summary refs log tree commit diff
path: root/synapse
diff options
context:
space:
mode:
authorPatrick Cloke <clokep@users.noreply.github.com>2021-10-27 10:48:02 -0400
committerGitHub <noreply@github.com>2021-10-27 14:48:02 +0000
commitb3e843be88d67633d11711ecc80d4e0390b1e723 (patch)
tree77f33ea6e023273ceee7bc2fa97d3551c103bfef /synapse
parentUpdate release date (diff)
downloadsynapse-b3e843be88d67633d11711ecc80d4e0390b1e723.tar.xz
Fix URL preview errors when previewing XML documents. (#11196)
Diffstat (limited to 'synapse')
-rw-r--r--synapse/rest/media/v1/preview_url_resource.py9
1 files changed, 6 insertions, 3 deletions
diff --git a/synapse/rest/media/v1/preview_url_resource.py b/synapse/rest/media/v1/preview_url_resource.py
index 278fd901e2..8ca97b5b18 100644
--- a/synapse/rest/media/v1/preview_url_resource.py
+++ b/synapse/rest/media/v1/preview_url_resource.py
@@ -718,9 +718,12 @@ def decode_body(
     if not body:
         return None
 
+    # The idea here is that multiple encodings are tried until one works.
+    # Unfortunately the result is never used and then LXML will decode the string
+    # again with the found encoding.
     for encoding in get_html_media_encodings(body, content_type):
         try:
-            body_str = body.decode(encoding)
+            body.decode(encoding)
         except Exception:
             pass
         else:
@@ -732,11 +735,11 @@ def decode_body(
     from lxml import etree
 
     # Create an HTML parser.
-    parser = etree.HTMLParser(recover=True, encoding="utf-8")
+    parser = etree.HTMLParser(recover=True, encoding=encoding)
 
     # Attempt to parse the body. Returns None if the body was successfully
     # parsed, but no tree was found.
-    return etree.fromstring(body_str, parser)
+    return etree.fromstring(body, parser)
 
 
 def _calc_og(tree: "etree.Element", media_uri: str) -> Dict[str, Optional[str]]: