Merge e5bfed6c77b47bdf07713d8d88c8b21247999ae7 into 2b4fbfce25902d557b86b003cf48f738129efce4

thx seproDev, bashonly: yt-dlp/yt-dlp#12748

test/test_youtube_signature.py

@@ -84,6 +84,21 @@ _SIG_TESTS = [
         '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
         '0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xxAj7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJ2OySqa0q',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'AAOAOq0QJ8wRAIgXmPlOPSBkkUs1bYFYlJCfe29xx8j7vgpDL0QwbdV06sCIEzpWqMGkFR20CFOS21Tp-7vj_EMu-m37KtXJoOy1',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        '0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpz2ICs6EVdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        '2aq0aqSyOoJXtK73m-uME_jv7-pT15gOFC02RFkGMqWpzEICs69VdbwQ0LDp1v7j8xx92efCJlYFYb1sUkkBSPOlPmXgIARw8JQ0qOAOAA',
+        'wAOAOq0QJ8ARAIgXmPlOPSBkkUs1bYFYlJCfe29xx8q7v1pDL0QwbdV96sCIEzpWqMGkFR20CFOg51Tp-7vj_EMu-m37KtXJoOySqa0',
+    ),
 ]
 
 _NSIG_TESTS = [
@@ -153,7 +168,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/c57c113c/player_ias.vflset/en_US/base.js',
-        '-Txvy6bT5R6LqgnQNx', 'dcklJCnRUHbgSg',
+        'M92UUMHa8PdvPd3wyM', '3hPqLJsiNZx7yA',
     ),
     (
         'https://www.youtube.com/s/player/5a3b6271/player_ias.vflset/en_US/base.js',
@@ -173,7 +188,7 @@ _NSIG_TESTS = [
     ),
     (
         'https://www.youtube.com/s/player/cfa9e7cb/player_ias.vflset/en_US/base.js',
-        'qO0NiMtYQ7TeJnfFG2', 'k9cuJDHNS5O7kQ',
+        'aCi3iElgd2kq0bxVbQ', 'QX1y8jGb2IbZ0w',
     ),
     (
         'https://www.youtube.com/s/player/8c7583ff/player_ias.vflset/en_US/base.js',
@@ -231,10 +246,6 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/f6e09c70/player_ias_tce.vflset/en_US/base.js',
         'W9HJZKktxuYoDTqW', 'jHbbkcaxm54',
     ),
-    (
-        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
-        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
-    ),
     (
         'https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js',
         'Sy4aDGc0VpYRR9ew_', '5UPOT1VhoZxNLQ',
@@ -259,6 +270,22 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js',
         'ir9-V6cdbCiyKxhr', '2PL7ZDYAALMfmA',
     ),
+    (
+        'https://www.youtube.com/s/player/643afba4/player_ias.vflset/en_US/base.js',
+        'W9HJZKktxuYoDTqW', 'larxUlagTRAcSw',
+    ),
+    (
+        'https://www.youtube.com/s/player/363db69b/player_ias.vflset/en_US/base.js',
+        'eWYu5d5YeY_4LyEDc', 'XJQqf-N7Xra3gg',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/player_ias.vflset/en_US/base.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
+    (
+        'https://www.youtube.com/s/player/4fcd6e4a/tv-player-ias.vflset/tv-player-ias.js',
+        'o_L251jm8yhZkWtBW', 'lXoxI3XvToqn6A',
+    ),
 ]
 
 
@@ -271,6 +298,8 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-en_US.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-phone-de_DE.vflset/base.js', '64dddad9'),
             ('https://www.youtube.com/s/player/64dddad9/player-plasma-ias-tablet-en_US.vflset/base.js', '64dddad9'),
+            ('https://www.youtube.com/s/player/e7567ecf/player_ias_tce.vflset/en_US/base.js', 'e7567ecf'),
+            ('https://www.youtube.com/s/player/643afba4/tv-player-ias.vflset/tv-player-ias.js', '643afba4'),
             # obsolete
             ('https://www.youtube.com/yts/jsbin/player_ias-vfle4-e03/en_US/base.js', 'vfle4-e03'),
             ('https://www.youtube.com/yts/jsbin/player_ias-vfl49f_g4/en_US/base.js', 'vfl49f_g4'),
@@ -280,8 +309,9 @@ class TestPlayerInfo(unittest.TestCase):
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflXGBaUN.js', 'vflXGBaUN'),
             ('https://s.ytimg.com/yts/jsbin/html5player-en_US-vflKjOTVq/html5player.js', 'vflKjOTVq'),
         )
+        ie = YoutubeIE(FakeYDL({'cachedir': False}))
         for player_url, expected_player_id in PLAYER_URLS:
-            player_id = YoutubeIE._extract_player_info(player_url)
+            player_id = ie._extract_player_info(player_url)
             self.assertEqual(player_id, expected_player_id)
 
 
@@ -301,8 +331,8 @@ class TestSignature(unittest.TestCase):
 def t_factory(name, sig_func, url_pattern):
     def make_tfunc(url, sig_input, expected_sig):
         m = url_pattern.match(url)
-        assert m, '%r should follow URL format' % url
+        assert m, '{0!r} should follow URL format'.format(url)
-        test_id = m.group('id')
+        test_id = re.sub(r'[/.-]', '_', m.group('id') or m.group('compat_id'))
 
         def test_func(self):
             basename = 'player-{0}-{1}.js'.format(name, test_id)
@@ -335,12 +365,16 @@ def n_sig(jscode, sig_input):
 
 
 make_sig_test = t_factory(
-    'signature', signature, re.compile(r'.*(?:-|/player/)(?P<id>[a-zA-Z0-9_-]+)(?:/.+\.js|(?:/watch_as3|/html5player)?\.[a-z]+)$'))
+    'signature', signature,
+    re.compile(r'''(?x)
+        .+/(?P<h5>html5)?player(?(h5)(?:-en_US)?-|/)(?P<id>[a-zA-Z0-9/._-]+)
+        (?(h5)/(?:watch_as3|html5player))?\.js$
+    '''))
 for test_spec in _SIG_TESTS:
     make_sig_test(*test_spec)
 
 make_nsig_test = t_factory(
-    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_-]+)/.+.js$'))
+    'nsig', n_sig, re.compile(r'.+/player/(?P<id>[a-zA-Z0-9_/.-]+)\.js$'))
 for test_spec in _NSIG_TESTS:
     make_nsig_test(*test_spec)
 

youtube_dl/extractor/pornhub.py

@@ -7,30 +7,35 @@ import operator
 import re
 
 from .common import InfoExtractor
+
 from ..compat import (
     compat_HTTPError,
-    compat_str,
-    compat_urllib_request,
 )
 from .openload import PhantomJSwrapper
 from ..utils import (
+    clean_html,
     determine_ext,
+    extract_attributes,
     ExtractorError,
+    get_element_by_class,
+    get_element_by_id,
     int_or_none,
     merge_dicts,
-    NO_DEFAULT,
+    parse_count,
-    orderedSet,
     remove_quotes,
-    str_to_int,
+    remove_start,
+    T,
+    traverse_obj,
     update_url_query,
-    urlencode_postdata,
     url_or_none,
+    urlencode_postdata,
+    urljoin,
 )
 
 
 class PornHubBaseIE(InfoExtractor):
     _NETRC_MACHINE = 'pornhub'
-    _PORNHUB_HOST_RE = r'(?:(?P<host>pornhub(?:premium)?\.(?:com|net|org))|pornhubthbh7ap3u\.onion)'
+    _PORNHUB_HOST_RE = r'(?:(?P<host>pornhub(?:premium)?\.(?:com|net|org))|pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd\.onion)'
 
     def _download_webpage_handle(self, *args, **kwargs):
         def dl(*args, **kwargs):
@@ -47,10 +52,7 @@ class PornHubBaseIE(InfoExtractor):
                 r'<body\b[^>]+\bonload=["\']go\(\)',
                 r'document\.cookie\s*=\s*["\']RNKEY=',
                 r'document\.location\.reload\(true\)')):
-            url_or_request = args[0]
+            url = urlh.geturl()
-            url = (url_or_request.get_full_url()
-                   if isinstance(url_or_request, compat_urllib_request.Request)
-                   else url_or_request)
             phantom = PhantomJSwrapper(self, required_version='2.0')
             phantom.get(url, html=webpage)
             webpage, urlh = dl(*args, **kwargs)
@@ -60,11 +62,17 @@ class PornHubBaseIE(InfoExtractor):
     def _real_initialize(self):
         self._logged_in = False
 
+    def _set_age_cookies(self, host):
+        self._set_cookie(host, 'age_verified', '1')
+        self._set_cookie(host, 'accessAgeDisclaimerPH', '1')
+        self._set_cookie(host, 'accessAgeDisclaimerUK', '1')
+        self._set_cookie(host, 'accessPH', '1')
+
     def _login(self, host):
         if self._logged_in:
             return
 
-        site = host.split('.')[0]
+        site = host.split('.', 1)[0]
 
         # Both sites pornhub and pornhubpremium have separate accounts
         # so there should be an option to provide credentials for both.
@@ -81,9 +89,9 @@ class PornHubBaseIE(InfoExtractor):
             login_url, None, 'Downloading %s login page' % site)
 
         def is_logged(webpage):
-            return any(re.search(p, webpage) for p in (
+            return bool(
-                r'class=["\']signOut',
+                get_element_by_id('profileMenuDropdown', webpage)
-                r'>Sign\s+[Oo]ut\s*<'))
+                or get_element_by_class('ph-icon-logout', webpage))
 
         if is_logged(login_page):
             self._logged_in = True
@@ -92,12 +100,12 @@ class PornHubBaseIE(InfoExtractor):
         login_form = self._hidden_inputs(login_page)
 
         login_form.update({
-            'username': username,
+            'email': username,
             'password': password,
         })
 
         response = self._download_json(
-            'https://www.%s/front/authenticate' % host, None,
+            'https://www.%s/front/authenticate' % host, 'login',
             'Logging in to %s' % site,
             data=urlencode_postdata(login_form),
             headers={
@@ -119,17 +127,12 @@ class PornHubBaseIE(InfoExtractor):
 
 
 class PornHubIE(PornHubBaseIE):
-    IE_DESC = 'PornHub and Thumbzilla'
+    IE_DESC = 'PornHub'  # Thumbzilla -> Redtube.com, Modelhub -> uviu.com
-    _VALID_URL = r'''(?x)
+    _PORNHUB_PATH_RE = r'/(?:(?:view_video\.php%s)\?(?:.+&)?viewkey=%s)(?P<id>[\da-z]+)'
-                    https?://
+    _VALID_URL = r'https?://(?:[^/]+\.)?%s%s' % (
-                        (?:
+        PornHubBaseIE._PORNHUB_HOST_RE, _PORNHUB_PATH_RE % ('|video/show', '|embed/'))
-                            (?:[^/]+\.)?
+    _PORNHUB_PATH_RE = _PORNHUB_PATH_RE % ('', '')
-                            %s
+    _EMBED_REGEX = [r'<iframe\s[^>]*?src=["\'](?P<url>(?:https?:)?//(?:www\.)?pornhub(?:premium)?\.(?:com|net|org)/embed/[\da-z]+)']
-                            /(?:(?:view_video\.php|video/show)\?viewkey=|embed/)|
-                            (?:www\.)?thumbzilla\.com/video/
-                        )
-                        (?P<id>[\da-z]+)
-                    ''' % PornHubBaseIE._PORNHUB_HOST_RE
     _TESTS = [{
         'url': 'http://www.pornhub.com/view_video.php?viewkey=648719015',
         'md5': 'a6391306d050e4547f62b3f485dd9ba9',
@@ -138,6 +141,7 @@ class PornHubIE(PornHubBaseIE):
             'ext': 'mp4',
             'title': 'Seductive Indian beauty strips down and fingers her pink pussy',
             'uploader': 'Babes',
+            'uploader_id': '/users/babes-com',
             'upload_date': '20130628',
             'timestamp': 1372447216,
             'duration': 361,
@@ -148,6 +152,10 @@ class PornHubIE(PornHubBaseIE):
             'age_limit': 18,
             'tags': list,
             'categories': list,
+            'cast': list,
+        },
+        'params': {
+            'format': '[format_id!^=hls]',
         },
     }, {
         # non-ASCII title
@@ -189,14 +197,27 @@ class PornHubIE(PornHubBaseIE):
             'categories': list,
             'subtitles': {
                 'en': [{
-                    "ext": 'srt'
+                    'ext': 'srt',
-                }]
+                }],
             },
         },
         'params': {
             'skip_download': True,
         },
         'skip': 'This video has been disabled',
+    }, {
+        'url': 'http://www.pornhub.com/view_video.php?viewkey=ph601dc30bae19a',
+        'info_dict': {
+            'id': 'ph601dc30bae19a',
+            'ext': 'mp4',
+            'timestamp': 1612564932,
+            'age_limit': 18,
+            'uploader': 'Projekt Melody',
+            'uploader_id': 'projekt-melody',
+            'upload_date': '20210205',
+            'title': '"Welcome to My Pussy Mansion" - CB Stream (02/03/21)',
+            'thumbnail': r're:https?://.+',
+        },
     }, {
         'url': 'http://www.pornhub.com/view_video.php?viewkey=ph557bbb6676d2d',
         'only_matching': True,
@@ -216,9 +237,6 @@ class PornHubIE(PornHubBaseIE):
         # private video
         'url': 'http://www.pornhub.com/view_video.php?viewkey=ph56fd731fce6b7',
         'only_matching': True,
-    }, {
-        'url': 'https://www.thumbzilla.com/video/ph56c6114abd99a/horny-girlfriend-sex',
-        'only_matching': True,
     }, {
         'url': 'http://www.pornhub.com/video/show?viewkey=648719015',
         'only_matching': True,
@@ -244,28 +262,36 @@ class PornHubIE(PornHubBaseIE):
         'url': 'https://www.pornhub.com/view_video.php?viewkey=ph5a9813bfa7156',
         'only_matching': True,
     }, {
-        'url': 'http://pornhubthbh7ap3u.onion/view_video.php?viewkey=ph5a9813bfa7156',
+        'url': 'http://pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd.onion/view_video.php?viewkey=ph5a9813bfa7156',
         'only_matching': True,
     }]
 
-    @staticmethod
+    @classmethod
-    def _extract_urls(webpage):
+    def _extract_urls(cls, webpage):
-        return re.findall(
+        def yield_urls():
-            r'<iframe[^>]+?src=["\'](?P<url>(?:https?:)?//(?:www\.)?pornhub(?:premium)?\.(?:com|net|org)/embed/[\da-z]+)',
+            for p in cls._EMBED_REGEX:
-            webpage)
+                for from_ in re.finditer(p, webpage):
+                    yield from_.group('url')
+
+        return list(yield_urls())
 
     def _extract_count(self, pattern, webpage, name):
-        return str_to_int(self._search_regex(
+        return parse_count(self._search_regex(
             pattern, webpage, '%s count' % name, fatal=False))
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        for _ in range(2):
+            mobj = self._match_valid_url(url)
+            video_id = mobj.group('id') if mobj else self._generic_id(url)
+            _, urlh = self._download_webpage_handle(url, video_id)
+            if url == urlh.geturl():
+                break
+            url = urlh.geturl()
+
         host = mobj.group('host') or 'pornhub.com'
-        video_id = mobj.group('id')
 
         self._login(host)
-
+        self._set_age_cookies(host)
-        self._set_cookie(host, 'age_verified', '1')
 
         def dl_webpage(platform):
             self._set_cookie(host, 'platform', platform)
@@ -276,7 +302,7 @@ class PornHubIE(PornHubBaseIE):
         webpage = dl_webpage('pc')
 
         error_msg = self._html_search_regex(
-            (r'(?s)<div[^>]+class=(["\'])(?:(?!\1).)*\b(?:removed|userMessageSection)\b(?:(?!\1).)*\1[^>]*>(?P<error>.+?)</div>',
+            (r'(?s)<div[^>]+class=("|\')(?:(?!\1).)*\b(?:removed|userMessageSection)\b(?:(?!\1).)*\1[^>]*>(?P<error>.+?)</div>',
              r'(?s)<section[^>]+class=["\']noVideo["\'][^>]*>(?P<error>.+?)</section>'),
             webpage, 'error message', default=None, group='error')
         if error_msg:
@@ -285,9 +311,9 @@ class PornHubIE(PornHubBaseIE):
                 'PornHub said: %s' % error_msg,
                 expected=True, video_id=video_id)
 
-        if any(re.search(p, webpage) for p in (
+        if bool(get_element_by_class('geoBlocked', webpage)
-                r'class=["\']geoBlocked["\']',
+                or self._search_regex(
-                r'>\s*This content is unavailable in your country')):
+                    r'>\s*This content is (unavailable) in your country', webpage, 'geo-restriction', default=False)):
             self.raise_geo_restricted()
 
         # video_title from flashvars contains whitespace instead of non-ASCII (see
@@ -304,36 +330,34 @@ class PornHubIE(PornHubBaseIE):
         video_urls_set = set()
         subtitles = {}
 
-        flashvars = self._parse_json(
+        def add_video_url(video_url, quality=None):
-            self._search_regex(
+            v_url = url_or_none(video_url)
-                r'var\s+flashvars_\d+\s*=\s*({.+?});', webpage, 'flashvars', default='{}'),
+            if not v_url:
-            video_id)
+                return
-        if flashvars:
+            if v_url in video_urls_set:
-            subtitle_url = url_or_none(flashvars.get('closedCaptionsFile'))
+                return
-            if subtitle_url:
+            video_urls.append((v_url, quality))
-                subtitles.setdefault('en', []).append({
+            video_urls_set.add(v_url)
-                    'url': subtitle_url,
-                    'ext': 'srt',
-                })
-            thumbnail = flashvars.get('image_url')
-            duration = int_or_none(flashvars.get('video_duration'))
-            media_definitions = flashvars.get('mediaDefinitions')
-            if isinstance(media_definitions, list):
-                for definition in media_definitions:
-                    if not isinstance(definition, dict):
-                        continue
-                    video_url = definition.get('videoUrl')
-                    if not video_url or not isinstance(video_url, compat_str):
-                        continue
-                    if video_url in video_urls_set:
-                        continue
-                    video_urls_set.add(video_url)
-                    video_urls.append(
-                        (video_url, int_or_none(definition.get('quality'))))
-        else:
-            thumbnail, duration = [None] * 2
 
-        def extract_js_vars(webpage, pattern, default=NO_DEFAULT):
+        flashvars = self._search_json(r'var\s+flashvars_\d+\s*=', webpage, 'flashvars', video_id)
+        flashvars = traverse_obj(flashvars, {
+            'closedCaptionsFile': ('closedCaptionsFile', T(url_or_none)),
+            'image_url': ('image_url', T(url_or_none)),
+            'video_duration': ('video_duration', T(int_or_none)),
+            'mediaDefinitions': ('mediaDefinitions', lambda _, v: v['videoUrl']),
+        }) or {}
+        subtitle_url = flashvars.get('closedCaptionsFile')
+        if subtitle_url:
+            subtitles.setdefault('en', []).append({
+                'url': subtitle_url,
+                'ext': 'srt',
+            })
+        thumbnail = flashvars.get('image_url')
+        duration = flashvars.get('video_duration')
+        for definition in flashvars.get('mediaDefinitions') or []:
+            add_video_url(definition['videoUrl'], int_or_none(definition.get('quality')))
+
+        def extract_js_vars(webpage, pattern, default=None):
             assignments = self._search_regex(
                 pattern, webpage, 'encoded url', default=default)
             if not assignments:
@@ -363,51 +387,33 @@ class PornHubIE(PornHubBaseIE):
                 js_vars[vname] = parse_js_value(value)
             return js_vars
 
-        def add_video_url(video_url):
-            v_url = url_or_none(video_url)
-            if not v_url:
-                return
-            if v_url in video_urls_set:
-                return
-            video_urls.append((v_url, None))
-            video_urls_set.add(v_url)
-
         def parse_quality_items(quality_items):
             q_items = self._parse_json(quality_items, video_id, fatal=False)
-            if not isinstance(q_items, list):
+            for v_url in traverse_obj(q_items, (Ellipsis, 'url')):
-                return
+                add_video_url(v_url)
-            for item in q_items:
-                if isinstance(item, dict):
-                    add_video_url(item.get('url'))
 
         if not video_urls:
             FORMAT_PREFIXES = ('media', 'quality', 'qualityItems')
             js_vars = extract_js_vars(
-                webpage, r'(var\s+(?:%s)_.+)' % '|'.join(FORMAT_PREFIXES),
+                webpage, r'(var\s+(?:%s)_.+)' % '|'.join(FORMAT_PREFIXES))
-                default=None)
+            for key, format_url in js_vars.items():
-            if js_vars:
+                if key.startswith(FORMAT_PREFIXES[-1]):
-                for key, format_url in js_vars.items():
+                    parse_quality_items(format_url)
-                    if key.startswith(FORMAT_PREFIXES[-1]):
+                elif any(key.startswith(p) for p in FORMAT_PREFIXES[:2]):
-                        parse_quality_items(format_url)
+                    add_video_url(format_url)
-                    elif any(key.startswith(p) for p in FORMAT_PREFIXES[:2]):
+            if not video_urls and get_element_by_id('lockedPlayer', webpage):
-                        add_video_url(format_url)
-            if not video_urls and re.search(
-                    r'<[^>]+\bid=["\']lockedPlayer', webpage):
                 raise ExtractorError(
                     'Video %s is locked' % video_id, expected=True)
 
         if not video_urls:
             js_vars = extract_js_vars(
                 dl_webpage('tv'), r'(var.+?mediastring.+?)</script>')
-            add_video_url(js_vars['mediastring'])
+            add_video_url(traverse_obj(js_vars, 'mediastring'))
 
         for mobj in re.finditer(
                 r'<a[^>]+\bclass=["\']downloadBtn\b[^>]+\bhref=(["\'])(?P<url>(?:(?!\1).)+)\1',
                 webpage):
-            video_url = mobj.group('url')
+            add_video_url(mobj.group('url'))
-            if video_url not in video_urls_set:
-                video_urls.append((video_url, None))
-                video_urls_set.add(video_url)
 
         upload_date = None
         formats = []
@@ -433,6 +439,13 @@ class PornHubIE(PornHubBaseIE):
                 'height': height,
             })
 
+        if not video_urls:
+            # import here to avoid mutually recursive dependency
+            from .generic import GenericIE
+            ret = GenericIE.generic_url_result(url, video_id=video_id, video_title=title, force_videoid=True)
+            ret['_type'] = 'url_transparent'
+            return ret
+
         for video_url, height in video_urls:
             if not upload_date:
                 upload_date = self._search_regex(
@@ -440,52 +453,55 @@ class PornHubIE(PornHubBaseIE):
                 if upload_date:
                     upload_date = upload_date.replace('/', '')
             if '/video/get_media' in video_url:
+                # self._set_cookie(host, 'platform', 'tv')
                 medias = self._download_json(video_url, video_id, fatal=False)
-                if isinstance(medias, list):
+                for media in traverse_obj(medias, lambda _, v: v['videoUrl']):
-                    for media in medias:
+                    video_url = url_or_none(media['videoUrl'])
-                        if not isinstance(media, dict):
+                    if not video_url:
-                            continue
+                        continue
-                        video_url = url_or_none(media.get('videoUrl'))
+                    height = int_or_none(media.get('quality'))
-                        if not video_url:
+                    add_format(video_url, height)
-                            continue
-                        height = int_or_none(media.get('quality'))
-                        add_format(video_url, height)
                 continue
             add_format(video_url)
+
         self._sort_formats(
             formats, field_preference=('height', 'width', 'fps', 'format_id'))
 
+        model_profile = self._search_json(
+            r'var\s+MODEL_PROFILE\s*=', webpage, 'model profile', video_id, fatal=False)
         video_uploader = self._html_search_regex(
             r'(?s)From:&nbsp;.+?<(?:a\b[^>]+\bhref=["\']/(?:(?:user|channel)s|model|pornstar)/|span\b[^>]+\bclass=["\']username)[^>]+>(.+?)<',
-            webpage, 'uploader', default=None)
+            webpage, 'uploader', default=None) or model_profile.get('username')
 
         def extract_vote_count(kind, name):
             return self._extract_count(
-                (r'<span[^>]+\bclass="votes%s"[^>]*>([\d,\.]+)</span>' % kind,
+                (r'<span[^>]+\bclass="votes%s"[^>]*>(\d[\d,\.]*[kKmM]?)</span>' % kind,
                  r'<span[^>]+\bclass=["\']votes%s["\'][^>]*\bdata-rating=["\'](\d+)' % kind),
                 webpage, name)
 
         view_count = self._extract_count(
-            r'<span class="count">([\d,\.]+)</span> [Vv]iews', webpage, 'view')
+            r'<span class="count">(\d[\d,\.]*[kKmM]?)</span> [Vv]iews', webpage, 'view')
         like_count = extract_vote_count('Up', 'like')
         dislike_count = extract_vote_count('Down', 'dislike')
         comment_count = self._extract_count(
-            r'All Comments\s*<span>\(([\d,.]+)\)', webpage, 'comment')
+            r'All Comments\s*<span>\((\d[\d,\.]*[kKmM]?)\)', webpage, 'comment')
 
         def extract_list(meta_key):
             div = self._search_regex(
                 r'(?s)<div[^>]+\bclass=["\'].*?\b%sWrapper[^>]*>(.+?)</div>'
                 % meta_key, webpage, meta_key, default=None)
             if div:
-                return re.findall(r'<a[^>]+\bhref=[^>]+>([^<]+)', div)
+                return [clean_html(x) for x in re.findall(r'(?s)<a[^>]+\bhref=[^>]+>.+?</a>', div)]
 
         info = self._search_json_ld(webpage, video_id, default={})
         # description provided in JSON-LD is irrelevant
-        info['description'] = None
+        for k in ('url', 'description'):
+            info.pop(k, None)
 
-        return merge_dicts({
+        return merge_dicts(info, {
             'id': video_id,
             'uploader': video_uploader,
+            'uploader_id': remove_start(model_profile.get('modelProfileLink'), '/model/'),
             'upload_date': upload_date,
             'title': title,
             'thumbnail': thumbnail,
@@ -498,8 +514,9 @@ class PornHubIE(PornHubBaseIE):
             'age_limit': 18,
             'tags': extract_list('tags'),
             'categories': extract_list('categories'),
+            'cast': extract_list('pornstars'),
             'subtitles': subtitles,
-        }, info)
+        })
 
 
 class PornHubPlaylistBaseIE(PornHubBaseIE):
@@ -512,65 +529,28 @@ class PornHubPlaylistBaseIE(PornHubBaseIE):
         # drop-down menu that uses similar pattern for videos (see
         # https://github.com/ytdl-org/youtube-dl/issues/11594).
         container = self._search_regex(
-            r'(?s)(<div[^>]+class=["\']container.+)', webpage,
+            r'(?s)(<div\s[^>]*class=["\']container.+)', webpage,
             'container', default=webpage)
 
+        def entries():
+            seen_ids = set()
+            for m in re.finditer(r'<\w+\s[^>]*(?<!-)\bhref\s*=\s*.("|\'|\b)%s\1[^>]*>' % (PornHubIE._PORNHUB_PATH_RE,), container):
+                video_id = m.group('id')
+                if video_id:
+                    if video_id in seen_ids:
+                        continue
+                    seen_ids.add(video_id)
+                elt = extract_attributes(m.group(0))
+                video_url = urljoin(host, elt.get('href'))
+                yield video_url, video_id, elt.get('title')
+
         return [
             self.url_result(
-                'http://www.%s/%s' % (host, video_url),
+                video_url, PornHubIE.ie_key(), video_title=title, video_id=video_id)
-                PornHubIE.ie_key(), video_title=title)
+            for video_url, video_id, title in entries()
-            for video_url, title in orderedSet(re.findall(
-                r'href="/?(view_video\.php\?.*\bviewkey=[\da-z]+[^"]*)"[^>]*\s+title="([^"]+)"',
-                container))
         ]
 
 
-class PornHubUserIE(PornHubPlaylistBaseIE):
-    _VALID_URL = r'(?P<url>https?://(?:[^/]+\.)?%s/(?:(?:user|channel)s|model|pornstar)/(?P<id>[^/?#&]+))(?:[?#&]|/(?!videos)|$)' % PornHubBaseIE._PORNHUB_HOST_RE
-    _TESTS = [{
-        'url': 'https://www.pornhub.com/model/zoe_ph',
-        'playlist_mincount': 118,
-    }, {
-        'url': 'https://www.pornhub.com/pornstar/liz-vicious',
-        'info_dict': {
-            'id': 'liz-vicious',
-        },
-        'playlist_mincount': 118,
-    }, {
-        'url': 'https://www.pornhub.com/users/russianveet69',
-        'only_matching': True,
-    }, {
-        'url': 'https://www.pornhub.com/channels/povd',
-        'only_matching': True,
-    }, {
-        'url': 'https://www.pornhub.com/model/zoe_ph?abc=1',
-        'only_matching': True,
-    }, {
-        # Unavailable via /videos page, but available with direct pagination
-        # on pornstar page (see [1]), requires premium
-        # 1. https://github.com/ytdl-org/youtube-dl/issues/27853
-        'url': 'https://www.pornhubpremium.com/pornstar/sienna-west',
-        'only_matching': True,
-    }, {
-        # Same as before, multi page
-        'url': 'https://www.pornhubpremium.com/pornstar/lily-labeau',
-        'only_matching': True,
-    }, {
-        'url': 'https://pornhubthbh7ap3u.onion/model/zoe_ph',
-        'only_matching': True,
-    }]
-
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        user_id = mobj.group('id')
-        videos_url = '%s/videos' % mobj.group('url')
-        page = self._extract_page(url)
-        if page:
-            videos_url = update_url_query(videos_url, {'page': page})
-        return self.url_result(
-            videos_url, ie=PornHubPagedVideoListIE.ie_key(), video_id=user_id)
-
-
 class PornHubPagedPlaylistBaseIE(PornHubPlaylistBaseIE):
     @staticmethod
     def _has_more(webpage):
@@ -617,23 +597,77 @@ class PornHubPagedPlaylistBaseIE(PornHubPlaylistBaseIE):
             page_entries = self._extract_entries(webpage, host)
             if not page_entries:
                 break
-            for e in page_entries:
+            for from_ in page_entries:
-                yield e
+                yield from_
             if not self._has_more(webpage):
                 break
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        mobj = self._match_valid_url(url)
         host = mobj.group('host')
         item_id = mobj.group('id')
 
         self._login(host)
+        self._set_age_cookies(host)
 
         return self.playlist_result(self._entries(url, host, item_id), item_id)
 
 
+class PornHubUserIE(PornHubPagedPlaylistBaseIE):
+    _VALID_URL = r'(?P<url>https?://(?:[^/]+\.)?%s/(?P<id>(?:(?:user|channel)s|model|pornstar)/[^/?#&]+))(?:[?#&]|/(?!videos)|$)' % PornHubBaseIE._PORNHUB_HOST_RE
+    _TESTS = [{
+        'url': 'https://www.pornhub.com/model/zoe_ph',
+        'info_dict': {
+            'id': 'zoe_ph',
+        },
+        'playlist_mincount': 118,
+    }, {
+        'url': 'https://www.pornhub.com/pornstar/liz-vicious',
+        'info_dict': {
+            'id': 'liz-vicious',
+        },
+        'playlist_mincount': 118,
+    }, {
+        'url': 'https://www.pornhub.com/users/russianveet69',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.pornhub.com/channels/povd',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.pornhub.com/model/zoe_ph?abc=1',
+        'only_matching': True,
+    }, {
+        # Unavailable via /videos page, but available with direct pagination
+        # on pornstar page (see [1]), requires premium
+        # 1. https://github.com/ytdl-org/youtube-dl/issues/27853
+        'url': 'https://www.pornhubpremium.com/pornstar/sienna-west',
+        'only_matching': True,
+    }, {
+        # Same as before, multi page
+        'url': 'https://www.pornhubpremium.com/pornstar/lily-labeau',
+        'only_matching': True,
+    }, {
+        'url': 'https://pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd.onion/model/zoe_ph',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        mobj = self._match_valid_url(url)
+        user_id, host = mobj.group('id', 'host')
+        videos_url = '%s/videos' % mobj.group('url')
+        page = self._extract_page(url)
+        if page:
+            videos_url = update_url_query(videos_url, {'page': page})
+
+        self._login(host)
+
+        return self.playlist_result(self._entries(videos_url, host, user_id), user_id.split('/')[-1])
+        # return self.url_result(
+        #     videos_url, ie=PornHubPagedVideoListIE.ie_key(), video_id=user_id)
+
+
 class PornHubPagedVideoListIE(PornHubPagedPlaylistBaseIE):
-    _VALID_URL = r'https?://(?:[^/]+\.)?%s/(?P<id>(?:[^/]+/)*[^/?#&]+)' % PornHubBaseIE._PORNHUB_HOST_RE
+    _VALID_URL = r'https?://(?:[^/]+\.)?%s/(?!playlist/|gif/)(?P<id>(?:[^/]+/)*[^/?#&]+)' % PornHubBaseIE._PORNHUB_HOST_RE
     _TESTS = [{
         'url': 'https://www.pornhub.com/model/zoe_ph/videos',
         'only_matching': True,
@@ -642,16 +676,20 @@ class PornHubPagedVideoListIE(PornHubPagedPlaylistBaseIE):
         'only_matching': True,
     }, {
         'url': 'https://www.pornhub.com/pornstar/jenny-blighe/videos',
-        'info_dict': {
+        'only_matching': True,
-            'id': 'pornstar/jenny-blighe/videos',
-        },
-        'playlist_mincount': 149,
     }, {
-        'url': 'https://www.pornhub.com/pornstar/jenny-blighe/videos?page=3',
+        'url': 'https://www.pornhub.com/pornstar/kylie-quinn/videos',
         'info_dict': {
-            'id': 'pornstar/jenny-blighe/videos',
+            'id': 'pornstar/kylie-quinn/videos',
         },
-        'playlist_mincount': 40,
+        'playlist_mincount': 80,
+    }, {
+        'url': 'https://www.pornhub.com/pornstar/kylie-quinn/videos?page=2',
+        'info_dict': {
+            'id': 'pornstar/kylie-quinn/videos',
+        },
+        # specific page: process just that page
+        'playlist_count': 40,
     }, {
         # default sorting as Top Rated Videos
         'url': 'https://www.pornhub.com/channels/povd/videos',
@@ -727,27 +765,14 @@ class PornHubPagedVideoListIE(PornHubPagedPlaylistBaseIE):
         'url': 'https://www.pornhub.com/video/incategories/60fps-1/hd-porn',
         'only_matching': True,
     }, {
-        'url': 'https://www.pornhub.com/playlist/44121572',
+        'url': 'https://pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd.onion/model/zoe_ph/videos',
-        'info_dict': {
-            'id': 'playlist/44121572',
-        },
-        'playlist_mincount': 132,
-    }, {
-        'url': 'https://www.pornhub.com/playlist/4667351',
-        'only_matching': True,
-    }, {
-        'url': 'https://de.pornhub.com/playlist/4667351',
-        'only_matching': True,
-    }, {
-        'url': 'https://pornhubthbh7ap3u.onion/model/zoe_ph/videos',
         'only_matching': True,
     }]
 
     @classmethod
     def suitable(cls, url):
-        return (False
+        return (not any(ph.suitable(url) for ph in (PornHubIE, PornHubUserIE, PornHubUserVideosUploadIE))
-                if PornHubIE.suitable(url) or PornHubUserIE.suitable(url) or PornHubUserVideosUploadIE.suitable(url)
+                and super(PornHubPagedVideoListIE, cls).suitable(url))
-                else super(PornHubPagedVideoListIE, cls).suitable(url))
 
 
 class PornHubUserVideosUploadIE(PornHubPagedPlaylistBaseIE):
@@ -762,6 +787,62 @@ class PornHubUserVideosUploadIE(PornHubPagedPlaylistBaseIE):
         'url': 'https://www.pornhub.com/model/zoe_ph/videos/upload',
         'only_matching': True,
     }, {
-        'url': 'http://pornhubthbh7ap3u.onion/pornstar/jenny-blighe/videos/upload',
+        'url': 'http://pornhubvybmsymdol4iibwgwtkpwmeyd6luq2gxajgjzfjvotyt5zhyd.onion/pornstar/jenny-blighe/videos/upload',
         'only_matching': True,
     }]
+
+
+class PornHubPlaylistIE(PornHubPlaylistBaseIE):
+    _VALID_URL = r'(?P<url>https?://(?:[^/]+\.)?%s/playlist/(?P<id>[^/?#&]+))' % PornHubBaseIE._PORNHUB_HOST_RE
+    _TESTS = [{
+        'url': 'https://www.pornhub.com/playlist/44121572',
+        'info_dict': {
+            'id': '44121572',
+        },
+        'playlist_mincount': 55,
+    }, {
+        'url': 'https://www.pornhub.com/playlist/4667351',
+        'only_matching': True,
+    }, {
+        'url': 'https://de.pornhub.com/playlist/4667351',
+        'only_matching': True,
+    }, {
+        'url': 'https://de.pornhub.com/playlist/4667351?page=2',
+        'only_matching': True,
+    }]
+
+    def _entries(self, url, host, item_id):
+        webpage = self._download_webpage(url, item_id, 'Downloading page 1')
+        playlist_id = self._search_regex(r'var\s+playlistId\s*=\s*"([^"]+)"', webpage, 'playlist_id')
+        video_count = int_or_none(
+            self._search_regex(r'var\s+itemsCount\s*=\s*([0-9]+)\s*\|\|', webpage, 'video_count'))
+        token = self._search_regex(r'var\s+token\s*=\s*"([^"]+)"', webpage, 'token')
+        page_count = (video_count - 36 + 39) // 40 + 1
+        page_entries = self._extract_entries(webpage, host)
+
+        def download_page(page_num):
+            note = 'Downloading page {0}'.format(page_num)
+            page_url = 'https://www.{0}/playlist/viewChunked'.format(host)
+            return self._download_webpage(page_url, item_id, note, query={
+                'id': playlist_id,
+                'page': page_num,
+                'token': token,
+            })
+
+        for page_num in range(1, page_count + 1):
+            if page_num > 1:
+                webpage = download_page(page_num)
+                page_entries = self._extract_entries(webpage, host)
+            if not page_entries:
+                break
+            for from_ in page_entries:
+                yield from_
+
+    def _real_extract(self, url):
+        mobj = self._match_valid_url(url)
+        host, item_id = mobj.group('host', 'id')
+
+        self._login(host)
+        self._set_age_cookies(host)
+
+        return self.playlist_result(self._entries(mobj.group('url'), host, item_id), item_id)

youtube_dl/extractor/youtube.py

@@ -692,9 +692,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         'invidious': '|'.join(_INVIDIOUS_SITES),
     }
     _PLAYER_INFO_RE = (
-        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})//(?:tv-)?player',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player',
-        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
+        r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$',
-        r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
+        r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$',
     )
     _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt')
 
@@ -1626,15 +1626,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         """ Return a string representation of a signature """
         return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 
-    @classmethod
+    def _extract_player_info(self, player_url):
-    def _extract_player_info(cls, player_url):
+        try:
-        for player_re in cls._PLAYER_INFO_RE:
+            return self._search_regex(
-            id_m = re.search(player_re, player_url)
+                self._PLAYER_INFO_RE, player_url, 'player info', group='id')
-            if id_m:
+        except ExtractorError as e:
-                break
+            raise ExtractorError(
-        else:
+                'Cannot identify player %r' % (player_url,), cause=e)
-            raise ExtractorError('Cannot identify player %r' % player_url)
-        return id_m.group('id')
 
     def _load_player(self, video_id, player_url, fatal=True, player_id=None):
         if not player_id:
@@ -1711,6 +1709,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 '    return %s\n') % (signature_id_tuple, expr_code)
         self.to_screen('Extracted signature function:\n' + code)
 
+    def _extract_sig_fn(self, jsi, funcname):
+        var_ay = self._search_regex(
+            r'''(?x)
+                (?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*)
+                    (var\s*[\w$]+\s*=\s*(?:
+                        ('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(\s*('|")\W+\3\s*\)|
+                        \[\s*(?:('|")(?:\\\4|(?!\4).)*\4\s*(?:(?=\])|,\s*))+\]
+                    ))(?=\s*[,;])
+            ''', jsi.code, 'useful values', default='')
+
+        sig_fn = jsi.extract_function_code(funcname)
+
+        if var_ay:
+            sig_fn = (sig_fn[0], ';\n'.join((var_ay, sig_fn[1])))
+
+        return sig_fn
+
     def _parse_sig_js(self, jscode):
         # Examples where `sig` is funcname:
         # sig=function(a){a=a.split(""); ... ;return a.join("")};
@@ -1736,8 +1751,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             jscode, 'Initial JS player signature function name', group='sig')
 
         jsi = JSInterpreter(jscode)
-        initial_function = jsi.extract_function(funcname)
+
-        return lambda s: initial_function([s])
+        initial_function = self._extract_sig_fn(jsi, funcname)
+
+        func = jsi.extract_function_from_code(*initial_function)
+
+        return lambda s: func([s])
 
     def _cached(self, func, *cache_id):
         def inner(*args, **kwargs):
@@ -1856,15 +1875,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
     def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None):
 
-        var_ay = self._search_regex(
-            r'(?:[;\s]|^)\s*(var\s*[\w$]+\s*=\s*"(?:\\"|[^"])+"\s*\.\s*split\("\W+"\))(?=\s*[,;])',
-            jsi.code, 'useful values', default='')
-
         func_name = self._extract_n_function_name(jsi.code)
 
-        func_code = jsi.extract_function_code(func_name)
+        func_code = self._extract_sig_fn(jsi, func_name)
-        if var_ay:
-            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
 
         if player_id:
             self.cache.store('youtube-nsig', player_id, func_code)
@@ -2136,7 +2149,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                     video_details = merge_dicts(*traverse_obj(
                         (player_response, api_player_response),
                         (Ellipsis, 'videoDetails', T(dict))))
-                    player_response.update(api_player_response or {})
+                    player_response.update(filter_dict(
+                        api_player_response or {}, cndn=lambda k, _: k != 'captions'))
                     player_response['videoDetails'] = video_details
 
         def is_agegated(playability):
@@ -2566,8 +2580,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         }
 
         pctr = traverse_obj(
-            player_response,
+            (player_response, api_player_response),
-            ('captions', 'playerCaptionsTracklistRenderer', T(dict)))
+            (Ellipsis, 'captions', 'playerCaptionsTracklistRenderer', T(dict)))
         if pctr:
             def process_language(container, base_url, lang_code, query):
                 lang_subs = []
@@ -2584,20 +2598,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             def process_subtitles():
                 subtitles = {}
                 for caption_track in traverse_obj(pctr, (
-                        'captionTracks', lambda _, v: v.get('baseUrl'))):
+                        Ellipsis, 'captionTracks', lambda _, v: (
+                            v.get('baseUrl') and v.get('languageCode')))):
                     base_url = self._yt_urljoin(caption_track['baseUrl'])
                     if not base_url:
                         continue
+                    lang_code = caption_track['languageCode']
                     if caption_track.get('kind') != 'asr':
-                        lang_code = caption_track.get('languageCode')
-                        if not lang_code:
-                            continue
                         process_language(
                             subtitles, base_url, lang_code, {})
                         continue
                     automatic_captions = {}
+                    process_language(
+                        automatic_captions, base_url, lang_code, {})
                     for translation_language in traverse_obj(pctr, (
-                            'translationLanguages', lambda _, v: v.get('languageCode'))):
+                            Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))):
                         translation_language_code = translation_language['languageCode']
                         process_language(
                             automatic_captions, base_url, translation_language_code,

youtube_dl/jsinterp.py

@@ -678,7 +678,7 @@ class JSInterpreter(object):
             return len(obj)
         try:
             return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)]
-        except (TypeError, KeyError, IndexError) as e:
+        except (TypeError, KeyError, IndexError, ValueError) as e:
             # allow_undefined is None gives correct behaviour
             if allow_undefined or (
                     allow_undefined is None and not isinstance(e, TypeError)):
@@ -1038,6 +1038,10 @@ class JSInterpreter(object):
                     left_val = self._index(left_val, idx)
             if isinstance(idx, float):
                 idx = int(idx)
+            if isinstance(left_val, list) and len(left_val) <= int_or_none(idx, default=-1):
+                # JS Array is a sparsely assignable list
+                # TODO: handle extreme sparsity without memory bloat, eg using auxiliary dict
+                left_val.extend((idx - len(left_val) + 1) * [JS_Undefined])
             left_val[idx] = self._operator(
                 m.group('op'), self._index(left_val, idx) if m.group('op') else None,
                 m.group('expr'), expr, local_vars, allow_recursion)
@@ -1204,9 +1208,10 @@ class JSInterpreter(object):
                 elif member == 'join':
                     assertion(isinstance(obj, list), 'must be applied on a list')
                     assertion(len(argvals) <= 1, 'takes at most one argument')
-                    return (',' if len(argvals) == 0 else argvals[0]).join(
+                    return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined)
-                        ('' if x in (None, JS_Undefined) else _js_toString(x))
+                            else argvals[0]).join(
-                        for x in obj)
+                                ('' if x in (None, JS_Undefined) else _js_toString(x))
+                                for x in obj)
                 elif member == 'reverse':
                     assertion(not argvals, 'does not take any arguments')
                     obj.reverse()
@@ -1364,19 +1369,21 @@ class JSInterpreter(object):
         code, _ = self._separate_at_paren(func_m.group('code'))  # refine the match
         return self.build_arglist(func_m.group('args')), code
 
-    def extract_function(self, funcname):
+    def extract_function(self, funcname, *global_stack):
         return function_with_repr(
-            self.extract_function_from_code(*self.extract_function_code(funcname)),
+            self.extract_function_from_code(*itertools.chain(
+                self.extract_function_code(funcname), global_stack)),
             'F<%s>' % (funcname,))
 
     def extract_function_from_code(self, argnames, code, *global_stack):
         local_vars = {}
 
+        start = None
         while True:
-            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code)
+            mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code[start:])
             if mobj is None:
                 break
-            start, body_start = mobj.span()
+            start, body_start = ((start or 0) + x for x in mobj.span())
             body, remaining = self._separate_at_paren(code[body_start - 1:])
             name = self._named_object(local_vars, self.extract_function_from_code(
                 [x.strip() for x in mobj.group('args').split(',')],