* release/0.7: (33 commits)
Update for 0.7.8
svq1dec: call avcodec_set_dimensions() after dimensions changed. Fixes NGS00148
vp3dec: Check coefficient index in vp3_dequant() Fixes NGS00145
qdm2dec: fix buffer overflow. Fixes NGS00144
h264: Fix invalid interlaced progressive MB combinations for direct mode prediction. Fixes Ticket312
mpegvideo: dont use ff_mspel_motion() for vc1 Fixes Ticket655
imgutils: Fix illegal read.
ac3probe: Detect Sonic Foundry Soft Encode AC3 as raw AC3. Our ac3 code chain can handle it fine. More ideal would be to write a demuxer that actually extracts what can be from the additional headers and uses it for whatever it can be used for.
mjpeg: support mpo Fixes stereoscopic_photo.mpo
Add a version bump and APIchanges entry for avcodec_open2 and avformat_find_stream_info.
lavf: fix multiplication overflow in avformat_find_stream_info()
lavf: fix invalid reads in avformat_find_stream_info()
lavf: add avformat_find_stream_info()
lavc: fix parentheses placement in avcodec_open2().
lavc: introduce avcodec_open2() as a replacement for avcodec_open().
rawdec: use a default sample rate if none is specified. Fixes "ffmpeg -f s16le -i /dev/zero"
rawdec: add check on sample_rate
qdm2dec: check remaining input bits in the mainloop of qdm2_fft_decode_tones() This is neccessary but likely not sufficient to prevent out of array reads.
cinepak: check strip_size
wma: Check channel number before init. Fixes Ticket240
...
Conflicts:
RELEASE
doc/APIchanges
libavcodec/avcodec.h
libavcodec/utils.c
libavcodec/version.h
libavdevice/v4l2.c
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* release/0.8: (31 commits)
svq1dec: call avcodec_set_dimensions() after dimensions changed. Fixes NGS00148
vp3dec: Check coefficient index in vp3_dequant() Fixes NGS00145
qdm2dec: fix buffer overflow. Fixes NGS00144
h264: Fix invalid interlaced progressive MB combinations for direct mode prediction. Fixes Ticket312
mpegvideo: dont use ff_mspel_motion() for vc1 Fixes Ticket655
imgutils: Fix illegal read.
ac3probe: Detect Sonic Foundry Soft Encode AC3 as raw AC3. Our ac3 code chain can handle it fine. More ideal would be to write a demuxer that actually extracts what can be from the additional headers and uses it for whatever it can be used for.
mjpeg: support mpo Fixes stereoscopic_photo.mpo
Add a version bump and APIchanges entry for avcodec_open2 and avformat_find_stream_info.
lavf: fix multiplication overflow in avformat_find_stream_info()
lavf: fix invalid reads in avformat_find_stream_info()
lavf: add avformat_find_stream_info()
lavc: fix parentheses placement in avcodec_open2().
lavc: introduce avcodec_open2() as a replacement for avcodec_open().
rawdec: use a default sample rate if none is specified. Fixes "ffmpeg -f s16le -i /dev/zero"
rawdec: add check on sample_rate
qdm2dec: check remaining input bits in the mainloop of qdm2_fft_decode_tones() This is neccessary but likely not sufficient to prevent out of array reads.
cinepak: check strip_size
wma: Check channel number before init. Fixes Ticket240
Do not try to read 16bit gray png files with alpha channel.
...
Conflicts:
libavcodec/version.h
libavformat/version.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
(cherry picked from commit 3fd0f6ed252e51ffaec7765a2637794366a513ba)
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
avfilter_draw_slice() is already called in the end_frame() callback,
this avoids multiple calls. This is done by adding a null draw_slice()
callback.
In particular fix crash occurring with -vf transpose=3,hflip, fix trac
issue #371.
(cherry picked from commit d9c23a0d5a56488b146eef17a19a9b47643be333)
avfilter_draw_slice() is already called in the end_frame() callback,
this avoids multiple calls. This is done by adding a null draw_slice()
callback.
In particular fix crash occurring with -vf transpose=3,hflip, fix trac
issue #371.
(cherry picked from commit d9c23a0d5a56488b146eef17a19a9b47643be333)
* release/0.7: (290 commits)
nuv: Fix combination of size changes and LZO compression.
av_lzo1x_decode: properly handle negative buffer length.
Do not call parse_keyframes_index with NULL stream.
update versions for 0.7 branch
Version numbers for 0.8.6
snow: emu edge support Fixes Ticket592
imc: validate channel count
imc: check for ff_fft_init() failure (cherry picked from commit 95fee70d6773fde1c34ff6422f48e5e66f37f263)
libgsmdec: check output buffer size before decoding (cherry picked from commit b03761b1309293bbf30edef767503875277b01cf)
configure: fix arch x86_32
mp3enc: avoid truncating id3v1 tags by one byte
asfdec: Check packet_replic_size earlier
cin audio: validate the channel count
binkaudio: add some buffer overread checks.
atrac1: validate number of channels (cherry picked from commit bff5b2c1ca1290ea30587ff2f76171f9e3854872)
atrac1: check output buffer size before decoding (cherry picked from commit 33684b9c12b74c0140fb91e8150263db4a48d55e)
vp3: fix oob read for negative tokens and memleaks on error. (cherry picked from commit 8370e426e42f2e4b9d14a1fb8107ecfe5163ce7f)
apedec: set s->currentframeblocks after validating nblocks
apedec: use unsigned int for 'nblocks' and make sure that it's within int range
apedec: check for data buffer realloc failure (cherry picked from commit 11ca8b2d7486e879926488404b3b79af774f0f2d)
...
Conflicts:
Changelog
Makefile
RELEASE
configure
libavcodec/error_resilience.c
libavcodec/mpegvideo.c
libavformat/matroskaenc.c
tests/ref/lavf/mxf
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* release/0.8: (185 commits)
h264: fix intra 16x16 mode check when using mbaff and constrained_intra_pred.
h264: check for invalid bit depth value.
h264: add entries for 11 and 12 bits in ff_h264_chroma_qp[][]
h264: fix the check for invalid SPS:num_ref_frames.
h264: do not let invalid values in h->ref_count on ff_h264_decode_ref_pic_list_reordering() errors.
Reject video with non multiple of 16 width/height in the 4xm decoder.
4xm decoder: fix data size for i2 frames.
4xm decoder: print some error messages in case of errors.
Check for out of bound accesses in the 4xm decoder.
Prevent block size from inreasing in the shorten decoder.
Check for out of bound reads in PTX decoder.
Clear FF_INPUT_BUFFER_PADDING_SIZE bytes at the end of the temporary buffers used in 4xm decoder.
Fix the check for missing references in ff_er_frame_end() for H264.
Prevent NULL dereference when the huffman table is invalid in the 4xm decoder.
Fix use of uninitialized memory in 4X Technologies demuxer.
h264: increase ref_poc size to 32 as it can be per field.
h264: set unused ref_counts to 0 as a precautionary meassure.
Remove Chnagelog it has nothing to do with reality
fate: fix motion pixels checksum change caused by backported bugfix
avienc: Add a limit on the number of skiped frames muxed in a row.
...
Conflicts:
Doxyfile
RELEASE
VERSION
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
The previous default value 0x0 was not good, since it is not even
valid.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 1ee20141900c98f9dc25eca121c66c3ff468c1e4)
In apply_unsharp(), when y is >= height, prevent out-of-buffer reading
from src, read from the last buffer line in src2 instead.
The check was implemented in the original unsharp libmpcodecs code and
lost in the port.
This also fixes output discrepancy between the two filters.
Signed-off-by: Anton Khirnov <anton@khirnov.net>
(cherry picked from commit 998e8519efbc772994c5ba19c0d39573998be9db)
* release/0.7: (296 commits)
Update version numbers for 0.7.5
vp6: partially propagate huffman tree building errors during coeff model parsing and fix misspelling
Check for huffman tree building error in vp6 decoder.
Release old pictures after a resolution change in vp5/6 decoder
Check for missing reference in vp5/6 decoder.
Check for invalid slices offsets in RV30/40 decoder.
Check output buffer size in nellymoser decoder.
Hack around gcc 4.6 breaking asm using call.
Hack around gcc 4.6 breaking asm using call.
Fix dxva2 decoding for some H264 samples. (cherry picked from commit bf7dc6b29d785f149f18c39db021413e08735546)
Fix dxva2 decoding for some H264 samples.
mp3demux: pass on error code on packet read.
Check for invalid slice offsets in real decoder.
rmdec: Reject invalid deinterleaving parameters
Use deinterleavers for demangling audio packets in RealMedia.
rv10: Reject slices that does not have the same type as the first one
rmdec: use the deinterleaving mode and not the codec when creating audio packets.
MAINTAINERS: add my GPG fingerprint. (cherry picked from commit 7882dc10f871bf25a848fe62a152f63814f9c7d1)
Support 3IVD in isom, produced by 3ivx DivX Doctor.
mpegpsdec: fix reading first mpegps packet (cherry picked from commit b2f230e23dd61112ac090b0c059d87b5f6bcb307)
...
Conflicts:
Changelog
Doxyfile
Makefile
RELEASE
configure
doc/general.texi
ffmpeg.c
ffplay.c
libavcodec/dxva2_h264.c
libavcodec/h264.c
libavcodec/h264_loopfilter.c
libavcodec/h264idct_template.c
libavcodec/kgv1dec.c
libavcodec/mpegvideo.c
libavcodec/tableprint.h
libavcodec/vp3.c
libavdevice/alsa-audio.h
libavformat/gxf.c
libavformat/mpegts.c
libavformat/segafilm.c
libavformat/utils.c
libavutil/dict.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* release/0.8: (154 commits)
vp6: partially propagate huffman tree building errors during coeff model parsing and fix misspelling
Check for huffman tree building error in vp6 decoder.
Release old pictures after a resolution change in vp5/6 decoder
Check for missing reference in vp5/6 decoder.
Check for invalid slices offsets in RV30/40 decoder.
Check output buffer size in nellymoser decoder.
Hack around gcc 4.6 breaking asm using call.
Fix dxva2 decoding for some H264 samples.
mp3demux: pass on error code on packet read.
Check for invalid slice offsets in real decoder.
rmdec: Reject invalid deinterleaving parameters
Use deinterleavers for demangling audio packets in RealMedia.
rv10: Reject slices that does not have the same type as the first one
rmdec: use the deinterleaving mode and not the codec when creating audio packets.
MAINTAINERS: add my GPG fingerprint. (cherry picked from commit 7882dc10f871bf25a848fe62a152f63814f9c7d1)
Support 3IVD in isom, produced by 3ivx DivX Doctor.
mpegpsdec: fix reading first mpegps packet (cherry picked from commit b2f230e23dd61112ac090b0c059d87b5f6bcb307)
Avoid NULL dereference on corrupted bitstream with real decoder.
Reject slices that does not have the same type than the first one in RV10/RV20 decoder.
check all svq3_get_ue_golomb() returns.
...
Conflicts:
Doxyfile
RELEASE
VERSION
libavcodec/rv34.c
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* release/0.8: (21 commits)
rtp: Fix integer underflow that could allow remote code execution.
cavsdec: avoid possible crash with crafted input
vf_scale: apply the same transform to the aspect during init that is applied per frame
Fix memory corruption in case of memory allocation failure in av_probe_input_buffer()
Make all option parsing functions match the function pointer type through which they are called.
mjpegdec; even better RSTn skiping Fixes Ticket426
jpegdec: better rst skiping Fixes Ticket426
mpeg4: fix another packed divx issue. Fixes getting_stuck.avi
mpeg4: adjust dummy frame threashold for packed divx. Fixes Ticket427
configure: add missing CFLAGS to fix building on the HURD
cavs: fix some crashes with invalid bitstreams
jpegdec: actually search for and parse RSTn
Fix compilation with --disable-avfilter. (cherry picked from commit 67a8251690a17f05630eb6f45a73db0f0e806c72)
libavfilter: fix --enable-small
0.8.2
cavs: fix oCERT #2011-002 FFmpeg/libavcodec insufficient boundary check
Fix possible crash when decoding mpeg streams.
Bink: clip AC coefficients during dequantization.
ffmpeg: fix passlogfile regression
Fix several security issues in matroskadec.c (MSVR-11-0080).
...
Conflicts:
Doxyfile
RELEASE
VERSION
Merged-by: Michael Niedermayer <michaelni@gmx.at>
* master: (780 commits)
ffmpeg: Fix doxygen comments for MetadataMap
filters.texi: fix wrong references in the "Filtergraph syntax" section
yadif: correct documentation on the parity parameter
mpegvideo.h: remove the 1 line difference to qatar
applehttp: fix variant discard logic
Fix possible crash when decoding mpeg streams.
h263dec: Fix asserts broken by the elimination of FF_COMMON_FRAME.
avidec: skip seek pos adjustment for non interleaved files. Fixes Ticket327
lsws: remove deprecated and unused stuff after the 0->1 major bump
cosmetics: remove some stray comments from AVCodec declarations
cosmetics: fix indentation/alignment in AVCodec declarations
Abort if command offset decreases, avoids potential endless loop.
Warn when falling back to unreliable UMF fps tag.
Detect NI-avi at playtime like mplayer. Fixes Ticket333
avidec: Fix XAN DPCM demuxing.
Fix a possible miscompilation of cabac with old (broken) compilers.
Fix -loop_input.
Set bits_per_coded_sample when encoding ADPCM.
vf_boxblur: call avfilter_draw_slice() at the end of draw_slice()
vf_boxblur: fix out-of-buffer access when h > w
...
Conflicts:
ffmpeg.c
libavcodec/avcodec.h
libavcodec/opt.h
libavcodec/version.h
libavdevice/avdevice.h
libavfilter/avfilter.h
libavformat/avformat.h
libavformat/aviobuf.c
libavformat/rtsp.c
libavformat/udp.c
libavformat/utils.c
libavformat/version.h
libavutil/avutil.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Previously, "dar" and "a" were documented like "input display aspect
ratio", but their actual value was "in_w/in_h".
In order to avoid to break scripts which rely on the "a" variable, the
patch keeps the same semantics but fixes the corresponding docs, and
fixes the semantics of the recently added "dar" variable, which
now correctly expresses the input Display Aspect Ratio value as
"(inw_w/in_h)*sar".
avfilter_draw_slice() is already called in the end_frame() callback,
this avoids multiple calls. This is done by adding a null draw_slice()
callback.
In particular fix crash occurring with -vf transpose=3,hflip, fix trac
issue #371.
* release/0.8: (82 commits)
Fix version numbers
rtp: disable udp fifos, the rtp code cannot work with the fifos in its current form as rtp bypasses the public API.
udp: allow fifo size to be tuned seperately
riff: Add mpgv MPEG-2 fourcc
Update Changelog
matroskadec: fix integer underflow if header length < probe length.
ffmpeg: fix operation with --disable-avfilter
vf_libopencv: replace opencv/cxtypes.h #include by opencv/cxcore.h
build: Create mlib optimization directories during out-of-tree builds.
changelog: misc typo and wording fixes (cherry picked from commit b047941d7da470ba0dcedb1fd0aa828075265ffc)
doc: Remove outdated comments about gcc 2.95 and gcc 3.3 support. (cherry picked from commit 5ccbf80963c1cc54aed97b1c81b1657ab91baf6a)
matroskadec: matroska_read_seek after after EBML_STOP leads to failure.
Update RELEASE file
update Changelog
mt: proper locking around release_buffer calls.
vp8/mt: flush worker thread, not application thread context, on seek.
docs: Mention the upstream bugzilla url about the dlltool vs MSVC issue
docs: Use proper markup for a literal command line option
docs: Don't recommend adding --enable-memalign-hack
docs: Remove needless configure options
...
Conflicts:
VERSION
libavcodec/opt.h
libavformat/utils.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
cxtypes.h works with version 2.1 and older, cxcore.h works with 2.2 and older.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
(cherry picked from commit 9bc8bcddbd4fc394e2268e9849dcbf3bad6de980)
The licence was changed from GPL to LGPL with explicit approval from
the original author.
See thread:
Subject: [FFmpeg-devel] [PATCH] lavfi: add rgbtestsrc source,
ported from MPlayer libmpcodecs
Date: Mon, 11 Jul 2011 16:32:41 +0200
In particular creates a special test_init function, which calls a
common init, and introduces a fill_picture_fn which points to the
filling function.
This is required by the pending patch for adding the rgbtestsrc
source.
* qatar/master:
APIchanges: fill in missing hashes and dates.
Add an APIChanges entry and bump minor versions for recent changes.
ffmpeg: print the low bitrate warning after the codec is openend.
doxygen: Move function documentation into the macro generating the function.
doxygen: Make sure parameter names match between .c and .h files.
h264: move fill_decode_neighbors()/fill_decode_caches() to h264_mvpred.h
H.264: Add more x86 assembly for 10-bit H.264 predict functions
lavf: fix invalid reads in avformat_find_stream_info()
cmdutils: replace opt_default with opt_default2() and remove set_context_opts
ffmpeg: use new avcodec_open2 and avformat_find_stream_info API.
ffplay: use new avcodec_open2 and avformat_find_stream_info API.
cmdutils: store all codec options in one dict instead of video/audio/sub
ffmpeg: check experimental flag after codec is opened.
ffmpeg: do not set GLOBAL_HEADER flag in the options context
Conflicts:
cmdutils.c
doc/APIchanges
ffmpeg.c
ffplay.c
libavcodec/version.h
libavformat/version.h
libswscale/swscale_unscaled.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>
With the following additions:
* support to gray format
* support to yuva420p format
* parametric luma/chroma/alpha radius
* consistency check on the radius values, avoid crashes with invalid values
* commit '142e76f1055de5dde44696e71a5f63f2cb11dedf':
swscale: fix crash with dithering due incorrect offset calculation.
matroskadec: fix stupid typo (!= -> ==)
build: remove duplicates from order-only directory prerequisite list
build: rework rules for things in the tools dir
configure: fix --cpu=host with gcc 4.6
ARM: use const macro to define constant data in asm
bitdepth: simplify FUNC/FUNCC macros
dsputil: remove ff_emulated_edge_mc macro used in one place
9/10-bit: simplify clipping macros
matroskadec: reindent
matroskadec: defer parsing of cues element until we seek.
lavc: add support for codec-specific defaults.
lavc: make avcodec_alloc_context3 officially public.
lavc: remove a half-working attempt at different defaults for audio/video codecs.
ac3dec: add a drc_scale private option
lavf: add avformat_find_stream_info()
lavc: introduce avcodec_open2() as a replacement for avcodec_open().
Conflicts:
Makefile
libavcodec/utils.c
libavformat/avformat.h
libswscale/swscale_internal.h
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Declaring tools associated with each library in their respective
makefiles allows these tools to easily depend on the correct
prerequisites and link against the libs they need.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Since avfilter_graph_parse() creates the "[in]" inout for the first
unlabelled input pad, it is expected that it will create an "[out]"
inout for last unlabelled output pad, even in the case where it cannot
find any open input pad with that name.
This change removes the check on the existence of an open input pad
named "out", so it simplifies the checked condition while implementing
a more intuitive behavior.
libavfilter/vf_drawtext.c:427:22: warning: variable 'dst_pixel' set but not used
libavfilter/vf_drawtext.c:463:22: warning: variable 'dst_pixel' set but not used
Actually I gave permission for LGPL long ago, but the original import
failed to update the license header.
(cherry picked from commit 082768f0b189b1706fdcd15b42dcca5fd0822315)
