aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl
diff options
context:
space:
mode:
Diffstat (limited to 'youtube_dl')
-rwxr-xr-xyoutube_dl/InfoExtractors.py13
1 files changed, 4 insertions, 9 deletions
diff --git a/youtube_dl/InfoExtractors.py b/youtube_dl/InfoExtractors.py
index 7545ae0b9..0860937ee 100755
--- a/youtube_dl/InfoExtractors.py
+++ b/youtube_dl/InfoExtractors.py
@@ -2869,8 +2869,7 @@ class StanfordOpenClassroomIE(InfoExtractor):
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
if mobj is None:
- self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
- return
+ raise ExtractorError(u'Invalid URL: %s' % url)
if mobj.group('course') and mobj.group('video'): # A specific video
course = mobj.group('course')
@@ -2907,12 +2906,9 @@ class StanfordOpenClassroomIE(InfoExtractor):
'upload_date': None,
}
- self.report_download_webpage(info['id'])
- try:
- coursepage = compat_urllib_request.urlopen(url).read()
- except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
- self._downloader.trouble(u'ERROR: unable to download course info page: ' + compat_str(err))
- return
+ coursepage = self._download_webpage(url, info['id'],
+ note='Downloading course info page',
+ errnote='Unable to download course info page')
m = re.search('<h1>([^<]+)</h1>', coursepage)
if m:
@@ -2936,7 +2932,6 @@ class StanfordOpenClassroomIE(InfoExtractor):
assert entry['type'] == 'reference'
results += self.extract(entry['url'])
return results
-
else: # Root page
info = {
'id': 'Stanford OpenClassroom',