Age | Commit message (Collapse) | Author |
|
CVS patchset: 3105
CVS date: 2002/10/31 05:23:02
|
|
CVS patchset: 3104
CVS date: 2002/10/31 02:22:58
|
|
CVS patchset: 3103
CVS date: 2002/10/30 21:45:21
|
|
CVS patchset: 3102
CVS date: 2002/10/30 17:51:13
|
|
xine_demux_stop_thread a chance to aquire the mutex, so that it can
change stream->demux_thread_running to 0.
Problem is, the demuxer thread has it's mutex blocked 99.9% of the time and
releases it only for an extremely short time, for the purpose of allowing
another thread to interrupt the demuxer.
On Solaris at least, the unlock()/lock() sequence will typically run in the
thread's allocated time slice and chances are mimimal that the demuxer
is preempted between the unlock()/lock(), which would allow another thread
to grab the mutex.
CVS patchset: 3101
CVS date: 2002/10/30 13:59:11
|
|
CVS patchset: 3100
CVS date: 2002/10/30 13:29:31
|
|
of doing so is to do it *silently* so the user is not aware of the
errors during every single frame decode
CVS patchset: 3099
CVS date: 2002/10/30 00:43:33
|
|
CVS patchset: 3098
CVS date: 2002/10/30 00:29:30
|
|
CVS patchset: 3097
CVS date: 2002/10/29 21:31:02
|
|
assember code)
CVS patchset: 3096
CVS date: 2002/10/29 20:56:32
|
|
xine-lib/intl/, when the system libraries do not have ngettext
CVS patchset: 3095
CVS date: 2002/10/29 17:19:49
|
|
CVS patchset: 3094
CVS date: 2002/10/29 17:16:29
|
|
CVS patchset: 3093
CVS date: 2002/10/29 17:08:54
|
|
CVS patchset: 3092
CVS date: 2002/10/29 16:30:53
|
|
CVS patchset: 3091
CVS date: 2002/10/29 16:30:07
|
|
CVS patchset: 3090
CVS date: 2002/10/29 16:29:16
|
|
- output fifo flushing
- more sophisticated discontinuity handling
- seek improvement by waiting for at least one frame
CVS patchset: 3089
CVS date: 2002/10/29 16:02:43
|
|
CVS patchset: 3088
CVS date: 2002/10/29 15:19:44
|
|
CVS patchset: 3087
CVS date: 2002/10/29 09:47:39
|
|
CVS patchset: 3086
CVS date: 2002/10/29 03:31:38
|
|
CVS patchset: 3085
CVS date: 2002/10/29 02:57:59
|
|
CVS patchset: 3084
CVS date: 2002/10/29 02:07:45
|
|
CVS patchset: 3083
CVS date: 2002/10/29 01:55:44
|
|
CVS patchset: 3082
CVS date: 2002/10/29 01:29:46
|
|
CVS patchset: 3081
CVS date: 2002/10/29 01:28:41
|
|
CVS patchset: 3080
CVS date: 2002/10/29 01:17:46
|
|
CVS patchset: 3079
CVS date: 2002/10/29 00:37:26
|
|
CVS patchset: 3078
CVS date: 2002/10/29 00:36:21
|
|
CVS patchset: 3077
CVS date: 2002/10/29 00:25:06
|
|
CVS patchset: 3076
CVS date: 2002/10/28 22:11:23
|
|
CVS patchset: 3075
CVS date: 2002/10/28 20:15:56
|
|
egrep.
CVS patchset: 3074
CVS date: 2002/10/28 14:03:59
|
|
to get a minimal dynamic linker symbol table
CVS patchset: 3073
CVS date: 2002/10/28 13:30:58
|
|
has to contain relative or absolute pathnames, in a build with separate
source and object directory hierarchies.
Simply install the headers into the destination directory, and let the
install program figure out the target filename from the source file name.
CVS patchset: 3072
CVS date: 2002/10/28 13:27:10
|
|
configure script
- xine-lib was installing static versions of the plugin modules in the default
configuration, which seems like a waste of time and space.
default to '--disable-static'
- Use the new AC_PROG_LIBTOOL_SANITYCHECK macro from _xine.m4, to work around
libtool problems on solaris with GCC + system linker
- build xine-plugins with a minimal dynamic linker symbol table, using a
libtool symbol filter. An xine plugin only needs to export
'xine_plugin_info', and nothing else
- xine-lib needs ngettext, for "plural forms" support. Check that with
AM_GNU_GETTEXT's need-ngettext parameter
- verify that the system's msgfmt command is able to compile message files
with "plural forms". Reject the one from GNU gettext 0.10.35.
CVS patchset: 3071
CVS date: 2002/10/28 13:25:31
|
|
be unnecessary. The second parameter for the AM_WITH_NLS macro
[ M_WITH_NLS(..., need-ngettext) ] already includes the necessary checks
for a working ngettext.
CVS patchset: 3070
CVS date: 2002/10/28 13:16:00
|
|
acceptable for an xine-lib compilation.
xine-lib needs a msgfmt with working "plural forms", and GNU gettext 0.10.35
cannot compile them.
- Add an ugly hack for libtool failures on solaris, when GNU CC is used with
the system's linker /usr/ccs/bin/ld
For some obscure reason, with this combination libtool prefers to use the
linker to build shared libraries, but it does not link them against
libgcc.{a,so}. Many xine-plugins cannot be loaded with dlopen, because
some subroutine from -lgcc is missing.
Hack the generated libtool to use "gcc -shared -mimpure-text" to build
xine plugin modules.
CVS patchset: 3069
CVS date: 2002/10/28 13:12:08
|
|
installed, so that GMSGFMT is defined as ':' (noop, in gettext.m4).
This causes problems in our build system, because now there are no .gmo
compiled message files available for installation, and the 'make install' fails.
Print a warning and continue.
CVS patchset: 3068
CVS date: 2002/10/28 13:03:37
|
|
shouldn't override CFLAGS and suggests we use AM_CFLAGS instead.
CFLAGS is not needed here, apparently - so away with it.
CVS patchset: 3067
CVS date: 2002/10/28 12:59:51
|
|
shouldn't override CFLAGS and suggest we use AM_CFLAGS instead.
CVS patchset: 3066
CVS date: 2002/10/28 12:58:07
|
|
has to contain relative or absolute pathnames, in a build with separate
source and object directory hierarchies.
Simply install the headers into the destination directory, and let the
install program figure out the source file name.
CVS patchset: 3065
CVS date: 2002/10/28 12:54:00
|
|
CVS patchset: 3064
CVS date: 2002/10/28 11:19:51
|
|
CVS patchset: 3063
CVS date: 2002/10/28 07:57:50
|
|
when frame_rate_code is invalid.
CVS patchset: 3062
CVS date: 2002/10/28 07:55:59
|
|
The concept in simple : compute duration using pts.
Here the formula : duration = ( pts - last_pts ) / (nb img between the 2 pts)
I've tested with libmpeg2 and ffmpeg. It works.
CVS patchset: 3061
CVS date: 2002/10/28 07:53:52
|
|
CVS patchset: 3060
CVS date: 2002/10/28 05:26:35
|
|
CVS patchset: 3059
CVS date: 2002/10/28 05:21:20
|
|
- time to cleanup demuxer mess... no more pthread stuff on any demuxer.
of course i haven't tested all demuxers, but they at least compile.
(sorry Mike, we will need to move a few variables around to make most of
your demuxers work again - not big deal i think)
- make api more consistent by providing xine_open/xine_close
and xine_start/xine_stop.
CVS patchset: 3058
CVS date: 2002/10/28 03:24:42
|
|
CVS patchset: 3057
CVS date: 2002/10/28 00:50:01
|
|
CVS patchset: 3056
CVS date: 2002/10/27 23:01:39
|