Uses the data-super-full-img attribute of thumbnails as first attempt to
get a full image URL. If that doesn't work (as is the case with mature
items), the JSON is used. thumbToFull is still broken in this commit,
but shouldn't be needed.
Finally added the support for newlines that I've been going on about in
the other commits. Also got rid of a comment that shouldn't have been
there, as well as fixing the fact that I just broke everything two
commits back and didn't notice until just now. Sorry about that!
Also, added some functions to AbstractHTMLRipper that should allow
description ripping to be added to other rippers as well.
hasDescriptionSupport() is a function that will be overridden by a
ripper that supports descriptions, and will trigger the description
ripper. getDescription will grab the description from a page, and must
be overridden if you want to grab a description with a ripper.
