[Ffmpeg-devel-irc] ffmpeg-devel.log.20120319

[burek]

[00:22] <CIA-17> ffmpeg: 03Derek Buitenhuis 07master * r6cabb679d6 10ffmpeg/ (9 files in 4 dirs): 
[00:22] <CIA-17> ffmpeg: ZeroCodec Decoder
[00:22] <CIA-17> ffmpeg: An obscure Japanese lossless video codec, originally intended
[00:22] <CIA-17> ffmpeg: for use with a remote desktop application.
[00:22] <CIA-17> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis at gmail.com>
[00:22] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:30] <michaelni> Daemon404, where can i find a sample zec0 file ?
[00:30] <Daemon404> i can make one
[00:30] <Daemon404> i was going to add a fate test
[00:30] <Daemon404> once it was pushed
[04:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * re8f2c0ca2e 10ffmpeg/libavcodec/h264.c: 
[04:18] <CIA-17> ffmpeg: h264: print the pps_id when its invalid.
[04:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:30] <CIA-17> ffmpeg: 03Joakim Plate 07release/0.8 * r4a22876675 10ffmpeg/libavfilter/vf_yadif.c: 
[05:30] <CIA-17> ffmpeg: vf_yadif: copy buffer properties like aspect for second frame as well
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: (cherry picked from commit 5feb67f8a1a17a4dd3cec0aa80ef0dc543fc7673)
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: Signed-off-by: Reinhard Tartler <siretart at tauware.de>
[05:30] <CIA-17> ffmpeg: 03Stefano Sabatini 07release/0.8 * r5c9ca599a7 10ffmpeg/ (doc/filters.texi libavfilter/vf_yadif.c): (log message trimmed)
[05:30] <CIA-17> ffmpeg: vf_yadif: correct documentation on the parity parameter
[05:30] <CIA-17> ffmpeg: 0 is top-field-first, 1 is bottom-field-first, not the other way
[05:30] <CIA-17> ffmpeg: around.
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: (cherry picked from commit 4703a7b50b098a53ec2f806bd41a00fd87ea9d8c)
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: 03Michael Niedermayer 07release/0.8 * r7f62cf120b 10ffmpeg/libavfilter/vf_scale.c: 
[05:30] <CIA-17> ffmpeg: vf_scale: apply the same transform to the aspect during init that is applied per frame
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: (cherry picked from commit d33e0c6bc819048b05c168d304fba7bdd75a80e1)
[05:30] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:30] <CIA-17> ffmpeg: Signed-off-by: Reinhard Tartler <siretart at tauware.de>
[05:30] <CIA-17> ffmpeg: 03Stefano Sabatini 07release/0.8 * r734a9bb05f 10ffmpeg/libavfilter/vf_unsharp.c: (log message trimmed)
[05:30] <CIA-17> ffmpeg: vf_unsharp: fix out-of-buffer read
[05:30] <CIA-17> ffmpeg: In apply_unsharp(), when y is >= height, prevent out-of-buffer reading
[05:30] <CIA-17> ffmpeg: from src, read from the last buffer line in src2 instead.
[05:30] <CIA-17> ffmpeg: The check was implemented in the original unsharp libmpcodecs code and
[05:30] <CIA-17> ffmpeg: lost in the port.
[05:30] <CIA-17> ffmpeg: This also fixes output discrepancy between the two filters.
[05:30] <CIA-17> ffmpeg: 03Stefano Sabatini 07release/0.8 * r958e0f705d 10ffmpeg/libavfilter/formats.c: (log message trimmed)
[05:30] <CIA-17> ffmpeg: lavfi: fix realloc size computation in avfilter_add_format()
[05:30] <CIA-17> ffmpeg: Replace sizeof((*avff)->formats)
[05:30] <CIA-17> ffmpeg: with sizeof(*(*avff)->formats)
[05:30] <CIA-17> ffmpeg: as the size of the array element is given by the pointed element
[05:31] <CIA-17> (115 lines omitted)
[05:31] <CIA-17> ffmpeg: 03Justin Ruggles 07release/0.8 * rfc0e151cdc 10ffmpeg/libavcodec/mpc8.c: 
[05:31] <CIA-17> ffmpeg: mpc8: check output buffer size before decoding
[05:31] <CIA-17> ffmpeg: (cherry picked from commit 5674d4b0a35a34b75e3533a8580e0b5a0a8895a7)
[05:31] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:31] <CIA-17> ffmpeg: 03Ronald S. Bultje 07release/0.8 * rbb7fd94eeb 10ffmpeg/libavcodec/mpegvideo.c: (log message trimmed)
[05:31] <CIA-17> ffmpeg: mpegvideo: fix position of bottom edge.
[05:31] <CIA-17> ffmpeg: It was wrong in colorspaces where horizontal and vertical chroma
[05:31] <CIA-17> ffmpeg: subsampling are not the same, e.g. 422.
[05:31] <CIA-17> ffmpeg: (cherry picked from commit 0884dd5a1b87aff6c8a06e6492dece5cef8f3978)
[05:31] <CIA-17> ffmpeg: Conflicts:
[05:31] <CIA-17> ffmpeg:  libavcodec/mpegvideo.c
[05:31] <CIA-17> ffmpeg: 03Justin Ruggles 07release/0.8 * r4562f95ba8 10ffmpeg/libavcodec/sipr.c: 
[05:31] <CIA-17> ffmpeg: sipr: fix the output data size check and only calculate it once.
[05:31] <CIA-17> ffmpeg: (cherry picked from commit 1b5a189f06879338088809b3049ea7620f4e7e78)
[05:31] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:31] <CIA-17> ffmpeg: 03Chris Rankin 07release/0.8 * rea311af23d 10ffmpeg/libavcodec/qcelpdec.c: 
[05:31] <CIA-17> ffmpeg: qcelpdec: fix the return value of qcelp_decode_frame().
[05:31] <CIA-17> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[05:31] <CIA-17> ffmpeg: (cherry picked from commit bde25700134b98068e2ad21c1f92955a4b489cdc)
[05:31] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:31] <CIA-17> (168 lines omitted)
[05:31] <CIA-17> ffmpeg: 03Laurent Aimar 07release/0.8 * rc98d7882d8 10ffmpeg/libavcodec/bink.c: 
[05:31] <CIA-17> ffmpeg: bink: Check for out of bound writes when building tree
[05:31] <CIA-17> ffmpeg: Signed-off-by: Janne Grunau <janne-libav at jannau.net>
[05:31] <CIA-17> ffmpeg: (cherry picked from commit 24adf7832b8370f3c1febbef6c686f574d360d32)
[05:31] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:31] <CIA-17> ffmpeg: 03Laurent Aimar 07release/0.8 * r9c78fe9360 10ffmpeg/libavcodec/bink.c: 
[05:31] <CIA-17> ffmpeg: bink: Check for various out of bound writes
[05:31] <CIA-17> ffmpeg: Signed-off-by: Janne Grunau <janne-libav at jannau.net>
[05:31] <CIA-17> ffmpeg: (cherry picked from commit a00676e48e49a3d794d6d2063ceca539e945a4a4)
[05:31] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:31] <CIA-17> ffmpeg: 03Mans Rullgard 07release/0.8 * r8c2ae575ad 10ffmpeg/libavcodec/aacdec.c: 
[05:32] <CIA-17> ffmpeg: aacdec: fix undefined shifts
[05:32] <CIA-17> ffmpeg: Since nnz can be zero, this is needed to avoid a shift by 32.
[05:32] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[05:32] <CIA-17> ffmpeg: (cherry picked from commit d12294304acd82cb219e3f66ca9cd6efb2194fa4)
[05:32] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:32] <CIA-17> ffmpeg: 03Mans Rullgard 07release/0.8 * r58afe6061a 10ffmpeg/libavcodec/dca.c: 
[05:32] <CIA-17> ffmpeg: dca: fix signed overflow in shift
[05:32] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[05:32] <CIA-17> ffmpeg: (cherry picked from commit 559c244d42be7a02c23976216b47fd63b80d6c7f)
[05:32] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[05:32] <CIA-17> (164 lines omitted)
[05:32] <CIA-17> ffmpeg: 03Laurent Aimar 07release/0.8 * r39fed2e95b 10ffmpeg/libavcodec/anm.c: 
[05:32] <CIA-17> ffmpeg: anm: prevent infinite loop
[05:32] <CIA-17> ffmpeg: Signed-off-by: Janne Grunau <janne-libav at jannau.net>
[05:32] <CIA-17> ffmpeg: (cherry picked from commit 2475f1a83ccf313d828b25f1769e3a37442ecf64)
[05:32] <CIA-17> ffmpeg: Signed-off-by: Anton Khirnov <anton at khirnov.net>
[09:51] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * r0d293880ba 10ffmpeg/libavcodec/iff.c: Match documentation for transparency in BGR32 iff images.
[09:58] <ubitux>     dts = avio_rb24(s->pb);
[09:58] <ubitux>     dts |= avio_r8(s->pb) << 24;
[09:58] <ubitux> any reason not to use avio_rb32()?
[09:58] <wbs> because it's not a 32 bit big-endian number
[09:58] <wbs> it's 24 bits in big endian and the upper 8 bits stored afterwards
[09:58] <wbs> you could see it as a little endian 32 bit number with the lower three bytes swapped
[09:59] <ubitux> oh, ok
[09:59] <ubitux> why is it so?
[10:00] <wbs> they probably only used 24 bits initially (iirc all other fields are big endian too), and at some point they realized "shit, we need larger timestamps" and chose to use some unused later byte of the packet header for that
[10:00] <ubitux> ah, i see :)
[10:34] <ubitux> is the "pts wrapping" officially documented in some iso or sth?
[10:35] <ubitux> it seems to be integrated quite early (916c80e94ab242489f06c2e578a792b56678134f) with pts_wrap_bits
[10:35] <ubitux> i'd like to understand this a bit more
[10:37] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * r78f2833f78 10ffmpeg/libavformat/mmf.c: Tell users about supported audio samplerates in mmf.
[13:37] <CIA-17> ffmpeg: 03Matthieu Bouron 07master * rf645132061 10ffmpeg/libavformat/gxf.c: 
[13:37] <CIA-17> ffmpeg: gxf: return AVERROR_EOF instead of AVERROR(EIO) at end of gxf_packet function
[13:37] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:33] <kriegerod> is there a function like usleep_with_check_interrupt(int amount_sleep, int sleep_step, AVIOInterruptCB interrupt_callback)? should it be nice to have?
[17:38] <michaelni> kriegerod, i think theres no such thing but it might be usefull
[18:55] <michaelni> Tjoppen, your help is needed on a not fully working mxf change on ffmpeg-dev
[18:55] <michaelni>  [FFmpeg-devel] [PATCH] mxf: micro patch - mxf parsing
[19:03] <ubitux> michaelni: what's different in the flv patch? FLV_VIDEO_FRAMETYPE_MASK=0xf0, and FLV_FRAME_VIDEO_INFO_CMD=5<<4=0x50
[19:03] <ubitux> did i miss sth?
[19:09] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r11642cd16f 10ffmpeg/libavcodec/sunrastenc.c: 
[19:09] <CIA-17> ffmpeg: lavc/sunrastenc: consider cases with linesize < 0
[19:09] <CIA-17> ffmpeg: Make sunrast_image_write_image() deal with cases when linesize is < 0.
[19:09] <CIA-17> ffmpeg: Fix trac ticket #1077.
[19:10] Action: michaelni sees no <<4 in the patch
[19:11] <ubitux> FLV_FRAME_VIDEO_INFO_CMD = 5 << FLV_VIDEO_FRAMETYPE_OFFSET,
[19:11] <ubitux> and libavformat/flv.h:#define FLV_VIDEO_FRAMETYPE_OFFSET   4
[19:12] <michaelni> ohh, i misread the patch
[19:12] <michaelni> its ok then
[19:12] <ubitux> ok thanks, will push soon :)
[20:10] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * re19ccc89ef 10ffmpeg/libavfilter/vf_aspect.c: 
[20:10] <CIA-17> ffmpeg: lavfi/aspect: fail with AVERROR(EINVAL) in case of invalid specified aspect
[20:10] <CIA-17> ffmpeg: Previously it was returning 0 in case the aspect was parsable but
[20:10] <CIA-17> ffmpeg: invalid.
[20:10] <CIA-17> ffmpeg: X-ColaCount: +10*PHIl
[21:03] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r2b07f572af 10ffmpeg/libavformat/mp3dec.c: 
[21:03] <CIA-17> ffmpeg: mp3_probe: consider id3 tags to be low scoring mp3.
[21:03] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:41] <Daemon404> lol
[22:41] <Daemon404> gpl troll on dev list
[22:42] <kierank> he's in the channel
[22:43] <durandal_1707> lol
[22:43] <Daemon404> o ok
[22:44] Action: gnafu wonders if that troll is Daemon404.
[22:45] Action: gnafu ducks.
[22:46] <durandal_1707> ^ troll
[23:00] <Tjoppen> michaelni: it's getting late -> I'll look at it tomorrow
[23:02] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rfc8ed1117e 10ffmpeg/libavcodec/utils.c: 
[23:02] <CIA-17> ffmpeg: avcodec_encode_audio2: Increase the audio buffer size.
[23:02] <CIA-17> ffmpeg: Fixes Ticket1104
[23:02] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:02] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r87c1783c77 10ffmpeg/libavcodec/ (snow.h snowenc.c): 
[23:02] <CIA-17> ffmpeg: snowenc: move runs from stack to heap.
[23:02] <CIA-17> ffmpeg: Fixes ticket1082
[23:02] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:18] <Compn> KSHawkEye : --enable-gpl
[23:36] Action: durandal_1707 more adpcm codecs...
[23:44] <KSHawkEye> Compn: I am compiling with --enable-gpl, so does that mean I ONLY need the gpl-3.0.txt license? Sorry for posting on the dev list, I'm under a time crunch and thought the most knowledgeable people with license terms would be on the dev list.
[23:49] <durandal_1707> most knowledgeable people with license terms are lawyers
[23:58] <durandal_1707> michaelni: i get segfault when using large -context for ffv1 when encoding
[23:59] <durandal_1707> guess alloc return value is not cheched...
[00:00] --- Tue Mar 20 2012