Merge tag 'v1.60.0rc2' into develop

Synapse 1.60.0rc2 (2022-05-27)
==============================

This release of Synapse adds a unique index to the `state_group_edges` table, in
order to prevent accidentally introducing duplicate information (for example,
because a database backup was restored multiple times). If your Synapse database
already has duplicate rows in this table, this could fail with an error and
require manual remediation.

Additionally, the signature of the `check_event_for_spam` module callback has changed.
The previous signature has been deprecated and remains working for now. Module authors
should update their modules to use the new signature where possible.

See [the upgrade notes](https://github.com/matrix-org/synapse/blob/develop/docs/upgrade.md#upgrading-to-v1600)
for more details.

Features
--------

- Add an option allowing users to use their password to reauthenticate for privileged actions even though password login is disabled. ([\#12883](#12883))

Bugfixes
--------

- Explicitly close `ijson` coroutines once we are done with them, instead of leaving the garbage collector to close them. ([\#12875](#12875))

Internal Changes
----------------

- Improve URL previews by not including the content of media tags in the generated description. ([\#12887](#12887))

Author: Sean Quah

CHANGES.md

@@ -1,4 +1,4 @@
-Synapse 1.60.0rc1 (2022-05-24)
+Synapse 1.60.0rc2 (2022-05-27)
 ==============================
 
 This release of Synapse adds a unique index to the `state_group_edges` table, in
@@ -17,6 +17,27 @@ for more details.
 Features
 --------
 
+- Add an option allowing users to use their password to reauthenticate for privileged actions even though password login is disabled. ([\#12883](https://github.com/matrix-org/synapse/issues/12883))
+
+
+Bugfixes
+--------
+
+- Explicitly close `ijson` coroutines once we are done with them, instead of leaving the garbage collector to close them. ([\#12875](https://github.com/matrix-org/synapse/issues/12875))
+
+
+Internal Changes
+----------------
+
+- Improve URL previews by not including the content of media tags in the generated description. ([\#12887](https://github.com/matrix-org/synapse/issues/12887))
+
+
+Synapse 1.60.0rc1 (2022-05-24)
+==============================
+
+Features
+--------
+
 - Measure the time taken in spam-checking callbacks and expose those measurements as metrics. ([\#12513](https://github.com/matrix-org/synapse/issues/12513))
 - Add a `default_power_level_content_override` config option to set default room power levels per room preset. ([\#12618](https://github.com/matrix-org/synapse/issues/12618))
 - Add support for [MSC3787: Allowing knocks to restricted rooms](https://github.com/matrix-org/matrix-spec-proposals/pull/3787). ([\#12623](https://github.com/matrix-org/synapse/issues/12623))

debian/changelog

@@ -1,3 +1,9 @@
+matrix-synapse-py3 (1.60.0~rc2) stable; urgency=medium
+
+  * New Synapse release 1.60.0rc2.
+
+ -- Synapse Packaging team <[email protected]>  Fri, 27 May 2022 11:04:55 +0100
+
 matrix-synapse-py3 (1.60.0~rc1) stable; urgency=medium
 
   * New Synapse release 1.60.0rc1.

docs/sample_config.yaml

@@ -2216,7 +2216,9 @@ sso:
 
 
 password_config:
-   # Uncomment to disable password login
+   # Uncomment to disable password login.
+   # Set to `only_for_reauth` to permit reauthentication for users that
+   # have passwords and are already logged in.
    #
    #enabled: false
 

docs/usage/configuration/config_documentation.md

@@ -2930,6 +2930,9 @@ Use this setting to enable password-based logins.
 
 This setting has the following sub-options:
 * `enabled`: Defaults to true.
+   Set to false to disable password authentication.
+   Set to `only_for_reauth` to allow users with existing passwords to use them
+   to log in and reauthenticate, whilst preventing new users from setting passwords.
 * `localdb_enabled`: Set to false to disable authentication against the local password
    database. This is ignored if `enabled` is false, and is only useful
    if you have other `password_providers`. Defaults to true. 

pyproject.toml

@@ -54,7 +54,7 @@ skip_gitignore = true
 
 [tool.poetry]
 name = "matrix-synapse"
-version = "1.60.0rc1"
+version = "1.60.0rc2"
 description = "Homeserver for the Matrix decentralised comms protocol"
 authors = ["Matrix.org Team and Contributors <[email protected]>"]
 license = "Apache-2.0"

synapse/config/auth.py

@@ -29,7 +29,18 @@ def read_config(self, config: JsonDict, **kwargs: Any) -> None:
         if password_config is None:
             password_config = {}
 
-        self.password_enabled = password_config.get("enabled", True)
+        passwords_enabled = password_config.get("enabled", True)
+        # 'only_for_reauth' allows users who have previously set a password to use it,
+        # even though passwords would otherwise be disabled.
+        passwords_for_reauth_only = passwords_enabled == "only_for_reauth"
+
+        self.password_enabled_for_login = (
+            passwords_enabled and not passwords_for_reauth_only
+        )
+        self.password_enabled_for_reauth = (
+            passwords_for_reauth_only or passwords_enabled
+        )
+
         self.password_localdb_enabled = password_config.get("localdb_enabled", True)
         self.password_pepper = password_config.get("pepper", "")
 
@@ -46,7 +57,9 @@ def read_config(self, config: JsonDict, **kwargs: Any) -> None:
     def generate_config_section(self, **kwargs: Any) -> str:
         return """\
         password_config:
-           # Uncomment to disable password login
+           # Uncomment to disable password login.
+           # Set to `only_for_reauth` to permit reauthentication for users that
+           # have passwords and are already logged in.
            #
            #enabled: false
 

synapse/federation/transport/client.py

@@ -1361,7 +1361,7 @@ class SendJoinParser(ByteParser[SendJoinResponse]):
     def __init__(self, room_version: RoomVersion, v1_api: bool):
         self._response = SendJoinResponse([], [], event_dict={})
         self._room_version = room_version
-        self._coros = []
+        self._coros: List[Generator[None, bytes, None]] = []
 
         # The V1 API has the shape of `[200, {...}]`, which we handle by
         # prefixing with `item.*`.
@@ -1409,6 +1409,9 @@ def write(self, data: bytes) -> int:
         return len(data)
 
     def finish(self) -> SendJoinResponse:
+        for c in self._coros:
+            c.close()
+
         if self._response.event_dict:
             self._response.event = make_event_from_dict(
                 self._response.event_dict, self._room_version
@@ -1431,7 +1434,7 @@ class _StateParser(ByteParser[StateRequestResponse]):
     def __init__(self, room_version: RoomVersion):
         self._response = StateRequestResponse([], [])
         self._room_version = room_version
-        self._coros = [
+        self._coros: List[Generator[None, bytes, None]] = [
             ijson.items_coro(
                 _event_list_parser(room_version, self._response.state),
                 "pdus.item",
@@ -1450,4 +1453,6 @@ def write(self, data: bytes) -> int:
         return len(data)
 
     def finish(self) -> StateRequestResponse:
+        for c in self._coros:
+            c.close()
         return self._response

synapse/handlers/auth.py

@@ -210,7 +210,8 @@ def __init__(self, hs: "HomeServer"):
 
         self.hs = hs  # FIXME better possibility to access registrationHandler later?
         self.macaroon_gen = hs.get_macaroon_generator()
-        self._password_enabled = hs.config.auth.password_enabled
+        self._password_enabled_for_login = hs.config.auth.password_enabled_for_login
+        self._password_enabled_for_reauth = hs.config.auth.password_enabled_for_reauth
         self._password_localdb_enabled = hs.config.auth.password_localdb_enabled
         self._third_party_rules = hs.get_third_party_event_rules()
 
@@ -387,13 +388,13 @@ def get_new_session_data() -> JsonDict:
         return params, session_id
 
     async def _get_available_ui_auth_types(self, user: UserID) -> Iterable[str]:
-        """Get a list of the authentication types this user can use"""
+        """Get a list of the user-interactive authentication types this user can use."""
 
         ui_auth_types = set()
 
         # if the HS supports password auth, and the user has a non-null password, we
         # support password auth
-        if self._password_localdb_enabled and self._password_enabled:
+        if self._password_localdb_enabled and self._password_enabled_for_reauth:
             lookupres = await self._find_user_id_and_pwd_hash(user.to_string())
             if lookupres:
                 _, password_hash = lookupres
@@ -402,7 +403,7 @@ async def _get_available_ui_auth_types(self, user: UserID) -> Iterable[str]:
 
         # also allow auth from password providers
         for t in self.password_auth_provider.get_supported_login_types().keys():
-            if t == LoginType.PASSWORD and not self._password_enabled:
+            if t == LoginType.PASSWORD and not self._password_enabled_for_reauth:
                 continue
             ui_auth_types.add(t)
 
@@ -710,7 +711,7 @@ async def _check_auth_dict(
             return res
 
         # fall back to the v1 login flow
-        canonical_id, _ = await self.validate_login(authdict)
+        canonical_id, _ = await self.validate_login(authdict, is_reauth=True)
         return canonical_id
 
     def _get_params_recaptcha(self) -> dict:
@@ -1064,7 +1065,7 @@ def can_change_password(self) -> bool:
         Returns:
             Whether users on this server are allowed to change or set a password
         """
-        return self._password_enabled and self._password_localdb_enabled
+        return self._password_enabled_for_login and self._password_localdb_enabled
 
     def get_supported_login_types(self) -> Iterable[str]:
         """Get a the login types supported for the /login API
@@ -1089,9 +1090,9 @@ def get_supported_login_types(self) -> Iterable[str]:
         # that comes first, where it's present.
         if LoginType.PASSWORD in types:
             types.remove(LoginType.PASSWORD)
-            if self._password_enabled:
+            if self._password_enabled_for_login:
                 types.insert(0, LoginType.PASSWORD)
-        elif self._password_localdb_enabled and self._password_enabled:
+        elif self._password_localdb_enabled and self._password_enabled_for_login:
             types.insert(0, LoginType.PASSWORD)
 
         return types
@@ -1100,6 +1101,7 @@ async def validate_login(
         self,
         login_submission: Dict[str, Any],
         ratelimit: bool = False,
+        is_reauth: bool = False,
     ) -> Tuple[str, Optional[Callable[["LoginResponse"], Awaitable[None]]]]:
         """Authenticates the user for the /login API
 
@@ -1110,6 +1112,9 @@ async def validate_login(
             login_submission: the whole of the login submission
                 (including 'type' and other relevant fields)
             ratelimit: whether to apply the failed_login_attempt ratelimiter
+            is_reauth: whether this is part of a User-Interactive Authorisation
+                flow to reauthenticate for a privileged action (rather than a
+                new login)
         Returns:
             A tuple of the canonical user id, and optional callback
                 to be called once the access token and device id are issued
@@ -1132,8 +1137,14 @@ async def validate_login(
         # special case to check for "password" for the check_password interface
         # for the auth providers
         password = login_submission.get("password")
+
         if login_type == LoginType.PASSWORD:
-            if not self._password_enabled:
+            if is_reauth:
+                passwords_allowed_here = self._password_enabled_for_reauth
+            else:
+                passwords_allowed_here = self._password_enabled_for_login
+
+            if not passwords_allowed_here:
                 raise SynapseError(400, "Password login has been disabled.")
             if not isinstance(password, str):
                 raise SynapseError(400, "Bad parameter: password", Codes.INVALID_PARAM)

synapse/http/matrixfederationclient.py

@@ -223,6 +223,7 @@ async def _handle_response(
 
     max_response_size = parser.MAX_RESPONSE_SIZE
 
+    finished = False
     try:
         check_content_type_is(response.headers, parser.CONTENT_TYPE)
 
@@ -231,6 +232,7 @@ async def _handle_response(
 
         length = await make_deferred_yieldable(d)
 
+        finished = True
         value = parser.finish()
     except BodyExceededMaxSize as e:
         # The response was too big.
@@ -281,6 +283,15 @@ async def _handle_response(
             e,
         )
         raise
+    finally:
+        if not finished:
+            # There was an exception and we didn't `finish()` the parse.
+            # Let the parser know that it can free up any resources.
+            try:
+                parser.finish()
+            except Exception:
+                # Ignore any additional exceptions.
+                pass
 
     time_taken_secs = reactor.seconds() - start_ms / 1000
 

synapse/rest/media/v1/preview_html.py

@@ -246,7 +246,9 @@ def parse_html_description(tree: "etree.Element") -> Optional[str]:
 
     Grabs any text nodes which are inside the <body/> tag, unless they are within
     an HTML5 semantic markup tag (<header/>, <nav/>, <aside/>, <footer/>), or
-    if they are within a <script/> or <style/> tag.
+    if they are within a <script/>, <svg/> or <style/> tag, or if they are within
+    a tag whose content is usually only shown to old browsers
+    (<iframe/>, <video/>, <canvas/>, <picture/>).
 
     This is a very very very coarse approximation to a plain text render of the page.
 
@@ -268,6 +270,12 @@ def parse_html_description(tree: "etree.Element") -> Optional[str]:
         "script",
         "noscript",
         "style",
+        "svg",
+        "iframe",
+        "video",
+        "canvas",
+        "img",
+        "picture",
         etree.Comment,
     )