| Age | Commit message (Collapse) | Author |
|---|
|
This reverts commit 5684e8ef15bc5a4ef2c407a4dafb33684e4afb95 and 33a066668f698c3875da717539fef7e61650945a
|
|
when playing MID files
|
|
|
|
|
|
User-provided data should never trigger an assert; return error instead.
Also fix an instance of get_bits where get_bits_long should have been used.
This eliminates the following warning:
libavcodec/vaapi_mpeg2.c:112:14: warning: variable 'start_code' set but not used
Signed-off-by: elupus <elupus@xbmc.org>
|
|
Fixes ffmpeg ticket #457.
Signed-off-by: elupus <elupus@xbmc.org>
|
|
|
|
if the last archive in a set was less than 256kb, the seek
code set an invalid offset leading to crashes
|
|
Gwenole Beauchesne
|
|
The filled in refs cause corruptions in the video frame
for a long time after it should have recovered.
|
|
|
|
corrupt archives.
|
|
|
|
The nVidia cards sometimes return E_PENDING and need time before they can start
decoding a new frame.
Helps nVidia cards with blocky pictures/pixellation artifacts after skip or
when CPU is busy.
|
|
first/second field state independant of them being reference or not. Fixes Ticket354
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
|
|
|
|
|
|
syslog.h
|
|
|
|
fixed: reinstate compressed rar dialogs
|
|
Set avctx->coded_width/height to uncropped h264 sizes
avctx->widht/height remain right/bottom cropped as previous behaviour.
Hardware decoders need to know the uncropped data to allocate surfaces
of correct height. Some hardware is picky and fails to decode properly
if a surface larger than needed is used during decode, so just aligning
up is not enough.
Written and submitted upstream by elupus.
|
|
this just fixes the crash, they still don't work properly.
see ticket #12178
|
|
closes ticket #4401
|
|
this backports upstream commits aedb87c57, 86a4e55e6, 8475ec190,
60fcc19bff (partially), 237bcaa27518, 33789862, 8af4ac5272, 26b64550d6
thanks a million to michael niedermayer for giving this priority!
|
|
|
|
not permitted in iOS5 due to ASLR and will panic the darwin kernel when ffmpeg is dyloaded
|
|
libcmyth bug fix
|
|
|
|
*Wrong debug callback function
*Reference counting broken on linux
*cmyth_mysql_testdb_connection returns a pointer to a stack variable
*ref_reallof derefence NULL pointer in debug build
|
|
|
|
backport of ffmpeg dc6c36ce46d4c4d7cb63503afc2ee44f00bf3725 on 2011-08-17
fixes a bitmap sub color problem exposed by mkclean --optimize foo.mkv.
The compressed private data could not be extracted, resulting in a bad color
palette and wrong colors on screen.
problem described in forum thread http://forum.xbmc.org/showthread.php?t=108942
|
|
|
|
GPUs with ATI UVDa and UVD+ have trouble when decoding doesn't start on an
I-Frame, and they don't recover on later I-Frames.
The variable to track the first I-Frame is in H264Context so that it can be
reset by code in h264 when initializing the context or flushing.
credits isidrogar, see ticket #11772.
sample on team ftp, samples/11772/CSI_ New York - TV3 - 2008-09-16_1.ts
|
|
see ticket #11643
|
|
See ticket #11643, sample on team ftp, samples/11643/vc-1 test.wmv
|
|
The number of slices is not very likely to change from frame to frame, so
at the beginning of a new frame, allocate memory for the amount of slices of
the previous frame. Saves a lot of reallocation, for some TV capture samples
there are over 200 slices.
There wasn't anywhere really appropriate to store last_slice_count (needs to
live from first frame to last frame), so this is likely to cause discussion to
merge upstream.
Adding members to dxva_context breaks ABI, which we don't care too much about
since on Windows we don't support external ffmpeg.
dxva mpeg2 code also has access to MpegEncContext, but adding there would
likely break ABI as well.
|
|
They used to result in images with a black bottom.
sample on team ftp samples/PR471/too_many_slices.ts
Inspired by the vaapi code to reallocate the slices array for each new slice.
Could be more efficient if the array could be preserved for all frames and
freed only at the end of the video, but there doesn't seem to be anywhere
appropriate to free the memory at the end.
Alternative is to allocate the proper size straight away for a new frame,
instead of realloc'ing for each slice.
|
|
|
|
to be pure convenence make and remove the internal/external bits, libcec is a system lib. treat it that way.
|
|
on darwin. sed will point to the native /usr/bin/sed which is not gnu sed, use this format so that it uses the correct sed as found by autotools
|
|
replaced some strcpy's that were causing crashes, with strncpy's in unrarxlib
|
|
two different ways resulting in different python HASH for the same module. If the HASH is different then imaging calls exit(1) and everything dies. Mod to ignore this. Note Imaging-1.2 alpha changes this behavor to use one namespace ie. PIL
|
|
--enable-external-libcec option to configure.
to keep a permanent copy, remove "make download" from bootstrap
to change the location of the tarball to another mirror, change lib/libcec/Makefile.in
|
|
|
|
|
|
|
|
#11342, thanks cg110.
Alpha array size is 256 in ffmpeg and patch was not updated at last ffmpeg
update, which caused memory corruption for some subs.
See ticket #11342 for details.
|
|
cross-build any python module
|
|
|
|
|
