Merge 45c034e262d2e7590cf3183a5e1c3bb2f565ef55 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
-        test_id = m.group('id')
+        assert m, '{0!r} should follow URL format'.format(url)
+        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/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'/(?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$',
-        r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$',
+        r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player',
+        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$',
+        r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$',
     )
     _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(cls, player_url):
-        for player_re in cls._PLAYER_INFO_RE:
-            id_m = re.search(player_re, player_url)
-            if id_m:
-                break
-        else:
-            raise ExtractorError('Cannot identify player %r' % player_url)
-        return id_m.group('id')
+    def _extract_player_info(self, player_url):
+        try:
+            return self._search_regex(
+                self._PLAYER_INFO_RE, player_url, 'player info', group='id')
+        except ExtractorError as e:
+            raise ExtractorError(
+                'Cannot identify player %r' % (player_url,), cause=e)
 
     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)
-        if var_ay:
-            func_code = (func_code[0], ';\n'.join((var_ay, func_code[1])))
+        func_code = self._extract_sig_fn(jsi, func_name)
 
         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,
-            ('captions', 'playerCaptionsTracklistRenderer', T(dict)))
+            (player_response, api_player_response),
+            (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/extractor/zoom.py

@@ -1,68 +1,215 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
-
 from .common import InfoExtractor
 from ..utils import (
     ExtractorError,
     int_or_none,
     js_to_json,
+    merge_dicts,
     parse_filesize,
+    parse_resolution,
+    strip_or_none,
+    T,
+    traverse_obj,
+    txt_or_none,
+    url_basename,
+    url_or_none,
     urlencode_postdata,
+    urljoin,
 )
 
+from functools import partial
+k_int_or_none = partial(int_or_none, scale=1000)
+
 
 class ZoomIE(InfoExtractor):
     IE_NAME = 'zoom'
-    _VALID_URL = r'(?P<base_url>https?://(?:[^.]+\.)?zoom.us/)rec(?:ording)?/(?:play|share)/(?P<id>[A-Za-z0-9_.-]+)'
-    _TEST = {
+    _VALID_URL = r'(?P<base_url>https?://(?:[^.]+\.)?zoom\.us/)rec(?:ording)?/(?P<type>play|share)/(?P<id>[\w.-]+)'
+    _TESTS = [{
         'url': 'https://economist.zoom.us/rec/play/dUk_CNBETmZ5VA2BwEl-jjakPpJ3M1pcfVYAPRsoIbEByGsLjUZtaa4yCATQuOL3der8BlTwxQePl_j0.EImBkXzTIaPvdZO5',
         'md5': 'ab445e8c911fddc4f9adc842c2c5d434',
         'info_dict': {
             'id': 'dUk_CNBETmZ5VA2BwEl-jjakPpJ3M1pcfVYAPRsoIbEByGsLjUZtaa4yCATQuOL3der8BlTwxQePl_j0.EImBkXzTIaPvdZO5',
             'ext': 'mp4',
             'title': 'China\'s "two sessions" and the new five-year plan',
+        },
+        'skip': 'Recording requires email authentication to access',
+    }, {
+        'url': 'https://us06web.zoom.us/rec/play/W1ctyErikzJ2CxtwlsTW3xNbiMHze6ZkU1adqeshzivi58DHEJ-7HX2Z8-nqK80a8d4CWHAhrSpsl9mG.OaL6JvfC1gAa1EvZ?canPlayFromShare=true&from=share_recording_detail&continueMode=true&componentName=rec-play&originRequestUrl=https%3A%2F%2Fus06web.zoom.us%2Frec%2Fshare%2F60THDorqjAyUm_IXKS88Z4KgfYRAER3wIG20jgrLqaSFBWJW14qBVBRkfHylpFrk.KXJxuNLN0sRBXyvf',
+        'md5': 'b180e7773a878e4799f194f2280648d5',
+        'info_dict': {
+            'id': 'W1ctyErikzJ2CxtwlsTW3xNbiMHze6ZkU1adqeshzivi58DHEJ-7HX2Z8-nqK80a8d4CWHAhrSpsl9mG.OaL6JvfC1gAa1EvZ',
+            'ext': 'mp4',
+            'title': 'Chipathon Bi-Weekly Meeting',
+            'description': 'Shared screen with speaker view',
+            'timestamp': 1715263581,
+            'upload_date': '20240509',
         }
-    }
+    }, {
+        # play URL
+        'url': 'https://ffgolf.zoom.us/rec/play/qhEhXbrxq1Zoucx8CMtHzq1Z_2YZRPVCqWK_K-2FkEGRsSLDeOX8Tu4P6jtjZcRry8QhIbvKZdtr4UNo.QcPn2debFskI9whJ',
+        'md5': '2c4b1c4e5213ebf9db293e88d9385bee',
+        'info_dict': {
+            'id': 'qhEhXbrxq1Zoucx8CMtHzq1Z_2YZRPVCqWK_K-2FkEGRsSLDeOX8Tu4P6jtjZcRry8QhIbvKZdtr4UNo.QcPn2debFskI9whJ',
+            'ext': 'mp4',
+            'title': 'Prépa AF2023 - Séance 5 du 11 avril - R20/VM/GO',
+        },
+        'skip': 'Recording expired',
+    }, {
+        # share URL
+        'url': 'https://us02web.zoom.us/rec/share/hkUk5Zxcga0nkyNGhVCRfzkA2gX_mzgS3LpTxEEWJz9Y_QpIQ4mZFOUx7KZRZDQA.9LGQBdqmDAYgiZ_8',
+        'md5': '90fdc7cfcaee5d52d1c817fc03c43c9b',
+        'info_dict': {
+            'id': 'hkUk5Zxcga0nkyNGhVCRfzkA2gX_mzgS3LpTxEEWJz9Y_QpIQ4mZFOUx7KZRZDQA.9LGQBdqmDAYgiZ_8',
+            'ext': 'mp4',
+            'title': 'Timea Andrea Lelik\'s Personal Meeting Room',
+        },
+        'skip': 'This recording has expired',
+    }, {
+        # view_with_share URL
+        'url': 'https://cityofdetroit.zoom.us/rec/share/VjE-5kW3xmgbEYqR5KzRgZ1OFZvtMtiXk5HyRJo5kK4m5PYE6RF4rF_oiiO_9qaM.UTAg1MI7JSnF3ZjX',
+        'md5': 'bdc7867a5934c151957fb81321b3c024',
+        'info_dict': {
+            'id': 'VjE-5kW3xmgbEYqR5KzRgZ1OFZvtMtiXk5HyRJo5kK4m5PYE6RF4rF_oiiO_9qaM.UTAg1MI7JSnF3ZjX',
+            'ext': 'mp4',
+            'title': 'February 2022 Detroit Revenue Estimating Conference',
+            'description': 'Speaker view',
+            'timestamp': 1645200510,
+            'upload_date': '20220218',
+            'duration': 7299,
+            'formats': 'mincount:3',
+        },
+    }, {
+        # ytdl-org/youtube-dl#32094
+        'url': 'https://us02web.zoom.us/rec/share/9pdVT4f2XWBaEOSqDJagSsvI0Yu2ixXW0YcJGIVhfV19Zr7E1q5gf0wTMZHnqrvq.Yoq1dDHeeKjaVcv3',
+        'md5': 'fdb6f8df7f5ee0c07ced5fae55c0ced4',
+        'info_dict': {
+            'id': '9pdVT4f2XWBaEOSqDJagSsvI0Yu2ixXW0YcJGIVhfV19Zr7E1q5gf0wTMZHnqrvq.Yoq1dDHeeKjaVcv3',
+            'ext': 'mp4',
+            'title': 'Untersuchungskurs Gruppe V',
+            'description': 'Shared screen with speaker view',
+            'timestamp': 1681889972,
+            'upload_date': '20230419',
+        },
+    }]
 
-    def _real_extract(self, url):
-        base_url, play_id = re.match(self._VALID_URL, url).groups()
-        webpage = self._download_webpage(url, play_id)
+    def _get_page_data(self, webpage, video_id):
+        return self._search_json(
+            r'window\.__data__\s*=', webpage, 'data', video_id, transform_source=js_to_json)
 
+    def _get_real_webpage(self, url, base_url, video_id, url_type):
+        webpage = self._download_webpage(url, video_id, note='Downloading {0} webpage'.format(url_type))
         try:
             form = self._form_hidden_inputs('password_form', webpage)
         except ExtractorError:
-            form = None
-        if form:
-            password = self._downloader.params.get('videopassword')
-            if not password:
-                raise ExtractorError(
-                    'This video is protected by a passcode, use the --video-password option', expected=True)
-            is_meeting = form.get('useWhichPasswd') == 'meeting'
-            validation = self._download_json(
-                base_url + 'rec/validate%s_passwd' % ('_meet' if is_meeting else ''),
-                play_id, 'Validating passcode', 'Wrong passcode', data=urlencode_postdata({
-                    'id': form[('meet' if is_meeting else 'file') + 'Id'],
-                    'passwd': password,
-                    'action': form.get('action'),
-                }))
-            if not validation.get('status'):
-                raise ExtractorError(validation['errorMessage'], expected=True)
-            webpage = self._download_webpage(url, play_id)
+            return webpage
 
-        data = self._parse_json(self._search_regex(
-            r'(?s)window\.__data__\s*=\s*({.+?});',
-            webpage, 'data'), play_id, js_to_json)
+        password = self.get_param('videopassword')
+        if not password:
+            raise ExtractorError(
+                'This video is protected by a passcode: use the --video-password option', expected=True)
+        is_meeting = form.get('useWhichPasswd') == 'meeting'
+        validation = self._download_json(
+            base_url + 'rec/validate%s_passwd' % ('_meet' if is_meeting else ''),
+            video_id, 'Validating passcode', 'Wrong passcode', data=urlencode_postdata({
+                'id': form[('meet' if is_meeting else 'file') + 'Id'],
+                'passwd': password,
+                'action': form.get('action'),
+            }))
+        if not validation.get('status'):
+            raise ExtractorError(validation['errorMessage'], expected=True)
+        return self._download_webpage(url, video_id, note='Re-downloading {0} webpage'.format(url_type))
 
-        return {
-            'id': play_id,
-            'title': data['topic'],
-            'url': data['viewMp4Url'],
-            'width': int_or_none(data.get('viewResolvtionsWidth')),
-            'height': int_or_none(data.get('viewResolvtionsHeight')),
+    def _real_extract(self, url):
+        base_url, url_type, video_id = self._match_valid_url(url).group('base_url', 'type', 'id')
+        query = {}
+
+        if url_type == 'share':
+            webpage = self._get_real_webpage(url, base_url, video_id, 'share')
+            meeting_id = self._get_page_data(webpage, video_id)['meetingId']
+            share_info = self._download_json(
+                '{0}nws/recording/1.0/play/share-info/{1}'.format(base_url, meeting_id),
+                video_id, note='Downloading share info JSON', fatal=False)
+            url = traverse_obj(share_info, (
+                'result', 'redirectUrl', T(lambda u: urljoin(base_url, u))))
+            if not url:
+                raise ExtractorError(traverse_obj(
+                    share_info, 'errorMessage') or 'No video found from share link')
+            query['continueMode'] = 'true'
+
+        webpage = self._get_real_webpage(url, base_url, video_id, 'play')
+        file_id = traverse_obj(webpage, T(lambda x: txt_or_none(self._get_page_data(x, video_id)['fileId'])))
+        if not file_id:
+            # When things go wrong, file_id can be empty string
+            raise ExtractorError('Unable to extract file ID')
+
+        data = self._download_json(
+            '{0}nws/recording/1.0/play/info/{1}'.format(base_url, file_id), video_id, query=query,
+            note='Downloading play info JSON')['result']
+
+        formats = []
+        subtitles = dict(
+            (s_type, [{'url': s_url, 'ext': 'vtt', }])
+            for s_type, s_url in traverse_obj(
+                ('transcript', 'cc', 'chapter'),
+                (Ellipsis,
+                 T(lambda t: (t, urljoin(base_url, txt_or_none(data['%sUrl' % (t,)])))),
+                 T(lambda x: x if x[1] else None)))) or None
+
+        def if_url(f):
+            return lambda x: f(x) if x.get('url') else None
+
+        formats.extend(traverse_obj(data, ((
+            ({
+                'url': ('viewMp4Url', T(url_or_none)),
+                'width': ('viewResolvtions', 0, T(int_or_none)),
+                'height': ('viewResolvtions', 1, T(int_or_none)),
+                'format_id': ('recording', 'id', T(txt_or_none)),
+                'filesize_approx': ('recording', 'fileSizeInMB', T(parse_filesize)),
+            }, T(if_url(lambda x: merge_dicts({
+                'format_note': 'Camera stream',
+                'ext': 'mp4',
+                'preference': 0,
+            }, x)))),
+            ({
+                'url': ('shareMp4Url', T(url_or_none)),
+                'width': ('shareResolvtions', 0, T(int_or_none)),
+                'height': ('shareResolvtions', 1, T(int_or_none)),
+                'format_id': ('shareVideo', 'id', T(txt_or_none)),
+                'filesize_approx': ('recording', 'fileSizeInMB', T(parse_filesize)),
+            }, T(if_url(lambda x: merge_dicts({
+                'format_note': 'Screen share stream',
+                'ext': 'mp4',
+                'preference': -1,
+            }, x)))),
+            ({
+                'url': ('viewMp4WithshareUrl', T(url_or_none)),
+            }, T(if_url(lambda x: merge_dicts({
+                'format_note': 'Screen share with camera',
+                'format_id': 'view_with_share',
+                'ext': 'mp4',
+                'preference': 1,
+            }, parse_resolution(self._search_regex(
+                r'_(\d+x\d+)\.mp4', url_basename(x['url']),
+                'resolution', default=None)), x)
+            )))), all)
+        ))
+
+        if not formats and data.get('message'):
+            raise ExtractorError('No media found; %s said "%s"' % (self.IE_NAME, data['message'],), expected=True)
+        self._sort_formats(formats)
+
+        return merge_dicts(traverse_obj(data, {
+            'title': ('meet', 'topic', T(strip_or_none)),
+            'description': ('recording', 'displayFileName', T(strip_or_none)),
+            'duration': ('duration', T(int_or_none)),
+            'timestamp': (('clipStartTime', 'fileStartTime'), T(k_int_or_none), any),
+        }), {
+            'id': video_id,
+            'subtitles': subtitles,
+            'formats': formats,
             'http_headers': {
                 'Referer': base_url,
             },
-            'filesize_approx': parse_filesize(data.get('fileSize')),
-        }
+        })

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(
-                        ('' if x in (None, JS_Undefined) else _js_toString(x))
-                        for x in obj)
+                    return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined)
+                            else argvals[0]).join(
+                                ('' 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(',')],