| Age | Commit message (Collapse) | Author | 
|---|
|  | CVS patchset: 5978
CVS date: 2004/01/03 14:50:13 | 
|  | Minor MSVC changes and "Xine" typos.
CVS patchset: 5977
CVS date: 2004/01/02 22:42:03 | 
|  | CVS patchset: 5976
CVS date: 2004/01/02 20:53:43 | 
|  | CVS patchset: 5975
CVS date: 2004/01/02 20:47:03 | 
|  | CVS patchset: 5974
CVS date: 2004/01/02 19:31:56 | 
|  | CVS patchset: 5973
CVS date: 2004/01/02 19:03:46 | 
|  | CVS patchset: 5972
CVS date: 2004/01/02 02:59:03 | 
|  | * use DirectX under CygWin and basic auto-stuff
 * move some my macros in configure.ac
 * I don't know why, but shared directx plugins can't be linked with
   libdxguid.a. Used cute-hack, idea is from mplayer. Also thanks to
   hexedit, because I haven't sources of w32api.
 * Fix some warnings in vo directx and ao directx.
 * Decrease version of directx. It seems it still works.
For compilation with DirectX under Cygwin you need header files from
directx-sdk. Autoconf macros are only basic, but it works. for example:
  DIRECTX_CFLAGS=-I/cygrive/c/Program\ Files/Directx/include ./autogen.sh
CVS patchset: 5971
CVS date: 2004/01/01 18:14:51 | 
|  | needed also under mingw.
CVS patchset: 5970
CVS date: 2004/01/01 18:06:59 | 
|  | CVS patchset: 5969
CVS date: 2004/01/01 15:55:24 | 
|  | CVS patchset: 5968
CVS date: 2004/01/01 14:20:52 | 
|  | CVS patchset: 5967
CVS date: 2003/12/31 23:29:06 | 
|  | CVS patchset: 5966
CVS date: 2003/12/31 18:31:39 | 
|  | CVS patchset: 5965
CVS date: 2003/12/31 14:45:54 | 
|  | endian audio samples
CVS patchset: 5964
CVS date: 2003/12/31 12:05:26 | 
|  | CVS patchset: 5963
CVS date: 2003/12/30 17:21:02 | 
|  | CVS patchset: 5962
CVS date: 2003/12/30 02:16:43 | 
|  | CVS patchset: 5961
CVS date: 2003/12/30 02:14:07 | 
|  | CVS patchset: 5960
CVS date: 2003/12/30 02:02:33 | 
|  | - some fixes to xine_decoder.c
CVS patchset: 5959
CVS date: 2003/12/30 02:00:10 | 
|  | CVS patchset: 5958
CVS date: 2003/12/29 18:13:56 | 
|  | to keep the tradition it looks like redhat is shipping broken
SDL libs again. mine (RH9) only works with hw_accel=0.
CVS patchset: 5957
CVS date: 2003/12/29 18:07:49 | 
|  | CVS patchset: 5956
CVS date: 2003/12/29 16:26:57 | 
|  | CVS patchset: 5955
CVS date: 2003/12/29 16:24:33 | 
|  | CVS patchset: 5954
CVS date: 2003/12/28 15:01:39 | 
|  | CVS patchset: 5953
CVS date: 2003/12/28 11:02:35 | 
|  | CVS patchset: 5952
CVS date: 2003/12/26 18:49:13 | 
|  | - external ffmpeg (with some patches) built by MINGW works, wow!
 - cleanups (only one definition file for all plugins)
 - working release build target
Notes to hybrid MINGW/MSVC build:
 - CVS version of ffmpeg is compiled by MINGW (small patch for compilation
   ffmpeg under MINGW, second for disabling flush because it crashes under
   windows)
 - libxine + xineui compiled by MSVC, additional patch into header files
   of ffmpeg is needed - export all data with '__declspec(dllimport)'
CVS patchset: 5951
CVS date: 2003/12/26 18:37:58 | 
|  | * add TODO items to provide a better solution
CVS patchset: 5950
CVS date: 2003/12/26 16:13:21 | 
|  | CVS patchset: 5949
CVS date: 2003/12/26 14:44:43 | 
|  | CVS patchset: 5948
CVS date: 2003/12/25 21:57:28 | 
|  | * fix compiler warnings
CVS patchset: 5947
CVS date: 2003/12/24 17:34:49 | 
|  | libw32 compiles here without a single compiler warning
CVS patchset: 5946
CVS date: 2003/12/24 16:55:35 | 
|  | CVS patchset: 5945
CVS date: 2003/12/24 16:40:01 | 
|  | are testing for nothing; additionally, access() checks against real UID,
but we actually want to check against effective UID
=> using open() is more straightforward.
CVS patchset: 5944
CVS date: 2003/12/24 15:48:24 | 
|  | CVS patchset: 5943
CVS date: 2003/12/24 13:36:13 | 
|  | CVS patchset: 5942
CVS date: 2003/12/24 13:34:17 | 
|  | CVS patchset: 5941
CVS date: 2003/12/24 13:27:47 | 
|  | CVS patchset: 5940
CVS date: 2003/12/24 13:27:09 | 
|  | CVS patchset: 5939
CVS date: 2003/12/24 13:07:23 | 
|  | pthread locks are _not_ fair (shocking news for me), so audio out has a
problem: the audio out loop holds the driver_lock almost always (since the
write() to the driver takes most of the time); other threads who want that
lock (like the frontend trying to change the volume) can starve;
the problem shows on kernel 2.6 with NPTL, but the potential is there in
all pthread implementations
CVS patchset: 5938
CVS date: 2003/12/24 13:06:10 | 
|  | CVS patchset: 5937
CVS date: 2003/12/24 12:58:58 | 
|  | - new helper function for demuxers (send data to fifo easily,
  provided it was already loaded to memory)
CVS patchset: 5936
CVS date: 2003/12/23 21:22:38 | 
|  | CVS patchset: 5935
CVS date: 2003/12/23 16:58:25 | 
|  | CVS patchset: 5934
CVS date: 2003/12/21 22:43:43 | 
|  | aliasing assumptions can be made (malloc()-returned pointers never alias
anything), which can leed to further optimizations
CVS patchset: 5933
CVS date: 2003/12/21 14:14:39 | 
|  | some additional sanity checks are made
CVS patchset: 5932
CVS date: 2003/12/21 14:11:09 | 
|  | CVS patchset: 5931
CVS date: 2003/12/21 14:09:46 | 
|  | CVS patchset: 5930
CVS date: 2003/12/21 00:22:10 | 
|  | CVS patchset: 5929
CVS date: 2003/12/20 23:42:46 |