platform/upstream/libav.git
12 years agoAVOptions: cosmetics, move and rename static av_set_number().
Anton Khirnov [Tue, 23 Aug 2011 11:03:42 +0000 (13:03 +0200)]
AVOptions: cosmetics, move and rename static av_set_number().

12 years agoAVOptions: split av_set_string3 into opt type-specific functions
Anton Khirnov [Tue, 23 Aug 2011 08:23:57 +0000 (10:23 +0200)]
AVOptions: split av_set_string3 into opt type-specific functions

Also stop attempting to honor the alloc parameter, as things break
horribly in that case.
It will be removed in upcoming successor to av_set_string3.

12 years agoavidec: fix signed overflow in avi_sync()
Mans Rullgard [Sat, 8 Oct 2011 00:51:37 +0000 (01:51 +0100)]
avidec: fix signed overflow in avi_sync()

Keeping byte values read from the file as unsigned is consistent
with how they are subsequently used and avoids an undefined left
shift by 24 when bit 7 is set.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agomxfdec: Fix some buffer overreads caused by the misuse of AVPacket related functions.
Alex Converse [Sat, 8 Oct 2011 01:41:06 +0000 (18:41 -0700)]
mxfdec: Fix some buffer overreads caused by the misuse of AVPacket related functions.

12 years agodnxhddec: handle cid change in bitstream.
Baptiste Coudurier [Sun, 26 Jun 2011 21:08:22 +0000 (14:08 -0700)]
dnxhddec: handle cid change in bitstream.

12 years agovc1: K&R reformatting cosmetics
Kostya Shishkov [Mon, 10 Oct 2011 14:04:58 +0000 (16:04 +0200)]
vc1: K&R reformatting cosmetics

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agora288: log an error message when output buffer is too small.
Justin Ruggles [Wed, 14 Sep 2011 18:46:50 +0000 (14:46 -0400)]
ra288: log an error message when output buffer is too small.

also return AVERROR(EINVAL) instead of -1.

12 years agora288: use a more descriptive calculation for output data size
Justin Ruggles [Wed, 14 Sep 2011 18:46:02 +0000 (14:46 -0400)]
ra288: use a more descriptive calculation for output data size

12 years agora144: use macro constants to make the code more understandable.
Justin Ruggles [Wed, 14 Sep 2011 18:42:31 +0000 (14:42 -0400)]
ra144: use macro constants to make the code more understandable.

12 years agora144dec: log an error message when output buffer is too small.
Justin Ruggles [Wed, 14 Sep 2011 18:24:59 +0000 (14:24 -0400)]
ra144dec: log an error message when output buffer is too small.

also return AVERROR(EINVAL) instead of -1.

12 years agora144dec: use a more descriptive calculation for output data size
Justin Ruggles [Wed, 14 Sep 2011 18:24:02 +0000 (14:24 -0400)]
ra144dec: use a more descriptive calculation for output data size

12 years agolavc: Micro version bump for VC1 interlaced decoding support
Mashiat Sarker Shakkhar [Thu, 6 Oct 2011 19:00:27 +0000 (00:00 +0500)]
lavc: Micro version bump for VC1 interlaced decoding support

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agovc1dec: interlaced stream decoding support 3/3
Mashiat Sarker Shakkhar [Thu, 6 Oct 2011 19:00:26 +0000 (00:00 +0500)]
vc1dec: interlaced stream decoding support 3/3

Cosmetics: break some lines and reformat TODOs

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agovc1dec: interlaced stream decoding support 2/3
Mashiat Sarker Shakkhar [Thu, 6 Oct 2011 19:00:25 +0000 (00:00 +0500)]
vc1dec: interlaced stream decoding support 2/3

Cosmetics: reformat some data tables in vc1data.c

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agovc1dec: interlaced stream decoding support 1/3
Mashiat Sarker Shakkhar [Thu, 6 Oct 2011 19:00:24 +0000 (00:00 +0500)]
vc1dec: interlaced stream decoding support 1/3

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agovaapi: Fix VC-1 decoding (reconstruct bitstream TTFRM correctly).
Gwenole Beauchesne [Mon, 10 Oct 2011 11:51:41 +0000 (13:51 +0200)]
vaapi: Fix VC-1 decoding (reconstruct bitstream TTFRM correctly).

Signed-off-by: Diego Biurrun <diego@biurrun.de>
12 years agoconfigure: Add -Wstrict-prototypes to CFLAGS if available.
Diego Biurrun [Mon, 10 Oct 2011 08:31:31 +0000 (10:31 +0200)]
configure: Add -Wstrict-prototypes to CFLAGS if available.

12 years agoavidec: simplify size parameter in memset()
Janne Grunau [Mon, 10 Oct 2011 11:28:18 +0000 (13:28 +0200)]
avidec: simplify size parameter in memset()

12 years agompeg4dec: use unsigned type for startcode in ff_mpeg4_decode_picture_header
Mans Rullgard [Sun, 9 Oct 2011 22:21:01 +0000 (23:21 +0100)]
mpeg4dec: use unsigned type for startcode in ff_mpeg4_decode_picture_header

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agompeg124: use sign_extend() function
Mans Rullgard [Sun, 9 Oct 2011 22:16:07 +0000 (23:16 +0100)]
mpeg124: use sign_extend() function

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoac3dec: use get_sbits() instead of manually sign-extending
Mans Rullgard [Sun, 9 Oct 2011 22:15:17 +0000 (23:15 +0100)]
ac3dec: use get_sbits() instead of manually sign-extending

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years ago4xm: fix signed overflow
Mans Rullgard [Sun, 9 Oct 2011 19:38:01 +0000 (20:38 +0100)]
4xm: fix signed overflow

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agowmavoice: fix a signed overflow
Mans Rullgard [Sun, 9 Oct 2011 19:46:22 +0000 (20:46 +0100)]
wmavoice: fix a signed overflow

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agompegvideo_enc: fix a signed overflow
Mans Rullgard [Sun, 9 Oct 2011 19:32:58 +0000 (20:32 +0100)]
mpegvideo_enc: fix a signed overflow

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agocrc: fix signed overflow
Mans Rullgard [Sat, 8 Oct 2011 00:59:51 +0000 (01:59 +0100)]
crc: fix signed overflow

This fixes a signed overflow from i << 24 when i == 255 by
making i unsigned.  The result of the shift is already
assigned to an variable of unsigned type.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agofate: run avconv with -nostats flag
Mans Rullgard [Fri, 7 Oct 2011 13:13:02 +0000 (14:13 +0100)]
fate: run avconv with -nostats flag

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoavtools: add -v as alias for -loglevel
Mans Rullgard [Fri, 7 Oct 2011 12:29:03 +0000 (13:29 +0100)]
avtools: add -v as alias for -loglevel

This saves a bit of typing.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoavconv: always print stats with AV_LOG_INFO
Anton Khirnov [Sun, 9 Oct 2011 14:32:54 +0000 (16:32 +0200)]
avconv: always print stats with AV_LOG_INFO

It's now possible to disable printing stats during encoding with
-nostats, so there's no reason to print them differently depending on
whether it's last report or not.

12 years agodoc/avconv: add forgotten output/per-stream info to -filter
Anton Khirnov [Sun, 9 Oct 2011 13:57:54 +0000 (15:57 +0200)]
doc/avconv: add forgotten output/per-stream info to -filter

12 years agoavconv: add -stats option to enable/disable printing encoding progress
Anton Khirnov [Sun, 9 Oct 2011 13:57:30 +0000 (15:57 +0200)]
avconv: add -stats option to enable/disable printing encoding progress

12 years agoavconv: in output_packet() don't set decoded_data_size for video.
Anton Khirnov [Tue, 27 Sep 2011 18:00:28 +0000 (20:00 +0200)]
avconv: in output_packet() don't set decoded_data_size for video.

It's only used for audio.

12 years agoavconv: remove pointless always true condition
Anton Khirnov [Tue, 27 Sep 2011 06:21:48 +0000 (08:21 +0200)]
avconv: remove pointless always true condition

12 years agoavconv: factorize common code in transcode_init()
Anton Khirnov [Tue, 27 Sep 2011 06:16:26 +0000 (08:16 +0200)]
avconv: factorize common code in transcode_init()

12 years agozmbv: remove memcpy() of decoded frame
Mans Rullgard [Sun, 9 Oct 2011 12:21:10 +0000 (13:21 +0100)]
zmbv: remove memcpy() of decoded frame

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agompeg12enc: use sign_extend() function
Mans Rullgard [Sat, 8 Oct 2011 11:56:54 +0000 (12:56 +0100)]
mpeg12enc: use sign_extend() function

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoh264pred: use unsigned types for pixel values, fix signed overflows
Mans Rullgard [Sat, 8 Oct 2011 20:22:06 +0000 (21:22 +0100)]
h264pred: use unsigned types for pixel values, fix signed overflows

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoh264: fix signed overflows in x*0x01010101 expressions
Mans Rullgard [Sat, 8 Oct 2011 20:19:51 +0000 (21:19 +0100)]
h264: fix signed overflows in x*0x01010101 expressions

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoh264pred: remove unused variables
Mans Rullgard [Sat, 8 Oct 2011 19:32:09 +0000 (20:32 +0100)]
h264pred: remove unused variables

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agolavf: fix signed overflow in avformat_find_stream_info()
Mans Rullgard [Sat, 8 Oct 2011 01:06:26 +0000 (02:06 +0100)]
lavf: fix signed overflow in avformat_find_stream_info()

On the first iteration through this code, last_dts is always
INT64_MIN (AV_NOPTS_VALUE) and the subtraction overflows in
an invalid manner.  Although the result is only used if the
input values are valid, performing the subtraction is still
not allowed in a strict environment.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agovp8: fix signed overflows
Mans Rullgard [Sat, 8 Oct 2011 14:03:51 +0000 (15:03 +0100)]
vp8: fix signed overflows

In addition to avoiding undefined behaviour, an unsigned type
makes more sense for packing multiple 8-bit values.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agomotion_est: fix some signed overflows
Mans Rullgard [Sat, 8 Oct 2011 12:52:44 +0000 (13:52 +0100)]
motion_est: fix some signed overflows

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agodca: fix signed overflow in shift
Mans Rullgard [Sat, 8 Oct 2011 12:49:42 +0000 (13:49 +0100)]
dca: fix signed overflow in shift

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agoaacdec: fix undefined shifts
Mans Rullgard [Sat, 8 Oct 2011 12:41:23 +0000 (13:41 +0100)]
aacdec: fix undefined shifts

Since nnz can be zero, this is needed to avoid a shift by 32.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agobink: Check for various out of bound writes
Laurent Aimar [Tue, 27 Sep 2011 12:16:41 +0000 (12:16 +0000)]
bink: Check for various out of bound writes

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agobink: Check for out of bound writes when building tree
Laurent Aimar [Tue, 27 Sep 2011 12:16:41 +0000 (12:16 +0000)]
bink: Check for out of bound writes when building tree

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoput_bits: fix invalid shift by 32 in flush_put_bits()
Mans Rullgard [Sat, 8 Oct 2011 01:09:42 +0000 (02:09 +0100)]
put_bits: fix invalid shift by 32 in flush_put_bits()

If flush_put_bits() is called when the 32-bit buffer is empty,
e.g. after writing a multiple of 32 bits, and invalid shift by
32 is performed.  Since flush_put_bits() is called infrequently,
this additional check should have negligible performance impact.

Signed-off-by: Mans Rullgard <mans@mansr.com>
12 years agompegps: Use av_get_packet() instead of poorly emulating it.
Alex Converse [Sat, 8 Oct 2011 00:02:36 +0000 (17:02 -0700)]
mpegps: Use av_get_packet() instead of poorly emulating it.

12 years agomotionpixels: decode only the 111 complete frames for fate
Janne Grunau [Fri, 7 Oct 2011 16:08:55 +0000 (18:08 +0200)]
motionpixels: decode only the 111 complete frames for fate

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agompc8: Check out of bound bands limit
Laurent Aimar [Sun, 25 Sep 2011 20:06:20 +0000 (20:06 +0000)]
mpc8: Check out of bound bands limit

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoxan: Prevent NULL dereference with missing palette
Laurent Aimar [Thu, 29 Sep 2011 03:12:07 +0000 (03:12 +0000)]
xan: Prevent NULL dereference with missing palette

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoxan: Check for out of bound reads in xan_huffman_decode()
Laurent Aimar [Thu, 29 Sep 2011 20:38:01 +0000 (20:38 +0000)]
xan: Check for out of bound reads in xan_huffman_decode()

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoxan: Fixed out of bound accesses in xan_unpack()
Laurent Aimar [Thu, 29 Sep 2011 03:12:07 +0000 (03:12 +0000)]
xan: Fixed out of bound accesses in xan_unpack()

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agomotionpixels: Prevent calling init_vlc() with invalid parameters
Laurent Aimar [Thu, 29 Sep 2011 03:12:07 +0000 (03:12 +0000)]
motionpixels: Prevent calling init_vlc() with invalid parameters

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoshorten: Fix out of bound writes in fix_bitshift()
Laurent Aimar [Fri, 30 Sep 2011 01:26:22 +0000 (01:26 +0000)]
shorten: Fix out of bound writes in fix_bitshift()

The data pointers s->decoded[*] already take into account s->nwrap.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agodsicinav: Check for out of bounds writes
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
dsicinav: Check for out of bounds writes

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agotiertexseqv: Check for out of bound reads
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
tiertexseqv: Check for out of bound reads

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoquickdraw: Check for out of bound reads
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
quickdraw: Check for out of bound reads

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agodsicinav: Check for out of bounds reads
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
dsicinav: Check for out of bounds reads

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agomotionpixels: Fix the size of workspace buffers
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
motionpixels: Fix the size of workspace buffers

Some buffers must be mod 4 in width and/or height.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agomotionpixels: Clear FF_INPUT_BUFFER_PADDING_SIZE bytes at the end of the temporary...
Laurent Aimar [Thu, 29 Sep 2011 23:13:35 +0000 (23:13 +0000)]
motionpixels: Clear FF_INPUT_BUFFER_PADDING_SIZE bytes at the end of the temporary buffer

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agowmavoice: Check for corrupted extra data
Laurent Aimar [Tue, 27 Sep 2011 22:15:31 +0000 (22:15 +0000)]
wmavoice: Check for corrupted extra data

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agowmavoice: Check for out of bound writes
Laurent Aimar [Tue, 27 Sep 2011 22:15:31 +0000 (22:15 +0000)]
wmavoice: Check for out of bound writes

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoxan: Prevent NULL dereferences with missing reference frame
Laurent Aimar [Tue, 27 Sep 2011 22:15:32 +0000 (22:15 +0000)]
xan: Prevent NULL dereferences with missing reference frame

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agobink: Prevent NULL dereferences with missing reference frame
Laurent Aimar [Tue, 27 Sep 2011 22:15:31 +0000 (22:15 +0000)]
bink: Prevent NULL dereferences with missing reference frame

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agowavpack: Reset internal state on corrupted blocks
Laurent Aimar [Mon, 26 Sep 2011 22:18:29 +0000 (22:18 +0000)]
wavpack: Reset internal state on corrupted blocks

wavpack_decode_block() supposes that it is called back with the exact
same buffer unless it has returned with an error. With multi-channels
files, wavpack_decode_frame() was breaking this assumption.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agowmapro: Validate the number of audio channels before using it
Laurent Aimar [Mon, 26 Sep 2011 22:18:29 +0000 (22:18 +0000)]
wmapro: Validate the number of audio channels before using it

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agompc8: Fix return value on EOF
Laurent Aimar [Sun, 25 Sep 2011 20:06:19 +0000 (20:06 +0000)]
mpc8: Fix return value on EOF

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoshorten: Prevent block size from increasing
Laurent Aimar [Sun, 2 Oct 2011 00:48:12 +0000 (00:48 +0000)]
shorten: Prevent block size from increasing

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoxan: Prevent out of bound accesses
Laurent Aimar [Tue, 27 Sep 2011 22:15:33 +0000 (22:15 +0000)]
xan: Prevent out of bound accesses

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoH264: Only wait before triggering ff_thread_setup_complete() until the next slice...
Michael Niedermayer [Mon, 3 Oct 2011 22:14:48 +0000 (00:14 +0200)]
H264: Only wait before triggering ff_thread_setup_complete() until the next slice that contains a start-of-field/frame macroblock

This allows concurrent decoding of the last field/frame, rather than
only the last slice, of data packets with multiple NAL units packed
together.

This will fix the slowdown reported in e.g. bug 52.

Signed-off-by: Anton Khirnov <anton@khirnov.net>
12 years agovp6: Reset the internal state when aborting key frames header parsing
Laurent Aimar [Fri, 23 Sep 2011 20:36:11 +0000 (22:36 +0200)]
vp6: Reset the internal state when aborting key frames header parsing

It prevents leaving the state only half initialized.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agovp56: Release old pictures after a resolution changes
Laurent Aimar [Wed, 21 Sep 2011 18:46:33 +0000 (20:46 +0200)]
vp56: Release old pictures after a resolution changes

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agovp6: Check for huffman tree build errors
Laurent Aimar [Wed, 21 Sep 2011 18:46:32 +0000 (20:46 +0200)]
vp6: Check for huffman tree build errors

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agovp56: Check for missing reference frame data
Laurent Aimar [Wed, 21 Sep 2011 18:46:30 +0000 (20:46 +0200)]
vp56: Check for missing reference frame data

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agocinepak: Fix invalid read access on extra data
Laurent Aimar [Sun, 11 Sep 2011 17:17:43 +0000 (19:17 +0200)]
cinepak: Fix invalid read access on extra data

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agovmd: fix segfaults on corruped streams
Laurent Aimar [Sun, 11 Sep 2011 17:17:45 +0000 (19:17 +0200)]
vmd: fix segfaults on corruped streams

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agocook: Fix js_vlc_bits value validation for joint stereo
Laurent Aimar [Sat, 17 Sep 2011 14:56:35 +0000 (16:56 +0200)]
cook: Fix js_vlc_bits value validation for joint stereo

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agosegafilm: Check for memory allocation failures in segafilm demuxer.
Laurent Aimar [Thu, 6 Oct 2011 20:53:41 +0000 (22:53 +0200)]
segafilm: Check for memory allocation failures in segafilm demuxer.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agosegafilm: Fix potential division by 0 on corrupted streams in the demuxer
Laurent Aimar [Mon, 12 Sep 2011 19:09:57 +0000 (21:09 +0200)]
segafilm: Fix potential division by 0 on corrupted streams in the demuxer

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoFixed segfault on corrupted sega streams in the demuxer.
Laurent Aimar [Sun, 11 Sep 2011 17:17:41 +0000 (19:17 +0200)]
Fixed segfault on corrupted sega streams in the demuxer.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agoFixed deference of NULL pointer in motionpixels decoder.
Laurent Aimar [Sat, 10 Sep 2011 11:28:13 +0000 (13:28 +0200)]
Fixed deference of NULL pointer in motionpixels decoder.

Some of the arguments given to init_vlc() come from the stream
and can be corrupted.

Signed-off-by: Janne Grunau <janne-libav@jannau.net>
12 years agolibx264: support 9- and 10-bit output.
Anton Khirnov [Wed, 5 Oct 2011 19:09:51 +0000 (21:09 +0200)]
libx264: support 9- and 10-bit output.

12 years agoh264: correct implicit_weight for field-interlaced pictures.
Ronald S. Bultje [Mon, 3 Oct 2011 15:41:51 +0000 (08:41 -0700)]
h264: correct implicit_weight for field-interlaced pictures.

12 years agompegvideo: set correct offset for edge emulation buffer.
Ronald S. Bultje [Mon, 3 Oct 2011 15:38:03 +0000 (08:38 -0700)]
mpegvideo: set correct offset for edge emulation buffer.

Using the old code, half of it was unused and the other half was too
small for e.g. >8bpp interlaced data, causing random buffer overruns.

12 years agompegvideo: fix position of bottom edge.
Ronald S. Bultje [Mon, 3 Oct 2011 14:37:24 +0000 (07:37 -0700)]
mpegvideo: fix position of bottom edge.

It was wrong in colorspaces where horizontal and vertical chroma
subsampling are not the same, e.g. 422.

12 years agoFix 'heigth' vs. 'height' typos.
Diego Biurrun [Wed, 5 Oct 2011 09:12:01 +0000 (11:12 +0200)]
Fix 'heigth' vs. 'height' typos.

12 years agolavc/lavf: use unique private classes.
Anton Khirnov [Mon, 3 Oct 2011 17:14:03 +0000 (19:14 +0200)]
lavc/lavf: use unique private classes.

This is needed by the new AVOptions API.

12 years agolavc: use designated initializers for av_codec_context_class
Anton Khirnov [Mon, 3 Oct 2011 12:13:17 +0000 (14:13 +0200)]
lavc: use designated initializers for av_codec_context_class

12 years agoqcelpdec: cosmetics: do not add line break before opening bracket in 'for',
Justin Ruggles [Wed, 14 Sep 2011 17:46:48 +0000 (13:46 -0400)]
qcelpdec: cosmetics: do not add line break before opening bracket in 'for',
'while', 'if/else', and 'switch' statements.

also fixes some spacing, but only if already changing a line

12 years agoqcelp: check output buffer size before decoding
Justin Ruggles [Wed, 14 Sep 2011 17:38:07 +0000 (13:38 -0400)]
qcelp: check output buffer size before decoding

12 years agoqcelpdec: fix the return value of qcelp_decode_frame().
Chris Rankin [Wed, 7 Sep 2011 09:17:30 +0000 (10:17 +0100)]
qcelpdec: fix the return value of qcelp_decode_frame().

Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
12 years agosipr: fix the output data size check and only calculate it once.
Justin Ruggles [Tue, 20 Sep 2011 19:27:44 +0000 (15:27 -0400)]
sipr: fix the output data size check and only calculate it once.

12 years agoSynchronize various 4CCs and codec tags from FFmpeg.
Diego Biurrun [Fri, 30 Sep 2011 15:28:28 +0000 (17:28 +0200)]
Synchronize various 4CCs and codec tags from FFmpeg.

12 years agoqdm2: check output buffer size before decoding
Justin Ruggles [Wed, 14 Sep 2011 17:57:04 +0000 (13:57 -0400)]
qdm2: check output buffer size before decoding

12 years agoFix out of bound reads in the QDM2 decoder.
Laurent Aimar [Fri, 30 Sep 2011 22:45:04 +0000 (00:45 +0200)]
Fix out of bound reads in the QDM2 decoder.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
12 years agoCheck for out of bound writes in the QDM2 decoder.
Laurent Aimar [Fri, 30 Sep 2011 22:45:05 +0000 (00:45 +0200)]
Check for out of bound writes in the QDM2 decoder.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Justin Ruggles <justin.ruggles@gmail.com>
12 years agoogg/celt: do not set sample_fmt in the demuxer
Justin Ruggles [Mon, 26 Sep 2011 19:06:33 +0000 (15:06 -0400)]
ogg/celt: do not set sample_fmt in the demuxer

12 years agoid3v2: remove pointless casts
Anton Khirnov [Sun, 2 Oct 2011 07:14:30 +0000 (09:14 +0200)]
id3v2: remove pointless casts

12 years agoid3v2: read TXXX frames with two calls to decode_str() instead of one.
Anton Khirnov [Sun, 2 Oct 2011 07:06:34 +0000 (09:06 +0200)]
id3v2: read TXXX frames with two calls to decode_str() instead of one.

Read the key in the first, value in the second.

This allows to avoid pointless strdups and simplify decode_str() by
dropping two of its parameters.

12 years agoid3v2: don't discard the whole tag when encountering empty frames.
Anton Khirnov [Sun, 2 Oct 2011 05:16:09 +0000 (07:16 +0200)]
id3v2: don't discard the whole tag when encountering empty frames.

While they're technically invalid, it's better to skip them and try to
read the rest of the tag.