X-Git-Url: http://git.cielonegro.org/gitweb.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fyoutube.py;h=06005f8d245eb975d67290cead3fb883c27f3f59;hb=b7df8f90a7d749ce609e7ff344464db73329117a;hp=9d542f893fc247a9b8a92a0010da02c66adb4fda;hpb=a61ce71468cb222338ccd8039dc631f3619dc585;p=youtube-dl.git

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 9d542f893..06005f8d2 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -16,6 +16,7 @@ from ..jsinterp import JSInterpreter
 from ..swfinterp import SWFInterpreter
 from ..compat import (
     compat_chr,
+    compat_HTTPError,
     compat_kwargs,
     compat_parse_qs,
     compat_urllib_parse_unquote,
@@ -288,10 +289,25 @@ class YoutubeEntryListBaseInfoExtractor(YoutubeBaseInfoExtractor):
             if not mobj:
                 break
 
-            more = self._download_json(
-                'https://youtube.com/%s' % mobj.group('more'), playlist_id,
-                'Downloading page #%s' % page_num,
-                transform_source=uppercase_escape)
+            count = 0
+            retries = 3
+            while count <= retries:
+                try:
+                    # Downloading page may result in intermittent 5xx HTTP error
+                    # that is usually worked around with a retry
+                    more = self._download_json(
+                        'https://youtube.com/%s' % mobj.group('more'), playlist_id,
+                        'Downloading page #%s%s'
+                        % (page_num, ' (retry #%d)' % count if count else ''),
+                        transform_source=uppercase_escape)
+                    break
+                except ExtractorError as e:
+                    if isinstance(e.cause, compat_HTTPError) and e.cause.code in (500, 503):
+                        count += 1
+                        if count <= retries:
+                            continue
+                    raise
+
             content_html = more['content_html']
             if not content_html.strip():
                 # Some webpages show a "Load more" button but they don't
@@ -1971,7 +1987,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
 
                     signature = self._decrypt_signature(
                         encrypted_sig, video_id, player_url, age_gate)
-                    url += '&signature=' + signature
+                    sp = try_get(url_data, lambda x: x['sp'][0], compat_str) or 'signature'
+                    url += '&%s=%s' % (sp, signature)
                 if 'ratebypass' not in url:
                     url += '&ratebypass=yes'
 
@@ -2035,8 +2052,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 url_or_none(try_get(
                     player_response,
                     lambda x: x['streamingData']['hlsManifestUrl'],
-                    compat_str)) or
-                url_or_none(try_get(
+                    compat_str))
+                or url_or_none(try_get(
                     video_info, lambda x: x['hlsvp'][0], compat_str)))
             if manifest_url:
                 formats = []
@@ -2084,8 +2101,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         else:
             self._downloader.report_warning('unable to extract uploader nickname')
 
-        channel_id = self._html_search_meta(
-            'channelId', video_webpage, 'channel id')
+        channel_id = (
+            str_or_none(video_details.get('channelId'))
+            or self._html_search_meta(
+                'channelId', video_webpage, 'channel id', default=None)
+            or self._search_regex(
+                r'data-channel-external-id=(["\'])(?P<id>(?:(?!\1).)+)\1',
+                video_webpage, 'channel id', default=None, group='id'))
         channel_url = 'http://www.youtube.com/channel/%s' % channel_id if channel_id else None
 
         # thumbnail image
@@ -2542,9 +2564,9 @@ class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor):
 
         search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
         title_span = (
-            search_title('playlist-title') or
-            search_title('title long-title') or
-            search_title('title'))
+            search_title('playlist-title')
+            or search_title('title long-title')
+            or search_title('title'))
         title = clean_html(title_span)
 
         return self.playlist_result(url_results, playlist_id, title)