[Ffmpeg-devel-irc] ffmpeg-devel.log.20120105

[burek]

[00:27] <pasteeater> that's what i told the op too
[00:48] <pasteeater> another spam account on trac: matildaharris
[00:48] <pasteeater> she's a witch!
[00:51] <ubitux> buuuurrrnnn
[01:00] <michaelni> pasteeater, how do u detect trac spamers ?
[01:01] <pasteeater> when i delete shit like "Phuket property" from the main page and then i look at page history.
[01:03] <michaelni> humpf, user deleted
[01:03] <pasteeater> thanks
[01:05] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r1171d938af 10ffmpeg/libavformat/utils.c: (log message trimmed)
[01:05] <CIA-101> ffmpeg: lavf: revert commit bb99ae3ae924c942a634bec7711ec7ee11c38eb9
[01:05] <CIA-101> ffmpeg:  Author: Michael Niedermayer <michaelni at gmx.at>
[01:05] <CIA-101> ffmpeg:  Date: Thu Nov 3 22:38:10 2011 +0100
[01:05] <CIA-101> ffmpeg:  lavf: fix null pointer dereference in rdt
[01:05] <CIA-101> ffmpeg:  Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:05] <CIA-101> ffmpeg: This is no longer needed and causes various problems with RTSP
[01:05] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * rad8e3304f7 10ffmpeg/libavformat/utils.c: 
[01:05] <CIA-101> ffmpeg: lavf: use avg_frame_rate and packet number to exit find_stream_info
[01:05] <CIA-101> ffmpeg: qatar does this too but clobbers the AVPacket.duration by approximate
[01:05] <CIA-101> ffmpeg: values.
[01:05] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:05] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * rc4eec85a1f 10ffmpeg/libavformat/rmdec.c: 
[01:05] <CIA-101> ffmpeg: Revert "rmdec: Avoid allocating huge packets"
[01:05] <CIA-101> ffmpeg: This reverts commit 66f71f3b5e2e7b6e5049bd9831c371e16aff8a53.
[01:05] <CIA-101> ffmpeg: This causes regressions with RDT.
[01:13] <CIA-101> ffmpeg: 03Oana Stratulat 07release/0.9 * r32b62c7d6a 10ffmpeg/libavformat/electronicarts.c: 
[01:13] <CIA-101> ffmpeg: electronicarts: Fix division by zero. Fixes Ticket #793
[01:13] <CIA-101> ffmpeg: Signed-off-by: Oana Stratulat <oanaandreeastratulat at gmail.com>
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: (cherry picked from commit 0d3a51e5d279dd2a56c81ba7a81a70128c5a7545)
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.9 * r9f9b6388a3 10ffmpeg/libavformat/electronicarts.c: 
[01:13] <CIA-101> ffmpeg: electronicarts: Pass error through ea_read_header().
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: (cherry picked from commit f37174bc1913d534913e8825f1609a637cc4b364)
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.9 * r66569b375c 10ffmpeg/libavcodec/mjpegdec.c: 
[01:13] <CIA-101> ffmpeg: jpegdec: Fix vlc table check for progressive jpegs.
[01:13] <CIA-101> ffmpeg: Fixes Ticket834
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: (cherry picked from commit 15b219fae9da1691dfb264f51637805e1ca63d1a)
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: 03Chris Evans 07release/0.9 * r49b8709870 10ffmpeg/libavformat/matroskadec.c: (log message trimmed)
[01:13] <CIA-101> ffmpeg: mkv: Fix a bug where a pointer was cached to an array that might later move due to
[01:13] <CIA-101> ffmpeg: a realloc()
[01:13] <CIA-101> ffmpeg: BUG=100492
[01:13] <CIA-101> ffmpeg: Review URL: http://codereview.chromium.org/8366004
[01:13] <CIA-101> ffmpeg: Fixes: 1 of 2 for CVE-2011-3893
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> ffmpeg: 03Chris Evans 07release/0.9 * r6a3fed8749 10ffmpeg/libavcodec/vorbis.c: (log message trimmed)
[01:13] <CIA-101> ffmpeg: ogg: Avoid the possibility to read out-of-bounds of a static global array in Vorbis
[01:13] <CIA-101> ffmpeg: Review URL: http://codereview.chromium.org/8365014
[01:13] <CIA-101> ffmpeg: This fixes 25% of CVE-2011-3893
[01:13] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:13] <CIA-101> (55 lines omitted)
[01:29] <OanaStratulat> michaelni: how can i get the latest src files without making clone again?
[01:33] <ubitux> like i said, you put your changes in a branch, go back to master, and run git pull
[01:38] <OanaStratulat> ubitux: have some time to discuss about https://ffmpeg.org/trac/ffmpeg/ticket/599 ?
[01:45] <OanaStratulat> michaelni: ping
[01:57] <ubitux> i'm not sure this will be easy to fix
[01:58] <OanaStratulat> ubitux: why is that?
[01:58] <OanaStratulat> ubitux: how is the duration calculated ?
[01:59] <ubitux> i guess the duration is written in the container (mov) but it should be read (calculated?) from the ts in the codec?
[01:59] <ubitux> i'm not familiar with that stuff at all
[02:00] <OanaStratulat> i cannot find another easy bug on trac :( 
[02:00] <OanaStratulat> suppose i must create one and then solve it
[02:05] <michaelni> OanaStratulat, yes thats an option, easy bugs tend to be fixed quickly
[02:06] <michaelni> about git, it was git reset --hard origin
[02:06] <OanaStratulat> so back to the drawing board, zzufing files
[02:06] <OanaStratulat> yes
[02:06] <OanaStratulat> found that
[02:07] <kierank> just don't write the duration in the mov
[02:08] <kierank> i mean don't use it for anything important
[02:09] <OanaStratulat> michaelni: can you give me an idea on how to automatically test some files with zzuf ?
[02:10] <OanaStratulat> michaelni: not to input each file
[02:19] <CIA-101> ffmpeg: 03Jonathan Nieder 07master * r96d8e18a98 10ffmpeg/doc/APIchanges: 
[02:19] <CIA-101> ffmpeg: APIchanges: mention cleaned up header includes
[02:19] <CIA-101> ffmpeg: Fixes bug 188.
[02:19] <CIA-101> ffmpeg: Signed-off-by: Jonathan Nieder <jrnieder at gmail.com>
[02:19] <CIA-101> ffmpeg: Signed-off-by: Janne Grunau <janne-libav at jannau.net>
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * rc98c6e18d8 10ffmpeg/libavcodec/adxenc.c: adxenc: remove unneeded log message
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * ra85ab8ad45 10ffmpeg/libavcodec/adxenc.c: adxenc: cosmetics: pretty-printing
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * re3d4f59f95 10ffmpeg/libavcodec/adxenc.c: adxenc: remove unneeded comments
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r25edfc88e3 10ffmpeg/libavcodec/adxenc.c: adxenc: change some data types
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r656e606cae 10ffmpeg/libavcodec/adxenc.c: adxenc: use BLOCK_SIZE and BLOCK_SAMPLES macros
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r6c117bd8e0 10ffmpeg/libavcodec/adxenc.c: 
[02:19] <CIA-101> ffmpeg: adxenc: remove unneeded loops
[02:19] <CIA-101> ffmpeg: avctx->frame_size is 32, so that is how many samples we process per call.
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r613668210b 10ffmpeg/libavcodec/adxenc.c: adxenc: avoid stereo deinterleaving
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * rf1be41c63d 10ffmpeg/libavcodec/adxenc.c: adxenc: use a loop to encode each channel
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r1fb47728cd 10ffmpeg/libavcodec/adxenc.c: 
[02:19] <CIA-101> ffmpeg: adxenc: use bytestream functions for header writing.
[02:19] <CIA-101> ffmpeg: 03Ronald S. Bultje 07master * r400ba1d735 10ffmpeg/libavcodec/h264.c: 
[02:19] <CIA-101> ffmpeg: h264: return index in buffer on end-of-sequence.
[02:19] <CIA-101> ffmpeg: Fixes hangs if the last packet contains an end-of-sequence NAL unit,
[02:19] <CIA-101> ffmpeg: bug 158.
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r3e57573fce 10ffmpeg/tests/ (codec-regression.sh ref/acodec/adpcm_adx): fate: add ADX encoding/decoding test
[02:19] <CIA-101> ffmpeg: 03Diego Biurrun 07master * r54e68fb3b8 10ffmpeg/libavcodec/ (dnxhdenc.c h264.c): 
[02:19] <CIA-101> ffmpeg: Merge some declarations and initializations.
[02:19] <CIA-101> ffmpeg: This fixes compilation failures related to START_TIMER/STOP_TIMER macros and
[02:19] <CIA-101> ffmpeg: -Werror=declaration-after-statement. START_TIMER declares variables and thus
[02:19] <CIA-101> ffmpeg: may not be placed after statements outside of a new block.
[02:19] <CIA-101> ffmpeg: 03Diego Biurrun 07master * r9dec55748c 10ffmpeg/libavcodec/ (cabac.c cabac.h): cabac: drop unused and disabled get_cabac_u() / get_cabac_ueg() functions
[02:19] <CIA-101> ffmpeg: 03Diego Biurrun 07master * re451c26c5f 10ffmpeg/libavcodec/ (cabac.c cabac.h): cabac: drop unused STRICT_LIMITS code branch
[02:19] <CIA-101> ffmpeg: 03Ronald S. Bultje 07master * r6b3995ccd1 10ffmpeg/libswscale/swscale.c: swscale: remove obsolete comment.
[02:19] <CIA-101> ffmpeg: 03Ronald S. Bultje 07master * r6ea64339c5 10ffmpeg/libswscale/ (Makefile x86/output.asm x86/scale.asm): 
[02:19] <CIA-101> ffmpeg: v4l2: simplify away io_method
[02:19] <CIA-101> (14 lines omitted)
[02:19] <CIA-101> ffmpeg: 03Anton Khirnov 07master * re8c04f6240 10ffmpeg/ (avconv.c doc/avconv.texi): avconv: add symbolic names for -vsync parameters
[02:19] <CIA-101> ffmpeg: 03Ronald S. Bultje 07master * rf910dbcdb0 10ffmpeg/libswscale/Makefile: swscale: add MMX files to MMX-OBJS instead of OBJS.
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * rcc40c056d0 10ffmpeg/libavcodec/adxenc.c: adxenc: log an error message and return AVERROR(EINVAL) for invalid channels
[02:19] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r7f83db3124 10ffmpeg/: (log message trimmed)
[02:19] <CIA-101> ffmpeg: Merge remote-tracking branch 'qatar/master'
[02:19] <CIA-101> ffmpeg: * qatar/master: (46 commits)
[02:19] <CIA-101> ffmpeg:  mtv: Make sure audio_subsegments is not 0
[02:19] <CIA-101> ffmpeg:  v4l2: use V4L2_FMT_FLAG_EMULATED only if it is defined
[02:19] <CIA-101> ffmpeg:  avconv: add symbolic names for -vsync parameters
[02:19] <CIA-101> ffmpeg:  flvdec: Fix compiler warning for uninitialized variables
[02:19] <CIA-101> ffmpeg: 03Shitiz Garg 07master * rfeb15cee5e 10ffmpeg/libavformat/mtv.c: 
[02:19] <CIA-101> ffmpeg: mtv: Make sure audio_subsegments is not 0
[02:19] <CIA-101> ffmpeg: audio_subsegments would be 0 and cause floating point exceptions
[02:19] <CIA-101> ffmpeg: Fixes bugzilla #144
[02:19] <CIA-101> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r779ef255e6 10ffmpeg/ (5 files in 3 dirs): adx: add an adx muxer
[02:19] <CIA-101> ffmpeg: 03Janne Grunau 07master * rf13a9ca906 10ffmpeg/libavdevice/v4l2.c: 
[02:19] <CIA-101> ffmpeg: v4l2: use V4L2_FMT_FLAG_EMULATED only if it is defined
[02:19] <CIA-101> ffmpeg: V4L2_FMT_FLAG_EMULATED was added in 2.6.32.
[02:19] <CIA-101> ffmpeg: 03Ronald S. Bultje 07master * r9ea3501d87 10ffmpeg/libswscale/ (swscale.c utils.c): 
[02:19] <CIA-101> ffmpeg: swscale: don't show full-chroma-int warning for non-RGB output.
[02:19] <CIA-101> ffmpeg: Non-RGB output always uses full chroma interpolation.
[02:19] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r754ebd1a5b 10ffmpeg/libavcodec/adxenc.c: adxenc: check output buffer size before writing
[02:19] <CIA-101> ffmpeg: 03Jean First 07master * r2df73eefb4 10ffmpeg/libavformat/flvdec.c: 
[02:19] <CIA-101> ffmpeg: flvdec: Fix compiler warning for uninitialized variables
[02:19] <CIA-101> ffmpeg: These can't be used uninitialized in practice, but the
[02:19] <CIA-101> ffmpeg: compiler doesn't realize it.
[02:20] <CIA-101> (35 lines omitted)
[02:22] <michaelni> OanaStratulat, you can write a shell script to run zzuf ove many files
[02:23] <OanaStratulat> michaelni: and how can i do when i have a crash to make a file with the output?
[02:23] <OanaStratulat> michaelni: not with all the files even if they succeeded the tests 
[02:31] <michaelni> zzuf has a different exit code depending on what happened
[02:32] <OanaStratulat> anyway 0 is for all tests passed i think no?
[02:33] <ubitux> don't we have trac issues for windows fate instance? :)
[02:34] <OanaStratulat> michaelni: do you use osx ?
[02:37] <michaelni> no osx here
[02:38] <OanaStratulat> i have a question not related to ffmpeg
[02:38] <OanaStratulat> i have downloaded zzuf and i have a patch how can i apply it ?
[02:38] <OanaStratulat> the patch fixes http://www.google.ro/search?aq=f&ix=hca&sourceid=chrome&ie=UTF-8&q=zzuf+memory+exceeded+osx
[02:43] <michaelni> if you have the zzuf source, "patch -p(some small number) <patchfile" should work
[02:44] <michaelni> you can try with --dry-run to see if it would work
[02:47] <OanaStratulat> ok
[02:47] <OanaStratulat> thx
[02:47] <OanaStratulat> anyway i`m closing the night here
[02:47] <OanaStratulat> good night everybody
[03:04] <CIA-101> ffmpeg: 03John Brooks 07release/0.9 * rfc9c5ad9ea 10ffmpeg/libavcodec/rv34.c: (log message trimmed)
[03:04] <CIA-101> ffmpeg: rv34: fix invalid memory access for small video dimensions
[03:04] <CIA-101> ffmpeg: For small video dimensions calculations of the upper bound for pixel
[03:04] <CIA-101> ffmpeg: access may result in negative value. Using an unsigned comparison
[03:04] <CIA-101> ffmpeg: works only if the greater operand is non-negative. This is fixed by
[03:04] <CIA-101> ffmpeg: doing edge emulation explicitly for such conditions.
[03:04] <CIA-101> ffmpeg: Signed-off-by: Kostya Shishkov <kostya.shishkov at gmail.com>
[03:04] <CIA-101> ffmpeg: 03Michael Niedermayer 07release/0.9 * r601e648458 10ffmpeg/ (Doxyfile RELEASE VERSION): 
[03:04] <CIA-101> ffmpeg: Update for 0.9.1
[03:04] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:04] <CIA-101> ffmpeg: 03John Brooks 07release/0.9 * re540446f6c 10ffmpeg/libavcodec/vc1dec.c: (log message trimmed)
[03:04] <CIA-101> ffmpeg: vc1dec: fix invalid memory access for small video dimensions
[03:04] <CIA-101> ffmpeg: For small video dimensions, these calculations of the upper bound
[03:04] <CIA-101> ffmpeg: for pixel access may have a negative result. Using an unsigned
[03:04] <CIA-101> ffmpeg: comparison to bound a potentially negative value only works if
[03:04] <CIA-101> ffmpeg: the greater operand is non-negative. Fixed by doing edge emulation
[03:04] <CIA-101> ffmpeg: when the upper bound is probably negative, everywhere that this
[06:41] <chrisballinger> hey guys
[06:42] <chrisballinger> does anyone have any experience with the libavformat code for decoding Core Audio (.caf) files that contain AAC?
[06:42] <chrisballinger> http://www.hydrogenaudio.org/forums/index.php?showtopic=89889
[06:43] <chrisballinger> sorry if this is the wrong channel but i figure i'd find a developer here who was perhaps somewhat familiar with that specific code
[11:36] <silverrocker> I have some questions concerning Google Summer of Code and ffmpeg project proposals. Is there someone specific I need to go to with my questions?
[11:38] <av500> you are a tad late
[11:38] <av500> or very early
[11:40] <silverrocker> av500: hah, well actaully I'm only partially intressted in Google SoC. I have my bachelor thesis coming up next month and I need to decide on what I'm going to do specifically. I have great intressed in multimedia compression and everything around that and my would-be-mentor suggested that I would take a look at the list of SoC to get some idea's. 
[11:41] <silverrocker> Also, the idea would be to contribute to the ffmpeg project but this would be my first time doing something like that so I don't really have a clue how to jump into the ffmpeg code which is actaully why I am here.
[11:46] <av500> git clone it
[11:52] <silverrocker> av500, But where do I start when I cloned it
[11:53] <av500> ./configure
[11:53] <av500> make
[11:53] <av500> where to start depends on what you want to do
[11:54] <av500> you are looking at: http://wiki.multimedia.cx/index.php?title=FFmpeg_/_Libav_Summer_Of_Code_2011
[11:54] <av500> i guess
[11:56] <silverrocker> av500, I have looked at that list but it is still unclear to me 
[11:58] <av500> what is your background?
[12:01] <av500> silverrocker: also, I would advise you find a project that your mentor can really mentor you with
[12:01] <av500> you should not rely on IRC to do that for you
[12:02] <av500> if your would be mentor is new to ffmpeg, I would not do it
[12:04] <silverrocker> av500, I'm currently in my 3rd bachelor year of computer science and I have had some video compression, math and of course C (and C++). I think these are the most relevant for what I would be doing
[12:08] <av500> thats fine
[12:08] <av500> still you need a mentor that is not IRC
[12:08] <av500> IRC will mentor you for GSOC
[12:09] <av500> I see a lot of students ending up in IRC with various projects where the mentor is unable to help
[12:09] <av500> most don't go far in the end
[12:30] <OanaStratulat> michaelni: ping
[12:33] <michaelni> OanaStratulat, pong
[12:33] <OanaStratulat> i have found a bug, can you check it if i supply the file?
[12:34] <michaelni> sure
[12:34] <OanaStratulat> ok just a moment
[12:36] <OanaStratulat> michaelni: http://ptp-roman.org/crash4XM here is the file, when i valgrind it i found seg11
[12:38] <michaelni> Invalid read of size 8 here
[12:39] <michaelni> so yes, its a bug
[12:39] <OanaStratulat> do you fink it is hard to fix?
[12:41] <michaelni> probably alot easier than the mov muxer one
[12:41] <michaelni> but i didnt look very carefull
[12:41] <OanaStratulat> think
[12:41] <silverrocker> av500, Thank you for the information. My promotor has experience with the code of ffmpeg and has done research related to that so I think I'm ok there but I'm not sure if I'm adequate for the task really.
[12:41] <av500> ok
[12:42] <av500> well, he should be able to tell you if you can do the task or not
[12:42] <av500> your mentor is somebody we know?
[12:42] <silverrocker> av500, ok thanks I will have to ask him for more information then before jumping in
[12:43] <OanaStratulat> michaelni: i see in gdb that the error is in memmove$VARIANT$sse42 () 
[12:43] <OanaStratulat> but it does not supply any .c file
[12:47] <michaelni> ==15701==    at 0x4C2A327: memcpy (mc_replace_strmem.c:635)
[12:47] <michaelni> ==15701==    by 0x84DE21: pcm_encode_frame (pcm.c:171)
[12:48] <OanaStratulat> michaelni: an i check if the var i want to memcpy are assigned?
[12:48] <OanaStratulat> can*
[12:56] <michaelni> OanaStratulat, not sure i understand you
[12:57] <OanaStratulat> michaelni: the segfault is because i want to access a part of memory that is not assigned to my program no?
[13:01] <michaelni> yes
[13:01] <OanaStratulat> michaelni: how can i verify that those parts of memory are assigned to my program ?
[13:02] <michaelni> access them, valgrind will tell ;)
[13:03] <OanaStratulat> michaelni: put i want to check them and if they are unassigned return an AVERR, like the bug with FPE i have fixed
[13:04] <OanaStratulat> michaelni: i want to implement this check in ffmpeg not to use external programs like valgrind
[13:06] <michaelni> its not easy to check that, also there must be something wrong if the part you read is not "assigned" to you
[13:20] <michaelni> maybe you should look at what parameters get passed into the pcm_encode_frame() function ?
[13:21] <michaelni> are the pointers non null, the size sane ?
[13:43] <ubitux> 13:26:17 <OanaStratulat> http://pastebin.com/W8BePW8k
[13:43] <ubitux> 13:26:28 <OanaStratulat> this is my output
[13:43] <ubitux> the encode is fed with an invalid buf_size
[13:43] <ubitux> do_audio_out in ffmpeg.c has a strange size value
[13:44] <ubitux> afaik this is where OanaStratulat is looking for
[13:44] <OanaStratulat> yes i am looking in that file
[13:49] <OanaStratulat> as i see in  do_audio_out audio_buf_size; is my buf size no?
[13:49] <ubitux> first, did you confirm the size_out value is also wrong when calling avcodec_encode_audio in ffmpeg.c?
[13:51] <OanaStratulat> yes
[13:51] <OanaStratulat> size is also too big
[13:51] <OanaStratulat> and then going negative
[13:52] <ubitux> you mean size at the beginning of do_audio_out?
[13:52] <OanaStratulat> http://pastebin.com/5bwqtRFC
[13:52] <ubitux> int size     = decoded_frame->nb_samples * dec->channels * isize;
[13:52] -OanaStratulat:#ffmpeg-devel- ret = avcodec_encode_audio(enc, audio_out, size_out,(short *)buftmp);
[13:52] <ubitux> if size is too big, check the 3 values here
[13:52] <OanaStratulat> the size there
[13:52] <OanaStratulat> i checked
[13:53] <ubitux> which one of the 3 values is wrong then?
[13:53] <OanaStratulat> wait what size are you talking about from avcodec_encode_audio ?
[13:53] <ubitux> "size"
[13:53] <OanaStratulat> or the int size you gave me?
[13:54] <ubitux> it seems there is a size_out = size
[13:54] <ubitux> so look at size at the beginning of the function
[13:54] <ubitux> and check if indeed size is wrong, and if so, which on the 3 values is wrong (nb samples? channels? or isize?)
[13:55] <OanaStratulat> wrong meaning very big or going negative ?
[13:55] <ubitux> i guess very big is not fine either
[13:55] <OanaStratulat> 20052 5884 46300 are good values for size
[13:55] <OanaStratulat> ?
[13:55] <ubitux> i guess yes
[13:55] <OanaStratulat> ok
[13:56] <OanaStratulat> at the end i have size : 268454888
[13:56] <OanaStratulat> this is bad
[13:56] <av500> not on your bank account
[13:56] <ubitux> thanks for your help av500 :D
[13:56] <ubitux> OanaStratulat: then look at int size     = decoded_frame->nb_samples * dec->channels * isize;
[13:56] <ubitux> and check the 3 variables here
[13:56] <ubitux> one or more may be wrong
[13:57] <OanaStratulat> yes
[13:57] <OanaStratulat> nb_samples what type is it?
[13:57] <av500> the number of samples for frame
[13:57] <av500> per frame
[13:57] <ubitux> OanaStratulat: %d should be fine to print all of them
[13:59] -OanaStratulat:#ffmpeg-devel- size : 268454888
[13:59] -OanaStratulat:#ffmpeg-devel- nb_samples : 67113722
[13:59] -OanaStratulat:#ffmpeg-devel- dec->channels : 2
[13:59] -OanaStratulat:#ffmpeg-devel- isize : 2
[13:59] <OanaStratulat> at the end ^
[14:00] <OanaStratulat> so nb_samples is too big 
[14:00] <OanaStratulat> in another output nb_sampels is 4850 and values like this 
[14:00] <ubitux> yes, so the decode of the frame gave an invalid number of samples :p
[14:00] <ubitux> not sure where this is defined&
[14:01] <OanaStratulat> mhmh
[14:01] <OanaStratulat> so don`t know where should i check next ?
[14:02] <ubitux> not yet precisely, i'm discovering the code in the same time ;)
[14:02] <OanaStratulat> aa ok
[14:03] <ubitux> avcodec_decode_audio4 will set the number of samples (obviously)
[14:03] <ubitux> now you have to check in the decode audio callback why it is set like this
[14:03] <OanaStratulat> as i see transcode_audio calls do_audio_out
[14:07] <ubitux> i guess the frame content is set in libavcodec/4xm.c
[14:08] <ubitux> but i don't know where the number of samples is specified&
[14:11] <OanaStratulat> michaelni: any idea for this?
[14:11] <av500> OanaStratulat: its called "staring at code"
[14:11] <av500> this value is being set somewhere
[14:11] <av500> add some printfs
[14:22] <OanaStratulat> ubitux: isn`t avcodec_decode_audio4 defined in avcodec ?
[14:26] <ubitux> avcodec_decode_audio4 should call the libavcodec/4xm.c decode callback
[14:27] <ubitux> OanaStratulat: why are you asking about avcodec?
[14:27] <OanaStratulat> i see that avcodec_decode_audio4 is in avcodec.h 
[14:28] <ubitux> yes the prototype is here, the code is in libavcodec/utils.c
[14:28] <michaelni> the size_out passed to avcodec_encode_audio() is negative
[14:29] <ubitux> michaelni: yes, but this is because the number of samples decoded is invalid
[14:29] <ubitux> afaiu
[14:29] <michaelni> it should not become negative ever 
[14:30] <OanaStratulat> so when the size_out is negative it should return invalid data ?
[14:30] <michaelni> id say whatever makes it negative is buggy
[14:32] <OanaStratulat> so size out is made by size_out = swr_convert(...)
[14:33] <OanaStratulat> is there where it is initialized ?
[14:43] <OanaStratulat> as i see it size_out is being made by int swr_convert
[15:33] <OanaStratulat> michaelni: ping
[15:35] <michaelni> OanaStratulat, pong
[15:36] <OanaStratulat> regarding the bug, as i see it size_out is being made by int swr_convert
[15:36] <michaelni> is it negative there ?
[15:36] <OanaStratulat> am i right?
[15:37] <OanaStratulat> yes size_out goes negative then segafult
[15:37] <michaelni> where does it go negative ?
[15:38] <OanaStratulat> http://pastebin.com/H8HPq5e9
[15:38] <OanaStratulat> it goes like 10 loops then it gives a negative value
[15:39] <michaelni> which assignment puts a negative value in it ?
[15:39] <OanaStratulat> http://pastebin.com/w9B8W5qx this is the output
[15:39] <OanaStratulat> let me av_log in different places and i`ll tell you
[15:39] <michaelni> ok
[15:43] <OanaStratulat> michaelni: if (coded_bps) size_out = size_out *coded_bps after this it goes negative
[15:45] <OanaStratulat> michaelni: int how big can it contain ?
[15:46] <OanaStratulat> if i av_log before the sizeout=size_out*coded_bps /8 it is positive and if i av_log after is negative
[15:46] <OanaStratulat> and coded_bps is positive
[15:48] <michaelni> you found where the bug is, now you have to fix it :)
[15:49] <OanaStratulat> if i make size_out bigger
[15:49] <OanaStratulat> change its type ?
[15:49] <OanaStratulat> is a good approach ?
[15:51] <OanaStratulat> i have changed it to int64_t 
[15:51] <OanaStratulat> and now it does not crash anymore
[15:55] <OanaStratulat> michaelni: ^
[16:46] <michaelni> OanaStratulat, making size_out int64_t should work, yes
[16:46] <OanaStratulat> is a good solution for this bug?
[16:47] <OanaStratulat> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7175252 please accept the claim
[16:48] <OanaStratulat> michaelni: so i must file the bug 1st no?
[16:50] <michaelni> yes, please open a bug report
[16:50] <michaelni> claim accepted
[16:50] <OanaStratulat> thx
[16:56] <OanaStratulat> michaelni: bug filed and patch submitted, http://google-melange.appspot.com/gci/task/view/google/gci2011/7175252 marked as completed
[17:03] <OanaStratulat> michaelni: do you have time to review now ?
[17:10] <michaelni> OanaStratulat, did you test the patch ?
[17:10] <OanaStratulat> yes
[17:10] <OanaStratulat> it does not crash anymore
[17:10] <michaelni> also the indention of the first line looks messed uü
[17:10] <michaelni> up
[17:11] <OanaStratulat> can you change that locally?
[17:11] <OanaStratulat> or how do i amend ?
[17:11] <OanaStratulat> git commit -amend ?
[17:13] <michaelni> git commit --amend -a
[17:14] <OanaStratulat> ok patch modified on trac
[17:15] <michaelni> looks ok
[17:16] <OanaStratulat> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7180407 please accept
[17:16] <OanaStratulat> if needed can you extend my deadline ?
[17:21] <OanaStratulat> ubitux: michaelni what does [avi @ 0x7f7f5c03d800] Too deeply nested ODML indexes mean ?
[17:25] <michaelni> OanaStratulat, i can extend as much as needed
[17:25] <OanaStratulat> michaelni: can you extend with 24 hours because 2morrow i am a bit busy
[17:25] <OanaStratulat> or you need some work to be submitted 1st ?
[17:26] <OanaStratulat> i have found a hang with a dxtory file
[17:26] <av500> OanaStratulat: it means the AVI file is beyond hope
[17:26] <michaelni> extended
[17:29] <OanaStratulat> av500: meaning it cannot be opened ?
[17:29] <av500> what file is that?
[17:29] <OanaStratulat> a dxtory file 
[17:29] <av500> what are you doing?
[17:29] <av500> whatever that is
[17:30] <av500> ah ok
[17:30] <av500> so its likely HUGE and therefore uses ODML
[17:31] <OanaStratulat> av500: ptp-roman.org/dxtory_crash here is the file
[17:31] <av500> #define MAX_ODML_DEPTH 1000
[17:31] <av500> 1000 nested indexes?
[17:32] <av500> that files is only 1.3MB
[17:32] <OanaStratulat> it is a fuzzed file
[17:32] <av500> ah
[17:32] <av500> so what do you expet?
[17:32] <av500> expect?
[17:32] <OanaStratulat> but i wanted it to crash, to have a bug to fix..
[17:33] <OanaStratulat> but i did not understand the output 
[17:33] <av500> well, I would check the return value of read_braindead_odml_indx
[17:33] <av500> inside read_braindead_odml_indx()
[17:34] <OanaStratulat> thx, ok i`m leaving see you later
[17:36] <ubitux> that was quick ;)
[17:54] <mkoch> hiho! sorry if it's a bit off on this channel. is it possible to compile an universal binary from ffmpeg on os x lion? I need the static libs  in a plugin which should also be compatible with both 32 and 64bit archs
[18:16] <j-b> bcoudurier: hi!
[18:16] <j-b> bcoudurier: http://mailman.videolan.org/pipermail/vlc-devel/2012-January/084719.html
[18:28] <bcoudurier> hey j-b
[18:28] <bcoudurier> happy new year
[18:28] <j-b> happy new year to you the same
[18:28] <j-b> bcoudurier: you might get interested for an invite to the above... or maybe not
[18:29] <bcoudurier> count me in
[18:31] <j-b> bcoudurier: mail me or courmisch then :)
[18:32] <bcoudurier> ok
[20:37] <ubitux> bcoudurier: are you ok with the current timecode patchset for now? i will think of a better public api in lavu ASAP; but i'd like to get done with the most of the features so we can move to a cleaner api then
[20:41] <bcoudurier> it's ok but this should have been done earlier
[20:41] <CIA-101> ffmpeg: 03Marton Balint 07master * r5d94f28e9e 10ffmpeg/ffplay.c: 
[20:41] <CIA-101> ffmpeg: ffplay: honor SDL_AUDIO_CHANNELS and make sure to use SDL supported number of audio channels
[20:41] <CIA-101> ffmpeg: Fixes ticket #838.
[20:41] <CIA-101> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[20:41] <CIA-101> ffmpeg: 03Marton Balint 07master * r8ee77fc8d4 10ffmpeg/ffplay.c: 
[20:41] <CIA-101> ffmpeg: ffplay: fix invalid wanted_channel_layout calculation
[20:41] <CIA-101> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[20:41] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r4d5bbbadd7 10ffmpeg/: 
[20:41] <CIA-101> ffmpeg: Merge remote-tracking branch 'cus/stable'
[20:41] <CIA-101> ffmpeg: * cus/stable:
[20:41] <CIA-101> ffmpeg:  ffplay: fix invalid wanted_channel_layout calculation
[20:41] <CIA-101> ffmpeg:  ffplay: honor SDL_AUDIO_CHANNELS and make sure to use SDL supported number of audio channels
[20:41] <CIA-101> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[20:41] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r02026d08a9 10ffmpeg/libavcodec/libvpxenc.c: 
[20:41] <CIA-101> ffmpeg: libvpxenc: Remove duplicate AVOption constants.
[20:41] <CIA-101> ffmpeg: Found-by: j at v2v.cc
[20:41] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:41] <ubitux> bcoudurier: ok thank you :)
[20:59] <CIA-101> ffmpeg: 03ami_stuff 07master * r871e2f4fac 10ffmpeg/libavcodec/pnm.c: 
[20:59] <CIA-101> ffmpeg: Support 16bpp grayscale pam decoding.
[20:59] <CIA-101> ffmpeg: Fixes ticket #881.
[20:59] <CIA-101> ffmpeg: 03ami_stuff 07master * ra44b63f6c8 10ffmpeg/libavcodec/pnm.c: 
[20:59] <CIA-101> ffmpeg: Support 48bpp pam decoding.
[20:59] <CIA-101> ffmpeg: Fixes ticket #882.
[21:00] <CIA-101> ffmpeg: 03Clément BSsch 07master * r2aa7375a98 10ffmpeg/libavcodec/ (timecode.c timecode.h version.h): timecode: add avpriv_check_timecode_rate().
[21:00] <CIA-101> ffmpeg: 03Clément BSsch 07master * radc278788b 10ffmpeg/libavformat/mov.c: mov: annotate read values in tmcd track.
[21:00] <CIA-101> ffmpeg: 03Clément BSsch 07master * r35da85562d 10ffmpeg/ (Changelog libavformat/mov.c): mov: support timecode extraction.
[21:14] <profoX`> hi.. does anyone know a good IRC channel or forum for generic question about video encoding? (currently developing small video encoder/decoder  for a small project which isn't related to ffmpeg, but I don't know any good resources)
[21:34] <ubitux> profoX`: doom9
[21:35] <ubitux> animemusicvideos.org too maybe
[21:37] <profoX`> ubitux: yea I was looking at doom9 but I couldn't really find a good place to post my question.. perhaps "New and alternative video codecs" I suppose.. thanks for the answer
[21:53] <CIA-101> ffmpeg: 03Clément BSsch 07master * r96fc1f0733 10ffmpeg/libavfilter/af_volume.c: volume: remove duplicated condition.
[22:25] <CIA-101> ffmpeg: 03Reimar Döffinger 07master * r1d0d63052b 10ffmpeg/libavcodec/sipr.c: (log message trimmed)
[22:25] <CIA-101> ffmpeg: sipr: fall back to setting mode based on bit_rate.
[22:25] <CIA-101> ffmpeg: Not all applications (e.g. MPlayer) set block_align, and
[22:25] <CIA-101> ffmpeg: when using a different demuxer it might not even be
[22:25] <CIA-101> ffmpeg: easily available.
[22:25] <CIA-101> ffmpeg: So fall back to selecting mode based on bit rate as before
[22:25] <CIA-101> ffmpeg: if block_align has not useful value.
[22:25] <CIA-101> ffmpeg: 03Reimar Döffinger 07master * r4278023240 10ffmpeg/libavcodec/vqavideo.c: 
[22:25] <CIA-101> ffmpeg: Check for overread in vqa video decoder.
[22:25] <CIA-101> ffmpeg: This issue was discovered while decoding the FATE sample vqa/ws_snd.vqa.
[22:25] <CIA-101> ffmpeg: For some unknown reason only audio decoding is tested by FATE for that file,
[22:25] <CIA-101> ffmpeg: but not video.
[22:25] <CIA-101> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[22:49] <durandal_1707> is -atag -vtag etc meant to be useful at all? it doesn't work with avi,mkv,mov
[23:02] <bcoudurier> it does work with mov
[23:02] <OanaStratulat> michaelni: ping
[23:02] <michaelni> durandal_1707, it should be usefull and working, if not open a bug report ill fix it
[23:02] <michaelni> OanaStratulat, pong
[23:03] <OanaStratulat> michaelni: ptp-roman.org/dxtory_crash this files makes a hang in ffmpeg
[23:04] <OanaStratulat> michaelni: can it be considered a bug?
[23:07] <michaelni> ffmpeg -i ~/videos/dxtory_crash  -f null -
[23:07] <michaelni> does not hang nor crash here
[23:07] <OanaStratulat> what is your output ?
[23:08] <OanaStratulat> can you pastebin it?
[23:11] <michaelni> http://pastebin.com/v1egeAtb
[23:12] <OanaStratulat> michaelni: i get [avi @ 0x7f7f5c03d800] Too deeply nested ODML indexes mean
[23:12] <OanaStratulat> now cloning again the ffmpeg
[23:12] <OanaStratulat> maybe some fixes were sent when i was away
[23:18] <OanaStratulat> michaelni: it hang on my ffmpeg and it is cloned right now..
[23:18] <OanaStratulat> ubitux: ping
[23:18] <ubitux> yup?
[23:19] <OanaStratulat> ptp-roman.org/dxtory_crash this files makes a hang in ffmpeg, can you confirm ?
[23:22] <ubitux> no issue there
[23:22] <OanaStratulat> mhmh
[23:22] <OanaStratulat> my ffmpeg goes wrong with it
[23:23] <OanaStratulat> av500: confirmed it this afternoon
[23:23] <ubitux> valgrind notices nothing with ./ffmpeg -i dxtory_crash -f null -
[23:24] <OanaStratulat> ubitux: try to download the file again
[23:24] <OanaStratulat> and check with it
[23:24] <OanaStratulat> i have modified it
[23:25] <ubitux> same
[23:25] <OanaStratulat> ok so need to find another one...
[23:25] <OanaStratulat> thx anyway
[23:51] <cbsrobot> ubitux: ping
[23:52] <ubitux> cbsrobot: yup?
[23:52] <cbsrobot> whas the difference between "frame number format" and QT timecode ?
[23:53] <OanaStratulat> ubitux: can you check ptp-roman.org/crash.mp4 if fails for you ?
[23:55] <ubitux> cbsrobot: http://developer.apple.com/library/mac/documentation/QuickTime/QTFF/qtff.pdf
[23:55] <ubitux> cbsrobot: check page 163 for QT timecode
[23:55] <ubitux> frame number format is just the frame number written on 32 bytes
[23:56] <ubitux> OanaStratulat: yes, i have an assert there
[23:56] <OanaStratulat> ubitux: can you say in more normal terms what an assert is ?
[23:56] <ubitux> "it should not happen, but let's check anyway"
[23:57] <ubitux> and actually here, it happens, so it doesn't crash but miserably fail
[23:58] <ubitux> michaelni: don't forget to update RELEASE :p
[23:58] <OanaStratulat> can it be considered a bug?
[23:58] <ubitux> michaelni: you should write a release checklist ;)
[23:58] <ubitux> OanaStratulat: yes
[23:59] <OanaStratulat> when i bt in gdb
[23:59] <OanaStratulat> i get 3 lines, the last line is where the error started?
[00:00] --- Fri Jan  6 2012