Fix XNXX in Python 3

This commit is contained in:
Philipp Hagemeister 2012-12-15 18:19:25 +01:00
parent 8f6f40d991
commit bec102a843

View File

@ -3424,13 +3424,14 @@ class XNXXIE(InfoExtractor):
if mobj is None: if mobj is None:
self._downloader.trouble(u'ERROR: invalid URL: %s' % url) self._downloader.trouble(u'ERROR: invalid URL: %s' % url)
return return
video_id = mobj.group(1).decode('utf-8') video_id = mobj.group(1)
self.report_webpage(video_id) self.report_webpage(video_id)
# Get webpage content # Get webpage content
try: try:
webpage = compat_urllib_request.urlopen(url).read() webpage_bytes = compat_urllib_request.urlopen(url).read()
webpage = webpage_bytes.decode('utf-8')
except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
self._downloader.trouble(u'ERROR: unable to download video webpage: %s' % err) self._downloader.trouble(u'ERROR: unable to download video webpage: %s' % err)
return return
@ -3439,19 +3440,19 @@ class XNXXIE(InfoExtractor):
if result is None: if result is None:
self._downloader.trouble(u'ERROR: unable to extract video url') self._downloader.trouble(u'ERROR: unable to extract video url')
return return
video_url = compat_urllib_parse.unquote(result.group(1).decode('utf-8')) video_url = compat_urllib_parse.unquote(result.group(1))
result = re.search(self.VIDEO_TITLE_RE, webpage) result = re.search(self.VIDEO_TITLE_RE, webpage)
if result is None: if result is None:
self._downloader.trouble(u'ERROR: unable to extract video title') self._downloader.trouble(u'ERROR: unable to extract video title')
return return
video_title = result.group(1).decode('utf-8') video_title = result.group(1)
result = re.search(self.VIDEO_THUMB_RE, webpage) result = re.search(self.VIDEO_THUMB_RE, webpage)
if result is None: if result is None:
self._downloader.trouble(u'ERROR: unable to extract video thumbnail') self._downloader.trouble(u'ERROR: unable to extract video thumbnail')
return return
video_thumbnail = result.group(1).decode('utf-8') video_thumbnail = result.group(1)
return [{ return [{
'id': video_id, 'id': video_id,