)))},f.guid=a.guid=a.guid||n.guid++,f):void 0},now:Date.now,support:k}),n.each("Boolean Number String Function Array Date RegExp Object Error".split(" "),function(a,b){h["[object "+b+"]"]=b.toLowerCase()});function s(a){var b=a.length,c=n.type(a);return"function"===c||n.isWindow(a)?!1:1===a.nodeType&&b?!0:"array"===c||0===b||"number"==typeof b&&b>0&&b-1 in a}var t=function(a){var b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u="sizzle"+1*new Date,v=a.document,w=0,x=0,y=hb(),z=hb(),A=hb(),B=function(a,b){return a===b&&(l=!0),0},C=1<<31,D={}.hasOwnProperty,E=[],F=E.pop,G=E.push,H=E.push,I=E.slice,J=function(a,b){for(var c=0,d=a.length;d>c;c++)if(a[c]===b)return c;return-1},K="checked|selected|async|autofocus|autoplay|controls|defer|disabled|hidden|ismap|loop|multiple|open|readonly|required|scoped",L="[\\x20\\t\\r\\n\\f]",M="(?:\\\\.|[\\w-]|[^\\x00-\\xa0])+",N=M.replace("w","w#"),O="\\["+L+"*("+M+")(?:"+L+"*([*^$|!~]?=)"+L+"*(?:'((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\"|("+N+"))|)"+L+"*\\]",P=":("+M+")(?:\\((('((?:\\\\.|[^\\\\'])*)'|\"((?:\\\\.|[^\\\\\"])*)\")|((?:\\\\.|[^\\\\()[\\]]|"+O+")*)|.*)\\)|)",Q=new RegExp(L+"+","g"),R=new RegExp("^"+L+"+|((?:^|[^\\\\])(?:\\\\.)*)"+L+"+$","g"),S=new RegExp("^"+L+"*,"+L+"*"),T=new RegExp("^"+L+"*([>+~]|"+L+")"+L+"*"),U=new RegExp("="+L+"*([^\\]'\"]*?)"+L+"*\\]","g"),V=new RegExp(P),W=new RegExp("^"+N+"$"),X={ID:new RegExp("^#("+M+")"),CLASS:new RegExp("^\\.("+M+")"),TAG:new RegExp("^("+M.replace("w","w*")+")"),ATTR:new RegExp("^"+O),PSEUDO:new RegExp("^"+P),CHILD:new RegExp("^:(only|first|last|nth|nth-last)-(child|of-type)(?:\\("+L+"*(even|odd|(([+-]|)(\\d*)n|)"+L+"*(?:([+-]|)"+L+"*(\\d+)|))"+L+"*\\)|)","i"),bool:new RegExp("^(?:"+K+")$","i"),needsContext:new RegExp("^"+L+"*[>+~]|:(even|odd|eq|gt|lt|nth|first|last)(?:\\("+L+"*((?:-\\d)?\\d*)"+L+"*\\)|)(?=[^-]|$)","i")},Y=/^(?:input|select|textarea|button)$/i,Z=/^h\d$/i,$=/^[^{]+\{\s*\[native \w/,_=/^(?:#([\w-]+)|(\w+)|\.([\w-]+))$/,ab=/[+~]/,bb=/'|\\/g,cb=new RegExp("\\\\([\\da-f]{1,6}"+L+"?|("+L+")|.)","ig"),db=function(a,b,c){var d="0x"+b-65536;return d!==d||c?b:0>d?String.fromCharCode(d+65536):String.fromCharCode(d>>10|55296,1023&d|56320)},eb=function(){m()};try{H.apply(E=I.call(v.childNodes),v.childNodes),E[v.childNodes.length].nodeType}catch(fb){H={apply:E.length?function(a,b){G.apply(a,I.call(b))}:function(a,b){var c=a.length,d=0;while(a[c++]=b[d++]);a.length=c-1}}}function gb(a,b,d,e){var f,h,j,k,l,o,r,s,w,x;if((b?b.ownerDocument||b:v)!==n&&m(b),b=b||n,d=d||[],k=b.nodeType,"string"!=typeof a||!a||1!==k&&9!==k&&11!==k)return d;if(!e&&p){if(11!==k&&(f=_.exec(a)))if(j=f[1]){if(9===k){if(h=b.getElementById(j),!h||!h.parentNode)return d;if(h.id===j)return d.push(h),d}else if(b.ownerDocument&&(h=b.ownerDocument.getElementById(j))&&t(b,h)&&h.id===j)return d.push(h),d}else{if(f[2])return H.apply(d,b.getElementsByTagName(a)),d;if((j=f[3])&&c.getElementsByClassName)return H.apply(d,b.getElementsByClassName(j)),d}if(c.qsa&&(!q||!q.test(a))){if(s=r=u,w=b,x=1!==k&&a,1===k&&"object"!==b.nodeName.toLowerCase()){o=g(a),(r=b.getAttribute("id"))?s=r.replace(bb,"\\$&"):b.setAttribute("id",s),s="[id='"+s+"'] ",l=o.length;while(l--)o[l]=s+rb(o[l]);w=ab.test(a)&&pb(b.parentNode)||b,x=o.join(",")}if(x)try{return H.apply(d,w.querySelectorAll(x)),d}catch(y){}finally{r||b.removeAttribute("id")}}}return i(a.replace(R,"$1"),b,d,e)}function hb(){var a=[];function b(c,e){return a.push(c+" ")>d.cacheLength&&delete b[a.shift()],b[c+" "]=e}return b}function ib(a){return a[u]=!0,a}function jb(a){var b=n.createElement("div");try{return!!a(b)}catch(c){return!1}finally{b.parentNode&&b.parentNode.removeChild(b),b=null}}function kb(a,b){var c=a.split("|"),e=a.length;while(e--)d.attrHandle[c[e]]=b}function lb(a,b){var c=b&&a,d=c&&1===a.nodeType&&1===b.nodeType&&(~b.sourceIndex||C)-(~a.sourceIndex||C);if(d)return d;if(c)while(c=c.nextSibling)if(c===b)return-1;return a?1:-1}function mb(a){return function(b){var c=b.nodeName.toLowerCase();return"input"===c&&b.type===a}}function nb(a){return function(b){var c=b.nodeName.toLowerCase();return("input"===c||"button"===c)&&b.type===a}}function ob(a){return ib(function(b){return b=+b,ib(function(c,d){var e,f=a([],c.length,b),g=f.length;while(g--)c[e=f[g]]&&(c[e]=!(d[e]=c[e]))})})}function pb(a){return a&&"undefined"!=typeof a.getElementsByTagName&&a}c=gb.support={},f=gb.isXML=function(a){var b=a&&(a.ownerDocument||a).documentElement;return b?"HTML"!==b.nodeName:!1},m=gb.setDocument=function(a){var b,e,g=a?a.ownerDocument||a:v;return g!==n&&9===g.nodeType&&g.documentElement?(n=g,o=g.documentElement,e=g.defaultView,e&&e!==e.top&&(e.addEventListener?e.addEventListener("unload",eb,!1):e.attachEvent&&e.attachEvent("onunload",eb)),p=!f(g),c.attributes=jb(function(a){return a.className="i",!a.getAttribute("className")}),c.getElementsByTagName=jb(function(a){return a.appendChild(g.createComment("")),!a.getElementsByTagName("*").length}),c.getElementsByClassName=$.test(g.getElementsByClassName),c.getById=jb(function(a){return o.appendChild(a).id=u,!g.getElementsByName||!g.getElementsByName(u).length}),c.getById?(d.find.ID=function(a,b){if("undefined"!=typeof b.getElementById&&p){var c=b.getElementById(a);return c&&c.parentNode?[c]:[]}},d.filter.ID=function(a){var b=a.replace(cb,db);return function(a){return a.getAttribute("id")===b}}):(delete d.find.ID,d.filter.ID=function(a){var b=a.replace(cb,db);return function(a){var c="undefined"!=typeof a.getAttributeNode&&a.getAttributeNode("id");return c&&c.value===b}}),d.find.TAG=c.getElementsByTagName?function(a,b){return"undefined"!=typeof b.getElementsByTagName?b.getElementsByTagName(a):c.qsa?b.querySelectorAll(a):void 0}:function(a,b){var c,d=[],e=0,f=b.getElementsByTagName(a);if("*"===a){while(c=f[e++])1===c.nodeType&&d.push(c);return d}return f},d.find.CLASS=c.getElementsByClassName&&function(a,b){return p?b.getElementsByClassName(a):void 0},r=[],q=[],(c.qsa=$.test(g.querySelectorAll))&&(jb(function(a){o.appendChild(a).innerHTML="",a.querySelectorAll("[msallowcapture^='']").length&&q.push("[*^$]="+L+"*(?:''|\"\")"),a.querySelectorAll("[selected]").length||q.push("\\["+L+"*(?:value|"+K+")"),a.querySelectorAll("[id~="+u+"-]").length||q.push("~="),a.querySelectorAll(":checked").length||q.push(":checked"),a.querySelectorAll("a#"+u+"+*").length||q.push(".#.+[+~]")}),jb(function(a){var b=g.createElement("input");b.setAttribute("type","hidden"),a.appendChild(b).setAttribute("name","D"),a.querySelectorAll("[name=d]").length&&q.push("name"+L+"*[*^$|!~]?="),a.querySelectorAll(":enabled").length||q.push(":enabled",":disabled"),a.querySelectorAll("*,:x"),q.push(",.*:")})),(c.matchesSelector=$.test(s=o.matches||o.webkitMatchesSelector||o.mozMatchesSelector||o.oMatchesSelector||o.msMatchesSelector))&&jb(function(a){c.disconnectedMatch=s.call(a,"div"),s.call(a,"[s!='']:x"),r.push("!=",P)}),q=q.length&&new RegExp(q.join("|")),r=r.length&&new RegExp(r.join("|")),b=$.test(o.compareDocumentPosition),t=b||$.test(o.contains)?function(a,b){var c=9===a.nodeType?a.documentElement:a,d=b&&b.parentNode;return a===d||!(!d||1!==d.nodeType||!(c.contains?c.contains(d):a.compareDocumentPosition&&16&a.compareDocumentPosition(d)))}:function(a,b){if(b)while(b=b.parentNode)if(b===a)return!0;return!1},B=b?function(a,b){if(a===b)return l=!0,0;var d=!a.compareDocumentPosition-!b.compareDocumentPosition;return d?d:(d=(a.ownerDocument||a)===(b.ownerDocument||b)?a.compareDocumentPosition(b):1,1&d||!c.sortDetached&&b.compareDocumentPosition(a)===d?a===g||a.ownerDocument===v&&t(v,a)?-1:b===g||b.ownerDocument===v&&t(v,b)?1:k?J(k,a)-J(k,b):0:4&d?-1:1)}:function(a,b){if(a===b)return l=!0,0;var c,d=0,e=a.parentNode,f=b.parentNode,h=[a],i=[b];if(!e||!f)return a===g?-1:b===g?1:e?-1:f?1:k?J(k,a)-J(k,b):0;if(e===f)return lb(a,b);c=a;while(c=c.parentNode)h.unshift(c);c=b;while(c=c.parentNode)i.unshift(c);while(h[d]===i[d])d++;return d?lb(h[d],i[d]):h[d]===v?-1:i[d]===v?1:0},g):n},gb.matches=function(a,b){return gb(a,null,null,b)},gb.matchesSelector=function(a,b){if((a.ownerDocument||a)!==n&&m(a),b=b.replace(U,"='$1']"),!(!c.matchesSelector||!p||r&&r.test(b)||q&&q.test(b)))try{var d=s.call(a,b);if(d||c.disconnectedMatch||a.document&&11!==a.document.nodeType)return d}catch(e){}return gb(b,n,null,[a]).length>0},gb.contains=function(a,b){return(a.ownerDocument||a)!==n&&m(a),t(a,b)},gb.attr=function(a,b){(a.ownerDocument||a)!==n&&m(a);var e=d.attrHandle[b.toLowerCase()],f=e&&D.call(d.attrHandle,b.toLowerCase())?e(a,b,!p):void 0;return void 0!==f?f:c.attributes||!p?a.getAttribute(b):(f=a.getAttributeNode(b))&&f.specified?f.value:null},gb.error=function(a){throw new Error("Syntax error, unrecognized expression: "+a)},gb.uniqueSort=function(a){var b,d=[],e=0,f=0;if(l=!c.detectDuplicates,k=!c.sortStable&&a.slice(0),a.sort(B),l){while(b=a[f++])b===a[f]&&(e=d.push(f));while(e--)a.splice(d[e],1)}return k=null,a},e=gb.getText=function(a){var b,c="",d=0,f=a.nodeType;if(f){if(1===f||9===f||11===f){if("string"==typeof a.textContent)return a.textContent;for(a=a.firstChild;a;a=a.nextSibling)c+=e(a)}else if(3===f||4===f)return a.nodeValue}else while(b=a[d++])c+=e(b);return c},d=gb.selectors={cacheLength:50,createPseudo:ib,match:X,attrHandle:{},find:{},relative:{">":{dir:"parentNode",first:!0}," ":{dir:"parentNode"},"+":{dir:"previousSibling",first:!0},"~":{dir:"previousSibling"}},preFilter:{ATTR:function(a){return a[1]=a[1].replace(cb,db),a[3]=(a[3]||a[4]||a[5]||"").replace(cb,db),"~="===a[2]&&(a[3]=" "+a[3]+" "),a.slice(0,4)},CHILD:function(a){return a[1]=a[1].toLowerCase(),"nth"===a[1].slice(0,3)?(a[3]||gb.error(a[0]),a[4]=+(a[4]?a[5]+(a[6]||1):2*("even"===a[3]||"odd"===a[3])),a[5]=+(a[7]+a[8]||"odd"===a[3])):a[3]&&gb.error(a[0]),a},PSEUDO:function(a){var b,c=!a[6]&&a[2];return X.CHILD.test(a[0])?null:(a[3]?a[2]=a[4]||a[5]||"":c&&V.test(c)&&(b=g(c,!0))&&(b=c.indexOf(")",c.length-b)-c.length)&&(a[0]=a[0].slice(0,b),a[2]=c.slice(0,b)),a.slice(0,3))}},filter:{TAG:function(a){var b=a.replace(cb,db).toLowerCase();return"*"===a?function(){return!0}:function(a){return a.nodeName&&a.nodeName.toLowerCase()===b}},CLASS:function(a){var b=y[a+" "];return b||(b=new RegExp("(^|"+L+")"+a+"("+L+"|$)"))&&y(a,function(a){return b.test("string"==typeof a.className&&a.className||"undefined"!=typeof a.getAttribute&&a.getAttribute("class")||"")})},ATTR:function(a,b,c){return function(d){var e=gb.attr(d,a);return null==e?"!="===b:b?(e+="","="===b?e===c:"!="===b?e!==c:"^="===b?c&&0===e.indexOf(c):"*="===b?c&&e.indexOf(c)>-1:"$="===b?c&&e.slice(-c.length)===c:"~="===b?(" "+e.replace(Q," ")+" ").indexOf(c)>-1:"|="===b?e===c||e.slice(0,c.length+1)===c+"-":!1):!0}},CHILD:function(a,b,c,d,e){var f="nth"!==a.slice(0,3),g="last"!==a.slice(-4),h="of-type"===b;return 1===d&&0===e?function(a){return!!a.parentNode}:function(b,c,i){var j,k,l,m,n,o,p=f!==g?"nextSibling":"previousSibling",q=b.parentNode,r=h&&b.nodeName.toLowerCase(),s=!i&&!h;if(q){if(f){while(p){l=b;while(l=l[p])if(h?l.nodeName.toLowerCase()===r:1===l.nodeType)return!1;o=p="only"===a&&!o&&"nextSibling"}return!0}if(o=[g?q.firstChild:q.lastChild],g&&s){k=q[u]||(q[u]={}),j=k[a]||[],n=j[0]===w&&j[1],m=j[0]===w&&j[2],l=n&&q.childNodes[n];while(l=++n&&l&&l[p]||(m=n=0)||o.pop())if(1===l.nodeType&&++m&&l===b){k[a]=[w,n,m];break}}else if(s&&(j=(b[u]||(b[u]={}))[a])&&j[0]===w)m=j[1];else while(l=++n&&l&&l[p]||(m=n=0)||o.pop())if((h?l.nodeName.toLowerCase()===r:1===l.nodeType)&&++m&&(s&&((l[u]||(l[u]={}))[a]=[w,m]),l===b))break;return m-=e,m===d||m%d===0&&m/d>=0}}},PSEUDO:function(a,b){var c,e=d.pseudos[a]||d.setFilters[a.toLowerCase()]||gb.error("unsupported pseudo: "+a);return e[u]?e(b):e.length>1?(c=[a,a,"",b],d.setFilters.hasOwnProperty(a.toLowerCase())?ib(function(a,c){var d,f=e(a,b),g=f.length;while(g--)d=J(a,f[g]),a[d]=!(c[d]=f[g])}):function(a){return e(a,0,c)}):e}},pseudos:{not:ib(function(a){var b=[],c=[],d=h(a.replace(R,"$1"));return d[u]?ib(function(a,b,c,e){var f,g=d(a,null,e,[]),h=a.length;while(h--)(f=g[h])&&(a[h]=!(b[h]=f))}):function(a,e,f){return b[0]=a,d(b,null,f,c),b[0]=null,!c.pop()}}),has:ib(function(a){return function(b){return gb(a,b).length>0}}),contains:ib(function(a){return a=a.replace(cb,db),function(b){return(b.textContent||b.innerText||e(b)).indexOf(a)>-1}}),lang:ib(function(a){return W.test(a||"")||gb.error("unsupported lang: "+a),a=a.replace(cb,db).toLowerCase(),function(b){var c;do if(c=p?b.lang:b.getAttribute("xml:lang")||b.getAttribute("lang"))return c=c.toLowerCase(),c===a||0===c.indexOf(a+"-");while((b=b.parentNode)&&1===b.nodeType);return!1}}),target:function(b){var c=a.location&&a.location.hash;return c&&c.slice(1)===b.id},root:function(a){return a===o},focus:function(a){return a===n.activeElement&&(!n.hasFocus||n.hasFocus())&&!!(a.type||a.href||~a.tabIndex)},enabled:function(a){return a.disabled===!1},disabled:function(a){return a.disabled===!0},checked:function(a){var b=a.nodeName.toLowerCase();return"input"===b&&!!a.checked||"option"===b&&!!a.selected},selected:function(a){return a.parentNode&&a.parentNode.selectedIndex,a.selected===!0},empty:function(a){for(a=a.firstChild;a;a=a.nextSibling)if(a.nodeType<6)return!1;return!0},parent:function(a){return!d.pseudos.empty(a)},header:function(a){return Z.test(a.nodeName)},input:function(a){return Y.test(a.nodeName)},button:function(a){var b=a.nodeName.toLowerCase();return"input"===b&&"button"===a.type||"button"===b},text:function(a){var b;return"input"===a.nodeName.toLowerCase()&&"text"===a.type&&(null==(b=a.getAttribute("type"))||"text"===b.toLowerCase())},first:ob(function(){return[0]}),last:ob(function(a,b){return[b-1]}),eq:ob(function(a,b,c){return[0>c?c+b:c]}),even:ob(function(a,b){for(var c=0;b>c;c+=2)a.push(c);return a}),odd:ob(function(a,b){for(var c=1;b>c;c+=2)a.push(c);return a}),lt:ob(function(a,b,c){for(var d=0>c?c+b:c;--d>=0;)a.push(d);return a}),gt:ob(function(a,b,c){for(var d=0>c?c+b:c;++db;b++)d+=a[b].value;return d}function sb(a,b,c){var d=b.dir,e=c&&"parentNode"===d,f=x++;return b.first?function(b,c,f){while(b=b[d])if(1===b.nodeType||e)return a(b,c,f)}:function(b,c,g){var h,i,j=[w,f];if(g){while(b=b[d])if((1===b.nodeType||e)&&a(b,c,g))return!0}else while(b=b[d])if(1===b.nodeType||e){if(i=b[u]||(b[u]={}),(h=i[d])&&h[0]===w&&h[1]===f)return j[2]=h[2];if(i[d]=j,j[2]=a(b,c,g))return!0}}}function tb(a){return a.length>1?function(b,c,d){var e=a.length;while(e--)if(!a[e](b,c,d))return!1;return!0}:a[0]}function ub(a,b,c){for(var d=0,e=b.length;e>d;d++)gb(a,b[d],c);return c}function vb(a,b,c,d,e){for(var f,g=[],h=0,i=a.length,j=null!=b;i>h;h++)(f=a[h])&&(!c||c(f,d,e))&&(g.push(f),j&&b.push(h));return g}function wb(a,b,c,d,e,f){return d&&!d[u]&&(d=wb(d)),e&&!e[u]&&(e=wb(e,f)),ib(function(f,g,h,i){var j,k,l,m=[],n=[],o=g.length,p=f||ub(b||"*",h.nodeType?[h]:h,[]),q=!a||!f&&b?p:vb(p,m,a,h,i),r=c?e||(f?a:o||d)?[]:g:q;if(c&&c(q,r,h,i),d){j=vb(r,n),d(j,[],h,i),k=j.length;while(k--)(l=j[k])&&(r[n[k]]=!(q[n[k]]=l))}if(f){if(e||a){if(e){j=[],k=r.length;while(k--)(l=r[k])&&j.push(q[k]=l);e(null,r=[],j,i)}k=r.length;while(k--)(l=r[k])&&(j=e?J(f,l):m[k])>-1&&(f[j]=!(g[j]=l))}}else r=vb(r===g?r.splice(o,r.length):r),e?e(null,g,r,i):H.apply(g,r)})}function xb(a){for(var b,c,e,f=a.length,g=d.relative[a[0].type],h=g||d.relative[" "],i=g?1:0,k=sb(function(a){return a===b},h,!0),l=sb(function(a){return J(b,a)>-1},h,!0),m=[function(a,c,d){var e=!g&&(d||c!==j)||((b=c).nodeType?k(a,c,d):l(a,c,d));return b=null,e}];f>i;i++)if(c=d.relative[a[i].type])m=[sb(tb(m),c)];else{if(c=d.filter[a[i].type].apply(null,a[i].matches),c[u]){for(e=++i;f>e;e++)if(d.relative[a[e].type])break;return wb(i>1&&tb(m),i>1&&rb(a.slice(0,i-1).concat({value:" "===a[i-2].type?"*":""})).replace(R,"$1"),c,e>i&&xb(a.slice(i,e)),f>e&&xb(a=a.slice(e)),f>e&&rb(a))}m.push(c)}return tb(m)}function yb(a,b){var c=b.length>0,e=a.length>0,f=function(f,g,h,i,k){var l,m,o,p=0,q="0",r=f&&[],s=[],t=j,u=f||e&&d.find.TAG("*",k),v=w+=null==t?1:Math.random()||.1,x=u.length;for(k&&(j=g!==n&&g);q!==x&&null!=(l=u[q]);q++){if(e&&l){m=0;while(o=a[m++])if(o(l,g,h)){i.push(l);break}k&&(w=v)}c&&((l=!o&&l)&&p--,f&&r.push(l))}if(p+=q,c&&q!==p){m=0;while(o=b[m++])o(r,s,g,h);if(f){if(p>0)while(q--)r[q]||s[q]||(s[q]=F.call(i));s=vb(s)}H.apply(i,s),k&&!f&&s.length>0&&p+b.length>1&&gb.uniqueSort(i)}return k&&(w=v,j=t),r};return c?ib(f):f}return h=gb.compile=function(a,b){var c,d=[],e=[],f=A[a+" "];if(!f){b||(b=g(a)),c=b.length;while(c--)f=xb(b[c]),f[u]?d.push(f):e.push(f);f=A(a,yb(e,d)),f.selector=a}return f},i=gb.select=function(a,b,e,f){var i,j,k,l,m,n="function"==typeof a&&a,o=!f&&g(a=n.selector||a);if(e=e||[],1===o.length){if(j=o[0]=o[0].slice(0),j.length>2&&"ID"===(k=j[0]).type&&c.getById&&9===b.nodeType&&p&&d.relative[j[1].type]){if(b=(d.find.ID(k.matches[0].replace(cb,db),b)||[])[0],!b)return e;n&&(b=b.parentNode),a=a.slice(j.shift().value.length)}i=X.needsContext.test(a)?0:j.length;while(i--){if(k=j[i],d.relative[l=k.type])break;if((m=d.find[l])&&(f=m(k.matches[0].replace(cb,db),ab.test(j[0].type)&&pb(b.parentNode)||b))){if(j.splice(i,1),a=f.length&&rb(j),!a)return H.apply(e,f),e;break}}}return(n||h(a,o))(f,b,!p,e,ab.test(a)&&pb(b.parentNode)||b),e},c.sortStable=u.split("").sort(B).join("")===u,c.detectDuplicates=!!l,m(),c.sortDetached=jb(function(a){return 1&a.compareDocumentPosition(n.createElement("div"))}),jb(function(a){return a.innerHTML="","#"===a.firstChild.getAttribute("href")})||kb("type|href|height|width",function(a,b,c){return c?void 0:a.getAttribute(b,"type"===b.toLowerCase()?1:2)}),c.attributes&&jb(function(a){return a.innerHTML="",a.firstChild.setAttribute("value",""),""===a.firstChild.getAttribute("value")})||kb("value",function(a,b,c){return c||"input"!==a.nodeName.toLowerCase()?void 0:a.defaultValue}),jb(function(a){return null==a.getAttribute("disabled")})||kb(K,function(a,b,c){var d;return c?void 0:a[b]===!0?b.toLowerCase():(d=a.getAttributeNode(b))&&d.specified?d.value:null}),gb}(a);n.find=t,n.expr=t.selectors,n.expr[":"]=n.expr.pseudos,n.unique=t.uniqueSort,n.text=t.getText,n.isXMLDoc=t.isXML,n.contains=t.contains;var u=n.expr.match.needsContext,v=/^<(\w+)\s*\/?>(?:<\/\1>|)$/,w=/^.[^:#\[\.,]*$/;function x(a,b,c){if(n.isFunction(b))return n.grep(a,function(a,d){return!!b.call(a,d,a)!==c});if(b.nodeType)return n.grep(a,function(a){return a===b!==c});if("string"==typeof b){if(w.test(b))return n.filter(b,a,c);b=n.filter(b,a)}return n.grep(a,function(a){return g.call(b,a)>=0!==c})}n.filter=function(a,b,c){var d=b[0];return c&&(a=":not("+a+")"),1===b.length&&1===d.nodeType?n.find.matchesSelector(d,a)?[d]:[]:n.find.matches(a,n.grep(b,function(a){return 1===a.nodeType}))},n.fn.extend({find:function(a){var b,c=this.length,d=[],e=this;if("string"!=typeof a)return this.pushStack(n(a).filter(function(){for(b=0;c>b;b++)if(n.contains(e[b],this))return!0}));for(b=0;c>b;b++)n.find(a,e[b],d);return d=this.pushStack(c>1?n.unique(d):d),d.selector=this.selector?this.selector+" "+a:a,d},filter:function(a){return this.pushStack(x(this,a||[],!1))},not:function(a){return this.pushStack(x(this,a||[],!0))},is:function(a){return!!x(this,"string"==typeof a&&u.test(a)?n(a):a||[],!1).length}});var y,z=/^(?:\s*(<[\w\W]+>)[^>]*|#([\w-]*))$/,A=n.fn.init=function(a,b){var c,d;if(!a)return this;if("string"==typeof a){if(c="<"===a[0]&&">"===a[a.length-1]&&a.length>=3?[null,a,null]:z.exec(a),!c||!c[1]&&b)return!b||b.jquery?(b||y).find(a):this.constructor(b).find(a);if(c[1]){if(b=b instanceof n?b[0]:b,n.merge(this,n.parseHTML(c[1],b&&b.nodeType?b.ownerDocument||b:l,!0)),v.test(c[1])&&n.isPlainObject(b))for(c in b)n.isFunction(this[c])?this[c](b[c]):this.attr(c,b[c]);return this}return d=l.getElementById(c[2]),d&&d.parentNode&&(this.length=1,this[0]=d),this.context=l,this.selector=a,this}return a.nodeType?(this.context=this[0]=a,this.length=1,this):n.isFunction(a)?"undefined"!=typeof y.ready?y.ready(a):a(n):(void 0!==a.selector&&(this.selector=a.selector,this.context=a.context),n.makeArray(a,this))};A.prototype=n.fn,y=n(l);var B=/^(?:parents|prev(?:Until|All))/,C={children:!0,contents:!0,next:!0,prev:!0};n.extend({dir:function(a,b,c){var d=[],e=void 0!==c;while((a=a[b])&&9!==a.nodeType)if(1===a.nodeType){if(e&&n(a).is(c))break;d.push(a)}return d},sibling:function(a,b){for(var c=[];a;a=a.nextSibling)1===a.nodeType&&a!==b&&c.push(a);return c}}),n.fn.extend({has:function(a){var b=n(a,this),c=b.length;return this.filter(function(){for(var a=0;c>a;a++)if(n.contains(this,b[a]))return!0})},closest:function(a,b){for(var c,d=0,e=this.length,f=[],g=u.test(a)||"string"!=typeof a?n(a,b||this.context):0;e>d;d++)for(c=this[d];c&&c!==b;c=c.parentNode)if(c.nodeType<11&&(g?g.index(c)>-1:1===c.nodeType&&n.find.matchesSelector(c,a))){f.push(c);break}return this.pushStack(f.length>1?n.unique(f):f)},index:function(a){return a?"string"==typeof a?g.call(n(a),this[0]):g.call(this,a.jquery?a[0]:a):this[0]&&this[0].parentNode?this.first().prevAll().length:-1},add:function(a,b){return this.pushStack(n.unique(n.merge(this.get(),n(a,b))))},addBack:function(a){return this.add(null==a?this.prevObject:this.prevObject.filter(a))}});function D(a,b){while((a=a[b])&&1!==a.nodeType);return a}n.each({parent:function(a){var b=a.parentNode;return b&&11!==b.nodeType?b:null},parents:function(a){return n.dir(a,"parentNode")},parentsUntil:function(a,b,c){return n.dir(a,"parentNode",c)},next:function(a){return D(a,"nextSibling")},prev:function(a){return D(a,"previousSibling")},nextAll:function(a){return n.dir(a,"nextSibling")},prevAll:function(a){return n.dir(a,"previousSibling")},nextUntil:function(a,b,c){return n.dir(a,"nextSibling",c)},prevUntil:function(a,b,c){return n.dir(a,"previousSibling",c)},siblings:function(a){return n.sibling((a.parentNode||{}).firstChild,a)},children:function(a){return n.sibling(a.firstChild)},contents:function(a){return a.contentDocument||n.merge([],a.childNodes)}},function(a,b){n.fn[a]=function(c,d){var e=n.map(this,b,c);return"Until"!==a.slice(-5)&&(d=c),d&&"string"==typeof d&&(e=n.filter(d,e)),this.length>1&&(C[a]||n.unique(e),B.test(a)&&e.reverse()),this.pushStack(e)}});var E=/\S+/g,F={};function G(a){var b=F[a]={};return n.each(a.match(E)||[],function(a,c){b[c]=!0}),b}n.Callbacks=function(a){a="string"==typeof a?F[a]||G(a):n.extend({},a);var b,c,d,e,f,g,h=[],i=!a.once&&[],j=function(l){for(b=a.memory&&l,c=!0,g=e||0,e=0,f=h.length,d=!0;h&&f>g;g++)if(h[g].apply(l[0],l[1])===!1&&a.stopOnFalse){b=!1;break}d=!1,h&&(i?i.length&&j(i.shift()):b?h=[]:k.disable())},k={add:function(){if(h){var c=h.length;!function g(b){n.each(b,function(b,c){var d=n.type(c);"function"===d?a.unique&&k.has(c)||h.push(c):c&&c.length&&"string"!==d&&g(c)})}(arguments),d?f=h.length:b&&(e=c,j(b))}return this},remove:function(){return h&&n.each(arguments,function(a,b){var c;while((c=n.inArray(b,h,c))>-1)h.splice(c,1),d&&(f>=c&&f--,g>=c&&g--)}),this},has:function(a){return a?n.inArray(a,h)>-1:!(!h||!h.length)},empty:function(){return h=[],f=0,this},disable:function(){return h=i=b=void 0,this},disabled:function(){return!h},lock:function(){return i=void 0,b||k.disable(),this},locked:function(){return!i},fireWith:function(a,b){return!h||c&&!i||(b=b||[],b=[a,b.slice?b.slice():b],d?i.push(b):j(b)),this},fire:function(){return k.fireWith(this,arguments),this},fired:function(){return!!c}};return k},n.extend({Deferred:function(a){var b=[["resolve","done",n.Callbacks("once memory"),"resolved"],["reject","fail",n.Callbacks("once memory"),"rejected"],["notify","progress",n.Callbacks("memory")]],c="pending",d={state:function(){return c},always:function(){return e.done(arguments).fail(arguments),this},then:function(){var a=arguments;return n.Deferred(function(c){n.each(b,function(b,f){var g=n.isFunction(a[b])&&a[b];e[f[1]](function(){var a=g&&g.apply(this,arguments);a&&n.isFunction(a.promise)?a.promise().done(c.resolve).fail(c.reject).progress(c.notify):c[f[0]+"With"](this===d?c.promise():this,g?[a]:arguments)})}),a=null}).promise()},promise:function(a){return null!=a?n.extend(a,d):d}},e={};return d.pipe=d.then,n.each(b,function(a,f){var g=f[2],h=f[3];d[f[1]]=g.add,h&&g.add(function(){c=h},b[1^a][2].disable,b[2][2].lock),e[f[0]]=function(){return e[f[0]+"With"](this===e?d:this,arguments),this},e[f[0]+"With"]=g.fireWith}),d.promise(e),a&&a.call(e,e),e},when:function(a){var b=0,c=d.call(arguments),e=c.length,f=1!==e||a&&n.isFunction(a.promise)?e:0,g=1===f?a:n.Deferred(),h=function(a,b,c){return function(e){b[a]=this,c[a]=arguments.length>1?d.call(arguments):e,c===i?g.notifyWith(b,c):--f||g.resolveWith(b,c)}},i,j,k;if(e>1)for(i=new Array(e),j=new Array(e),k=new Array(e);e>b;b++)c[b]&&n.isFunction(c[b].promise)?c[b].promise().done(h(b,k,c)).fail(g.reject).progress(h(b,j,i)):--f;return f||g.resolveWith(k,c),g.promise()}});var H;n.fn.ready=function(a){return n.ready.promise().done(a),this},n.extend({isReady:!1,readyWait:1,holdReady:function(a){a?n.readyWait++:n.ready(!0)},ready:function(a){(a===!0?--n.readyWait:n.isReady)||(n.isReady=!0,a!==!0&&--n.readyWait>0||(H.resolveWith(l,[n]),n.fn.triggerHandler&&(n(l).triggerHandler("ready"),n(l).off("ready"))))}});function I(){l.removeEventListener("DOMContentLoaded",I,!1),a.removeEventListener("load",I,!1),n.ready()}n.ready.promise=function(b){return H||(H=n.Deferred(),"complete"===l.readyState?setTimeout(n.ready):(l.addEventListener("DOMContentLoaded",I,!1),a.addEventListener("load",I,!1))),H.promise(b)},n.ready.promise();var J=n.access=function(a,b,c,d,e,f,g){var h=0,i=a.length,j=null==c;if("object"===n.type(c)){e=!0;for(h in c)n.access(a,b,h,c[h],!0,f,g)}else if(void 0!==d&&(e=!0,n.isFunction(d)||(g=!0),j&&(g?(b.call(a,d),b=null):(j=b,b=function(a,b,c){return j.call(n(a),c)})),b))for(;i>h;h++)b(a[h],c,g?d:d.call(a[h],h,b(a[h],c)));return e?a:j?b.call(a):i?b(a[0],c):f};n.acceptData=function(a){return 1===a.nodeType||9===a.nodeType||!+a.nodeType};function K(){Object.defineProperty(this.cache={},0,{get:function(){return{}}}),this.expando=n.expando+K.uid++}K.uid=1,K.accepts=n.acceptData,K.prototype={key:function(a){if(!K.accepts(a))return 0;var b={},c=a[this.expando];if(!c){c=K.uid++;try{b[this.expando]={value:c},Object.defineProperties(a,b)}catch(d){b[this.expando]=c,n.extend(a,b)}}return this.cache[c]||(this.cache[c]={}),c},set:function(a,b,c){var d,e=this.key(a),f=this.cache[e];if("string"==typeof b)f[b]=c;else if(n.isEmptyObject(f))n.extend(this.cache[e],b);else for(d in b)f[d]=b[d];return f},get:function(a,b){var c=this.cache[this.key(a)];return void 0===b?c:c[b]},access:function(a,b,c){var d;return void 0===b||b&&"string"==typeof b&&void 0===c?(d=this.get(a,b),void 0!==d?d:this.get(a,n.camelCase(b))):(this.set(a,b,c),void 0!==c?c:b)},remove:function(a,b){var c,d,e,f=this.key(a),g=this.cache[f];if(void 0===b)this.cache[f]={};else{n.isArray(b)?d=b.concat(b.map(n.camelCase)):(e=n.camelCase(b),b in g?d=[b,e]:(d=e,d=d in g?[d]:d.match(E)||[])),c=d.length;while(c--)delete g[d[c]]}},hasData:function(a){return!n.isEmptyObject(this.cache[a[this.expando]]||{})},discard:function(a){a[this.expando]&&delete this.cache[a[this.expando]]}};var L=new K,M=new K,N=/^(?:\{[\w\W]*\}|\[[\w\W]*\])$/,O=/([A-Z])/g;function P(a,b,c){var d;if(void 0===c&&1===a.nodeType)if(d="data-"+b.replace(O,"-$1").toLowerCase(),c=a.getAttribute(d),"string"==typeof c){try{c="true"===c?!0:"false"===c?!1:"null"===c?null:+c+""===c?+c:N.test(c)?n.parseJSON(c):c}catch(e){}M.set(a,b,c)}else c=void 0;return c}n.extend({hasData:function(a){return M.hasData(a)||L.hasData(a)},data:function(a,b,c){return M.access(a,b,c) -},removeData:function(a,b){M.remove(a,b)},_data:function(a,b,c){return L.access(a,b,c)},_removeData:function(a,b){L.remove(a,b)}}),n.fn.extend({data:function(a,b){var c,d,e,f=this[0],g=f&&f.attributes;if(void 0===a){if(this.length&&(e=M.get(f),1===f.nodeType&&!L.get(f,"hasDataAttrs"))){c=g.length;while(c--)g[c]&&(d=g[c].name,0===d.indexOf("data-")&&(d=n.camelCase(d.slice(5)),P(f,d,e[d])));L.set(f,"hasDataAttrs",!0)}return e}return"object"==typeof a?this.each(function(){M.set(this,a)}):J(this,function(b){var c,d=n.camelCase(a);if(f&&void 0===b){if(c=M.get(f,a),void 0!==c)return c;if(c=M.get(f,d),void 0!==c)return c;if(c=P(f,d,void 0),void 0!==c)return c}else this.each(function(){var c=M.get(this,d);M.set(this,d,b),-1!==a.indexOf("-")&&void 0!==c&&M.set(this,a,b)})},null,b,arguments.length>1,null,!0)},removeData:function(a){return this.each(function(){M.remove(this,a)})}}),n.extend({queue:function(a,b,c){var d;return a?(b=(b||"fx")+"queue",d=L.get(a,b),c&&(!d||n.isArray(c)?d=L.access(a,b,n.makeArray(c)):d.push(c)),d||[]):void 0},dequeue:function(a,b){b=b||"fx";var c=n.queue(a,b),d=c.length,e=c.shift(),f=n._queueHooks(a,b),g=function(){n.dequeue(a,b)};"inprogress"===e&&(e=c.shift(),d--),e&&("fx"===b&&c.unshift("inprogress"),delete f.stop,e.call(a,g,f)),!d&&f&&f.empty.fire()},_queueHooks:function(a,b){var c=b+"queueHooks";return L.get(a,c)||L.access(a,c,{empty:n.Callbacks("once memory").add(function(){L.remove(a,[b+"queue",c])})})}}),n.fn.extend({queue:function(a,b){var c=2;return"string"!=typeof a&&(b=a,a="fx",c--),arguments.lengthx",k.noCloneChecked=!!b.cloneNode(!0).lastChild.defaultValue}();var U="undefined";k.focusinBubbles="onfocusin"in a;var V=/^key/,W=/^(?:mouse|pointer|contextmenu)|click/,X=/^(?:focusinfocus|focusoutblur)$/,Y=/^([^.]*)(?:\.(.+)|)$/;function Z(){return!0}function $(){return!1}function _(){try{return l.activeElement}catch(a){}}n.event={global:{},add:function(a,b,c,d,e){var f,g,h,i,j,k,l,m,o,p,q,r=L.get(a);if(r){c.handler&&(f=c,c=f.handler,e=f.selector),c.guid||(c.guid=n.guid++),(i=r.events)||(i=r.events={}),(g=r.handle)||(g=r.handle=function(b){return typeof n!==U&&n.event.triggered!==b.type?n.event.dispatch.apply(a,arguments):void 0}),b=(b||"").match(E)||[""],j=b.length;while(j--)h=Y.exec(b[j])||[],o=q=h[1],p=(h[2]||"").split(".").sort(),o&&(l=n.event.special[o]||{},o=(e?l.delegateType:l.bindType)||o,l=n.event.special[o]||{},k=n.extend({type:o,origType:q,data:d,handler:c,guid:c.guid,selector:e,needsContext:e&&n.expr.match.needsContext.test(e),namespace:p.join(".")},f),(m=i[o])||(m=i[o]=[],m.delegateCount=0,l.setup&&l.setup.call(a,d,p,g)!==!1||a.addEventListener&&a.addEventListener(o,g,!1)),l.add&&(l.add.call(a,k),k.handler.guid||(k.handler.guid=c.guid)),e?m.splice(m.delegateCount++,0,k):m.push(k),n.event.global[o]=!0)}},remove:function(a,b,c,d,e){var f,g,h,i,j,k,l,m,o,p,q,r=L.hasData(a)&&L.get(a);if(r&&(i=r.events)){b=(b||"").match(E)||[""],j=b.length;while(j--)if(h=Y.exec(b[j])||[],o=q=h[1],p=(h[2]||"").split(".").sort(),o){l=n.event.special[o]||{},o=(d?l.delegateType:l.bindType)||o,m=i[o]||[],h=h[2]&&new RegExp("(^|\\.)"+p.join("\\.(?:.*\\.|)")+"(\\.|$)"),g=f=m.length;while(f--)k=m[f],!e&&q!==k.origType||c&&c.guid!==k.guid||h&&!h.test(k.namespace)||d&&d!==k.selector&&("**"!==d||!k.selector)||(m.splice(f,1),k.selector&&m.delegateCount--,l.remove&&l.remove.call(a,k));g&&!m.length&&(l.teardown&&l.teardown.call(a,p,r.handle)!==!1||n.removeEvent(a,o,r.handle),delete i[o])}else for(o in i)n.event.remove(a,o+b[j],c,d,!0);n.isEmptyObject(i)&&(delete r.handle,L.remove(a,"events"))}},trigger:function(b,c,d,e){var f,g,h,i,k,m,o,p=[d||l],q=j.call(b,"type")?b.type:b,r=j.call(b,"namespace")?b.namespace.split("."):[];if(g=h=d=d||l,3!==d.nodeType&&8!==d.nodeType&&!X.test(q+n.event.triggered)&&(q.indexOf(".")>=0&&(r=q.split("."),q=r.shift(),r.sort()),k=q.indexOf(":")<0&&"on"+q,b=b[n.expando]?b:new n.Event(q,"object"==typeof b&&b),b.isTrigger=e?2:3,b.namespace=r.join("."),b.namespace_re=b.namespace?new RegExp("(^|\\.)"+r.join("\\.(?:.*\\.|)")+"(\\.|$)"):null,b.result=void 0,b.target||(b.target=d),c=null==c?[b]:n.makeArray(c,[b]),o=n.event.special[q]||{},e||!o.trigger||o.trigger.apply(d,c)!==!1)){if(!e&&!o.noBubble&&!n.isWindow(d)){for(i=o.delegateType||q,X.test(i+q)||(g=g.parentNode);g;g=g.parentNode)p.push(g),h=g;h===(d.ownerDocument||l)&&p.push(h.defaultView||h.parentWindow||a)}f=0;while((g=p[f++])&&!b.isPropagationStopped())b.type=f>1?i:o.bindType||q,m=(L.get(g,"events")||{})[b.type]&&L.get(g,"handle"),m&&m.apply(g,c),m=k&&g[k],m&&m.apply&&n.acceptData(g)&&(b.result=m.apply(g,c),b.result===!1&&b.preventDefault());return b.type=q,e||b.isDefaultPrevented()||o._default&&o._default.apply(p.pop(),c)!==!1||!n.acceptData(d)||k&&n.isFunction(d[q])&&!n.isWindow(d)&&(h=d[k],h&&(d[k]=null),n.event.triggered=q,d[q](),n.event.triggered=void 0,h&&(d[k]=h)),b.result}},dispatch:function(a){a=n.event.fix(a);var b,c,e,f,g,h=[],i=d.call(arguments),j=(L.get(this,"events")||{})[a.type]||[],k=n.event.special[a.type]||{};if(i[0]=a,a.delegateTarget=this,!k.preDispatch||k.preDispatch.call(this,a)!==!1){h=n.event.handlers.call(this,a,j),b=0;while((f=h[b++])&&!a.isPropagationStopped()){a.currentTarget=f.elem,c=0;while((g=f.handlers[c++])&&!a.isImmediatePropagationStopped())(!a.namespace_re||a.namespace_re.test(g.namespace))&&(a.handleObj=g,a.data=g.data,e=((n.event.special[g.origType]||{}).handle||g.handler).apply(f.elem,i),void 0!==e&&(a.result=e)===!1&&(a.preventDefault(),a.stopPropagation()))}return k.postDispatch&&k.postDispatch.call(this,a),a.result}},handlers:function(a,b){var c,d,e,f,g=[],h=b.delegateCount,i=a.target;if(h&&i.nodeType&&(!a.button||"click"!==a.type))for(;i!==this;i=i.parentNode||this)if(i.disabled!==!0||"click"!==a.type){for(d=[],c=0;h>c;c++)f=b[c],e=f.selector+" ",void 0===d[e]&&(d[e]=f.needsContext?n(e,this).index(i)>=0:n.find(e,this,null,[i]).length),d[e]&&d.push(f);d.length&&g.push({elem:i,handlers:d})}return h]*)\/>/gi,bb=/<([\w:]+)/,cb=/<|&#?\w+;/,db=/<(?:script|style|link)/i,eb=/checked\s*(?:[^=]|=\s*.checked.)/i,fb=/^$|\/(?:java|ecma)script/i,gb=/^true\/(.*)/,hb=/^\s*\s*$/g,ib={option:[1,""],thead:[1,"","
"],col:[2,"","
"],tr:[2,"","
"],td:[3,"","
"],_default:[0,"",""]};ib.optgroup=ib.option,ib.tbody=ib.tfoot=ib.colgroup=ib.caption=ib.thead,ib.th=ib.td;function jb(a,b){return n.nodeName(a,"table")&&n.nodeName(11!==b.nodeType?b:b.firstChild,"tr")?a.getElementsByTagName("tbody")[0]||a.appendChild(a.ownerDocument.createElement("tbody")):a}function kb(a){return a.type=(null!==a.getAttribute("type"))+"/"+a.type,a}function lb(a){var b=gb.exec(a.type);return b?a.type=b[1]:a.removeAttribute("type"),a}function mb(a,b){for(var c=0,d=a.length;d>c;c++)L.set(a[c],"globalEval",!b||L.get(b[c],"globalEval"))}function nb(a,b){var c,d,e,f,g,h,i,j;if(1===b.nodeType){if(L.hasData(a)&&(f=L.access(a),g=L.set(b,f),j=f.events)){delete g.handle,g.events={};for(e in j)for(c=0,d=j[e].length;d>c;c++)n.event.add(b,e,j[e][c])}M.hasData(a)&&(h=M.access(a),i=n.extend({},h),M.set(b,i))}}function ob(a,b){var c=a.getElementsByTagName?a.getElementsByTagName(b||"*"):a.querySelectorAll?a.querySelectorAll(b||"*"):[];return void 0===b||b&&n.nodeName(a,b)?n.merge([a],c):c}function pb(a,b){var c=b.nodeName.toLowerCase();"input"===c&&T.test(a.type)?b.checked=a.checked:("input"===c||"textarea"===c)&&(b.defaultValue=a.defaultValue)}n.extend({clone:function(a,b,c){var d,e,f,g,h=a.cloneNode(!0),i=n.contains(a.ownerDocument,a);if(!(k.noCloneChecked||1!==a.nodeType&&11!==a.nodeType||n.isXMLDoc(a)))for(g=ob(h),f=ob(a),d=0,e=f.length;e>d;d++)pb(f[d],g[d]);if(b)if(c)for(f=f||ob(a),g=g||ob(h),d=0,e=f.length;e>d;d++)nb(f[d],g[d]);else nb(a,h);return g=ob(h,"script"),g.length>0&&mb(g,!i&&ob(a,"script")),h},buildFragment:function(a,b,c,d){for(var e,f,g,h,i,j,k=b.createDocumentFragment(),l=[],m=0,o=a.length;o>m;m++)if(e=a[m],e||0===e)if("object"===n.type(e))n.merge(l,e.nodeType?[e]:e);else if(cb.test(e)){f=f||k.appendChild(b.createElement("div")),g=(bb.exec(e)||["",""])[1].toLowerCase(),h=ib[g]||ib._default,f.innerHTML=h[1]+e.replace(ab,"<$1>")+h[2],j=h[0];while(j--)f=f.lastChild;n.merge(l,f.childNodes),f=k.firstChild,f.textContent=""}else l.push(b.createTextNode(e));k.textContent="",m=0;while(e=l[m++])if((!d||-1===n.inArray(e,d))&&(i=n.contains(e.ownerDocument,e),f=ob(k.appendChild(e),"script"),i&&mb(f),c)){j=0;while(e=f[j++])fb.test(e.type||"")&&c.push(e)}return k},cleanData:function(a){for(var b,c,d,e,f=n.event.special,g=0;void 0!==(c=a[g]);g++){if(n.acceptData(c)&&(e=c[L.expando],e&&(b=L.cache[e]))){if(b.events)for(d in b.events)f[d]?n.event.remove(c,d):n.removeEvent(c,d,b.handle);L.cache[e]&&delete L.cache[e]}delete M.cache[c[M.expando]]}}}),n.fn.extend({text:function(a){return J(this,function(a){return void 0===a?n.text(this):this.empty().each(function(){(1===this.nodeType||11===this.nodeType||9===this.nodeType)&&(this.textContent=a)})},null,a,arguments.length)},append:function(){return this.domManip(arguments,function(a){if(1===this.nodeType||11===this.nodeType||9===this.nodeType){var b=jb(this,a);b.appendChild(a)}})},prepend:function(){return this.domManip(arguments,function(a){if(1===this.nodeType||11===this.nodeType||9===this.nodeType){var b=jb(this,a);b.insertBefore(a,b.firstChild)}})},before:function(){return this.domManip(arguments,function(a){this.parentNode&&this.parentNode.insertBefore(a,this)})},after:function(){return this.domManip(arguments,function(a){this.parentNode&&this.parentNode.insertBefore(a,this.nextSibling)})},remove:function(a,b){for(var c,d=a?n.filter(a,this):this,e=0;null!=(c=d[e]);e++)b||1!==c.nodeType||n.cleanData(ob(c)),c.parentNode&&(b&&n.contains(c.ownerDocument,c)&&mb(ob(c,"script")),c.parentNode.removeChild(c));return this},empty:function(){for(var a,b=0;null!=(a=this[b]);b++)1===a.nodeType&&(n.cleanData(ob(a,!1)),a.textContent="");return this},clone:function(a,b){return a=null==a?!1:a,b=null==b?a:b,this.map(function(){return n.clone(this,a,b)})},html:function(a){return J(this,function(a){var b=this[0]||{},c=0,d=this.length;if(void 0===a&&1===b.nodeType)return b.innerHTML;if("string"==typeof a&&!db.test(a)&&!ib[(bb.exec(a)||["",""])[1].toLowerCase()]){a=a.replace(ab,"<$1>");try{for(;d>c;c++)b=this[c]||{},1===b.nodeType&&(n.cleanData(ob(b,!1)),b.innerHTML=a);b=0}catch(e){}}b&&this.empty().append(a)},null,a,arguments.length)},replaceWith:function(){var a=arguments[0];return this.domManip(arguments,function(b){a=this.parentNode,n.cleanData(ob(this)),a&&a.replaceChild(b,this)}),a&&(a.length||a.nodeType)?this:this.remove()},detach:function(a){return this.remove(a,!0)},domManip:function(a,b){a=e.apply([],a);var c,d,f,g,h,i,j=0,l=this.length,m=this,o=l-1,p=a[0],q=n.isFunction(p);if(q||l>1&&"string"==typeof p&&!k.checkClone&&eb.test(p))return this.each(function(c){var d=m.eq(c);q&&(a[0]=p.call(this,c,d.html())),d.domManip(a,b)});if(l&&(c=n.buildFragment(a,this[0].ownerDocument,!1,this),d=c.firstChild,1===c.childNodes.length&&(c=d),d)){for(f=n.map(ob(c,"script"),kb),g=f.length;l>j;j++)h=c,j!==o&&(h=n.clone(h,!0,!0),g&&n.merge(f,ob(h,"script"))),b.call(this[j],h,j);if(g)for(i=f[f.length-1].ownerDocument,n.map(f,lb),j=0;g>j;j++)h=f[j],fb.test(h.type||"")&&!L.access(h,"globalEval")&&n.contains(i,h)&&(h.src?n._evalUrl&&n._evalUrl(h.src):n.globalEval(h.textContent.replace(hb,"")))}return this}}),n.each({appendTo:"append",prependTo:"prepend",insertBefore:"before",insertAfter:"after",replaceAll:"replaceWith"},function(a,b){n.fn[a]=function(a){for(var c,d=[],e=n(a),g=e.length-1,h=0;g>=h;h++)c=h===g?this:this.clone(!0),n(e[h])[b](c),f.apply(d,c.get());return this.pushStack(d)}});var qb,rb={};function sb(b,c){var d,e=n(c.createElement(b)).appendTo(c.body),f=a.getDefaultComputedStyle&&(d=a.getDefaultComputedStyle(e[0]))?d.display:n.css(e[0],"display");return e.detach(),f}function tb(a){var b=l,c=rb[a];return c||(c=sb(a,b),"none"!==c&&c||(qb=(qb||n("