pepepper/youtube-dl
1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-12-22 12:10:10 +09:00
Code Issues Projects Releases Wiki Activity

[YouTube] Fix nsig processing for player b22ef6e7

Browse Source 
* improve extraction of function name (like yt-dlp/yt-dlp#10390)
* always use JSInterp to extract function code (yt-dlp/yt-dlp#10396, thx seproDev, pukkandan)
This commit is contained in:
dirkf 2024-07-10 18:20:59 +01:00
parent d35ce6ce95
commit 16f5bbc464
2 changed files with 14 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
test/test_youtube_signature.py
View File

@ -162,6 +162,10 @@ _NSIG_TESTS = [
'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js', 'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
'1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg', '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
), ),
(
'https://www.youtube.com/s/player/b22ef6e7/player_ias.vflset/en_US/base.js',
'b6HcntHGkvBLk_FRf', 'kNPW6A7FyP2l8A',
),
] ]

24
youtube_dl/extractor/youtube.py
View File

@ -1636,7 +1636,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
try: try:
jsi, player_id, func_code = self._extract_n_function_code(video_id, player_url) jsi, player_id, func_code = self._extract_n_function_code(video_id, player_url)
except ExtractorError as e: except ExtractorError as e:
raise ExtractorError('Unable to extract nsig jsi, player_id, func_codefunction code', cause=e) raise ExtractorError('Unable to extract nsig function code', cause=e)
if self.get_param('youtube_print_sig_code'): if self.get_param('youtube_print_sig_code'):
self.to_screen('Extracted nsig function from {0}:\n{1}\n'.format( self.to_screen('Extracted nsig function from {0}:\n{1}\n'.format(
player_id, func_code[1])) player_id, func_code[1]))
@ -1658,8 +1658,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
def _extract_n_function_name(self, jscode): def _extract_n_function_name(self, jscode):
func_name, idx = self._search_regex( func_name, idx = self._search_regex(
r'\.get\("n"\)\)&&\(b=(?P<nfunc>[a-zA-Z_$][\w$]*)(?:\[(?P<idx>\d+)\])?\([\w$]+\)', # new: (b=String.fromCharCode(110),c=a.get(b))&&c=nfunc[idx](c)
jscode, 'Initial JS player n function name', group=('nfunc', 'idx')) # old: .get("n"))&&(b=nfunc[idx](b)
# older: .get("n"))&&(b=nfunc(b)
r'''(?x)
(?:\(\s*(?P<b>[a-z])\s*=\s*String\s*\.\s*fromCharCode\s*\(\s*110\s*\)\s*,(?P<c>[a-z])\s*=\s*[a-z]\s*)?
\.\s*get\s*\(\s*(?(b)(?P=b)|"n")(?:\s*\)){2}\s*&&\s*\(\s*(?(c)(?P=c)|b)\s*=\s*
(?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\)
''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'))
if not idx: if not idx:
return func_name return func_name
@ -1679,17 +1685,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
func_name = self._extract_n_function_name(jscode) func_name = self._extract_n_function_name(jscode)
# For redundancy func_code = jsi.extract_function_code(func_name)
func_code = self._search_regex(
r'''(?xs)%s\s*=\s*function\s*\((?P<var>[\w$]+)\)\s*
# NB: The end of the regex is intentionally kept strict
{(?P<code>.+?}\s*return\ [\w$]+.join\(""\))};''' % func_name,
jscode, 'nsig function', group=('var', 'code'), default=None)
if func_code:
func_code = ([func_code[0]], func_code[1])
else:
self.write_debug('Extracting nsig function with jsinterp')
func_code = jsi.extract_function_code(func_name)
self.cache.store('youtube-nsig', player_id, func_code) self.cache.store('youtube-nsig', player_id, func_code)
return jsi, player_id, func_code return jsi, player_id, func_code