Merge branch 'yt-dlp:master' into master

Author: observeroftime01

.github/workflows/build.yml

@@ -240,7 +240,7 @@ jobs:
     permissions:
       contents: read
       actions: write  # For cleaning up cache
-    runs-on: macos-12
+    runs-on: macos-13
 
     steps:
       - uses: actions/checkout@v4
@@ -346,7 +346,7 @@ jobs:
   macos_legacy:
     needs: process
     if: inputs.macos_legacy
-    runs-on: macos-12
+    runs-on: macos-13
 
     steps:
       - uses: actions/checkout@v4

yt_dlp/extractor/adobepass.py

@@ -1355,6 +1355,7 @@
 class AdobePassIE(InfoExtractor):  # XXX: Conventionally, base classes should end with BaseIE/InfoExtractor
     _SERVICE_PROVIDER_TEMPLATE = 'https://sp.auth.adobe.com/adobe-services/%s'
     _USER_AGENT = 'Mozilla/5.0 (X11; Linux i686; rv:47.0) Gecko/20100101 Firefox/47.0'
+    _MODERN_USER_AGENT = 'Mozilla/5.0 (Windows NT 10.0; rv:131.0) Gecko/20100101 Firefox/131.0'
     _MVPD_CACHE = 'ap-mvpd'
 
     _DOWNLOADING_LOGIN_PAGE = 'Downloading Provider Login Page'
@@ -1454,7 +1455,11 @@ def extract_redirect_url(html, url=None, fatal=False):
                             'no_iframe': 'false',
                             'domain_name': 'adobe.com',
                             'redirect_url': url,
-                        })
+                        }, headers={
+                            # yt-dlp's default user-agent is usually too old for Comcast_SSO
+                            # See: https://github.com/yt-dlp/yt-dlp/issues/10848
+                            'User-Agent': self._MODERN_USER_AGENT,
+                        } if mso_id == 'Comcast_SSO' else None)
                 elif not self._cookies_passed:
                     raise_mvpd_required()
 

yt_dlp/extractor/substack.py

@@ -2,7 +2,13 @@
 import urllib.parse
 
 from .common import InfoExtractor
-from ..utils import js_to_json, str_or_none, traverse_obj
+from ..networking import HEADRequest
+from ..utils import (
+    determine_ext,
+    js_to_json,
+    str_or_none,
+)
+from ..utils.traversal import traverse_obj
 
 
 class SubstackIE(InfoExtractor):
@@ -43,6 +49,19 @@ class SubstackIE(InfoExtractor):
             'uploader': "Andrew Zimmern's Spilled Milk ",
             'uploader_id': '577659',
         },
+    }, {
+        # Podcast that needs its file extension resolved to mp3
+        'url': 'https://persuasion1.substack.com/p/summers',
+        'md5': '1456a755d46084744facdfac9edf900f',
+        'info_dict': {
+            'id': '141970405',
+            'ext': 'mp3',
+            'title': 'Larry Summers on What Went Wrong on Campus',
+            'description': 'Yascha Mounk and Larry Summers also discuss the promise and perils of artificial intelligence.',
+            'thumbnail': r're:https://substackcdn\.com/image/.+\.jpeg',
+            'uploader': 'Persuasion',
+            'uploader_id': '61579',
+        },
     }]
 
     @classmethod
@@ -89,7 +108,15 @@ def _real_extract(self, url):
         post_type = webpage_info['post']['type']
         formats, subtitles = [], {}
         if post_type == 'podcast':
-            formats, subtitles = [{'url': webpage_info['post']['podcast_url']}], {}
+            fmt = {'url': webpage_info['post']['podcast_url']}
+            if not determine_ext(fmt['url'], default_ext=None):
+                # The redirected format URL expires but the original URL doesn't,
+                # so we only want to extract the extension from this request
+                fmt['ext'] = determine_ext(self._request_webpage(
+                    HEADRequest(fmt['url']), display_id,
+                    'Resolving podcast file extension',
+                    'Podcast URL is invalid').url)
+            formats.append(fmt)
         elif post_type == 'video':
             formats, subtitles = self._extract_video_formats(webpage_info['post']['videoUpload']['id'], canonical_url)
         else:

yt_dlp/extractor/twitter.py

@@ -934,14 +934,13 @@ class TwitterIE(TwitterBaseIE):
             'uploader_id': 'MoniqueCamarra',
             'live_status': 'was_live',
             'release_timestamp': 1658417414,
-            'description': 'md5:acce559345fd49f129c20dbcda3f1201',
+            'description': r're:Twitter Space participated by Sergej Sumlenny.+',
             'timestamp': 1658407771,
             'release_date': '20220721',
             'upload_date': '20220721',
         },
         'add_ie': ['TwitterSpaces'],
         'params': {'skip_download': 'm3u8'},
-        'skip': 'Requires authentication',
     }, {
         # URL specifies video number but --yes-playlist
         'url': 'https://twitter.com/CTVJLaidlaw/status/1600649710662213632/video/1',
@@ -1856,8 +1855,6 @@ def _build_graphql_query(self, space_id):
 
     def _real_extract(self, url):
         space_id = self._match_id(url)
-        if not self.is_logged_in:
-            self.raise_login_required('Twitter Spaces require authentication')
         space_data = self._call_graphql_api('HPEisOmj1epUNLCWTYhUWw/AudioSpaceById', space_id)['audioSpace']
         if not space_data:
             raise ExtractorError('Twitter Space not found', expected=True)

yt_dlp/extractor/youtube.py

@@ -4701,11 +4701,12 @@ def process_language(container, base_url, lang_code, sub_name, query):
                 headers=self.generate_api_headers(ytcfg=master_ytcfg),
                 note='Downloading initial data API JSON')
 
+        COMMENTS_SECTION_IDS = ('comment-item-section', 'engagement-panel-comments-section')
         info['comment_count'] = traverse_obj(initial_data, (
             'contents', 'twoColumnWatchNextResults', 'results', 'results', 'contents', ..., 'itemSectionRenderer',
             'contents', ..., 'commentsEntryPointHeaderRenderer', 'commentCount',
         ), (
-            'engagementPanels', lambda _, v: v['engagementPanelSectionListRenderer']['panelIdentifier'] == 'comment-item-section',
+            'engagementPanels', lambda _, v: v['engagementPanelSectionListRenderer']['panelIdentifier'] in COMMENTS_SECTION_IDS,
             'engagementPanelSectionListRenderer', 'header', 'engagementPanelTitleHeaderRenderer', 'contextualInfo',
         ), expected_type=self._get_count, get_all=False)
 

yt_dlp/utils/_utils.py

@@ -824,14 +824,18 @@ class Popen(subprocess.Popen):
         _startupinfo = None
 
     @staticmethod
-    def _fix_pyinstaller_ld_path(env):
-        """Restore LD_LIBRARY_PATH when using PyInstaller
-            Ref: https://github.com/pyinstaller/pyinstaller/blob/develop/doc/runtime-information.rst#ld_library_path--libpath-considerations
-                 https://github.com/yt-dlp/yt-dlp/issues/4573
-        """
+    def _fix_pyinstaller_issues(env):
         if not hasattr(sys, '_MEIPASS'):
             return
 
+        # Force spawning independent subprocesses for exes bundled with PyInstaller>=6.10
+        # Ref: https://pyinstaller.org/en/v6.10.0/CHANGES.html#incompatible-changes
+        #      https://github.com/yt-dlp/yt-dlp/issues/11259
+        env['PYINSTALLER_RESET_ENVIRONMENT'] = '1'
+
+        # Restore LD_LIBRARY_PATH when using PyInstaller
+        # Ref: https://pyinstaller.org/en/v6.10.0/runtime-information.html#ld-library-path-libpath-considerations
+        #      https://github.com/yt-dlp/yt-dlp/issues/4573
         def _fix(key):
             orig = env.get(f'{key}_ORIG')
             if orig is None:
@@ -845,7 +849,7 @@ def _fix(key):
     def __init__(self, args, *remaining, env=None, text=False, shell=False, **kwargs):
         if env is None:
             env = os.environ.copy()
-        self._fix_pyinstaller_ld_path(env)
+        self._fix_pyinstaller_issues(env)
 
         self.__text_mode = kwargs.get('encoding') or kwargs.get('errors') or text or kwargs.get('universal_newlines')
         if text is True: