Age | Commit message (Collapse) | Author |
|
(This assumes that if the compiler isn't gcc, __inline__ isn't supported.)
CVS patchset: 7540
CVS date: 2005/05/14 23:17:10
|
|
- Better extension handling:
Distributing own, mangled, glext.h (now myglext.h).
Resolving glXgetProcAddress[ARB] during runtime now
(original patch by Robert Osfield).
Should compile now virtually anywhere.
- Test for existence / working GLU (original patch by Alien).
- Windows often were opened with the wrong visual ID depending on
the single/double buffering flag. Fixed.
CVS patchset: 7478
CVS date: 2005/04/21 17:37:40
|
|
fix xine-lib builds without X11: HAVE_XV must not be defined conditionally,
otherwise configure will complain
also fix the Xv test for old shells with broken handling of empty variables
CVS patchset: 7307
CVS date: 2004/12/27 17:22:30
|
|
CVS patchset: 7250
CVS date: 2004/12/13 21:35:21
|
|
try some fallbacking (libXv.so <-> libXv.a) if prefered was not found
CVS patchset: 7244
CVS date: 2004/12/13 17:50:20
|
|
used, xv_lib and XV_LIB. The problem with this is that XV_LIB is used
to set the Xv library to use in the makefiles. I chose to use XV_LIB
throughout the m4 file.
CVS patchset: 7229
CVS date: 2004/12/12 04:25:05
|
|
CVS patchset: 7228
CVS date: 2004/12/12 03:08:35
|
|
Add --enable-shared-xv to force use of libXv.so over libXv.a. The
default behavior is to use libXv.a if configure option is not used.
CVS patchset: 7227
CVS date: 2004/12/12 03:04:58
|
|
CVS patchset: 7181
CVS date: 2004/12/01 20:26:45
|
|
OpenGL support
CVS patchset: 7158
CVS date: 2004/11/24 16:29:32
|
|
(you never know who might want to override it)
CVS patchset: 7019
CVS date: 2004/10/08 20:27:21
|
|
CVS patchset: 6999
CVS date: 2004/09/24 19:24:27
|
|
Added a warning.
CVS patchset: 6981
CVS date: 2004/09/20 19:02:47
|
|
CVS patchset: 6945
CVS date: 2004/09/07 07:46:59
|
|
- usability MSVC for frontends
- remove unistd.h from public headers
- add configure option "--with-dxheaders"
- install .def file under WIN32
- MSVC port update (some people might want it)
CVS patchset: 6944
CVS date: 2004/09/06 18:34:37
|
|
Fix what I perceive to be a bug only on platforms that have a automake
installed that contains multiple version. Gentoo releases a package of
automake that contains all the latest stable version (1.4 to 1.8). In
order to use a specific one besides the default you need to declare
WANT_AUTOMAKE=1.6 in your environment varibles.
------------------------------------
doc/faq/Makefile.am
doc/hackersguide/Makefile.am
Applied Frantisek Dvorak changes for the documentation.
CVS patchset: 6939
CVS date: 2004/09/05 02:14:38
|
|
List of some changes:
- replaced some _MSC_VER by more common WIN32
- define INTLDIR, remove -static flag for included intl
- shared more common CFLAGS with DEBUG_CFLAGS
- use WIN32_CFLAGS for all building
- separate some flags into THREAD_CFLAGS_CONFIG,
THREAD_CFLAGS_CONFIG and ZLIB_LIB_CONFIG for public xine-config,
automatically use internal libs if necessary
- don't warn about missing X for mingw and cygwin
- libw32dll disabled for WIN32 (making native loader would be
interesting, or porting wine code to Windows? :->)
- DVB and RTP disabled for WIN32, not ported yet
- fix build and fix a warning in cdda
- fix build for nosefart and libfaad
- implement configure option --disable-freetype
- sync libxine.pc and xine-config.in
- add -liberty to goom under WIN32
- move original build files from included phread and zlib into archives
and replace them by autotools
CVS patchset: 6910
CVS date: 2004/08/27 19:33:17
|
|
Separate opengl configure tests into opengl.m4
Fix a segfault in debug build
CVS patchset: 6905
CVS date: 2004/08/26 17:57:50
|
|
* use FFMPEG_CPPFLAGS instead of FFMPEG_CFLAGS
CVS patchset: 6882
CVS date: 2004/08/16 15:16:51
|
|
CVS patchset: 6872
CVS date: 2004/08/02 18:41:30
|
|
CVS patchset: 6865
CVS date: 2004/07/31 18:57:45
|
|
AC_TRY_COMPILE ensure that second call is wrapped in brackets ([ ACE_TRY_COMPILE ...]). Apparently without them autoheader and automake get confused. Replaced the AC_LANG_CPLUSPLUS, AC_LANG_SAVE and AC_LANG_RESTORE with the recommended calls of AC_LANG_PUSH(C++) and AC_LANG_POP([C++]). Let me know if the last two things cause problems. I can revert this change.
CVS patchset: 6842
CVS date: 2004/07/24 02:35:47
|
|
#992824)
CVS patchset: 6840
CVS date: 2004/07/23 12:42:14
|
|
FC2).
CVS patchset: 6800
CVS date: 2004/07/16 19:12:27
|
|
CVS patchset: 6478
CVS date: 2004/05/02 19:49:52
|
|
CVS patchset: 6217
CVS date: 2004/03/05 16:34:32
|
|
CVS patchset: 6191
CVS date: 2004/02/28 19:56:41
|
|
http://cvs.mandrakesoft.com/cgi-bin/cvsweb.cgi/SPECS/xine-lib/
via Goetz Waschk who reports:
The amd64 patch (xine-lib-1-rc0a-amd64.patch) sets some conservative
CFLAGS for amd64,
the lib64 patch (xine-lib-1-rc0a-lib64.patch) replaces hardcoded
/lib to support the lib64 library dir on amd64,
the directfb patch (xine-lib-1-rc2-no-directfb.patch) adds a
configure option to disable directfb,
the linuxfb patch (xine-lib-1-rc3a-no-linuxfb.patch) does the same
for linux framebuffer and
the 64bit fixes patch (xine-lib-1-rc3-64bit-fixes.patch) doesn't
apply at the moment against the CVS -- demux_ogg.c was not applied.
it includes some 64 bit pointer and other fixes for 64bit architectures.
from Gwenole Beauchesne
I haven't tested other than apply and compile.
CVS patchset: 6174
CVS date: 2004/02/19 02:50:24
|
|
because linux/compiler.h defines inline already
CVS patchset: 6132
CVS date: 2004/02/12 17:59:02
|
|
CVS patchset: 5995
CVS date: 2004/01/04 20:30:20
|
|
this problem)
CVS patchset: 5776
CVS date: 2003/11/23 23:29:27
|
|
"inline __attribute__ ((__always_inline__))" should have the same effect,
but my gcc 3.3.1 emits much fewer warnings with the latter
(That is because static inline functions for example in ffmpeg are implemented
in the header, but not used in every file that includes this header. Now
gcc reports unused static functions with a "defined but not used" warning
unless they are static inline. The former define would have dropped the inline
and therefore triggers those warnings.)
CVS patchset: 5749
CVS date: 2003/11/16 15:31:05
|
|
(now my 0.9.0rc7 is accepted again :) )
CVS patchset: 5345
CVS date: 2003/09/05 13:51:27
|
|
now works for libtool-1.4.x and libtool-1.5.
Changed the commands for building the assembly files in libw32dll to
match that suggested by Chris Rankin <rankincj@yahoo.com>.
CVS patchset: 5203
CVS date: 2003/07/20 20:48:40
|
|
CVS patchset: 5120
CVS date: 2003/07/02 12:43:37
|
|
encoder lib is found is not true any longer -> we have libavcodec's encoder now
CVS patchset: 4991
CVS date: 2003/05/31 10:44:27
|
|
CVS patchset: 4926
CVS date: 2003/05/25 13:39:13
|
|
CVS patchset: 4913
CVS date: 2003/05/23 20:12:42
|
|
extracted some common targets and variables into a common Makefile
which is then included
CVS patchset: 4844
CVS date: 2003/05/14 16:21:44
|
|
CVS patchset: 4717
CVS date: 2003/04/30 08:49:38
|
|
CVS patchset: 4529
CVS date: 2003/04/01 16:52:37
|
|
CVS patchset: 4524
CVS date: 2003/03/31 16:22:23
|
|
CVS patchset: 4518
CVS date: 2003/03/31 09:38:33
|
|
CVS patchset: 4182
CVS date: 2003/02/17 14:58:40
|
|
looking for the directory of where to install the xine m4 script. Checking
the gnu.org web page and mailing list I found out that we are already doing
what we need without do anything special like determining if root is
running the configure script. If someone does not define ACLOCAL_DIR we use the
$(datadir)/aclocal directory. Since this is a comestic change and affects
the general quality of our builds I have only commented it out without
removing it.
Affects: Users configurating xine-lib as 'root' using one version of automake
and autoconf on one machine and building on another with different versions
of automake and autoconf. For example, ran autogen.sh --noconfig on system
with automake 1.7.2 and autoconf 2.57 while build system had automake 1.6.3
and autoconf 2.53.
Thanks to Chris Purnell <cjp@lost.org.uk> for helping me understand the
nature of the problem.
CVS patchset: 3987
CVS date: 2003/01/22 01:35:04
|
|
CVS patchset: 3976
CVS date: 2003/01/20 02:30:01
|
|
CVS patchset: 3919
CVS date: 2003/01/14 21:55:06
|
|
CVS patchset: 3891
CVS date: 2003/01/13 12:38:08
|
|
because it is not entirely free; therefore we include the em8300.h header in our
tree, because this is the only thing we need
Please note that the dxr3 plugins will now be built for everyone unless you
specify --disable-dxr3. But this should be no problem, since they detect dxr3
presence at runtime.
CVS patchset: 3866
CVS date: 2003/01/11 11:18:00
|
|
CVS patchset: 3738
CVS date: 2002/12/31 22:17:14
|