GitLab

It wasn't detecting when there aren't more videos

Also detect videos that come from Youtube

Now it downloads the video over http in one file, it doesn't downloads in fragments
Added a test and use the methods in InfoExtractor for downloading webpages

The new extractor and YoutubeSubscriptionsIE are subclasses of YoutubeFeedsInfoExtractor, which allows to fetch videos from http://www.youtube.com/feed_ajax

Also extract a Sina url that doesn't require to follow a redirection.

It just embed video from other sites.
Modified the _VALID_URL of Youku to catch embed urls.

The page title is now 'Instagram', so we build it.
Also extract the description

It's already covered by the test for length 87

Order the cases from higher to lower length.

[youtube] Add generator for signature 92

It supports some of the websites of the Condé Nast group: WIRED, GQ, Vogue, Glamour, W Magazine and Vanity Fair.

This reverts commit fe6fad12 and closes #1066.