Fix trac ticket #2247.

17% faster
Note the speedup is smaller if its done before
storing in the context (gcc 4.6.3 core i7)
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This makes the code about 7% faster
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes Ticket2229
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reference:
commit 3615e2be
Author: Michael Niedermayer <michaelni@gmx.at>
Date:   Tue Dec 2 22:02:57 2003 +0000

    h263_h_loop_filter_mmx

    Originally committed as revision 2553 to svn://svn.ffmpeg.org/ffmpeg/trunk

commit 359f98de
Author: Michael Niedermayer <michaelni@gmx.at>
Date:   Tue Dec 2 20:28:10 2003 +0000

    h263_v_loop_filter_mmx

    Originally committed as revision 2552 to svn://svn.ffmpeg.org/ffmpeg/trunkSigned-off-by: Michael Niedermayer <michaelni@gmx.at>

* qatar/master:
  dsputil: x86: Fix compile error
  dsputil: x86: Convert h263 loop filter to yasm
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit '12b54a1f':
  mpegvideo: remove an unused function parameter
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit '605b047b':
  rv10: improve buffer size check.
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

* commit 'c5fcdb44':
  error_resilience: remove a useless if() and FIXME
  h264: remove silly macros
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit 'ca1fe6c0':
  h263: remove an unused parameter from ff_h263_decode_init_vlc
  flac: add channel layout masks for streams with 7 or 8 channels.
  flac: don't check the number of channels before setting the channel layout.

Conflicts:
	libavcodec/version.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit 'a846dccb':
  h264chroma: x86: Fix building with yasm disabled
  rv34: Drop now unnecessary dsputil dependencies

Conflicts:
	libavcodec/x86/Makefile
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit '79dad2a9':
  dsputil: Separate h264chroma

Conflicts:
	libavcodec/dsputil_template.c
	libavcodec/ppc/dsputil_ppc.c
	libavcodec/vc1dec.c
	libavcodec/vc1dsp.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This is unnecessary, as ff_flac_set_channel_layout can handle any number of channels.

Accidentally prefixed ff_ with cextern.
Signed-off-by: Martin Storsjö <martin@martin.st>

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Non-ascii chars are not correctly rendered in both MAN and HTML output or
may even cause compilation failure. Thus avoid to use them until we find
a satisfying way to deal with them.

Fix trac ticket #2243.

The DTS needs to be resynched against the segment start PTS, or the
resulting DTS may result < PTS.
Reported-by: Owen Jones <riots6@gmail.com>

See thread:
Subject: [FFmpeg-user] pts/dts error using reset_timestamps while splitting a DVD
Date: Sat, 19 Jan 2013 08:58:27 +0000

Fixes ticket #1764.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Check slice count and input buffer size before constructing a possibly
invalid pointer, not after.

pp_time is never set for h264

They serve no useful purpose and wreak all kind of havoc when h264.h is
included elsewhere.

They were added to the latest FLAC specification:
https://git.xiph.org/?p=flac-website.git;a=commit;h=65c199a2Signed-off-by: Anton Khirnov <anton@khirnov.net>

This is unnecessary, as ff_flac_set_channel_layout can handle any number of channels.
Signed-off-by: Anton Khirnov <anton@khirnov.net>

* rdp/combined2:
  dshow: allow for more codec ID's
  dshow: better error reporting
  dshow: fix log message
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Modified the fate test crc generator to print the side_data's
crc if side_data is present.
Signed-off-by: Vignesh Venkatasubramanian <vigneshv@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

* qatar/master:
  mpegvideo: initialize dummy reference frames.

Conflicts:
	libavcodec/mpegvideo.c
	tests/ref/fate/filter-yadif-mode1

The luma init is left at 0x80 as it was previously, and tested and
working since a long time. Its also mandated to be 0x80 by at least
one specification. Cases that need different values already had
code in place to handle them.
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Found-by: aballier
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

* commit 'dff6197d':
  nuv: do not rely on get_buffer() initializing the frame.
  yop: initialize palette to 0

Conflicts:
	libavcodec/nuv.c
	tests/ref/fate/nuv-rtjpeg-fh
	tests/ref/fate/yop
Merged-by: Michael Niedermayer <michaelni@gmx.at>

* commit '8136f234':
  yop: check for input overreads.
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Martin Storsjö <martin@martin.st>

* commit '06cf597c':
  yop: check that extradata is large enough.

Conflicts:
	libavcodec/yop.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>