[Ffmpeg-devel-irc] ffmpeg-devel.log.20120618

[burek]

[02:37] <michaelni> haiku doesnt support ulimit -t 1234 ? or am i stupid ?
[02:45] Action: Daemon404 wonders if haiku is actually use by anyone but its devs
[02:45] <Daemon404> used*
[09:03] <CIA-119> ffmpeg: 03Carl Eugen Hoyos 07master * r3b3150ec9b 10ffmpeg/libavcodec/utvideo.c: 
[09:03] <CIA-119> ffmpeg: Fix interlaced prediction for RGB utvideo.
[09:03] <CIA-119> ffmpeg: Fixes ticket #1215
[09:58] <CIA-119> ffmpeg: 03Michael Bradshaw 07master * r2243f0d078 10ffmpeg/libavformat/ (avformat.h options.c utils.c version.h): 
[09:58] <CIA-119> ffmpeg: lavf: add field for how duration is estimated
[09:58] <CIA-119> ffmpeg: Signed-off-by: Michael Bradshaw <mbradshaw at sorensonmedia.com>
[09:58] <CIA-119> ffmpeg: 03Nicolas George 07master * r8c3514647e 10ffmpeg/ffmpeg.c: 
[09:58] <CIA-119> ffmpeg: ffmpeg.c: count subtitles in total size.
[09:58] <CIA-119> ffmpeg: This avoids the "Output file is empty" warning
[09:58] <CIA-119> ffmpeg: with subtitles-only files.
[09:58] <CIA-119> ffmpeg: 03Nicolas George 07master * rc49e960a60 10ffmpeg/ (doc/ffprobe.xsd ffprobe.c): ffprobe: dump AVPacket.convergence_duration.
[09:58] <CIA-119> ffmpeg: 03Nicolas George 07master * r9ae3e45521 10ffmpeg/ (Changelog doc/ffprobe.texi ffprobe.c): ffprobe: add -show_data option.
[16:02] <Vardan> hi all
[16:03] <Vardan> people does ffmpeg has frame buffering utils?
[16:03] <Vardan> I want to buffer my AVFrames
[16:08] <av500> no
[16:10] <Vardan> where can I found usage documentation about libavfilter/bufferqueue.h ?
[16:11] <av500> in there
[16:14] <Compn> the source and developer documentation
[16:14] <ubitux> there is a fifo system
[16:34] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r3a0a2f33a6 10ffmpeg/libavutil/qsort.h: 
[16:34] <CIA-119> ffmpeg: libavutil: add AV_QSORT()
[16:34] <CIA-119> ffmpeg: about 2-5 times faster than gnu libcs qsort()
[16:34] <CIA-119> ffmpeg: And should be 100% binary identical across platforms.
[16:34] <CIA-119> ffmpeg: I will bump the version once the API is certainly stable and
[16:34] <CIA-119> ffmpeg: everyone is happy with the API.
[16:34] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:40] <ubitux> michaelni: heh, nice :)
[16:45] <ubitux> michaelni: so in case of equality, is the original order kept?
[16:45] <michaelni> ubitux, not at the moment :(
[16:45] <ubitux> ok :)
[16:46] <durandal_1707> ???
[16:46] <ubitux> michaelni: and you benchmarked it with how many elements?
[16:48] <ubitux> michaelni: any reason to call it AV_QSORT instead of just AV_SORT?
[16:49] <ubitux> (nit: we use "Copyright" instead of "copyright")
[16:51] <durandal_1707> how to handle audio frame in middle of video frame?
[16:52] <durandal_1707> seek or copy ?
[16:53] <durandal_1707> this is for demuxer (obviously)
[16:53] <av500> so video frame has 2 parts?
[16:54] <av500> well, formats like TS have audio and video in small chunks
[16:54] <durandal_1707> nope, just video related chunks may be before and after audio chunks...
[16:54] <michaelni> ubitux, 10000
[16:55] <ubitux> ok
[16:55] <ubitux> do you think it's really worth writing a macro?
[16:55] <ubitux> i mean it avoids the need of a cast for the cmp function
[16:55] <ubitux> but that's all
[16:56] <ubitux> the code gets quite big, maybe it's not a good idea to have it duplicated too much
[16:56] <Vardan> people how to get the duplication of AVFrame?
[16:58] <durandal_1707> AVFrame 1 = AVFrame 2
[16:59] <Vardan> I have pointer of AVFrame
[17:05] <michaelni> ubitux, if speed matters it has to be a macro if it doesnt matter just use glibcs qsort
[17:06] <michaelni> we could try a inline function too instead of a macro but that would make it less predictable what gcc will do
[17:08] <ubitux> huh, the macro helps the perf here?
[17:12] <michaelni> there are lots of calls to that comparission function if its not a macro
[17:45] <durandal_1707> so do you guys have any comments for smush patches?
[17:46] <ubitux> michaelni: oh indeed the cmp func is most likely inlined, i forgot this
[18:11] <Daemon404> nevcairiel, hows that upstreaming going ;)
[19:07] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r88f87ebaa3 10ffmpeg/libavutil/qsort.h: 
[19:07] <CIA-119> ffmpeg: qsort: add some forgotten ()
[19:07] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:07] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r096db654af 10ffmpeg/libavutil/qsort.h: 
[19:07] <CIA-119> ffmpeg: qsort doxy
[19:07] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:07] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rf87dacb27d 10ffmpeg/libavutil/qsort.h: 
[19:07] <CIA-119> ffmpeg: libavutil: add a merge sort.
[19:07] <CIA-119> ffmpeg: compared to qsort this is slower but its stable and doesnt have a O(n^2) worst
[19:07] <CIA-119> ffmpeg: case
[19:07] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * r634e874de5 10ffmpeg/libavformat/amr.c: 
[20:25] <CIA-119> ffmpeg: amr: More space cleanup
[20:25] <CIA-119> ffmpeg: This was missed in the previous cleanup patch.
[20:25] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * re9ef88fbd2 10ffmpeg/libavformat/rtpenc.c: 
[20:25] <CIA-119> ffmpeg: rtpenc: Fix memory leaks in the muxer open function
[20:25] <CIA-119> ffmpeg: Also return a proper error code in these cases.
[20:25] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * reb564b23a3 10ffmpeg/libavformat/http.c: 
[20:25] <CIA-119> ffmpeg: http: Fail reading if the connection has gone away
[20:25] <CIA-119> ffmpeg: This can happen if doing a new request using the same socket,
[20:25] <CIA-119> ffmpeg: but the new request failed, which clears the urlcontext.
[20:25] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * reb3918c1fd 10ffmpeg/libavformat/amr.c: 
[20:25] <CIA-119> ffmpeg: amr: Mark an array const
[20:25] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:25] <CIA-119> ffmpeg: 03Ronald S. Bultje 07master * rd9669eab0b 10ffmpeg/libavcodec/ (7 files in 2 dirs): 
[20:25] <CIA-119> ffmpeg: dwt: remove variable-length arrays
[20:25] <CIA-119> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * r35127bf156 10ffmpeg/libavformat/http.c: 
[20:25] <CIA-119> ffmpeg: http: Properly handle chunked transfer-encoding for replies to post data
[20:25] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:25] <CIA-119> ffmpeg: 03Martin Storsjö 07master * rbecc3c629b 10ffmpeg/configure: 
[20:25] <CIA-119> ffmpeg: configure: Sort the library listings in the help text alphabetically
[20:26] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[20:26] <CIA-119> ffmpeg: 03Mans Rullgard 07master * r7457ebee3a 10ffmpeg/tests/fate/vorbis.mak: 
[20:26] <CIA-119> ffmpeg: fate: vorbis: add 5.1 surround test
[20:26] <CIA-119> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[20:26] <CIA-119> ffmpeg: 03Martin Storsjö 07master * rd77f4afa98 10ffmpeg/libavformat/ (rtpenc.h rtpenc_h264.c sdp.c version.h): 
[20:26] <CIA-119> ffmpeg: rtpenc: Allow requesting H264 RTP packetization mode 0
[20:26] <CIA-119> ffmpeg: This requires all NAL units to fit within single RTP packets. It
[20:26] <CIA-119> ffmpeg: doesn't change the actual packetization for packets that fit, but
[20:26] <CIA-119> ffmpeg: errors out and gives a helpful hint if the NAL units would have to
[20:26] <CIA-119> ffmpeg: be split, and signals the right packetization mode in the SDP.
[20:26] <CIA-119> (27 lines omitted)
[21:51] Action: Daemon404 gives nevcairiel a big ol' prod
[22:03] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rda42f56985 10ffmpeg/libavcodec/asv1.c: 
[22:03] <CIA-119> ffmpeg: asv1: switch to av_assert
[22:03] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:03] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r4b5c7d4d9a 10ffmpeg/libavcodec/huffyuv.c: 
[22:03] <CIA-119> ffmpeg: huffyuv: switch to av_assert
[22:03] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:15] <ubitux> does anyone know why Ronald isn't sending his patches to libav-devel?
[22:15] <ubitux> it looks like it's Måns cherry-picking from his repo or something
[22:18] <iive> ubitux:  hum?
[22:19] <ubitux> http://lists.libav.org/pipermail/libav-devel/2012-June/029280.html
[22:25] <Compn> ubitux : possibly because ronald isnt working on msvc at the moment
[22:25] <Compn> but that he asked mans to review them ?
[22:25] <Compn> all guesses...
[22:25] <ubitux> it doesn't look like that way
[22:25] <ubitux> Ronald looked pretty angry in the other thread :/
[22:26] <Compn> which mail ?
[22:26] Action: Compn not reading
[22:26] <llogan> bcoudurier: do you still admin the mailing lists on occasion?
[22:28] <ubitux> Compn: starts here: http://lists.libav.org/pipermail/libav-devel/2012-June/029138.html
[22:28] <Compn> oh i just read the mail above that , so close :)
[22:28] <Compn> hes replying to his own post
[22:29] <Compn> so he is working on msvc
[22:29] <Compn> maybe patch review is going on irc ?
[22:29] <ubitux> and looking at: http://lists.libav.org/pipermail/libav-devel/2012-June/029159.html and http://lists.libav.org/pipermail/libav-devel/2012-June/029135.html
[22:29] <Compn> or wherever they discuss things
[22:30] <ubitux> i'm wondering if he stopped contributing to libav concerning msvc stuff because of this
[22:30] <ubitux> and Måns might be trying to get things smooth by picking and reviewing his stuff in his private branch
[22:30] <ubitux> or something like that
[22:30] <ubitux> but i don't want to make assumptions :)
[22:30] <ubitux> that's why i'm asking :p
[22:30] <Compn> heh
[22:30] <Compn> why not ask ronald ?
[22:30] <Compn> like /q bbb
[22:31] <ubitux> i don't want to annoy busy men ;)
[22:31] Action: Compn bangs head against wall
[22:31] <ubitux> :D
[22:31] Action: ubitux stops annoying Compn 
[22:31] <ubitux> you look busy
[22:32] <Compn> whoa master thesis h264-mvc
[22:32] <Compn> sounds complicated 
[22:33] <Compn> ubitux : well you say you dont want to guess, but then you ask people who can only guess...
[22:33] <Compn> not important
[22:33] <Compn> just funny
[22:33] <ubitux> :)
[22:41] <kierank> Compn: i think that's a guy i spoke to a few days ago
[22:44] <kierank> oh it's not apparently
[00:00] --- Tue Jun 19 2012