aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor
diff options
context:
space:
mode:
authorPhilipp Hagemeister <phihag@phihag.de>2014-08-24 06:58:11 +0200
committerPhilipp Hagemeister <phihag@phihag.de>2014-08-24 06:58:11 +0200
commitebab4520ff85cb5ccd5a3a6ee8c4bec213adda43 (patch)
treef834399af60785844e0693503a1725ea3ca64281 /youtube_dl/extractor
parenta71d1414eb531a84dd19cd8ce5b68b3e947fe48f (diff)
[generic] Use default opener for HEAD request (Fixes #3528)
Diffstat (limited to 'youtube_dl/extractor')
-rw-r--r--youtube_dl/extractor/generic.py64
1 files changed, 6 insertions, 58 deletions
diff --git a/youtube_dl/extractor/generic.py b/youtube_dl/extractor/generic.py
index 5a6fe65bc..8b11f7f7a 100644
--- a/youtube_dl/extractor/generic.py
+++ b/youtube_dl/extractor/generic.py
@@ -356,58 +356,6 @@ class GenericIE(InfoExtractor):
"""Report information extraction."""
self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
- def _send_head(self, url):
- """Check if it is a redirect, like url shorteners, in case return the new url."""
-
- class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
- """
- Subclass the HTTPRedirectHandler to make it use our
- HEADRequest also on the redirected URL
- """
- def redirect_request(self, req, fp, code, msg, headers, newurl):
- if code in (301, 302, 303, 307):
- newurl = newurl.replace(' ', '%20')
- newheaders = dict((k,v) for k,v in req.headers.items()
- if k.lower() not in ("content-length", "content-type"))
- try:
- # This function was deprecated in python 3.3 and removed in 3.4
- origin_req_host = req.get_origin_req_host()
- except AttributeError:
- origin_req_host = req.origin_req_host
- return HEADRequest(newurl,
- headers=newheaders,
- origin_req_host=origin_req_host,
- unverifiable=True)
- else:
- raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
-
- class HTTPMethodFallback(compat_urllib_request.BaseHandler):
- """
- Fallback to GET if HEAD is not allowed (405 HTTP error)
- """
- def http_error_405(self, req, fp, code, msg, headers):
- fp.read()
- fp.close()
-
- newheaders = dict((k,v) for k,v in req.headers.items()
- if k.lower() not in ("content-length", "content-type"))
- return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
- headers=newheaders,
- origin_req_host=req.get_origin_req_host(),
- unverifiable=True))
-
- # Build our opener
- opener = compat_urllib_request.OpenerDirector()
- for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
- HTTPMethodFallback, HEADRedirectHandler,
- compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
- opener.add_handler(handler())
-
- response = opener.open(HEADRequest(url))
- if response is None:
- raise ExtractorError('Invalid URL protocol')
- return response
-
def _extract_rss(self, url, video_id, doc):
playlist_title = doc.find('./channel/title').text
playlist_desc_el = doc.find('./channel/description')
@@ -511,9 +459,13 @@ class GenericIE(InfoExtractor):
self.to_screen('%s: Requesting header' % video_id)
- try:
- response = self._send_head(url)
+ head_req = HEADRequest(url)
+ response = self._request_webpage(
+ head_req, video_id,
+ note=False, errnote='Could not send HEAD request to %s' % url,
+ fatal=False)
+ if response is not False:
# Check for redirect
new_url = response.geturl()
if url != new_url:
@@ -541,10 +493,6 @@ class GenericIE(InfoExtractor):
'upload_date': upload_date,
}
- except compat_urllib_error.HTTPError:
- # This may be a stupid server that doesn't like HEAD, our UA, or so
- pass
-
try:
webpage = self._download_webpage(url, video_id)
except ValueError: