summary refs log tree commit diff
path: root/synapse/rest
diff options
context:
space:
mode:
authorPatrick Cloke <clokep@users.noreply.github.com>2021-10-14 14:51:44 -0400
committerGitHub <noreply@github.com>2021-10-14 18:51:44 +0000
commitefd0074ab76a9449087e38afadcc5a1b4d5a2813 (patch)
tree31b9aac04d2adb2bd6728edb8834b902bb27ffd9 /synapse/rest
parentAttempt different character encodings when previewing a URL. (#11077) (diff)
downloadsynapse-efd0074ab76a9449087e38afadcc5a1b4d5a2813.tar.xz
Ensure each charset is attempted only once during media preview. (#11089)
There's no point in trying more than once since it is guaranteed to
continually fail.
Diffstat (limited to 'synapse/rest')
-rw-r--r--synapse/rest/media/v1/preview_url_resource.py34
1 files changed, 28 insertions, 6 deletions
diff --git a/synapse/rest/media/v1/preview_url_resource.py b/synapse/rest/media/v1/preview_url_resource.py
index 7ee91a0c05..278fd901e2 100644
--- a/synapse/rest/media/v1/preview_url_resource.py
+++ b/synapse/rest/media/v1/preview_url_resource.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 codecs
 import datetime
 import errno
 import fnmatch
@@ -22,7 +23,7 @@ import re
 import shutil
 import sys
 import traceback
-from typing import TYPE_CHECKING, Dict, Generator, Iterable, Optional, Tuple, Union
+from typing import TYPE_CHECKING, Dict, Generator, Iterable, Optional, Set, Tuple, Union
 from urllib import parse as urlparse
 
 import attr
@@ -631,6 +632,14 @@ class PreviewUrlResource(DirectServeJsonResource):
             logger.debug("No media removed from url cache")
 
 
+def _normalise_encoding(encoding: str) -> Optional[str]:
+    """Use the Python codec's name as the normalised entry."""
+    try:
+        return codecs.lookup(encoding).name
+    except LookupError:
+        return None
+
+
 def get_html_media_encodings(body: bytes, content_type: Optional[str]) -> Iterable[str]:
     """
     Get potential encoding of the body based on the (presumably) HTML body or the content-type header.
@@ -652,30 +661,43 @@ def get_html_media_encodings(body: bytes, content_type: Optional[str]) -> Iterab
     Returns:
         The character encoding of the body, as a string.
     """
+    # There's no point in returning an encoding more than once.
+    attempted_encodings: Set[str] = set()
+
     # Limit searches to the first 1kb, since it ought to be at the top.
     body_start = body[:1024]
 
     # Check if it has an encoding set in a meta tag.
     match = _charset_match.search(body_start)
     if match:
-        yield match.group(1).decode("ascii")
+        encoding = _normalise_encoding(match.group(1).decode("ascii"))
+        if encoding:
+            attempted_encodings.add(encoding)
+            yield encoding
 
     # TODO Support <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/>
 
     # Check if it has an XML document with an encoding.
     match = _xml_encoding_match.match(body_start)
     if match:
-        yield match.group(1).decode("ascii")
+        encoding = _normalise_encoding(match.group(1).decode("ascii"))
+        if encoding and encoding not in attempted_encodings:
+            attempted_encodings.add(encoding)
+            yield encoding
 
     # Check the HTTP Content-Type header for a character set.
     if content_type:
         content_match = _content_type_match.match(content_type)
         if content_match:
-            yield content_match.group(1)
+            encoding = _normalise_encoding(content_match.group(1))
+            if encoding and encoding not in attempted_encodings:
+                attempted_encodings.add(encoding)
+                yield encoding
 
     # Finally, fallback to UTF-8, then windows-1252.
-    yield "utf-8"
-    yield "windows-1252"
+    for fallback in ("utf-8", "cp1252"):
+        if fallback not in attempted_encodings:
+            yield fallback
 
 
 def decode_body(