[extractor/common] Support HTML media elements without child nodes

This commit is contained in:
Yen Chi Hsuan 2016-10-12 01:40:28 +08:00
parent 55642487f0
commit cea364f70c
No known key found for this signature in database
GPG Key ID: 3FDDD575826C5C30
2 changed files with 6 additions and 1 deletions

View File

@ -1,6 +1,7 @@
version <unreleased>
Core
+ Support HTML media elements without child nodes
* [Makefile] Support for GNU make < 4 is fixed; BSD make dropped (#9387)
Extractors

View File

@ -1802,7 +1802,11 @@ def _media_formats(src, cur_media_type):
return is_plain_url, formats
entries = []
for media_tag, media_type, media_content in re.findall(r'(?s)(<(?P<tag>video|audio)[^>]*>)(.*?)</(?P=tag)>', webpage):
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))
for media_tag, media_type, media_content in media_tags:
media_info = {
'formats': [],
'subtitles': {},