[Ffmpeg-devel-irc] ffmpeg-devel.log.20120612

burek burek021 at gmail.com
Wed Jun 13 02:05:04 CEST 2012


[00:58] <CIA-119> ffmpeg: 03Stefano Sabatini 07master * r5683de00e9 10ffmpeg/libavutil/error.c: 
[00:58] <CIA-119> ffmpeg: lavu/error: fix sign error in av_sterror() error code
[00:58] <CIA-119> ffmpeg: The value returned by strerror_r may be positive depending on the system,
[00:58] <CIA-119> ffmpeg: so it is necessary to use AVERROR() in order to get a negative
[00:58] <CIA-119> ffmpeg: number. This way the successive negative check will not fail, and the
[00:58] <CIA-119> ffmpeg: function will print a meaningful error message on the buffer.
[00:58] <CIA-119> ffmpeg: 03Stefano Sabatini 07master * r3ecaba677e 10ffmpeg/ (libavutil/parseutils.c tests/ref/fate/parseutils): 
[00:58] <CIA-119> ffmpeg: tests/parseutils: change test to make it platform independent
[00:58] <CIA-119> ffmpeg: Make the test independent from the actual error string used to describe a
[00:58] <CIA-119> ffmpeg: certain error code.
[00:58] <CIA-119> ffmpeg: 03Stefano Sabatini 07master * r359abb18cf 10ffmpeg/ (4 files in 2 dirs): 
[00:58] <CIA-119> ffmpeg: lavu/error: add av_make_error_string() and av_err2str() convenience utilities
[00:58] <CIA-119> ffmpeg: These functions are modeled after the corresponding utilities in
[00:58] <CIA-119> ffmpeg: libavutil/timestamp.h.
[00:58] <CIA-119> ffmpeg: 03Stefano Sabatini 07master * r4299fd23d3 10ffmpeg/libavutil/ (Makefile error.c): 
[00:58] <CIA-119> ffmpeg: lavu/error: use a table to store error codes and strings, add test
[00:58] <CIA-119> ffmpeg: The table is mostly useful to enumerate the available AVERROR* in the
[00:58] <CIA-119> ffmpeg: test output.
[04:27] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r4289b6690d 10ffmpeg/libswresample/audioconvert.c: 
[04:27] <CIA-119> ffmpeg: swr: replace /16 by >>4
[04:27] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:34] <ubitux> michaelni: isn't this the same as off = len & ~0xf?
[07:34] <ubitux> ah, reimar pointed it out
[10:31] <michaelni> ubitux, indeed
[10:31] <michaelni> thx
[13:56] <burek> is there any particular reason why ffprobe was implemented as a separate cmd line tool and not integrated in ffmpeg?
[14:13] <ubitux> it does something else
[14:13] <ubitux> is there a problem with that?
[14:14] <ubitux> ffmpeg transcodes and transmux
[14:14] <ubitux> ffprobe is an introspection tool
[14:14] <burek> not quite, but it's a waste of resources I guess
[14:14] <burek> it does similar analysis like ffmpeg, so..
[14:17] <av500> patches welcome I guess
[14:21] <ubitux> i'm not sure it's a good idea
[14:22] <ubitux> ffprobe is already 2k+ lines of code
[14:22] <av500> cat ffprobe.c >> ffmpeg.c
[14:22] <ubitux> it has a very specific task
[14:22] <ubitux> av500: great idea
[14:22] <ubitux> :)
[14:40] <burek> :)))
[16:36] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rc8553bea9e 10ffmpeg/libswresample/audioconvert.c: 
[16:36] <CIA-119> ffmpeg: swr: use &~15
[16:36] <CIA-119> ffmpeg: Idea-by: reimar (and ubitux on IRC)
[16:36] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:36] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r634d539c12 10ffmpeg/libswresample/swresample.c: 
[16:36] <CIA-119> ffmpeg: swresample.c: update copyleft year
[16:36] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:36] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r9abbbf7530 10ffmpeg/libswresample/swresample_internal.h: 
[16:36] <CIA-119> ffmpeg: swr: add function pointers for SIMD rematrixing
[16:36] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:36] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r48a45f819f 10ffmpeg/libswresample/ (rematrix.c swresample_internal.h): 
[16:36] <CIA-119> ffmpeg: swr: add native matrix for SIMD
[16:36] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:36] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rbeb0cd6acf 10ffmpeg/libswresample/ (5 files in 2 dirs): 
[16:39] <rhlee> how do I compile ffmpeg with debug enabled for the libav libs? --enable-debug only enables symbols for the ffmpeg executables
[16:41] <rhlee> think i found it, disable stripping
[16:45] <michaelni> rhlee, its ffmpeg_g that has the debug symbols, but you wont hear that awnser as you are smart enough to log off 2 minutues after asking 
[16:45] <michaelni> iam fast but not that fast ...
[16:46] <Compn> hehe
[16:47] <Compn> good try tho michaelni
[16:47] <Compn> its a race to answer questions from those people
[16:47] <Compn> so i've learned to say their name first, then hit enter
[16:47] <Compn> then they have to wait for you to finish the sentence :)
[16:48] <av500> Compn: most of them use a browser client and dont realize they are closing a tab
[16:48] <Compn> true
[17:18] <Daemon404> michaelni, ping-o
[17:19] <Daemon404> or maybe ubitux ... someone who isnt scared to touch ffmpeg.c
[17:29] <ubitux> Daemon404: it's not your habit to ask to ask
[17:30] <Daemon404> true
[17:30] <Daemon404> basically, ffmpeg cant handle files with negative pts
[17:30] <Daemon404> avconv can
[17:46] <michaelni> Daemon404, do you have a testcase ?
[17:46] <Daemon404> michaelni, i do
[17:46] <Daemon404> let me grab it
[17:51] <Daemon404> sent
[17:51] <Daemon404> since im using my own code for this, and not ffmpeg.c, im currently dropping frames with negative pts
[17:51] <Daemon404> im unsure if this is the proper thing to do
[17:52] <Daemon404> also for some reason the dts larger than the pts in this
[17:53] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rd504266cef 10ffmpeg/libswresample/x86/ (rematrix.asm swresample_x86.c): 
[17:53] <CIA-119> ffmpeg: swr: mix_1_1_int16_sse
[17:53] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:53] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r728f86edfc 10ffmpeg/libswresample/ (rematrix.c x86/rematrix.asm x86/swresample_x86.c): 
[17:53] <CIA-119> ffmpeg: swr: mix_2_1_int16_mmx/sse
[17:53] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:19] <michaelni> Daemon404, are you aware that that file contains a data track starting an hour before the video and audio ?
[18:19] <Daemon404> i hadnt looked into the data
[18:19] <Daemon404> since we ignore it
[18:21] <Daemon404> only looked at the video packets
[19:20] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r47695b1d65 10ffmpeg/libavformat/utils.c: 
[19:20] <CIA-119> ffmpeg: lavf: handle data like subtitles in start time calculation.
[19:20] <CIA-119> ffmpeg: This fixes an issue with a crazy data track starting with a large
[19:20] <CIA-119> ffmpeg: negative timestamp.
[19:20] <CIA-119> ffmpeg: It could as well be solved in all user apps, but this is looking
[19:20] <CIA-119> ffmpeg: attractively simpler ...
[19:20] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:20] <michaelni> Daemon404, your file should be working now
[19:20] <ohsix> is there an accompanying warning so it would possibly be fixed in user apps? :D
[19:21] <michaelni> making ffmpeg consider only the starttimes from the "active" streams would be interresting too but its trickier
[19:21] <michaelni> as most of the code is on the input file side and there may be 2 output files one which uses the data stream and one that doesnt
[19:23] <michaelni> so ill take the lazy approuch here and see if someone moves the code over to the output side ...
[19:30] <Daemon404> michaelni, neat
[19:32] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r2a622c2369 10ffmpeg/libavformat/utils.c: 
[19:32] <CIA-119> ffmpeg: lavf: add a "warning" when discarding a oddly backward subtitle or data starttime
[19:32] <CIA-119> ffmpeg: Idea-by: ohsix
[19:32] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:59] <CIA-119> ffmpeg: 03Samuel Pitoiset 07master * r0a9a225733 10ffmpeg/libavformat/rtmpproto.c: 
[20:59] <CIA-119> ffmpeg: rtmp: Fix a possible access to invalid memory location when the playpath is too short.
[20:59] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:59] <CIA-119> ffmpeg: 03Anton Khirnov 07master * r47b812e9ce 10ffmpeg/avconv.c: 
[20:59] <CIA-119> ffmpeg: avconv: support only native pthreads.
[20:59] <CIA-119> ffmpeg: Our w32pthreads wrapper has various issues and is only supposed to be
[20:59] <CIA-119> ffmpeg: used in libavcodec.
[20:59] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rdc1c79be1b 10ffmpeg/: (log message trimmed)
[20:59] <CIA-119> ffmpeg: Merge remote-tracking branch 'qatar/master'
[20:59] <CIA-119> ffmpeg: * qatar/master:
[20:59] <CIA-119> ffmpeg:  avfilter: Log an error if avfilter fails to configure a link.
[20:59] <CIA-119> ffmpeg:  avconv: support only native pthreads.
[20:59] <CIA-119> ffmpeg:  rtmp: Fix a possible access to invalid memory location when the playpath is too short.
[20:59] <CIA-119> ffmpeg: Conflicts:
[20:59] <CIA-119> ffmpeg: 03Alex Converse 07master * r65a80ee1ec 10ffmpeg/libavfilter/avfilter.c: avfilter: Log an error if avfilter fails to configure a link.
[21:52] <CIA-119> ffmpeg: 03Paul B Mahol 07master * ra7fbc5d004 10ffmpeg/libavcodec/libvorbis.c: 
[21:52] <CIA-119> ffmpeg: libvorbis: cosmetics: remove empty line at EOF
[21:52] <CIA-119> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[22:56] <CIA-119> ffmpeg: 03Paul B Mahol 07master * rf06269dda5 10ffmpeg/libavcodec/ (8 files): 
[22:56] <CIA-119> ffmpeg: cosmetics: align more codecs declarations
[22:56] <CIA-119> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[23:09] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r2942431a0f 10ffmpeg/libavutil/arm/intmath.h: 
[23:09] <CIA-119> ffmpeg: arm/intmath: disable av_clip_uintp2_arm with config_small
[23:09] <CIA-119> ffmpeg: The code otherwise breaks compilation (gcc 4.5 tested but probably any compiler)
[23:09] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Wed Jun 13 2012


More information about the Ffmpeg-devel-irc mailing list