Merge commit 'd5c62122a7b26704bf867a1262df358623bf5edf'

* commit 'd5c62122a7b26704bf867a1262df358623bf5edf':
  Move av_reverse table to libavcodec

Conflicts:
libavcodec/asvenc.c
libavcodec/vble.c
libavutil/common.h
libavutil/mathematics.c

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

Merge commit '930c9d4373e0f3cb7c64fcfc129127a309f6d066'

* commit '930c9d4373e0f3cb7c64fcfc129127a309f6d066':
  avutil: Duplicate ff_log2_tab instead of sharing it across libs

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

Merge commit 'f6c38c5f4ed6683a6a61db2ed418a68bbe5f5507'

* commit 'f6c38c5f4ed6683a6a61db2ed418a68bbe5f5507':
  avfilter: call x86 init functions under if (ARCH_X86), not if (HAVE_MMX)
  rtspdec: Set the default port for listen mode, if none is specified
  tscc2: Fix an out of array access
  rtmpproto: Fix an out of array write
  rtspdec: Fix use of uninitialized byte
  vp8: reset loopfilter delta values at keyframes.
  avutil: add yuva422p and yuva444p formats

Conflicts:
libavutil/pixdesc.c
libavutil/pixfmt.h
tests/ref/lavfi/pixdesc
tests/ref/lavfi/pixfmts_copy
tests/ref/lavfi/pixfmts_null
tests/ref/lavfi/pixfmts_scale
tests/ref/lavfi/pixfmts_vflip

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

vp8dec: reset loopfilter delta values at keyframes

while testing a VP8 encoder I found a mismatch between FFMPEG and libvpx
VP8 decoders. The reason for this mismatch is that FFMPEG doesn't reset
loopfilter delta values at keyframes. Patch that fixes the issue is below.
I've verified that the output of ffmpeg after this patch matches libvpx.

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

Allow autodetection of some dnxhd files that can be decoded correctly.

Fixes ticket #1807.

fate: update pcx reference

It changed because encoder now writes 0/1 for
sample aspect ratio instead of 0/0.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

pcxenc: store sample aspect ratio

Signed-off-by: Paul B Mahol <onemda@gmail.com>

pcx: read sample aspect ratio

Signed-off-by: Paul B Mahol <onemda@gmail.com>

bethsoftvideo: return meaningfull error codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

truemotion2: use more meaningful return codes

Signed-off-by: Paul B Mahol <onemda@gmail.com>

ffserver: fix return value of add_codec()

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

jpegls: increase run_index to 4

Fixes part of CID717913
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

jpegls: fix off limit

Fixes part of CID717913
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

rv30_decode_intra_types: make check tighter

This makes no differnce as the affected values cannot occur
But it will likely help static analyzers

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

ff_celp_lp_synthesis_filterf: change loop end check

This makes no difference for supported input but avoids a
out of array read with input that is never passed in currently

Fixes CID717919
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ff_celp_lp_synthesis_filterf: check that filter_length is within the supported range

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

rtpdec_qdm2: change one assert to av_assert0

Failure of this assert could lead to hard to debug behavior

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

wtvdec: fix memleak on error

Fixes CID718002
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avio: redesign ffio_rewind_with_probe_data()

This prevents a double free

Fixes CID718285
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffv1: avoid checking a double for equality

if 0.0 != 0.0 a out of array read would occur, equal checks
with floating point may behave in such odd ways, though
this is very unlikely in a real implementation of a compiler

Fixes: CID718936
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

af_pan: fix memleak of arg

Fixes CID718988
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vf_fade: fix memleaks of args

Fixes: CID718989
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffserver: prevent nb_streams from becoming too large

Fixes CID732249

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

aacsbr: change order of operation to prevent out of array read

Fixes CID732250
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fix writing 12 bit DPX

The DPX encoder now writes 12 DPX that open correctly (tested with
ImageMagick), and also correspond to the 12 bit sample files at
http://samples.ffmpeg.org/image-samples/dpx_samples.zip

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

Fix DPX decoder

Rewrite the DPX decoder to work with provided sample DPXs at
http://samples.ffmpeg.org/image-samples/dpx_samples.zip

The decoder could only decode 8 and 10 bit without alpha correctly,
failing or even crashing at other flavors. This patch aims to fix
these issues, properly decoding all variants of DPX provided in the
referenced DPX sample zip. For 10 and 12 bit, the alpha channel
is ignored, but decoding is still possible.

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

lavf/swf: transform the swf tags define in an enum.

lavf/swf: re-align after previous commit.

lavf/swf: define more tags.

Based on swf specs v10, page 271-273.

lavf/swfdec: return more meaningful error codes.

mpeg4videodec: disable frame multithreading for GMC, its not implemented at all

This fixes race conditions in decoding MPEG-4 GMC files

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

lavf/swfdec: fix flushing with compressed swf.

We now only return in the middle of the refill in case of read error, so
inflate can be re-called if zlib needs an empty (z.avail_in=0) inflate
call for flushing.

Move av_reverse table to libavcodec

It is only used in that library.

avutil: Duplicate ff_log2_tab instead of sharing it across libs

The table is so small that the space gain is not worth the
performance overhead of cross-library access.

movenc: fix timescale for timecode track

regression since b02493e476

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

avfilter: call x86 init functions under if (ARCH_X86), not if (HAVE_MMX)

add dshow maintainer

Signed-off-by: rogerdpack <rogerpack2005@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

motion_est: more complete SAB diamond size check

This makes no difference with the current #defines

Fixes CID732255
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

caca: fix 10l typo

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

sws: do not use av_pix_fmt_descriptors directly

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavd: do not use av_pix_fmt_descriptors directly

Signed-off-by: Paul B Mahol <onemda@gmail.com>

exr: fix twin ;

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

avidec: Use sample size in case incorrect timestamps for aac in AVI (Ticket #1755)

In some case for aac in AVI, avidec extracts wrong PTS value.
(www.ffmpeg.org/trac/ffmpeg/ticket/1755)

I found additional case(ss=4096) and add condition.

Problematic file link : https://docs.google.com/open?id=0B6r7ZfWFIypCOTdZQUtGVEdJUUE

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

imgutils: do not use av_pix_fmt_descriptors directly

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Parse DEFINESOUND tags in swf (fix ticket 1638)

Signed-off-by: Michael Bradshaw <mbradshaw@sorensonmedia.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc: do not use av_pix_fmt_descriptors directly

Signed-off-by: Paul B Mahol <onemda@gmail.com>

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  lavu,lavd: do not use av_pix_fmt_descriptors directly.

Conflicts:
libavutil/imgutils.c

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

rtspdec: Set the default port for listen mode, if none is specified

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

tscc2: Fix an out of array access

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

rtmpproto: Fix an out of array write

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

rtspdec: Fix use of uninitialized byte

ffurl_read_complete can return 0 as well as negative error codes.

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

Merge commit '59ee9f78b0cc4fb84ae606fa317d8102ad32a627'

* commit '59ee9f78b0cc4fb84ae606fa317d8102ad32a627':
  lavfi: do not use av_pix_fmt_descriptors directly.

Conflicts:
libavfilter/buffersrc.c
libavfilter/drawutils.c
libavfilter/filtfmts.c
libavfilter/vf_ass.c
libavfilter/vf_boxblur.c
libavfilter/vf_drawtext.c
libavfilter/vf_lut.c
libavfilter/vf_pad.c
libavfilter/vf_scale.c
libavfilter/vf_showinfo.c
libavfilter/vf_transpose.c

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

Merge commit '50ba57e0ce63d9904269ea0728936a0c79f8bfb5'

* commit '50ba57e0ce63d9904269ea0728936a0c79f8bfb5':
  lavc: do not use av_pix_fmt_descriptors directly.

Conflicts:
libavcodec/imgconvert.c
libavcodec/libopenjpegdec.c
libavcodec/libopenjpegenc.c
libavcodec/mpegvideo.c
libavcodec/rawdec.c
libavcodec/rawenc.c
libavcodec/tiffenc.c

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

yop: check return value of avformat_new_stream()

Fixes null pointer dereference, fixes CID703729.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

vp8: reset loopfilter delta values at keyframes.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Merge commit '9953ff3cd844eb5f6d8dfce98cad94b78a0fc7dc'

* commit '9953ff3cd844eb5f6d8dfce98cad94b78a0fc7dc':
  mpegvideo: fix indentation
  sws: do not use av_pix_fmt_descriptors directly.

Conflicts:
libavcodec/mpegvideo.c
libswscale/swscale_internal.h
libswscale/swscale_unscaled.c
libswscale/utils.c

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

Merge commit 'b7f1010c8fce09096057528f7cd29589ea1ae7df'

* commit 'b7f1010c8fce09096057528f7cd29589ea1ae7df':
  tools: do not use av_pix_fmt_descriptors directly.
  pixdesc: add functions for accessing pixel format descriptors.
  build: add support for Tru64 (OSF/1)
  md5: Allocate a normal private context for the opaque md5 context pointer

Conflicts:
cmdutils.c
doc/APIchanges
ffprobe.c
libavformat/md5enc.c
libavutil/version.h
tools/graph2dot.c

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

avutil: add yuva422p and yuva444p formats

Merge commit 'e002e3291e6dc7953f843abf56fc14f08f238b21'

* commit 'e002e3291e6dc7953f843abf56fc14f08f238b21':
  Use the new aes/md5/sha/tree allocation functions
  avutil: Add functions for allocating opaque contexts for algorithms
  svq3: fix pointer type warning
  svq3: replace unsafe pointer casting with intreadwrite macros
  parseutils-test: various cleanups

Conflicts:
doc/APIchanges
libavcodec/svq3.c
libavutil/parseutils.c
libavutil/version.h

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

Merge commit '15ba7f6525c0f56f0c8e3e3e0c0c5129de054f41'

* commit '15ba7f6525c0f56f0c8e3e3e0c0c5129de054f41':
  parseutils: fix const removal warning
  prepare 9_beta1 release

Conflicts:
Changelog
RELEASE
libavutil/parseutils.c

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

Merge commit '9734b8ba56d05e970c353dfd5baafa43fdb08024'

* commit '9734b8ba56d05e970c353dfd5baafa43fdb08024':
  Move avutil tables only used in libavcodec to libavcodec.

Conflicts:
libavcodec/mathtables.c
libavutil/intmath.h

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

Merge commit '0a75d1da23b8659ec49391469bb592da12760077'

* commit '0a75d1da23b8659ec49391469bb592da12760077':
  options_table: refs option is not snow-only
  random_seed: Support using CryptGenRandom on windows
  doc: update the faq entry about custom I/O

Conflicts:
doc/faq.texi
libavcodec/options_table.h

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

truemotion2: remove unreachable code

Fixes CID610345.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavu,lavd: do not use av_pix_fmt_descriptors directly.

lavfi: do not use av_pix_fmt_descriptors directly.

lavc: do not use av_pix_fmt_descriptors directly.

mpegvideo: fix indentation

tools: do not use av_pix_fmt_descriptors directly.

sws: do not use av_pix_fmt_descriptors directly.

pixdesc: add functions for accessing pixel format descriptors.

Make av_pix_fmt_descriptors table static on next major bump.

Making the table public is dangerous, since the caller has no way to
know how large it actually is. It also prevents adding new fields to
AVPixFmtDescriptor without a major bump.

takdec: stop decoding in case of unknown bps

Signed-off-by: Paul B Mahol <onemda@gmail.com>

takdec: s/bits_per_coded_sample/bits_per_raw_sample

bits_per_coded_sample should be set from demuxer and
not from decoder.

Prior to this change value set from demuxer would get
overwritten in decoder.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

nellymoserenc: fix array element ordering

Fixes assertion failures in trellis encoder
Fixes valgrind warnings in trellis encoder
Fixes CID732257
Fixes CID732256
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

build: add support for Tru64 (OSF/1)

Signed-off-by: Mans Rullgard <mans@mansr.com>

tscc2: fix out of array access

Fixes CID732260
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

rtmpproto: fix out of array write

Fixes CID732263
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

motion-test: zero dsp context

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

dv: zero dsp before init, this fixes use of uninitialized dct_bits

Fixes CID732292
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavc: Fix use of uninitialized field.

Fixes CID732297
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

rtspdec:read_line: fix use of uninitialized byte

Fixes CID732301
Fixes CID723300
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

sanm: fix off by 1 error in draw_glyph()

This fix changes nothing as nothing passed a so large value

Fixes CID733785
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

update_initial_timestamps: increase pts_buffer size

Fixes CID733786
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

md5: Allocate a normal private context for the opaque md5 context pointer

This avoids having to overestimate the md5 context size, which
isn't known beforehand, allowing us to use the new allocate functions
instead.

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

Use the new aes/md5/sha/tree allocation functions

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

avutil: Add functions for allocating opaque contexts for algorithms

The current API where the plain size is exposed is not of much
use - in most cases it is allocated dynamically anyway.

If allocated e.g. on the stack via an uint8_t array, there's no
guarantee that the struct's members are aligned properly (unless
the array is overallocated and the opaque pointer within it
manually aligned to some unspecified alignment).

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

bethsoftvid: check return value of av_packet_new_side_data()

Fixes null pointer dereference, fixes CID733777.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

svq3: fix pointer type warning

Fixes:
libavcodec/svq3.c:661:9: warning: passing argument 2 of 'svq3_decode_block' from incompatible pointer type
libavcodec/svq3.c:208:19: note: expected 'DCTELEM *' but argument is of type 'DCTELEM (*)[32]'

Signed-off-by: Mans Rullgard <mans@mansr.com>

svq3: replace unsafe pointer casting with intreadwrite macros

Signed-off-by: Mans Rullgard <mans@mansr.com>

bmp: unbreak non BMP_RGB compression for v4 and v5

Fixes CID733728 & CID733729.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

parseutils: fix const removal warning

The const qualifier is still removed although it happens inside
the strtol() function so no warning is generated.

Fixes:
libavutil/parseutils.c:110:11: warning: assignment discards qualifiers from pointer target type

Signed-off-by: Mans Rullgard <mans@mansr.com>

parseutils-test: various cleanups

- make tables static const
- remove useless use of compound literal
- break long lines
- fix a comma/semicolon typo

Signed-off-by: Mans Rullgard <mans@mansr.com>

lavfi/ebur128: fix typo in condition.

Fixes CID733727.

prepare 9_beta1 release

Move avutil tables only used in libavcodec to libavcodec.

random_seed: fix digest size

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

random_seed: fix out of array read

Fixes CID733787
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

options_table: refs option is not snow-only

yuv2rgb: fix declared array sizes, so they match actuals.

Fixes CID733789, and others
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

random_seed: Support using CryptGenRandom on windows

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

jpeglsenc: favor av_freep() for saftey over av_free() when a variable is still accessible afterwards

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