| Age | Commit message (Collapse) | Author |
|---|
|
ao_loop() called sched_yield() to give other threads a chance to acquire
driver->lock. But on multi-core systems, it takes too long to wake up the
acquiring thread so that ao_loop() takes driver->lock again before the
other threads had a chance to acquire it.
Therefore a cond var is introduced which ao_loop() can wait for. The
cond var will be signalled when one of the other threads has acquired
driver->lock. This prevents starvation.
|
|
|
|
pause/resume freeze with pulseaudio).
|
|
Don't give up immediately if demux_unstick_ao_loop() reports that xine might
be stuck, because it's not necessarily so.
According to my tests, this fixes
http://bugs.kde.org/show_bug.cgi?id=180339#c42 and
http://bugs.debian.org/514114.
This has been tested with Amarok and kde 4.1.x (with phonon) and kaffeine.
|
|
|
|
(If this causes xine-lib to hang on certain broken streams, we have a problem.)
|
|
This requires that many other files include config.h themselves.
Also convert <config.h> to "config.h".
|
|
(Bug 182)
|
|
|
|
|
|
The test stream sometimes causes the ASF demuxer to report "unknown GUID"
several times then act as if the headers have been received, with the audio
output thread waiting forever for data to be passed to it by the decoder, in
this case the ffmpeg audio decoder.
This particular hang occurs if playback is stopped before the demuxer has
decided that headers have been received (later, and we'd have the problem
which the parent cset of this cset fixes); having the demuxer control insert
empty audio buffers where it would otherwise wait forever for the audio
output thread to receive some data.
Test stream: mmsh://213.92.19.8:80/radiodeejay?MSWMExt=.asf
|
|
Add warning flags to the DEBUG_CFLAGS too.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
When passing XINE_STREAM_INFO_SKIPPED_FRAMES or
XINE_STREAM_INFO_DISCARDED_FRAMES to xine_get_stream_info(), the return
value is always 0, even if there are skipped or discarded frames.
|
|
dlopen of a static library causes a Windows error dialog otherwise.
just ignore *.dll.a in plugins folder
|
|
<xine-internals.h> includes <xine-utils/refcounter.h> which includes
<pthread.h>. As all xine sources include <xine-internal.h>, the
PTHREAD_CFLAGS are always needed. This patch removed it from two Makefile.am
and adds them to CPPFLAGS
|
|
|
|
|
|
This should be simpler and faster for the compiler to optimise rather
than a _X_[BLM]E_32 macro and a comparison.
--HG--
extra : transplant_source : %C7%1C%16%17N%3Bh%9B%EB%AA%00%A9%F1%15%C8%CB%8A%99%EE%7D
|
|
--HG--
extra : transplant_source : %99%9B%D5%B3Ro%87%BFV%E9%2BY%AA%83QE/%CD%5D%ED
|
|
--HG--
extra : transplant_source : T6%E7%60%7F%D4%60%C8l4%9AX%97%86hrR%AD%13%C0
|
|
|
|
This overrides the normal content/extension detection, but can be overridden
by the user.
This is an input plugin API extension; ABI is unchanged.
The version is not bumped (we can't bump it due to 1.2).
|
|
text/plain is specifically filtered out because it could cause lots of false
positives, at least with http; web servers tend to default to text/plain for
unknown types.
|
|
|
|
These occur where the MIME type used as the key is a substring of a type in
some plugin's MIME type list but is not an exact match.
|
|
Now that the macro are either imported from the system or defined by
configure. don't define them in every source file.
|
|
All the initialisation functions returning a new object instance that
was allocated through malloc() or calloc() can get the malloc
attribute so that the compiler can optimise their call.
|
|
|
|
uint32_t).
|
|
|
|
|
|
The new structure xine_current_frame_data_t additionally contains cropping
and interlacing information, which both are required for proper conversion
of the image. The existing functions have been adopted to use the code of
the new function. The changeset should be ABI compatible.
|
|
When using a variable to store the result of strlen(), make sure its
type is size_t just like the function returns, rather than using a
smaller type like int.
--HG--
extra : transplant_source : %B0%D5%3B%D7%1Dy%0A%7E%7D%7C%023%08%B2%CE%D3t%0B3%D7
|
|
Rewrite xine_get_file_extensions() and xine_get_mime_types() so that
they share the joining code, and do one less loop over the list of
plugins.
Create a new _x_concatenate_with_string() function to convert an array
of strings into a single string, which is called by both functions.
--HG--
extra : transplant_source : %BC7%3D%C7%A5%AA%A6%BB%11%9A%94%F0%E9%F1V%B7%29%C68h
|
|
Instead of duplicating a buffer after using sn?printf() on it, use
asprintf() directly on the final variable.
|
|
|
|
Use the proper function for common memory operations (memset() for
zeroing, memcpy() for copying, memmove() for moving), instead of
looping through arrays.
By extension, remove loops to reset arrays when they were allocated
with calloc() and thus already zeroed.
|
|
Also don't reset memory after it's allocated if we do that with
calloc.
|
|
- Add a configure test for the attribute, during xine build process.
- Define the attribute as supported when using GCC 2.95 or later
outside xine build process.
- Use the new XINE_PACKED define instead of the attribute directly.
- Check for SUPPORT_ATTRIBUTE_PACKED rather than doing strange
subdefines.
|
|
|
