Merge branch 'rav/url_preview_limit_title_2' into matrix-org-hotfixes

AUTHORS.rst

@@ -1,34 +1,8 @@
-Erik Johnston <erik at matrix.org>
- * HS core
- * Federation API impl
+The following is an incomplete list of people outside the core team who have
+contributed to Synapse. It is no longer maintained: more recent contributions
+are listed in the `changelog <CHANGES.md>`_.
 
-Mark Haines <mark at matrix.org>
- * HS core
- * Crypto
- * Content repository
- * CS v2 API impl
-
-Kegan Dougal <kegan at matrix.org>
- * HS core
- * CS v1 API impl
- * AS API impl
-
-Paul "LeoNerd" Evans <paul at matrix.org>
- * HS core
- * Presence
- * Typing Notifications
- * Performance metrics and caching layer
-
-Dave Baker <dave at matrix.org>
- * Push notifications
- * Auth CS v2 impl
-
-Matthew Hodgson <matthew at matrix.org>
- * General doc & housekeeping
- * Vertobot/vertobridge matrix<->verto PoC
-
-Emmanuel Rohee <manu at matrix.org>
- * Supporting iOS clients (testability and fallback registration)
+----
 
 Turned to Dust <dwinslow86 at gmail.com>
  * ArchLinux installation instructions
@@ -62,16 +36,13 @@ Christoph Witzany <christoph at web.crofting.com>
  * Add LDAP support for authentication
 
 Pierre Jaury <pierre at jaury.eu>
-* Docker packaging
+ * Docker packaging
 
 Serban Constantin <serban.constantin at gmail dot com>
  * Small bug fix
 
-Jason Robinson <jasonr at matrix.org>
- * Minor fixes
-
 Joseph Weston <joseph at weston.cloud>
- + Add admin API for querying HS version
+ * Add admin API for querying HS version
 
 Benjamin Saunders <ben.e.saunders at gmail dot com>
  * Documentation improvements

CHANGES.md

@@ -1,3 +1,17 @@
+Synapse 1.5.0 (2019-10-29)
+==========================
+
+Security updates
+----------------
+
+This release includes a security fix ([\#6262](https://github.com/matrix-org/synapse/issues/6262), below). Administrators are encouraged to upgrade as soon as possible.
+
+Bugfixes
+--------
+
+- Fix bug where room directory search was case sensitive. ([\#6268](https://github.com/matrix-org/synapse/issues/6268))
+
+
 Synapse 1.5.0rc2 (2019-10-28)
 =============================
 
@@ -19,13 +33,6 @@ Internal Changes
 Synapse 1.5.0rc1 (2019-10-24)
 ==========================
 
-This release includes a database migration step **which may take a long time to complete**:
-
-- Allow devices to be marked as hidden, for use by features such as cross-signing.
-  This adds a new field with a default value to the devices field in the database,
-  and so the database upgrade may take a long time depending on how many devices
-  are in the database. ([\#5759](https://github.com/matrix-org/synapse/issues/5759))
-
 Features
 --------
 
@@ -69,6 +76,10 @@ Internal Changes
 ----------------
 
 - Update `user_filters` table to have a unique index, and non-null columns. Thanks to @pik for contributing this. ([\#1172](https://github.com/matrix-org/synapse/issues/1172), [\#6175](https://github.com/matrix-org/synapse/issues/6175), [\#6184](https://github.com/matrix-org/synapse/issues/6184))
+- Allow devices to be marked as hidden, for use by features such as cross-signing.
+  This adds a new field with a default value to the devices field in the database,
+  and so the database upgrade may take a long time depending on how many devices
+  are in the database. ([\#5759](https://github.com/matrix-org/synapse/issues/5759))
 - Move lookup-related functions from RoomMemberHandler to IdentityHandler. ([\#5978](https://github.com/matrix-org/synapse/issues/5978))
 - Improve performance of the public room list directory. ([\#6019](https://github.com/matrix-org/synapse/issues/6019), [\#6152](https://github.com/matrix-org/synapse/issues/6152), [\#6153](https://github.com/matrix-org/synapse/issues/6153), [\#6154](https://github.com/matrix-org/synapse/issues/6154))
 - Edit header dicts docstrings in `SimpleHttpClient` to note that `str` or `bytes` can be passed as header keys. ([\#6077](https://github.com/matrix-org/synapse/issues/6077))

CONTRIBUTING.rst

@@ -114,17 +114,6 @@ directory, you will need both a regular newsfragment *and* an entry in the
 debian changelog. (Though typically such changes should be submitted as two
 separate pull requests.)
 
-Attribution
-~~~~~~~~~~~
-
-Everyone who contributes anything to Matrix is welcome to be listed in the
-AUTHORS.rst file for the project in question. Please feel free to include a
-change to AUTHORS.rst in your pull request to list yourself and a short
-description of the area(s) you've worked on. Also, we sometimes have swag to
-give away to contributors - if you feel that Matrix-branded apparel is missing
-from your life, please mail us your shipping address to matrix at matrix.org and
-we'll try to fix it :)
-
 Sign off
 ~~~~~~~~
 

UPGRADE.rst

@@ -2,7 +2,7 @@ Upgrading Synapse
 =================
 
 Before upgrading check if any special steps are required to upgrade from the
-what you currently have installed to current version of Synapse. The extra
+version you currently have installed to the current version of Synapse. The extra
 instructions that may be required are listed later in this document.
 
 * If Synapse was installed using `prebuilt packages
@@ -29,7 +29,7 @@ instructions that may be required are listed later in this document.
      running:
 
      .. code:: bash
-     
+
        git pull
        pip install --upgrade .
 
@@ -75,6 +75,16 @@ 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 v1.5.0
+===================
+
+This release includes a database migration which may take several minutes to
+complete if there are a large number (more than a million or so) of entries in
+the ``devices`` table. This is only likely to a be a problem on very large
+installations.
+
+
 Upgrading to v1.4.0
 ===================
 

changelog.d/6286.bugfix

@@ -1 +0,0 @@
-Fix bug where room directory search was case sensitive.

changelog.d/6331.feature

@@ -0,0 +1 @@
+Limit the length of data returned by url previews, to prevent DoS attacks.

changelog.d/6334.feature

@@ -0,0 +1 @@
+Limit the length of data returned by url previews, to prevent DoS attacks.

debian/changelog

@@ -1,3 +1,9 @@
+matrix-synapse-py3 (1.5.0) stable; urgency=medium
+
+  * New synapse release 1.5.0.
+
+ -- Synapse Packaging team <packages@matrix.org>  Tue, 29 Oct 2019 14:28:41 +0000
+
 matrix-synapse-py3 (1.4.1) stable; urgency=medium
 
   * New synapse release 1.4.1.

synapse/__init__.py

@@ -36,7 +36,7 @@ try:
 except ImportError:
     pass
 
-__version__ = "1.5.0rc2"
+__version__ = "1.5.0"
 
 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

synapse/rest/media/v1/preview_url_resource.py

@@ -56,6 +56,9 @@ logger = logging.getLogger(__name__)
 _charset_match = re.compile(br"<\s*meta[^>]*charset\s*=\s*([a-z0-9-]+)", flags=re.I)
 _content_type_match = re.compile(r'.*; *charset="?(.*?)"?(;|$)', flags=re.I)
 
+OG_TAG_NAME_MAXLEN = 50
+OG_TAG_VALUE_MAXLEN = 1000
+
 
 class PreviewUrlResource(DirectServeResource):
     isLeaf = True
@@ -169,7 +172,7 @@ class PreviewUrlResource(DirectServeResource):
             ts (int):
 
         Returns:
-            Deferred[str]: json-encoded og data
+            Deferred[bytes]: json-encoded og data
         """
         # check the URL cache in the DB (which will also provide us with
         # historical previews, if we have any)
@@ -270,6 +273,18 @@ class PreviewUrlResource(DirectServeResource):
             logger.warn("Failed to find any OG data in %s", url)
             og = {}
 
+        # filter out any stupidly long values
+        keys_to_remove = []
+        for k, v in og.items():
+            # values can be numeric as well as strings, hence the cast to str
+            if len(k) > OG_TAG_NAME_MAXLEN or len(str(v)) > OG_TAG_VALUE_MAXLEN:
+                logger.warning(
+                    "Pruning overlong tag %s from OG data", k[:OG_TAG_NAME_MAXLEN]
+                )
+                keys_to_remove.append(k)
+        for k in keys_to_remove:
+            del og[k]
+
         logger.debug("Calculated OG for %s as %s", url, og)
 
         jsonog = json.dumps(og)
@@ -504,6 +519,10 @@ def _calc_og(tree, media_uri):
     og = {}
     for tag in tree.xpath("//*/meta[starts-with(@property, 'og:')]"):
         if "content" in tag.attrib:
+            # if we've got more than 50 tags, someone is taking the piss
+            if len(og) >= 50:
+                logger.warning("Skipping OG for page with too many 'og:' tags")
+                return {}
             og[tag.attrib["property"]] = tag.attrib["content"]
 
     # TODO: grab article: meta tags too, e.g.:

tests/rest/media/v1/test_url_preview.py

@@ -247,6 +247,41 @@ class URLPreviewTests(unittest.HomeserverTestCase):
         self.assertEqual(channel.code, 200)
         self.assertEqual(channel.json_body["og:title"], "\u0434\u043a\u0430")
 
+    def test_overlong_title(self):
+        self.lookups["matrix.org"] = [(IPv4Address, "8.8.8.8")]
+
+        end_content = (
+            b"<html><head>"
+            b"<title>" + b"x" * 2000 + b"</title>"
+            b'<meta property="og:description" content="hi" />'
+            b"</head></html>"
+        )
+
+        request, channel = self.make_request(
+            "GET", "url_preview?url=http://matrix.org", shorthand=False
+        )
+        request.render(self.preview_url)
+        self.pump()
+
+        client = self.reactor.tcpClients[0][2].buildProtocol(None)
+        server = AccumulatingProtocol()
+        server.makeConnection(FakeTransport(client, self.reactor))
+        client.makeConnection(FakeTransport(server, self.reactor))
+        client.dataReceived(
+            (
+                b"HTTP/1.0 200 OK\r\nContent-Length: %d\r\n"
+                b'Content-Type: text/html; charset="windows-1251"\r\n\r\n'
+            )
+            % (len(end_content),)
+            + end_content
+        )
+
+        self.pump()
+        self.assertEqual(channel.code, 200)
+        res = channel.json_body
+        # We should only see the `og:description` field, as `title` is too long and should be stripped out
+        self.assertCountEqual(["og:description"], res.keys())
+
     def test_ipaddr(self):
         """
         IP addresses can be previewed directly.