[Ffmpeg-devel-irc] ffmpeg-devel.log.20111213

[burek]

[00:02] <Compn> lol
[00:02] <Compn> luke-jr : wheres it crash ?
[00:03] <luke-jr> Compn: I think it exhausted the system memory.
[00:03] <luke-jr> enabled distcc and it went OK
[00:03] <luke-jr> now trying again with extra-cflags for NEON
[00:42] <durandal_1707> awfull: http://pastebin.com/W1gawcKc
[00:43] <michaelni> did someone forget to use a loop or a macro ;)
[00:43] <kierank> i'd bet that's amiga l33t coder stuff
[00:47] <michaelni> its just permuting bits around
[00:57] <michaelni> also this can be optimized by working on multiple bits at once
[00:58] <michaelni> not that it would matter for us with todays desktop systems
[00:59] <CIA-49> ffmpeg: 03Mans Rullgard 07master * r5695ae46f8 10ffmpeg/libavcodec/h264.c: 
[00:59] <CIA-49> ffmpeg: h264: remove useless cast
[00:59] <CIA-49> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[00:59] <CIA-49> ffmpeg: 03Mans Rullgard 07master * r179ae53fed 10ffmpeg/libavcodec/proresdec.c: 
[00:59] <CIA-49> ffmpeg: proresdec: fix decode_slice() prototype
[00:59] <CIA-49> ffmpeg: Make the function prototype match the argument of
[00:59] <CIA-49> ffmpeg: AVCodecCntext.execute() and remove the cast hiding
[00:59] <CIA-49> ffmpeg: this mismatch.
[00:59] <CIA-49> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[00:59] <CIA-49> ffmpeg: 03Mans Rullgard 07master * r5b0d35eaed 10ffmpeg/ (configure libavutil/x86/bswap.h): 
[00:59] <CIA-49> ffmpeg: x86: bswap: remove test for bswap instruction
[00:59] <CIA-49> ffmpeg: Firstly, this test never worked as intended, always reporting
[00:59] <CIA-49> ffmpeg: success. Secondly, bswap is available from 486 onward and can
[00:59] <CIA-49> ffmpeg: thus be assumed present.
[00:59] <CIA-49> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[00:59] <CIA-49> ffmpeg: 03Martin Storsjö 07master * rf1dba9e498 10ffmpeg/libavcodec/x86/cabac.h: 
[00:59] <CIA-49> ffmpeg: x86: Require 7 registers for the cabac asm
[01:00] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r3a7f7678eb 10ffmpeg/ (6 files in 2 dirs): 
[01:00] <CIA-49> ffmpeg: lavf: deprecate av_close_input_stream().
[01:00] <CIA-49> ffmpeg: And remove all its uses.
[01:00] <CIA-49> ffmpeg: 03Anton Khirnov 07master * rcd3716b9aa 10ffmpeg/ (14 files in 4 dirs): Replace all uses of av_close_input_file() with avformat_close_input().
[01:00] <CIA-49> ffmpeg: 03Paul B Mahol 07master * r8b494b7b27 10ffmpeg/libavcodec/vcr1.c: 
[01:00] <CIA-49> ffmpeg: vcr1: drop unnecessary emms_c() calls without MMX code
[01:00] <CIA-49> ffmpeg: Signed-off-by: Diego Biurrun <diego at biurrun.de>
[01:00] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rc83ef636a6 10ffmpeg/ (Makefile configure): 
[01:00] <CIA-49> ffmpeg: build: Disable avconv more completely while leaving the user the option to enable it
[01:00] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:01] <CIA-49> ffmpeg: 03Diego Biurrun 07master * r58c42af722 10ffmpeg/ (51 files in 7 dirs): doxygen: misc consistency, spelling and wording fixes
[01:01] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r8bc7fe4daf 10ffmpeg/: (log message trimmed)
[01:01] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/master'
[01:01] <CIA-49> ffmpeg: * qatar/master:
[01:01] <CIA-49> (16 lines omitted)
[01:37] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r376ee20614 10ffmpeg/libavcodec/ (h263dec.c ituh263dec.c): 
[01:37] <CIA-49> ffmpeg: h263dec: restore error concealment functionality after merge
[01:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:37] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rc9cab022a3 10ffmpeg/libavcodec/ (mpeg4video.h mpeg4videodec.c): 
[01:37] <CIA-49> ffmpeg: mpeg4dec: restore error concealment functionality after merge
[01:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:37] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rb95565074f 10ffmpeg/libavcodec/msmpeg4.c: 
[01:37] <CIA-49> ffmpeg: msmpeg4dec: restore error concealment functionality after merge
[01:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:37] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r41c2bbaac6 10ffmpeg/libavcodec/utils.c: 
[01:37] <CIA-49> ffmpeg: lavc: use hexadecimal debug print for showing err_recognization values.
[01:37] <CIA-49> ffmpeg: They are flags which makes them easier to read as hex than decimal.
[01:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:19] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r4a6855343f 10ffmpeg/libavformat/movenc.c: 
[03:19] <CIA-49> ffmpeg: mp4enc: dont store cts in trun if they are unneeded.
[03:19] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:19] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rd9a41695ed 10ffmpeg/libavformat/ (movenc.c movenc.h): 
[03:19] <CIA-49> ffmpeg: mp4enc: Only store flags in trun when they differ from the default and set a sane default.
[03:19] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:19] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r1f83f04609 10ffmpeg/libavformat/movenc.c: 
[03:19] <CIA-49> ffmpeg: mp4enc: dont store durations in trun if they are unneeded.
[03:19] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:14] <Daemon404> [22:14] <+Daemon404> whats the difference betwene fate, and codec-regerssion.sh?
[04:17] <Compn> fate is automated ?
[04:17] Action: Compn acts dumb
[04:18] <Compn> where is codec-regression.sh ?
[04:18] <Daemon404> in tests/
[04:18] <Daemon404> and its not clear where it gets its data from, for one.
[04:19] <Daemon404> im going to ignore it and use fate
[04:19] <Daemon404> >_>
[04:19] <Compn> is that what you get when you 'make test' ?
[04:19] <Daemon404> make fate does it too, but... it's the vsynth1 etc stuff
[04:20] <Daemon404> sure is no documentation
[04:20] <Compn> maybe fate does it 
[04:20] <Compn> heh
[04:20] <Compn> http://git.videolan.org/?p=ffmpeg.git;a=history;f=tests/codec-regression.sh;h=7676074d1fc00878953ffaf3ce9955da86c96e87;hb=HEAD
[04:20] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * ra6c4c0f274 10ffmpeg/libavformat/ (movenc.c movenc.h): 
[04:20] <CIA-49> ffmpeg: mp4enc: Avoid storing sizes in trun when they all match.
[04:20] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:23] <Daemon404> hmm i think i see how it worksish
[04:27] <Compn> did you read the fate docs ?
[04:31] <Daemon404> yes
[04:31] <Daemon404> fate docs do not cover any of that stuff
[04:31] <Daemon404> or how it works, or how to add tests
[04:31] <Daemon404> and the difference between the different files you can add tests in
[04:33] <Asakura> fmm, build failed with ffmpeg 0.9, yasm 1.2, centos 6.1, latest x264 at libavcodec/x86/dsputil_yasm
[04:42] <Asakura> ffmpeg 0.8.7 is build on same environment. 
[04:54] <michaelni> Asakura, if it hasnt been reported yet to our bug tracker, please report it there with a full bugreport
[04:55] <Asakura> I will try with git version code again before report.
[05:04] <Asakura> YASM	libavcodec/x86/dsputil_yasm.o / git version is pass.
[05:06] <michaelni> Asakura, what is the error yasm prints ?
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1120: error: instruction expected after label
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1120: error: undefined symbol `CLIPD_MMX.loop' (first use)
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1120: error:  (Each undefined symbol is reported only once.)
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1125: error: redefinition of `SPLATD_SSE2.loop'
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1123: error: `SPLATD_SSE2.loop' previously defined here
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1125: error: undefined symbol `CLIPD_SSE2.loop' (first use)
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1129: error: redefinition of `SPLATD_SSE2.loop'
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1123: error: `SPLATD_SSE2.loop' previously defined here
[05:27] -Asakura:#ffmpeg-devel- libavcodec/x86/dsputil_yasm.asm:1129: error: undefined symbol `CLIPD_SSE41.loop' (first use)
[05:28] <Asakura> deleted some same error messages though.
[05:29] <Asakura> I try same configure options on git version, and I will replace file to 0.9 version to use.
[05:29] <Chikuzen> Asakura: use pastebin
[05:31] <Asakura> http://pastebin.com/nK7XcyRM
[05:31] <Asakura> is it ok ?
[05:34] <relaxed> Asakura: you're sure yasm is in your path?
[05:35] <Asakura> yes, build passed other source, x264 too
[05:37] <Asakura> fmm, no difference between git version and 0.9 version...
[05:37] <Asakura> at dsputil_yasm.asm
[07:15] <burek> is there a way to give ffmpeg a file with all the cmd line options, to avoid escaping the options because of the shell?
[07:15] <burek> something like conf file just like ffserver uses
[07:18] <drv> xargs?
[07:20] <burek> hmh, let me take a look at what's that :)
[07:22] <burek> i think it doesn't do escaping, only splits long cmd line params into sublists
[07:23] <burek> but this also would be a problem because
[07:23] <burek> find /path -type f -print0 | xargs -0 rm
[07:23] <burek> In the above example, the find utility feeds the input of xargs with a long list of file names. xargs then splits this list into sublists and calls rm once for every sublist.
[07:23] <burek> note "once for every sublist."
[07:24] <drv> oh, sorry, missed the part where you mentioned escaping
[07:44] <jasta> hi folks
[08:49] <kcm1700> i'm curious about why libavcodec/libx264.c:687 has "threads" value by default AV_STRINGIFY(X264_THREADS_AUTO). it's checked out from n0.9
[08:50] <kcm1700> and couldn't find a definition of X264_THREADS_AUTO
[08:58] <Chikuzen> X264_THREADS_AUTO is defined by x264.h
[08:59] <kcm1700> aha, thanks a lot.
[13:55] <ubitux> mmh is it possible to use the lavfi device as input in ffmpeg?
[13:56] <ubitux> i'm not able to find the ffplay -f lavfi sourcefilter=... equivalent for ffmpeg
[15:40] <ubitux> i have a pretty weird problem with the amovie filter
[15:41] <ubitux> i have a mp3 that can't be played with "./ffplay -f lavfi amovie=1.mp3"
[15:41] <ubitux> but works with ./ffplay 1.mp3
[15:42] <ubitux> i thought using the new api might solve the issue, but actually no :)
[15:44] <ubitux> the decode value is pretty weird
[15:45] <ubitux> (-1094995529)
[15:49] <michaelni> ubitux, maybe a MKTAG error code
[15:50] <ubitux> mmh the error may not be handled correctly
[15:51] <ubitux> actually, the error handling in ffplay makes no sense to me
[15:52] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=ffplay.c;h=c84dc2fd9a85681babf18352dd20e1564f9028f0;hb=HEAD#l2063
[15:52] <ubitux> oh my bad.
[15:53] <ubitux> ok found it.
[15:59] <ubitux> the audio decode example is wrong btw
[16:00] <ubitux> i also wonder if the way ffmpeg deals with it is correct
[16:00] <ubitux> but well.
[16:09] <michaelni> ubitux, how is your valgrind fate client ?
[16:09] <michaelni> ive just fixed one issue valgrind of fate  pointed at
[16:10] <michaelni> and another looks like a libc bug 
[16:10] <ubitux> i just bought the box yesterday night
[16:10] <ubitux> i might setup things tonight, or this week
[16:10] <michaelni> you bought a new box for this ?
[16:10] <ubitux> i needed one for me :)
[16:10] <michaelni> ahh
[16:10] <ubitux> quake3 server mainly :D
[16:10] <michaelni> :)
[16:11] <ubitux> i'll try to setup it asap
[16:11] <ubitux> i'm playing with Nicolas amerge filter right now
[16:11] <ubitux> and i'm unfortunately not able to make it works
[16:11] <ubitux> (he sent a WIP last month)
[16:12] <ubitux> i'd like to check if i can use it to merge more than two inputs
[16:12] <ubitux> like say you have a media with 8 streams and want only one out of them
[16:12] <ubitux> it will basically ends the limitation of one input stream per output stream in ffmpeg
[16:13] <michaelni> cool
[16:59] <ubitux> mateo`: nice ;)
[17:10] <ubitux> heh nicolas mentionning the merge filter
[18:07] <ubitux> michaelni: any idea why ffmpeg isn't setting pkt.size to 0 in case of avcodec_decode_audio4() failure?
[18:08] <ubitux> also i'm not sure to get what you are asking about dts/pts/pos and how it applies here
[18:29] <michaelni> ubitux, the packet with the size is/should not be reused in ffmpeg.c
[18:29] <michaelni> the dts/pts/pos should be reset when out of a lets say 1000byte packet only 500 bytes are used
[18:29] <michaelni> as they are not correct for the next 500 byte
[18:57] Action: Daemon404 pokes michaelni to take a quick look at his simple patch, so he can v2 it and send updated patches to both ML
[18:58] Action: Daemon404 runs
[19:11] <michaelni> Daemon404, your review is ready
[19:11] <Daemon404> thank you :)
[19:12] <ubitux> <@michaelni> ubitux, the packet with the size is/should not be reused in ffmpeg.c // how is it reused in ffplay/filters then?
[19:12] <ubitux> i mean, what's the difference?
[19:22] <michaelni> ubitux, not sure its actually neccessary in ffplay.c
[19:31] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r53be37e368 10ffmpeg/ (libavcodec/msrledec.c tests/ref/fate/aasc): 
[19:31] <CIA-49> ffmpeg: msrledec: Check for overreads
[19:31] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:31] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rbe5db7004f 10ffmpeg/libavcodec/aasc.c: 
[19:31] <CIA-49> ffmpeg: aascdec: Check input buffer size on raw data.
[19:31] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:32] <ubitux> <@michaelni> the dts/pts/pos should be reset when out of a lets say 1000byte packet only 500 bytes are used // do we already have some similar code?
[19:32] <ubitux> what could be the invalid data btw?
[19:33] <ubitux> some metadata interlaced in the audio stream or such?
[19:40] <michaelni> ubitux,  see *pkt_p/dts  = AV_NOPTS_VALUE; in ffmpeg.c
[19:41] <ubitux> ok this is done only for video
[19:41] <ubitux> i added this: +    avcodec_get_frame_defaults(movie->frame);
[19:41] <ubitux> which should reset the pts/dts
[19:42] <ubitux> it's done before the decode
[19:42] <ubitux> (just like the example)
[19:42] <michaelni> the AVPacket must be reset not the AVFrame
[19:43] <ubitux> well that's what is done in the example
[19:43] <ubitux> doc/examples/decoding_encoding.c
[19:43] <ubitux> afaiu
[19:43] <ubitux> not the first time though
[19:44] <ubitux> and as i mentionned the example does not deal correctly with len < 0 :(
[19:45] <GordonFreeman> hi
[19:51] <michaelni> hi
[19:54] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r6c39b3c278 10ffmpeg/ffmpeg.c: 
[19:54] <CIA-49> ffmpeg: ffmpeg: reset dts/pts after decoding the first subpacket
[19:54] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:55] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r0e187fd13c 10ffmpeg/doc/examples/decoding_encoding.c: 
[19:55] <CIA-49> ffmpeg: decoding example: reset pts/dts after subpacket.
[19:55] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:57] <ubitux> michaelni: why not changing "pic->pts = pic->best_effort_timestamp = AV_NOPTS_VALUE" into "pic->pts = pic->dts = pic->best_effort_timestamp = AV_NOPTS_VALUE" in avcodec_get_frame_defaults() instead?
[20:14] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r9aeaf62903 10ffmpeg/ (3 files in 2 dirs): 
[20:14] <CIA-49> ffmpeg: Fix all GNU %Ld
[20:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:14] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r35f707abe1 10ffmpeg/libavcodec/twinvq.c: 
[20:14] <CIA-49> ffmpeg: twinnvq: remove ;;
[20:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:14] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rf39258d587 10ffmpeg/libavcodec/utils.c: 
[20:14] <CIA-49> ffmpeg: lavc: set pkt_?ts to NOPTS in get_frame_defaults.
[20:14] <CIA-49> ffmpeg: Idea-by: ubitux
[20:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:19] <Daemon404> lol i love tags.
[20:19] <Daemon404> i love the new tags*
[20:26] <CIA-49> ffmpeg: 03Clément BSsch 07master * r220481e1d6 10ffmpeg/libavfilter/src_movie.c: 
[20:26] <CIA-49> ffmpeg: lavfi: skip the frame in case of error.
[20:26] <CIA-49> ffmpeg: This fixes playback with some audio files through filters.
[20:26] <CIA-49> ffmpeg: 03Clément BSsch 07master * r0b4e5451ba 10ffmpeg/libavfilter/src_movie.c: lavfi: use the new audio API in amovie source filter.
[20:28] <evil_andy> I'm trying to update the mpegts module so that it can handle private data streams embedded in mpegts feeds.  I had a couple questions though as to how I should go about it.
[20:29] <evil_andy> at this point, I've added a new registered type for the private data stream, and a dummy codec to go with it.  The problem is that the type of the stream is AVMEDIA_TYPE_DATA
[20:30] <evil_andy> and I've noticed that my codec never gets called, and I'm pretty sure it's due to the fact that almost everywhere codecs are initialized, there's an if/else block for video, audio, and subtitle codecs.  my question is, should I be adding support for data codecs, or should I be calling my codec a subtitle codec (which it isn't)
[20:31] <kierank> what stream is it?
[20:31] <evil_andy> KLV
[20:31] <kierank> ah
[20:33] <evil_andy> I'm trying to do it 'properly'. I've been given an opportunity to do it the right way and return my work to the community (yeah, I know that sounds like a no-brainer, but I know a lot of times it gets modified and never returned)
[20:33] <kierank> bcoudurier may have some suggestions
[20:34] <evil_andy> Ok, cool. thanks. yeah, it's taken a few months (to get the argument right), but my company agreed to let me fix the code and contribute it back :)
[20:38] <ubitux> that's a good news :)
[20:38] <ubitux> thank you in advance :)
[20:39] <evil_andy> no problem. I've always wanted to do it, never had an opportunity
[20:55] <ubitux> is there anyone handling a fate box here?
[20:56] <ubitux> i'm wondering if i should run a make clean/distclean in the script
[20:56] <ubitux> or well at least re-run the ./configure
[20:56] <ubitux> (is it already done by the fate.sh script?)
[21:10] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rb39e01a6e8 10ffmpeg/ (libavcodec/rv34_parser.c libavformat/bit.c): 
[21:10] <CIA-49> ffmpeg: clenaup: #ifdec CONFIG*
[21:10] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:10] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r16b771b1f4 10ffmpeg/ (4 files in 4 dirs): 
[21:10] <CIA-49> ffmpeg: cleanup: for(int
[21:10] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:11] <michaelni> ubitux, IIRC its all automatic
[21:12] <michaelni> also when you run valgrind, you will need to exclude some libc string functions that overread arrays
[21:13] <ubitux> oh?
[21:14] <ubitux> never expected that with valgrind
[21:14] <Daemon404> btw michaelni, it's a no-go on the win32/clang fate box... the dang thing has a metal case and it gets dangerously hot under load
[21:14] <Daemon404> and i dont want to burn my house down...
[22:09] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.9 * r7881dad2b2 10ffmpeg/libavformat/seek-test.c: 
[22:09] <CIA-49> ffmpeg: seek-test: use standard format specifiers in snprintf()
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: (cherry picked from commit cd2f7ed0007f4803b6bd845366b2398abb32c355)
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.9 * rb8c41e5c08 10ffmpeg/tests/Makefile: 
[22:09] <CIA-49> ffmpeg: fate: replace forgotten avconv by ffmpeg in vsynth / asynth stuff
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: (cherry picked from commit 02a01c2fa168fcf3943c81aa0360076dd4856427)
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: 03Tomas Härdin 07release/0.9 * r57111c558e 10ffmpeg/libavformat/mxfdec.c: (log message trimmed)
[22:09] <CIA-49> ffmpeg: mxfdec: Bring back accumulated_offset that was removed in 5e67e3e
[22:09] <CIA-49> ffmpeg: I thought it had to do with file offsets, but's actually the offset inside
[22:09] <CIA-49> ffmpeg: the essence container.
[22:09] <CIA-49> ffmpeg: In other words, unbreak multiple EditUnitByteCounts.
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: (cherry picked from commit c0ee522df6c9ef09cdcb5bce7adb4b745aa6485a)
[22:09] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.9 * r849ee7ee82 10ffmpeg/ (Makefile configure): 
[22:09] <CIA-49> ffmpeg: build: Disable avconv more completely while leaving the user the option to enable it
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: (cherry picked from commit c83ef636a665b2e17f1b6745a98ec7a5719264df)
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:09] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.9 * r6289b6c811 10ffmpeg/libavcodec/twinvq.c: 
[22:09] <CIA-49> ffmpeg: twinnvq: remove ;;
[22:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:10] <CIA-49> (22 lines omitted)
[22:26] <cbsrobot> hmmm - if the decoder and the encoder are multithreaded - who gets how many threads ?
[22:32] <michaelni> they are independant of each other ...
[22:32] <Daemon404> that doesnt seem very optimal
[22:34] <michaelni> no
[22:35] <michaelni> improvments are welcome :)
[22:38] <durandal_1707> iff from libavc apparently already have code which do ham(hold and modify) stuff, doing this again for cdxl would just duplicate code
[22:38] <durandal_1707> there is also planar to chunk conversion
[22:49] <durandal_1707> how should i resolve this?
[22:50] <durandal_1707> make cdxl only demuxer and add custom codec to libavcodec/iff.c ?
[22:52] <michaelni> possible
[22:52] <ubitux> ok the valgrind fate instance seems to be running fine
[22:52] <michaelni> or export the needed functions in iff.c and call them from a cdxl.c decoder
[22:55] <durandal_1707> creating new header?
[22:56] <Daemon404> er
[22:57] <Daemon404> a cleaner way would be to export ham-related functions to something liek ham.c/h
[22:57] <Daemon404> and make iff and cdxl use it
[22:57] <Daemon404> s/export/move/
[22:58] <durandal_1707> planar 2 chunk conversion would need separate files too
[23:06] <ubitux> heh the valgrind tests are quite slow :)
[23:08] <CIA-49> ffmpeg: 03Reimar Döffinger 07master * r36be045ed7 10ffmpeg/ (libavcodec/aac.h libavcodec/aacdec.c libavformat/nsvdec.c): 
[23:08] <CIA-49> ffmpeg: Support "VLB " audio in NSV files.
[23:08] <CIA-49> ffmpeg: See samples/nsv/vlb_audio.
[23:08] <CIA-49> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[23:11] <michaelni> ubitux, i know i had them running through today
[23:11] <ubitux> :)
[23:11] <michaelni> __strspn_sse42 & __GI___strncasecmp_l seem overreading arrays in them
[23:12] <michaelni> and my valgrind got then stuck on an invalid instruction somewhere
[23:13] <ubitux> atm i have these tests failing: http://213.251.184.135/test.log
[23:13] <michaelni> thats not very informative
[23:13] <michaelni> the valgrind output would be needed ...
[23:14] <ubitux> yes sure :)
[23:14] <ubitux> do you have any option to share?
[23:15] <michaelni> sadly no
[23:16] <ubitux> bcoudurier: when you have a minute, would you mind adding http://blackhole.ubitux.fr/fate.pub ?
[23:16] <michaelni> i just know V=2 likely wont show the output
[23:17] <ubitux> michaelni: i could add a --log-file option
[23:17] <ubitux> and somehow share the url
[23:17] <ubitux> in the comment maybe
[23:18] <ubitux> but that's a bit hacky, there is certainly a better way to deal with this
[23:19] <ubitux> heh it won't work actually
[23:20] <ubitux> i need to create a log file per test
[23:20] <michaelni> ubitux, fix fate to print the full output for each failing test, this would be usefull for local fate too
[23:21] <michaelni> :)
[23:22] <Daemon404> michaelni, YES
[23:22] <Daemon404> fates crappy errors drove me nuts when learning it
[23:23] <ubitux> :(
[23:23] <Daemon404> Error 1 and no other info
[23:23] <ubitux> you have a diff ;)
[23:24] <Daemon404> no
[23:24] <Daemon404> i was trying to figure out how to add an encoder test
[23:24] <Daemon404> and lots of times thats all it output.
[23:24] <Daemon404> no diff. no nothing.
[23:33] <ubitux> michaelni: do you mind backporting the lavfi audio playback fix to 0.9?
[23:33] <CIA-49> ffmpeg: 03Alex Converse 07master * r5cd56e193f 10ffmpeg/libavcodec/aacdec.c: aacdec: Use intfloat.h rather than local punning union.
[23:33] <CIA-49> ffmpeg: 03Mans Rullgard 07master * r40901fc14e 10ffmpeg/libavcodec/ (rv34.c rv34dsp.c rv34dsp.h): 
[23:33] <CIA-49> ffmpeg: rv34: move 4x4 dequant to RV34DSPContext
[23:33] <CIA-49> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[23:33] <CIA-49> ffmpeg: 03Mans Rullgard 07master * r878dda5db1 10ffmpeg/ (8 files in 8 dirs): 
[23:33] <CIA-49> ffmpeg: build: move inclusion of subdir.mak to main subdir loop
[23:33] <CIA-49> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[23:33] <CIA-49> ffmpeg: 03Luca Barbato 07master * r5bf2ac2b37 10ffmpeg/libavcodec/ (12 files): 
[23:33] <CIA-49> ffmpeg: error_resilience: use the ER_ namespace
[23:33] <CIA-49> ffmpeg: Add the namespace to {AC_,DC_,MV_}{END,ERROR} macros
[23:33] <CIA-49> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[23:33] <CIA-49> ffmpeg: 03Luca Barbato 07master * r63ccd46687 10ffmpeg/libavcodec/ (8 files): 
[23:33] <CIA-49> ffmpeg: lavc: introduce ER_MB_END and ER_MB_ERROR
[23:33] <CIA-49> ffmpeg: Simplify a little error resilience calls
[23:33] <CIA-49> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[23:34] <CIA-49> ffmpeg:  yadif: support 10bit YUV
[23:34] <CIA-49> ffmpeg:  vc1: mark with ER_MB_ERROR bits overconsumption
[23:34] <CIA-49> (3 lines omitted)
[23:41] <CIA-49> ffmpeg: 03Clément BSsch 07release/0.9 * r9aaf250fb3 10ffmpeg/libavfilter/src_movie.c: 
[23:41] <CIA-49> ffmpeg: lavfi: skip the frame in case of error.
[23:41] <CIA-49> ffmpeg: This fixes playback with some audio files through filters.
[23:41] <CIA-49> ffmpeg: (cherry picked from commit 220481e1d6582277ae4c02b05f87251a8670e6cc)
[23:41] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:41] <ubitux> thx :)
[23:43] <durandal_1707> is Jai 
[23:43] <durandal_1707> Menon still available here
[23:49] <Daemon404> BSsch <-- i know a lot of people with that last name
[23:49] <Daemon404> ubitux, dutch?
[23:50] <ubitux> nop, french
[23:50] <Daemon404> oic
[23:50] <ubitux> but alsatian origins afaik
[23:50] <Daemon404> weird
[23:50] <microchip_> ubitux: you eat frogs too? :p
[23:51] <Daemon404> thats just ripe for a racist joke >.>
[23:51] <microchip_> no, it's a question
[23:51] <ubitux> this frogs obsession is pretty strange actually
[23:51] <ubitux> and no, i never ate frogs :p
[23:51] <microchip_> good
[23:52] Action: Daemon404 has
[23:52] <ubitux> Daemon404: so well, i guess it has some german origins at some point
[23:52] Action: microchip_ knows 3 french peeps who do
[23:52] <ubitux> but i really have no clue and don't really care about :p
[23:52] <Daemon404> ;p
[23:52] <ubitux> i just enjoy puting weird char in my name
[23:52] <ubitux> even though it might not be the correct way of writing it ;)
[23:53] <Daemon404> well my dutch and german friends spell it boesche
[23:53] <Daemon404> without the funny char
[23:53] <ubitux> it can't be pronounced well without the S in french
[23:54] <Daemon404> oic
[23:54] <ubitux> also i think there is another guy with the exact same first & last name in France
[23:54] <ubitux> writing it S makes a difference with him ;)
[23:56] <ubitux> i wonder if i reached 50% of the tests&
[23:57] <ubitux> so ~12 failing tests so far mmh
[00:00] --- Wed Dec 14 2011