ffmpeg.git
9 years agovp8: use av_clip_uintp2() where possible
Mans Rullgard [Sat, 28 May 2011 16:20:49 +0000 (17:20 +0100)]
vp8: use av_clip_uintp2() where possible

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoac3enc: fix LOCAL_ALIGNED usage in count_mantissa_bits()
Mans Rullgard [Sat, 28 May 2011 19:27:09 +0000 (20:27 +0100)]
ac3enc: fix LOCAL_ALIGNED usage in count_mantissa_bits()

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoac3dsp: do not use the ff_* prefix when referencing ff_ac3_bap_bits.
Justin Ruggles [Sat, 28 May 2011 18:39:18 +0000 (14:39 -0400)]
ac3dsp: do not use the ff_* prefix when referencing ff_ac3_bap_bits.

this should fix the windows builds

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agoac3dsp: fix loop condition in ac3_update_bap_counts_c()
Justin Ruggles [Sat, 28 May 2011 18:40:16 +0000 (14:40 -0400)]
ac3dsp: fix loop condition in ac3_update_bap_counts_c()

9 years agoARM: unbreak build
Mans Rullgard [Sat, 28 May 2011 17:36:07 +0000 (18:36 +0100)]
ARM: unbreak build

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoac3enc: modify mantissa bit counting to keep bap counts for all values of bap
Justin Ruggles [Thu, 26 May 2011 19:53:25 +0000 (15:53 -0400)]
ac3enc: modify mantissa bit counting to keep bap counts for all values of bap
instead of just 0 to 4.

This does all the actual bit counting as a final step.

9 years agoac3enc: split mantissa bit counting into a separate function.
Justin Ruggles [Thu, 26 May 2011 16:31:31 +0000 (12:31 -0400)]
ac3enc: split mantissa bit counting into a separate function.

No speed difference. This is to allow for more flexible bit counting.

9 years agoac3enc: store per-block/channel bap pointers by reference block in a 2D array
Justin Ruggles [Thu, 26 May 2011 19:57:21 +0000 (15:57 -0400)]
ac3enc: store per-block/channel bap pointers by reference block in a 2D array
rather than in the AC3Block struct.

This will make it easier to access the bap values without having to chase
the reference block pointers each time.

9 years agoget_bits: add av_unused tag to cache variable
Mans Rullgard [Sat, 28 May 2011 14:46:55 +0000 (15:46 +0100)]
get_bits: add av_unused tag to cache variable

This silences numerous compiler warnings from skip_bits(),
where the cache variable is not used.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agosws: replace all long with int.
Anton Khirnov [Thu, 26 May 2011 16:57:26 +0000 (18:57 +0200)]
sws: replace all long with int.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
9 years agoARM: aacdec: fix constraints on inline asm
Mans Rullgard [Fri, 27 May 2011 21:52:49 +0000 (22:52 +0100)]
ARM: aacdec: fix constraints on inline asm

This adds output operands for modified memory allowing the
volatile qualifiers to be dropped.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: remove unnecessary volatile from inline asm
Mans Rullgard [Fri, 27 May 2011 21:49:28 +0000 (22:49 +0100)]
ARM: remove unnecessary volatile from inline asm

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: add "cc" clobbers to inline asm where needed
Mans Rullgard [Fri, 27 May 2011 20:59:46 +0000 (21:59 +0100)]
ARM: add "cc" clobbers to inline asm where needed

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: improve FASTDIV asm
Mans Rullgard [Fri, 27 May 2011 20:43:15 +0000 (21:43 +0100)]
ARM: improve FASTDIV asm

This uses one register less.  Also add missing "cc" clobber.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoac3enc: use LOCAL_ALIGNED macro
Mans Rullgard [Fri, 27 May 2011 19:31:57 +0000 (20:31 +0100)]
ac3enc: use LOCAL_ALIGNED macro

Aligned local variables must use the LOCAL_ALIGNED macro to ensure
correct alignment on all systems.  The unusual size argument to
memset is required since the implementation of LOCAL_ALIGNED may
use a pointer to an array.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoAPIchanges: fill in git hash for av_get_pix_fmt_name (0420bd7).
Ronald S. Bultje [Sat, 28 May 2011 13:46:42 +0000 (09:46 -0400)]
APIchanges: fill in git hash for av_get_pix_fmt_name (0420bd7).

9 years agolavu: add av_get_pix_fmt_name() convenience function
Stefano Sabatini [Sat, 28 May 2011 00:39:20 +0000 (02:39 +0200)]
lavu: add av_get_pix_fmt_name() convenience function

Also deprecate avcodec_get_pix_fmt_name() in its favor.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
9 years agocmdutils: remove OPT_FUNC2
Stefano Sabatini [Fri, 27 May 2011 23:53:00 +0000 (01:53 +0200)]
cmdutils: remove OPT_FUNC2

Make ff* tools only accept opt_* functions taking two arguments.

The distinction between functions with one and two arguments is quite
pointless. Simplify parse_options() code.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
9 years agoswscale: fix crash in bilinear scaling.
Ronald S. Bultje [Fri, 27 May 2011 23:01:47 +0000 (19:01 -0400)]
swscale: fix crash in bilinear scaling.

9 years agovpxenc: add VP8E_SET_STATIC_THRESHOLD mapping
James Zern [Tue, 12 Apr 2011 02:40:47 +0000 (04:40 +0200)]
vpxenc: add VP8E_SET_STATIC_THRESHOLD mapping

via the equivalent AVCodecContext::mb_threshold

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
9 years agowebm: support stereo videos in matroska/webm muxer
Alok Ahuja [Sat, 28 May 2011 06:12:22 +0000 (08:12 +0200)]
webm: support stereo videos in matroska/webm muxer

Create a stereo_mode metadata tag to specify the stereo 3d video layout
using the StereoMode tag in a matroska/webm video track.

9 years agolavf: deprecate AVFormatParameters.pix_fmt.
Anton Khirnov [Fri, 27 May 2011 05:38:49 +0000 (07:38 +0200)]
lavf: deprecate AVFormatParameters.pix_fmt.

9 years agorawdec: add a pixel_format private option.
Anton Khirnov [Tue, 24 May 2011 07:02:11 +0000 (09:02 +0200)]
rawdec: add a pixel_format private option.

9 years agov4l2: add a pixel_format private option.
Anton Khirnov [Tue, 24 May 2011 07:02:11 +0000 (09:02 +0200)]
v4l2: add a pixel_format private option.

9 years agolibdc1394: add a pixel_format private option.
Anton Khirnov [Tue, 24 May 2011 07:02:11 +0000 (09:02 +0200)]
libdc1394: add a pixel_format private option.

9 years agocosmetics: indentation and alignment after previous commit
Justin Ruggles [Tue, 24 May 2011 23:38:30 +0000 (19:38 -0400)]
cosmetics: indentation and alignment after previous commit

9 years agoac3enc: add support for E-AC-3 encoding.
Justin Ruggles [Tue, 24 May 2011 19:20:56 +0000 (15:20 -0400)]
ac3enc: add support for E-AC-3 encoding.

This adds basic stream format support and allows for arbitrary bit rates
rather than just those supported in AC-3.

9 years agoac3enc: Move AC-3 AVOptions array to a separate file to make it easier to
Justin Ruggles [Tue, 24 May 2011 19:14:36 +0000 (15:14 -0400)]
ac3enc: Move AC-3 AVOptions array to a separate file to make it easier to
use only selected options for the different AC-3 encoder types.

9 years agoARM: disable ff_vector_fmul_vfp on VFPv3 systems
Mans Rullgard [Fri, 27 May 2011 19:18:26 +0000 (20:18 +0100)]
ARM: disable ff_vector_fmul_vfp on VFPv3 systems

This function uses old-style vector operations deprecated in VFPv3.
Some implementations, e.g. Cortex-A9, support them only through
slow software emulation.  Cortex-A8 does have this functionality
in hardware, but as it also has NEON, this function is not used
there regardless.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: check for VFPv3
Mans Rullgard [Fri, 27 May 2011 19:12:01 +0000 (20:12 +0100)]
ARM: check for VFPv3

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoswscale: Remove unused variables in x86 code.
Diego Biurrun [Fri, 27 May 2011 17:46:39 +0000 (19:46 +0200)]
swscale: Remove unused variables in x86 code.

libswscale/x86/swscale_template.c:2072: warning: unused variable ‘canMMX2BeUsed’
libswscale/x86/swscale_template.c:2145: warning: unused variable ‘canMMX2BeUsed’
libswscale/x86/swscale_template.c:2209: warning: unused variable ‘chrVPixBuf’
libswscale/x86/swscale_template.c:2237: warning: unused variable ‘chrVSrcPtr’

9 years agodoc: Drop DJGPP section, Libav now compiles out-of-the-box on FreeDOS.
Diego Biurrun [Thu, 26 May 2011 12:39:37 +0000 (14:39 +0200)]
doc: Drop DJGPP section, Libav now compiles out-of-the-box on FreeDOS.

9 years agox86: Add appropriate ifdefs around certain AVX functions.
Diego Biurrun [Fri, 27 May 2011 19:18:12 +0000 (21:18 +0200)]
x86: Add appropriate ifdefs around certain AVX functions.

nasm versions prior to 2.09 have trouble assembling some of our AVX code.
Protect these sections by preprocessor macros to allow compilation to pass.

9 years agocmdutils: use sws_freeContext() instead of av_freep().
Ronald S. Bultje [Fri, 27 May 2011 15:57:39 +0000 (11:57 -0400)]
cmdutils: use sws_freeContext() instead of av_freep().

av_freep(swsContext) will leak all memory potentially
allocated within the swsContext.

9 years agoswscale: delay allocation of formatConvBuffer().
Ronald S. Bultje [Fri, 27 May 2011 16:23:32 +0000 (12:23 -0400)]
swscale: delay allocation of formatConvBuffer().

That means it won't be allocated when not needed. Alongside
this, it fixes valgrind/fate-detected memory leaks.

9 years agoswscale: fix build with --disable-swscale-alpha.
Ronald S. Bultje [Fri, 27 May 2011 15:36:43 +0000 (11:36 -0400)]
swscale: fix build with --disable-swscale-alpha.

9 years agomovenc: Deprecate the global RTP hinting flag, use a private AVOption instead
Martin Storsjö [Fri, 20 May 2011 09:27:02 +0000 (12:27 +0300)]
movenc: Deprecate the global RTP hinting flag, use a private AVOption instead

Instead of -fflags rtphint, set -movflags rtphint instead.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agomovenc: Add an AVClass for setting muxer specific options
Martin Storsjö [Sat, 21 May 2011 11:57:04 +0000 (14:57 +0300)]
movenc: Add an AVClass for setting muxer specific options

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agoswscale: fix non-bitexact yuv2yuv[X2]() MMX/MMX2 functions.
Ronald S. Bultje [Fri, 27 May 2011 13:28:38 +0000 (09:28 -0400)]
swscale: fix non-bitexact yuv2yuv[X2]() MMX/MMX2 functions.

9 years agoconfigure: report yasm/nasm presence properly
Luca Barbato [Thu, 26 May 2011 18:03:37 +0000 (20:03 +0200)]
configure: report yasm/nasm presence properly

If the secondary assembler is in use report the proper name

9 years agotcp: make connect() timeout properly
Luca Barbato [Fri, 20 May 2011 12:09:21 +0000 (14:09 +0200)]
tcp: make connect() timeout properly

The connect() timeout can take minutes, gets misreported as EIO and
isn't interruptible.

9 years agorawdec: factor video demuxer definitions into a macro.
Anton Khirnov [Wed, 25 May 2011 06:14:13 +0000 (08:14 +0200)]
rawdec: factor video demuxer definitions into a macro.

9 years agortspdec: add initial_pause private option.
Anton Khirnov [Tue, 24 May 2011 06:44:10 +0000 (08:44 +0200)]
rtspdec: add initial_pause private option.

Deprecate corresponding AVFormatParameters field.

9 years agolavf: deprecate AVFormatParameters.width/height.
Anton Khirnov [Thu, 26 May 2011 06:14:03 +0000 (08:14 +0200)]
lavf: deprecate AVFormatParameters.width/height.

9 years agotty: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
tty: add video_size private option.

9 years agorawdec: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
rawdec: add video_size private option.

9 years agox11grab: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
x11grab: add video_size private option.

9 years agox11grab: factorize returning error codes.
Anton Khirnov [Thu, 26 May 2011 19:17:05 +0000 (21:17 +0200)]
x11grab: factorize returning error codes.

9 years agovfwcap: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
vfwcap: add video_size private option.

9 years agov4l2: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
v4l2: add video_size private option.

9 years agov4l2: factorize returning error codes.
Anton Khirnov [Thu, 26 May 2011 18:37:08 +0000 (20:37 +0200)]
v4l2: factorize returning error codes.

This will be useful in the following commit.

9 years agolibdc1394: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
libdc1394: add video_size private option.

9 years agolibdc1394: return meaninful error codes.
Anton Khirnov [Thu, 26 May 2011 15:13:09 +0000 (17:13 +0200)]
libdc1394: return meaninful error codes.

9 years agobktr: add video_size private option.
Anton Khirnov [Tue, 24 May 2011 05:43:01 +0000 (07:43 +0200)]
bktr: add video_size private option.

9 years agobktr: factorize returning error codes.
Anton Khirnov [Thu, 26 May 2011 18:14:53 +0000 (20:14 +0200)]
bktr: factorize returning error codes.

This will be useful in the following commit.

9 years agoid3v2: Check malloc result. ID3v2 tags can be very large.
Alex Converse [Thu, 26 May 2011 00:57:33 +0000 (17:57 -0700)]
id3v2: Check malloc result. ID3v2 tags can be very large.

9 years agoid3v2: Initialize tflags for version 2.2.
Alex Converse [Thu, 26 May 2011 00:03:12 +0000 (17:03 -0700)]
id3v2: Initialize tflags for version 2.2.

9 years agoARM: add ARMv6 optimised av_clip_uintp2
Mans Rullgard [Tue, 24 May 2011 14:46:28 +0000 (15:46 +0100)]
ARM: add ARMv6 optimised av_clip_uintp2

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: remove volatile from asm statements in libavutil/intmath
Mans Rullgard [Tue, 24 May 2011 14:45:50 +0000 (15:45 +0100)]
ARM: remove volatile from asm statements in libavutil/intmath

The volatile qualifiers are not needed on these statements as
their effects are fully specified by constraints.

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agoARM: fix av_clipl_int32_arm()
Mans Rullgard [Wed, 25 May 2011 17:43:28 +0000 (18:43 +0100)]
ARM: fix av_clipl_int32_arm()

Signed-off-by: Mans Rullgard <mans@mansr.com>
9 years agov4l: include avdevice.h
Anton Khirnov [Thu, 26 May 2011 18:32:05 +0000 (20:32 +0200)]
v4l: include avdevice.h

Fixes build.

9 years agoffserver: move close_connection() call to avoid a temporary string and copy.
Mike Williams [Wed, 18 May 2011 13:14:22 +0000 (09:14 -0400)]
ffserver: move close_connection() call to avoid a temporary string and copy.

Signed-off-by: Mike Williams <mike@mikebwilliams.com>
Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years agolavf: initialize demuxer private options.
Anton Khirnov [Thu, 26 May 2011 04:59:08 +0000 (06:59 +0200)]
lavf: initialize demuxer private options.

9 years agoAVOptions: set string default values.
Anton Khirnov [Thu, 26 May 2011 06:48:31 +0000 (08:48 +0200)]
AVOptions: set string default values.

9 years agolavdevice: mark v4l for removal on next major bump.
Anton Khirnov [Thu, 26 May 2011 15:48:52 +0000 (17:48 +0200)]
lavdevice: mark v4l for removal on next major bump.

9 years agoswscale: fix compile on ppc.
Ronald S. Bultje [Thu, 26 May 2011 15:32:32 +0000 (11:32 -0400)]
swscale: fix compile on ppc.

9 years agoswscale: fix compile on x86-32.
Ronald S. Bultje [Thu, 26 May 2011 14:36:47 +0000 (10:36 -0400)]
swscale: fix compile on x86-32.

9 years agobuild: Remove generated .version file on distclean.
Diego Biurrun [Thu, 26 May 2011 13:48:31 +0000 (15:48 +0200)]
build: Remove generated .version file on distclean.

9 years agoconfigure: Add -D_GNU_SOURCE to CPPFLAGS on OS/2.
Diego Biurrun [Thu, 26 May 2011 12:21:04 +0000 (14:21 +0200)]
configure: Add -D_GNU_SOURCE to CPPFLAGS on OS/2.

The flag is required for some C99 math functions to be declared.

9 years agodoc: Drop hint at --enable-memalign-hack for MinGW, it is now autodetected.
Diego Biurrun [Thu, 26 May 2011 12:40:29 +0000 (14:40 +0200)]
doc: Drop hint at --enable-memalign-hack for MinGW, it is now autodetected.

9 years agoffplay: Remove disabled code.
Diego Biurrun [Tue, 26 Apr 2011 12:29:14 +0000 (14:29 +0200)]
ffplay: Remove disabled code.

9 years agoMark parameterless function declarations as 'void'.
Diego Biurrun [Thu, 26 May 2011 11:37:13 +0000 (13:37 +0200)]
Mark parameterless function declarations as 'void'.

9 years agoswscale: use av_clip_uint8() in yuv2yuv1_c().
Ronald S. Bultje [Wed, 25 May 2011 19:28:12 +0000 (15:28 -0400)]
swscale: use av_clip_uint8() in yuv2yuv1_c().

9 years agoswscale: remove VOF/VOFW.
Ronald S. Bultje [Thu, 26 May 2011 13:17:52 +0000 (09:17 -0400)]
swscale: remove VOF/VOFW.

9 years agoswscale: split chroma buffers into separate U/V planes.
Ronald S. Bultje [Wed, 25 May 2011 18:30:09 +0000 (14:30 -0400)]
swscale: split chroma buffers into separate U/V planes.

Preparatory step to implement support for sizes > VOFW.

9 years agoswscale: replace formatConvBuffer[VOF] by allocated array.
Ronald S. Bultje [Thu, 26 May 2011 13:15:38 +0000 (09:15 -0400)]
swscale: replace formatConvBuffer[VOF] by allocated array.

This allows to convert between formats of arbitrary width,
regardless of the value of VOF/VOFW.

9 years agorgb2rgb: remove duplicate mmx/mmx2/3dnow/sse2 functions.
Ronald S. Bultje [Tue, 24 May 2011 22:28:40 +0000 (18:28 -0400)]
rgb2rgb: remove duplicate mmx/mmx2/3dnow/sse2 functions.

Many functions have such a prefix, but do not actually use any
instructions or features from that set, thus giving the false
impression that swscale is highly optimized for a particular
system, whereas in reality it is not.

9 years agoswscale: reindent h[cy]scale_fast() and updateDitherTables().
Ronald S. Bultje [Tue, 24 May 2011 21:50:03 +0000 (17:50 -0400)]
swscale: reindent h[cy]scale_fast() and updateDitherTables().

9 years agoswscale: reformat x86/swscale_template.c.
Ronald S. Bultje [Thu, 26 May 2011 13:11:29 +0000 (09:11 -0400)]
swscale: reformat x86/swscale_template.c.

Interleave macros and code so that it's easier to find the
actual code that belongs to a function. Also reindent where
appropriate and remove dead code.

9 years agoswscale: remove duplicate mmx/mmx2 functions if they are identical.
Ronald S. Bultje [Tue, 24 May 2011 20:45:41 +0000 (16:45 -0400)]
swscale: remove duplicate mmx/mmx2 functions if they are identical.

9 years agoswscale: remove if (c->dstFormat) branch from yuv2packed[12X]().
Ronald S. Bultje [Tue, 24 May 2011 20:02:47 +0000 (16:02 -0400)]
swscale: remove if (c->dstFormat) branch from yuv2packed[12X]().

This allows cutting up the function in much smaller and easier-
to-maintain chunks.

9 years agoswscale: remove if(full_chr_int) from yuv2packed1().
Ronald S. Bultje [Tue, 24 May 2011 19:18:40 +0000 (15:18 -0400)]
swscale: remove if(full_chr_int) from yuv2packed1().

If that flag is set, swScale() already proxies the call to
yuv2rgbXinC_full(). Therefore, this flag is never set when
yuv2packed1() is called.

9 years agoswscale: remove if(accurate_rnd) branch from functions.
Ronald S. Bultje [Tue, 24 May 2011 19:16:19 +0000 (15:16 -0400)]
swscale: remove if(accurate_rnd) branch from functions.

9 years agoffserver: Fix a null pointer dereference as a result of the FF_API_MAX_STREAMS cleanup.
Mike Williams [Wed, 18 May 2011 15:03:10 +0000 (11:03 -0400)]
ffserver: Fix a null pointer dereference as a result of the FF_API_MAX_STREAMS cleanup.

Fixed another dereference in the RTSP code.

Removed a useless variable.

Changed an unnecessary looping assignment to a simple assignment suggested by
Maksym.

Added fixes and tweaks suggested by Maksym Veremeyenko [verem@m1stereo.tv] and
Clément B.

9 years agolibdc1394: fix compilation.
Anton Khirnov [Thu, 26 May 2011 05:25:41 +0000 (07:25 +0200)]
libdc1394: fix compilation.

Add a forgotten comma and move options/class definition outside of
HAVE_LIBDC1394_1.

9 years agoswscale: revive SWS_CPU_CAPS until next major bump.
Anton Khirnov [Wed, 25 May 2011 19:47:31 +0000 (21:47 +0200)]
swscale: revive SWS_CPU_CAPS until next major bump.

9 years agoswscale: Remove commented-out printf cruft.
Diego Biurrun [Wed, 25 May 2011 23:40:56 +0000 (01:40 +0200)]
swscale: Remove commented-out printf cruft.

9 years agoac3enc: initialize all coefficients to zero.
Justin Ruggles [Thu, 26 May 2011 00:12:57 +0000 (20:12 -0400)]
ac3enc: initialize all coefficients to zero.

Uninitialized coefficients were being used to generate exponents, some
of which actually ended up in the final stream.  Even though, they were
just extra exponents that are not used by any decoder, it is still
better to have consistent output for testing. This also fixes valgrind
errors.

9 years agodoc: create separate section for audio encoders
Stefano Sabatini [Thu, 19 May 2011 17:41:19 +0000 (19:41 +0200)]
doc: create separate section for audio encoders

Having a separate section for audio encoders simplifies navigation
and is slightly more consistent with the rest of the manual.

Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agoswscale: Remove orphaned, commented-out function declaration.
Diego Biurrun [Wed, 25 May 2011 20:07:55 +0000 (22:07 +0200)]
swscale: Remove orphaned, commented-out function declaration.

9 years agoswscale: Eliminate rgb24toyv12_c() duplication.
Michael Niedermayer [Wed, 25 May 2011 18:49:52 +0000 (20:49 +0200)]
swscale: Eliminate rgb24toyv12_c() duplication.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Diego Biurrun <diego@biurrun.de>
9 years agoRemove h263_msmpeg4 from MpegEncContext.
Alex Converse [Tue, 24 May 2011 23:28:45 +0000 (16:28 -0700)]
Remove h263_msmpeg4 from MpegEncContext.

It was long ago superseded by msmpeg4_version.

9 years agoAPIchanges: Fill in git hash for fps_probe_size (30315a8)
Alex Converse [Wed, 25 May 2011 19:36:41 +0000 (12:36 -0700)]
APIchanges: Fill in git hash for fps_probe_size (30315a8)

9 years agoavformat: Add fpsprobesize as an AVOption.
Alex Converse [Mon, 9 May 2011 21:34:23 +0000 (14:34 -0700)]
avformat: Add fpsprobesize as an AVOption.

9 years agoavoptions: Return explicitly NAN or {0,0} if the option isn't found
Martin Storsjö [Wed, 25 May 2011 13:47:09 +0000 (16:47 +0300)]
avoptions: Return explicitly NAN or {0,0} if the option isn't found

This actually matches what av_get_double did earlier, the
0.0/0.0 division was intentional, for producing NAN.

Still keeping the check for the return value from
av_get_number, for clarity.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortmp: Reindent
Martin Storsjö [Wed, 25 May 2011 19:01:15 +0000 (22:01 +0300)]
rtmp: Reindent

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agortmp: Don't try to do av_malloc(0)
Martin Storsjö [Wed, 25 May 2011 16:08:29 +0000 (19:08 +0300)]
rtmp: Don't try to do av_malloc(0)

Some received packets can have size 0. The return value from
av_malloc(0) may be NULL, which is ok if the size was 0. On
OS X, however, the returned pointer is non-null but leads to
crashes when trying to free it.

Signed-off-by: Martin Storsjö <martin@martin.st>
9 years agotty: replace AVFormatParameters.sample_rate abuse with a private option.
Anton Khirnov [Mon, 23 May 2011 17:23:55 +0000 (19:23 +0200)]
tty: replace AVFormatParameters.sample_rate abuse with a private option.

9 years agoFix end time of last chapter in compute_chapters_end
John Stebbins [Mon, 23 May 2011 21:26:54 +0000 (14:26 -0700)]
Fix end time of last chapter in compute_chapters_end

Parenthesis are misplaced in calculation of max_time.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
9 years agoffmpeg: get rid of useless AVInputStream.nb_streams.
Anton Khirnov [Mon, 23 May 2011 15:19:54 +0000 (17:19 +0200)]
ffmpeg: get rid of useless AVInputStream.nb_streams.

It's a duplicate of AVFormatContext.nb_streams.