0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-23 02:25:11 +00:00

[generic] Generalize redirect regex

This commit is contained in:
Sergey M․ 2015-03-18 00:05:40 +06:00
parent 84f8101606
commit ed9a25dd61

View file

@ -1268,16 +1268,16 @@ def filter_video(urls):
# HTML5 video
found = re.findall(r'(?s)<video[^<]*(?:>.*?<source[^>]*)?\s+src=["\'](.*?)["\']', webpage)
if not found:
REDIRECT_REGEX = r'[0-9]{,2};\s*(?:URL|url)=\'?([^\'"]+)'
found = re.search(
r'(?i)<meta\s+(?=(?:[a-z-]+="[^"]+"\s+)*http-equiv="refresh")'
r'(?:[a-z-]+="[^"]+"\s+)*?content="[0-9]{,2};\s*(?:URL|url)=\'?([^\'"]+)',
r'(?:[a-z-]+="[^"]+"\s+)*?content="%s' % REDIRECT_REGEX,
webpage)
if not found:
# Look also in Refresh HTTP header
refresh_header = head_response.headers.get('Refresh')
if refresh_header:
found = re.search(
r'[0-9]{,2};\s*(?:URL|url)=(.+)', refresh_header)
found = re.search(REDIRECT_REGEX, refresh_header)
if found:
new_url = found.group(1)
self.report_following_redirect(new_url)