[extractor/common] Speed-up media tags regex (closes #11979)

This commit is contained in:
Sergey M․ 2017-02-06 00:20:30 +07:00
parent 815482d4eb
commit 2aec7256ae
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -1959,7 +1959,12 @@ def _media_formats(src, cur_media_type):
media_tags = [(media_tag, media_type, '')
for media_tag, media_type
in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
media_tags.extend(re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', 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
# https://github.com/rg3/youtube-dl/issues/11979, example URL:
# http://www.porntrex.com/maps/videositemap.xml).
r'(?s)(<(?P<tag>video|audio)(?:\s+[^>]*)?>)(.*?)</(?P=tag)>', webpage))
for media_tag, media_type, media_content in media_tags:
media_info = {
'formats': [],