aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--youtube_dl/extractor/gdcvault.py42
1 files changed, 37 insertions, 5 deletions
diff --git a/youtube_dl/extractor/gdcvault.py b/youtube_dl/extractor/gdcvault.py
index 89d5994ee..de14ae1fb 100644
--- a/youtube_dl/extractor/gdcvault.py
+++ b/youtube_dl/extractor/gdcvault.py
@@ -8,6 +8,7 @@ from ..utils import (
compat_urllib_request,
)
+
class GDCVaultIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?gdcvault\.com/play/(?P<id>\d+)/(?P<name>(\w|-)+)'
_TESTS = [
@@ -31,6 +32,15 @@ class GDCVaultIE(InfoExtractor):
'skip_download': True, # Requires rtmpdump
}
},
+ {
+ 'url': 'http://www.gdcvault.com/play/1015301/Thexder-Meets-Windows-95-or',
+ 'md5': 'a5eb77996ef82118afbbe8e48731b98e',
+ 'info_dict': {
+ 'id': '1015301',
+ 'ext': 'flv',
+ 'title': 'Thexder Meets Windows 95, or Writing Great Games in the Windows 95 Environment',
+ }
+ }
]
def _parse_mp4(self, xml_description):
@@ -103,18 +113,40 @@ class GDCVaultIE(InfoExtractor):
webpage_url = 'http://www.gdcvault.com/play/' + video_id
start_page = self._download_webpage(webpage_url, video_id)
- xml_root = self._html_search_regex(r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root', None, False)
+ direct_url = self._search_regex(
+ r's1\.addVariable\("file",\s*encodeURIComponent\("(/[^"]+)"\)\);',
+ start_page, 'url', default=None)
+ if direct_url:
+ video_url = 'http://www.gdcvault.com/' + direct_url
+ title = self._html_search_regex(
+ r'<td><strong>Session Name</strong></td>\s*<td>(.*?)</td>',
+ start_page, 'title')
+
+ return {
+ 'id': video_id,
+ 'url': video_url,
+ 'ext': 'flv',
+ 'title': title,
+ }
+ xml_root = self._html_search_regex(
+ r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>',
+ start_page, 'xml root', default=None)
if xml_root is None:
# Probably need to authenticate
- start_page = self._login(webpage_url, video_id)
- if start_page is None:
+ login_res = self._login(webpage_url, video_id)
+ if login_res is None:
self.report_warning('Could not login.')
else:
+ start_page = login_res
# Grab the url from the authenticated page
- xml_root = self._html_search_regex(r'<iframe src="(?P<xml_root>.*?)player.html.*?".*?</iframe>', start_page, 'xml root')
+ xml_root = self._html_search_regex(
+ r'<iframe src="(.*?)player.html.*?".*?</iframe>',
+ start_page, 'xml root')
- xml_name = self._html_search_regex(r'<iframe src=".*?\?xml=(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename', None, False)
+ xml_name = self._html_search_regex(
+ r'<iframe src=".*?\?xml=(.+?\.xml).*?".*?</iframe>',
+ start_page, 'xml filename', default=None)
if xml_name is None:
# Fallback to the older format
xml_name = self._html_search_regex(r'<iframe src=".*?\?xmlURL=xml/(?P<xml_file>.+?\.xml).*?".*?</iframe>', start_page, 'xml filename')