Browse Source

[extractor/common] Handle malformed f4m manifests

Sergey M․ 10 years ago
parent
commit
97f4aecfc1
1 changed files with 5 additions and 1 deletions
  1. 5 1
      youtube_dl/extractor/common.py

+ 5 - 1
youtube_dl/extractor/common.py

@@ -28,6 +28,7 @@ from ..utils import (
     clean_html,
     clean_html,
     compiled_regex_type,
     compiled_regex_type,
     ExtractorError,
     ExtractorError,
+    fix_xml_ampersands,
     float_or_none,
     float_or_none,
     int_or_none,
     int_or_none,
     RegexNotFoundError,
     RegexNotFoundError,
@@ -837,7 +838,10 @@ class InfoExtractor(object):
     def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
     def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None):
         manifest = self._download_xml(
         manifest = self._download_xml(
             manifest_url, video_id, 'Downloading f4m manifest',
             manifest_url, video_id, 'Downloading f4m manifest',
-            'Unable to download f4m manifest')
+            'Unable to download f4m manifest',
+            # Some manifests may be malformed, e.g. prosiebensat1 generated manifests
+            # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244)
+            transform_source=lambda s: fix_xml_ampersands(s).strip())
 
 
         formats = []
         formats = []
         manifest_version = '1.0'
         manifest_version = '1.0'