[Ffmpeg-devel-irc] ffmpeg-devel.log.20120519

Sun May 20 02:05:03 CEST 2012

[00:17] <michaelni> ubitux, thx, ill push a fix for this in a moment
[00:24] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r48a40e09c4 10ffmpeg/ffmpeg.c: 
[00:24] <CIA-105> ffmpeg: ffmpeg: fix warning & bug for keep_pix_fmt
[00:24] <CIA-105> ffmpeg: Found-by: ubitux
[00:24] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:57] <ubitux> saste!
[00:57] <saste> hey clement
[00:57] <ubitux> :)
[00:57] <saste> i'm almost posting my showwaves  transmedia filter
[00:58] <ubitux> haha i was going to say i'm again on the spectrum filter
[00:58] <saste> :)
[00:58] <ubitux> btw, did you see Luca's avprobe json patch?
[00:58] <saste> then i'll tackle again the pesky projectm wrapper
[00:58] <ubitux> ah good idea:)
[00:59] <saste> unless i'll fancy to do some cool math and devise some freak transform on my own
[01:00] <ubitux> :)
[01:00] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r01e98b1b00 10ffmpeg/avconv.c: 
[01:00] <CIA-105> ffmpeg: avconv: fix behavior with -ss as an output option.
[01:00] <CIA-105> ffmpeg: Don't return from poll_filters() immediately, there may be other frames
[01:00] <CIA-105> ffmpeg: and/or other streams to handle.
[01:00] <CIA-105> ffmpeg: Fixes a memleak.
[01:00] <CIA-105> ffmpeg: 03Martin Storsjö 07master * r4b7304e80d 10ffmpeg/libavformat/rtmpproto.c: 
[01:00] <CIA-105> ffmpeg: rtmp: Don't assume path points to a string of nonzero length
[01:00] <CIA-105> ffmpeg: If using the new -rtmp_app and -rtmp_playpath parameters,
[01:00] <CIA-105> ffmpeg: one can in many cases set the main url to just rtmp://server/.
[01:00] <CIA-105> ffmpeg: If the trailing slash is omitted, path is a string of zero length,
[01:00] <CIA-105> ffmpeg: and using path+1 will end up reading uninitialized data.
[01:00] <CIA-105> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:00] <CIA-105> ffmpeg: 03Mans Rullgard 07master * rab5c38ec45 10ffmpeg/tests/regression-funcs.sh: 
[01:00] <CIA-105> ffmpeg: fate: remove unused code from regressions-funcs.sh
[01:00] <CIA-105> ffmpeg: The only uses of the do_avconv_nomd5() function were removed in
[01:00] <CIA-105> ffmpeg: 85cf49f.
[01:00] <CIA-105> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[01:00] <CIA-105> ffmpeg: 03Diego Biurrun 07master * r4cc8e15371 10ffmpeg/tests/fate/ac3.mak: 
[01:00] <CIA-105> ffmpeg: fate: Set FUZZ factor of (e)ac3-encode test to 3.
[01:00] <CIA-105> ffmpeg: This is necessary to account for different rounding done in the 3DNow!
[01:00] <CIA-105> ffmpeg: Assembly optimizations.
[01:00] <CIA-105> ffmpeg: 03Diego Biurrun 07master * r1afc278994 10ffmpeg/tests/fate/vorbis.mak: 
[01:00] <CIA-105> ffmpeg: fate: Set FUZZ factor of vorbis-13 test to 2.
[01:00] <CIA-105> ffmpeg: This is necessary to account for different rounding done in the 3DNow!
[01:00] <CIA-105> ffmpeg: Assembly optimizations.
[01:00] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r31dfe20dd2 10ffmpeg/: (log message trimmed)
[01:00] <CIA-105> ffmpeg: Merge remote-tracking branch 'qatar/master'
[01:00] <CIA-105> ffmpeg: * qatar/master:
[01:00] <CIA-105> ffmpeg:  doc: Replace some @file tags by more suitable markup.
[01:01] <ubitux> saste: should we consider adding a quick test suite for audio filters (until Nicolas or someone else write a atestsrc)?
[01:01] <CIA-105> ffmpeg:  fate: Set FUZZ factor of vorbis-13 test to 2.
[01:01] <CIA-105> ffmpeg:  fate: Set FUZZ factor of (e)ac3-encode test to 3.
[01:01] <CIA-105> ffmpeg:  fate: remove unused code from regressions-funcs.sh
[01:01] <CIA-105> ffmpeg: 03Diego Biurrun 07master * r90c9edba58 10ffmpeg/doc/ (general.texi platform.texi protocols.texi): doc: Replace some @file tags by more suitable markup.
[01:02] <saste> uhm, as nicolas pointed out aevalsrc uses floating point arithmetic, so its resuts are not portable
[01:02] <saste> so again maybe atestsrc is an option to consider
[01:02] <ubitux> well, adding a aformat asking for u8 should make outputs more likely the same everywhere :)
[01:02] <ubitux> oh and btw about that
[01:03] <ubitux> did you see my note about aevalsrc outputing by default u8 instead of s16?
[01:03] <ubitux> (most likely not related to aevalsrc btw)
[01:03] <saste> yes, i just need to explore that
[01:03] <saste> that area (format negotiation) changed a lot especially with the recent merges and i lost track of it
[01:04] <ubitux> ok
[01:19] <ubitux> saste: question about source filters: should we really have options for video sizes?
[01:19] <ubitux> for instance vsrc color, life and testsrc
[01:19] <saste> ubitux: why not?
[01:19] <ubitux> should we just request a size in the outlink?
[01:21] <saste> i suppose this could be possible as well
[01:21] <ubitux> it would be easier for example for ffplay to reconfigure the filterchain
[01:21] <ubitux> like "hey the user just resized the window, let's reconfigure every filter with this new resolution i'm changing in the format filter"
[01:21] <ubitux> it should be easier to do that
[01:22] <saste> what if the input changes?
[01:22] <saste> this is what happens with mpeg ts many times
[01:22] <saste> we should reconfigure even in that case
[01:23] <ubitux> how would that be related?
[01:27] <saste> also in case the user resized the window, what's the current behavior to implement?
[01:27] <saste> resize the *output* (like SDL already does, but with an inaccurate scaler)
[01:28] <saste> or resize the input source (e.g. life/testsrc)
[01:28] <ubitux> i'd say reconfigure the filtergraph to request a new resolution?
[01:28] <ubitux> (auto insert vf scale etc.)
[01:28] <ubitux> isn't it already the case?
[01:28] <saste> also the second option is not always available, for example when the input is fixed
[01:29] <saste> no right now rescaling is completely handled by SDL
[01:29] <ubitux> oh, ok.
[01:29] <saste> btw i'd really like to play with opengl or another toolkit
[01:30] <ubitux> well then reconfiguring the filtergraph with a new bufferskink configuration looks sane (maybe a bit too slow to do that as fast)
[01:31] <ubitux> arg sorry i'm too tired i don't think i'll be able to keep enough neurons to keep this conversation up
[01:31] <ubitux> see you soon saste, 'night everyone :)
[01:32] <saste> ubitux: good night!
[11:10] <CIA-105> ffmpeg: 03Nicolas George 07master * r7997fc49ef 10ffmpeg/libavfilter/video.c: lavfi: remove stray reference to planar/packed.
[11:10] <CIA-105> ffmpeg: 03Nicolas George 07master * r00be2f5726 10ffmpeg/ffmpeg.c: ffmpeg: adapt to the buffersrc API changes.
[11:10] <CIA-105> ffmpeg: 03Nicolas George 07master * raa1246ea42 10ffmpeg/libavfilter/ (Makefile buffersrc.c buffersrc.h src_buffer.c): (log message trimmed)
[11:10] <CIA-105> ffmpeg: lavfi: unify src_buffer and buffersrc.
[11:10] <CIA-105> ffmpeg: Most of the code is moved to buffersrc.c
[11:10] <CIA-105> ffmpeg: to help Git see the similarities.
[11:10] <CIA-105> ffmpeg: src_buffer.c still contains the obsolete compatibility functions.
[11:10] <CIA-105> ffmpeg: The format change detection code was removed as it has been
[11:10] <CIA-105> ffmpeg: implemented directly in ffmpeg.
[11:10] <CIA-105> ffmpeg: 03Nicolas George 07master * r9567f410aa 10ffmpeg/libavfilter/avfilter.h: 
[11:10] <CIA-105> ffmpeg: avfilter: move some fields in AVFilterLink.
[11:10] <CIA-105> ffmpeg: It restores ABI compatibility with libav.
[11:10] <CIA-105> ffmpeg: The fields that move were very recent or morally private.
[12:10] <CIA-105> ffmpeg: 03Clément BSsch 07master * reb945ab484 10ffmpeg/ffmpeg.c: 
[12:10] <CIA-105> ffmpeg: ffmpeg: misc cosmetics.
[12:10] <CIA-105> ffmpeg: Reduce diff against avconv, and change some doxycomments into normal
[12:10] <CIA-105> ffmpeg: comments (no point in doxifying ffmpeg internal fields).
[12:16] <ubitux> saste: 'seems my brain is up again
[12:16] <ubitux> in the doc/examples, the filtering codes are using a init_filters() with a parameter they ignore
[12:17] <ubitux> (and fallback on the global)
[12:17] <ubitux> should we just drop the parameter
[12:17] <ubitux> or make use of it?
[12:17] <saste> both examples?
[12:17] <ubitux> yep, the audio was copied from the video :)
[12:20] <saste> ouch was a typo
[12:20] <saste> but not the idea is that is better to use parameters rather than globals
[12:21] <ubitux> yeah i agree
[12:21] <saste> but in this case we have several globals (e.g. avfilter_graph)
[12:21] <saste> so i don't know
[12:21] <ubitux> it's not that important though :p
[12:21] <saste> the simpler the better I suppose, that in this case would mean to drop the parameter
[12:21] <ubitux> well, fixing the typo gives a smaller diff :)
[12:22] <ubitux> (because you don't need to change the caller)
[12:26] <saste> ubitux: whatever, and a direct commit should be ok
[12:26] <ubitux> ok
[12:26] <ubitux> i'm stacking it on top of the examples patchset
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r7aedfeebfb 10ffmpeg/doc/examples/Makefile: doc/examples: add libavr to libraries to avoid link failures.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r06e4b98382 10ffmpeg/Makefile: build: remove non-existant examples rule.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r136ca0e5f8 10ffmpeg/doc/examples/Makefile: 
[12:32] <CIA-105> ffmpeg: doc/examples: remove explicit rules.
[12:32] <CIA-105> ffmpeg: These rules are already in gmake builtins.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * rc6bbe3c2a7 10ffmpeg/doc/examples/Makefile: doc/examples: add debug compilation flag.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r50df68fd50 10ffmpeg/doc/examples/decoding_encoding.c: doc/examples: print operation filename in decoding_encoding.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r5e82ec9b42 10ffmpeg/doc/examples/metadata.c: 
[12:32] <CIA-105> ffmpeg: doc/examples: properly close input in metadata.
[12:32] <CIA-105> ffmpeg: Avoid some memleaks.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r60ea6b0996 10ffmpeg/libavfilter/buffersink.h: lavfi/buffersink: typo fix in a comment.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r4fd573970a 10ffmpeg/doc/examples/ (filtering_audio.c filtering_video.c): doc/examples: make use of the parameter filters_descr parameter in init_filters().
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r0cc9ee5fab 10ffmpeg/Makefile: doc/examples: install them with the shared data.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r6ad974ae1c 10ffmpeg/libavformat/id3v2.c: 
[12:32] <CIA-105> ffmpeg: lavf/id3v2: do not export empty fields.
[12:32] <CIA-105> ffmpeg: This also avoids a memleak.
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * re5fcf3646a 10ffmpeg/libavformat/id3v2.c: 
[12:32] <CIA-105> ffmpeg: lavf/id3v2: always strdup the value.
[12:32] <CIA-105> ffmpeg: This simplifies the code but also fix a warning: ff_id3v1_genre_str
[12:32] <CIA-105> ffmpeg: array contains const strings so do the string dup now instead of in
[12:32] <CIA-105> ffmpeg: av_dict_set().
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * re692b9d564 10ffmpeg/doc/examples/filtering_video.c: 
[12:32] <CIA-105> ffmpeg: doc/examples: fix pts heuristic in filtering_video.
[12:32] <CIA-105> ffmpeg: Note: the condition looked wrong in the first place ("if DTS is not
[12:32] <CIA-105> ffmpeg: defined, then set PTS to that undefined value [...]").
[12:32] <CIA-105> ffmpeg: 03Clément BSsch 07master * r6c211085bb 10ffmpeg/doc/examples/ (decoding_encoding.c muxing.c): 
[12:32] <CIA-105> ffmpeg: doc/examples: use system headers in decoding_encoding and muxing.
[12:33] <ubitux> heh fun, CIA-105 is reordering the commit by time
[13:28] <CIA-105> ffmpeg: 03Stefano Sabatini 07master * r150227e8ed 10ffmpeg/ (5 files in 2 dirs): 
[13:28] <CIA-105> ffmpeg: lavfi/asplit: move asplit code to vf_split.c, and make it support N outputs
[13:28] <CIA-105> ffmpeg: The move allows to share the init code already used by split.
[13:29] <CIA-105> ffmpeg: 03Stefano Sabatini 07master * r0e82d317eb 10ffmpeg/libavfilter/ (Makefile vf_split.c): 
[13:29] <CIA-105> ffmpeg: lavfi: rename vf_split.c -> f_split.c
[13:29] <CIA-105> ffmpeg: The file now contains both an audio and a video split filter.
[13:35] <pisto> I think there's a bug with image2pipe. After 2000 frames or so, it outputs an error (error decoding stream) and goes misaligned, and in subsequent frames I can see pieces of different images into one
[13:35] <pisto> the input codec is pbm
[13:35] <pisto> I can dump the input stream and send over, is somebody wants to take a look
[13:37] <pisto> if*
[14:15] <michaelni> pisto, if this is with latest ffmpeg then please open a issue at https://ffmpeg.org/trac/ffmpeg
[14:16] <pisto> ok
[14:19] <pisto> by the way, is image2pipe supposed to work with streamed png?
[14:19] <pisto> if so, I have another bug to report
[14:21] <michaelni> if by streamed png you mean concatenated pngs then it should yes
[14:29] <pisto> opened
[14:30] <pisto> #1321
[14:57] <pisto> it seems to fail when I stream more than exactly 1989 frames
[15:04] <pisto> oh, I just noticed that png support for image2pipe has been added 5 days after I compiled ffmpeg :D
[16:14] <pisto> I have been asked to run ffmpeg without external codec for output. is yuv4 ok?
[16:16] <michaelni> pisto, should be you could also use "-f null -" if the output is not needed
[16:36] <pisto> I compiled ffmpeg from git, it doesn't install any man page, is that normal?
[17:06] <Compn> pisto : did you make install it ?
[17:06] <pisto> sure
[17:32] <CIA-105> ffmpeg: 03Nicolas George 07master * r5ac9ef6493 10ffmpeg/libavfilter/ (avfilter.h formats.c): 
[17:32] <CIA-105> ffmpeg: avfilter: make AVFilterFormats compatible with libav.
[17:32] <CIA-105> ffmpeg: The list goes back to being simple integers,
[17:32] <CIA-105> ffmpeg: and avfilter_make_format64_list is fixed to work
[17:32] <CIA-105> ffmpeg: with the correct structure directly.
[19:28] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r186be49fd5 10ffmpeg/libswresample/swresample.c: 
[19:28] <CIA-105> ffmpeg: swr: fill_audio_data() support NULL
[19:28] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r75e25f8540 10ffmpeg/libswresample/swresample.c: 
[19:29] <CIA-105> ffmpeg: swr: add reversefill_audiodata()
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r5027e3daf7 10ffmpeg/libswresample/swresample.c: 
[19:29] <CIA-105> ffmpeg: swr: fix buf_set() so it works with input == output
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rc5c355546d 10ffmpeg/libswresample/swresample.c: 
[19:29] <CIA-105> ffmpeg: swr: add some asserts to resample()
[19:29] <CIA-105> ffmpeg: These check that the internal and input buffers match in format
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r7a59964ba9 10ffmpeg/libswresample/ (swresample.c swresample.h): 
[19:29] <CIA-105> ffmpeg: swr: add swr_inject_silence()
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * re17a039647 10ffmpeg/libavfilter/af_aresample.c: 
[19:29] <CIA-105> ffmpeg: af_aresample: use extended_data
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rf88f705abc 10ffmpeg/libswresample/ (swresample.c swresample.h swresample_internal.h): 
[19:29] <CIA-105> ffmpeg: swr: add swr_drop_output()
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r72a242c998 10ffmpeg/libswresample/ (swresample.c swresample.h swresample_internal.h): 
[19:29] <CIA-105> ffmpeg: swr: add swr_next_pts()
[19:29] <CIA-105> ffmpeg: parameter descriptions partly reuse text from af_asyncts
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r72872d0201 10ffmpeg/libswresample/swresample.c: 
[19:29] <CIA-105> ffmpeg: swr: Fix issue with injecting more data than can be used up in the next call.
[19:29] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:29] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r087d09b6d5 10ffmpeg/libavfilter/af_aresample.c: 
[19:30] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:30] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r9d30a96aa1 10ffmpeg/ffmpeg.c: 
[19:30] <CIA-105> ffmpeg: ffmpeg: inject a aresample filter for -async
[19:30] <CIA-105> ffmpeg: Based on similar code from anton
[19:30] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:30] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rbdbfd4569d 10ffmpeg/libswresample/swresample.c: 
[19:30] <CIA-105> ffmpeg: swr: make literal 32 a named constant in realloc_audio()
[19:30] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:30] <CIA-105> ffmpeg: 03Nicolas George 07master * r3d2515a8f3 10ffmpeg/libavfilter/f_split.c: 
[19:30] <CIA-105> ffmpeg: f_split: fix output type.
[19:30] <CIA-105> ffmpeg: The previous version checked the filter context name,
[19:31] <CIA-105> (3 lines omitted)
[19:56] <ubitux> michaelni: any reason not to use AUTO_INSERT_FILTER macro for async?
[19:56] <ubitux> the only missing thing i see is the value of -async not being in the av_log
[19:56] <ubitux> but that could be fixed using __VA_ARGS__ in the macro
[19:57] <ubitux> (and an extra fmt)
[19:58] <michaelni> ubitux, no reason, feel free to change it
[19:59] <ubitux> ok
[20:08] <CIA-105> ffmpeg: 03Carl Eugen Hoyos 07master * r2596d7261d 10ffmpeg/libavcodec/avuidec.c: Fix interlaced AVUI decoding.
[20:08] <CIA-105> ffmpeg: 03Carl Eugen Hoyos 07master * rf1892348c2 10ffmpeg/libavcodec/avuidec.c: avuidec.c: Fix indentation.
[20:08] <CIA-105> ffmpeg: 03Carl Eugen Hoyos 07master * rd9b2410f75 10ffmpeg/libavcodec/avuidec.c: Simplify AVUI decoding, one skip value is enough.
[20:08] <CIA-105> ffmpeg: 03Carl Eugen Hoyos 07master * r612abe2773 10ffmpeg/libavcodec/avuidec.c: Use a variable instead of a repeated calculation when decoding AVUI.
[20:17] <Daemon404> michaelni, yay new async!
[20:32] <ubitux> seems libav is now up to rewrite the amerge filter
[20:32] <ubitux> seriously..
[20:37] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r0ad26cdf24 10ffmpeg/avconv.c: avconv: add support for audio in complex filtergraphs.
[20:37] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r7b3b24a8ee 10ffmpeg/libavfilter/ (audio.c video.c): lavfi: initialize pts to AV_NOPTS_VALUE when creating new buffer refs.
[20:37] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r54c5dd89e3 10ffmpeg/ (6 files in 3 dirs): 
[20:37] <CIA-105> ffmpeg: lavfi: Add fps filter.
[20:37] <CIA-105> ffmpeg: Partially based on a patch by Robert Nagy <ronag89 at gmail.com>
[20:38] <CIA-105> ffmpeg: 03Anton Khirnov 07master * r755cd4197d 10ffmpeg/ (libavformat/mov.c tests/ref/fate/vc1-ism): 
[20:38] <CIA-105> ffmpeg: mov: enable parsing for VC-1.
[20:38] <CIA-105> ffmpeg: This makes lavf discard broken timestamps for non-B frames in
[20:38] <CIA-105> ffmpeg: samples/isom/vc1-wmapro.ism.
[20:38] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r040a796dab 10ffmpeg/: (log message trimmed)
[20:38] <CIA-105> ffmpeg: Merge commit '755cd4197d53946208e042f095b930dca18d9430'
[20:38] <CIA-105> ffmpeg: * commit '755cd4197d53946208e042f095b930dca18d9430':
[20:38] <CIA-105> ffmpeg:  mov: enable parsing for VC-1.
[20:38] <CIA-105> ffmpeg:  lavfi: Add fps filter.
[20:38] <CIA-105> ffmpeg:  lavfi: initialize pts to AV_NOPTS_VALUE when creating new buffer refs.
[20:38] <CIA-105> ffmpeg:  avconv: add support for audio in complex filtergraphs.
[20:53] <ubitux> it seems we will need to add a real INI output
[20:54] <ubitux> it looks like libav doesn't care again about users parsing the current ffprobe output
[21:24] Action: Daemon404 gives ubitux a cookie
[21:26] <ubitux> e
[21:26] <ubitux> Daemon404: sorry but i find that totally insane :p
[21:27] <Daemon404> of course.
[21:27] <Daemon404> as far as theyre concerned ffmpeg is nto somethign to take itno consideration
[21:27] <Daemon404> i dotn really have a good argument for why they should.
[21:27] <ubitux> users should be taken into consideration
[21:28] <ubitux> and it's more like "we don't want to look at ffmpeg" instead of "we ignore ffmpeg"
[21:28] <ubitux> ffprobe has writers support since a long long time ago
[21:28] <ubitux> we support json, xml, csv, etc etc
[21:28] <ubitux> and again, they're rewritting everything in a poorly form
[21:28] <ubitux> anyway, i'm sorry but right now they're just stupid
[21:29] <ubitux> they could just pick stuff and be done, but no, NIH syndrom rise back again
[21:30] <Daemon404> dotn get me started on NIH.
[21:31] <ubitux> :)
[21:31] <ubitux> if at least they were doing it in a better shape
[21:32] <ubitux> well, whatever
[21:32] <Daemon404> =p
[21:32] Action: Daemon404 is currently looking at a different type of insanity: http://japland.org/a.txt
[21:32] <ubitux> nice ascii art
[21:32] <Daemon404> clang's errors are nice
[21:32] <Daemon404> c++ is not
[21:47] <ubitux> wth is the code duplication between configure_audio_filters and configure_output_audio_filters...
[22:22] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * r43bc5cf9f4 10ffmpeg/libavcodec/mpegaudiodec.c: 
[22:22] <CIA-105> ffmpeg: mpegaudiodec: skip initial zeros.
[22:22] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:25] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rc5371f775c 10ffmpeg/libavcodec/mpegaudiodec.c: 
[22:25] <CIA-105> ffmpeg: mpegaudidec: 10l forgot buf_size
[22:25] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:31] <CIA-105> ffmpeg: 03Michael Niedermayer 07master * rd1384c00d4 10ffmpeg/libswresample/swresample.c: 
[23:31] <CIA-105> ffmpeg: swr: skip soft compensation when its disabled.
[23:31] <CIA-105> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Sun May 20 2012