diff options
| author | Sergey M․ <dstftw@gmail.com> | 2020-12-09 00:05:21 +0700 |
|---|---|---|
| committer | Sergey M․ <dstftw@gmail.com> | 2020-12-09 00:05:21 +0700 |
| commit | 5a1fbbf8b7215aab0e6382e93eaa1561093352cf (patch) | |
| tree | cf405ecd230d2259465ed2d65be86190b1d6ef3f /youtube_dl | |
| parent | e2bdf8bf4f3de7698d1d2844687e3acc760b34e7 (diff) | |
[extractor/common] Fix inline HTML5 media tags processing and add test (closes #27345)
Diffstat (limited to 'youtube_dl')
| -rw-r--r-- | youtube_dl/extractor/common.py | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index dd07a1cae..74e40fabb 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -2515,9 +2515,9 @@ class InfoExtractor(object): # https://www.ampproject.org/docs/reference/components/amp-video) # For dl8-* tags see https://delight-vr.com/documentation/dl8-video/ _MEDIA_TAG_NAME_RE = r'(?:(?:amp|dl8(?:-live)?)-)?(video|audio)' - media_tags = [(media_tag, media_type, '') - for media_tag, media_type - in re.findall(r'(?s)(<%s[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)] + media_tags = [(media_tag, media_tag_name, media_type, '') + for media_tag, media_tag_name, media_type + in re.findall(r'(?s)(<(%s)[^>]*/>)' % _MEDIA_TAG_NAME_RE, webpage)] media_tags.extend(re.findall( # We only allow video|audio followed by a whitespace or '>'. # Allowing more characters may end up in significant slow down (see |