This is wrt the bug at:
https://bugzilla.redhat.com/show_bug.cgi?id=639672
Please see the thread at:
http://lists.fedoraproject.org/pipermail/devel/2010-October/143688.html
This could also be helpful:
http://lists.fedoraproject.org/pipermail/devel/2010-October/143701.html
Adding support for just decoding arithmetic coded files should not stand in anyone's way due to philosophical reasons. Ideally I'd like support for encoding too, but we can argue on that later.
This is a patch against libjpeg-turbo HEAD to add support for decoding arithmetic coded content. It is based on a patch by Guido Vollbeding guivol@esc.de, which included code to both encode and decode arithmetic coded content. This patch only adds the decoding portion. No new arithmetic coded content can be produced, but existing arithmetic coded content can be decoded.
libjpeg-turbo with this patch has been tested with some arithmetic coded images (digital photographs).
(git-svn) Patch against libjpeg-turbo HEAD to add support for decoding arithmetic coded content
I have no philosophical objections to it. My concerns are strictly legal, and I am following the Fedora Project's lead on the matter. As soon as they decide arithmetic coding is legally safe to include in Fedora, it will be included in libjpeg-turbo (which Fedora is now using as their primary JPEG codec.)
Both encoding and decoding have been fully implemented in the libjpeg-turbo SVN trunk and will be in LJT 1.1. These features can be disabled individually using configure or CMake options.