Age | Commit message (Collapse) | Author |
|
be built with automake but instead imported inline and built using its own build system. This is an import of a slightly modified FFmpeg current tree. xine-lib builds, install and run fine with it, but there are of course plenty of things that needs to be fixed before it can even be considered for a 1.2.x series. Work will continue in the next days of course.
CVS patchset: 8397
CVS date: 2006/12/02 01:19:48
|
|
CVS patchset: 8146
CVS date: 2006/08/02 07:02:37
|
|
CVS patchset: 7872
CVS date: 2006/02/05 14:11:16
|
|
it is likely to break, so please keep both pieces
CVS patchset: 7767
CVS date: 2005/10/23 02:11:16
|
|
CVS patchset: 7463
CVS date: 2005/04/19 05:16:45
|
|
CVS patchset: 6437
CVS date: 2004/04/25 18:57:04
|
|
CVS patchset: 5615
CVS date: 2003/10/27 15:24:38
|
|
know if i overlooked something.
as usual, preliminary QA: tested non debug builds and several codecs including
divx3/4/5, mpeg4, xvid, msmpeg4v3, svq1, wmv7, dv (video/audio), wma
i also enabled wmv8 by default since it worked fine with the streams i have. i'm not
sure about current state of that so we might enable it only for non-x86 users in
case of trouble.
CVS patchset: 4488
CVS date: 2003/03/26 14:43:46
|
|
- fixes wma bugs
- mace, huffyuv and mp3 decoders imported (but not enabled)
tested: wma (v1 and v2), mpeg4, msmpeg4 v1, v2 and v3, divx3, divx4, divx5, xvid and
dv decoders. everything looks fine.
CVS patchset: 3828
CVS date: 2003/01/08 13:18:42
|
|
CVS patchset: 3414
CVS date: 2002/12/02 21:58:56
|
|
something must be broken at ffmpeg, also happens with mplayer)
- added priority sorted lists, now autoprobing should work again.
- fixed infinite loop in plugin loader.
obs: latest ffmpeg contains ppc optimizations, someone will have to enable these
though.
CVS patchset: 2676
CVS date: 2002/09/16 21:49:34
|
|
diff_to_ffmpeg_cvs.txt updated.
tested only on x86, please report any problems, compilation errors, etc.
alpha architecture added but makefiles were not updated.
CVS patchset: 2058
CVS date: 2002/06/10 14:51:26
|
|
xine-lib should really never do an exit or abort, but instead pass back nice error values to the calling application, but until that happens, use abort() as that is tracable with gdb, whereas exit(1) is not backtraceable.
CVS patchset: 1811
CVS date: 2002/04/29 23:31:59
|
|
CVS patchset: 1315
CVS date: 2001/12/27 21:06:08
|