[Ffmpeg-devel-irc] ffmpeg-devel.log.20120407

Sun Apr 8 02:05:03 CEST 2012

[00:12] <ubitux> it seems threading failed again
[00:12] <ubitux> (http://fate.ffmpeg.org/report.cgi?time=20120406212225&slot=x86_64-archlinux-gcc-threads)
[02:03] <michaelni> ubitux, can you bisect this easily/automatically in case its a regression ? (some bug IIRC also pointed to a thread regression that only happens sometimes)
[02:03] <michaelni> like maybe running the failing test 10 times and if all pass consider it a pass for git bisect
[02:04] <michaelni> anyway, i need to sleep a bit, ill read read the irclog tomorrow
[02:05] <ubitux> it will take quite a while to run this i guess
[02:05] <ubitux> i don't know if it's a regression, i setup the threads instances recentrly
[02:05] <ubitux> recently*
[02:06] <iive> is there something that could be done to easy tracking threading problems?
[02:06] <seriousman> valgrind's tools?
[02:07] <iive> no... i mean, during normal operation. like lock holding blocking for a second or more...
[02:07] <iive> something that would give useful debug information even when not specifically debugging. 
[02:07] <seriousman> looks like a good idea
[02:08] <seriousman> divert pthread_mutex_lock() with a variant that calls usleep with random values?
[02:09] <iive> i'm not familiar with pthread functions, isn't there some with a timeout?
[02:10] <seriousman> not for locks
[02:10] <seriousman> i meant to use dlopen tricks
[02:11] <seriousman> e.g. pthread_mutex_lock(pthread_mutex_t *m) { real_pthread_mutex_lock(m); usleep(random()); }
[02:14] <kierank> just write a macro to measure how long each lock takes
[02:14] <kierank> and then printf something
[02:18] <iive> from the man page PTHREAD_MUTEX_ERRORCHECK looks like nice thing to try
[02:25] <iive> seriousman: i think i was more like using try_lock() and a sleep if it failed, so a check could be done after few iterations.
[02:26] <iive> hum, there is timedlock() what does exactly what I want.
[02:27] <iive> what/that
[02:27] <iive> time to sleep.
[02:27] <iive> have fun.
[07:07] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rb717f35bfb 10ffmpeg/tests/lavf-regression.sh: 
[07:07] <CIA-17> ffmpeg: fate: try to fix >8bit dpx reg test
[07:07] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[08:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rab553612e7 10ffmpeg/libavcodec/put_bits.h: 
[08:18] <CIA-17> ffmpeg: put_bits: switch assert to av_assert2()
[08:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[08:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3fb945a935 10ffmpeg/libavformat/mpegenc.c: 
[08:18] <CIA-17> ffmpeg: mpegpsenc: fix user specified mux rate
[08:18] <CIA-17> ffmpeg: Fixes Ticket1175
[08:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:27] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rfe5b7db07a 10ffmpeg/configure: 
[11:27] <CIA-17> ffmpeg: build system: support setting avassert() assert_level
[11:27] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:27] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rd40132dab3 10ffmpeg/libavcodec/proresenc_kostya.c: 
[11:27] <CIA-17> ffmpeg: proresenc_kostya: fix encoded mantissa.
[11:27] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:28] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * re3b7079ddf 10ffmpeg/ (libavcodec/adpcmenc.c tests/ref/acodec/adpcm_ima_qt): 
[11:28] <CIA-17> ffmpeg: adpcmenc: fix encoded s_0 value.
[11:28] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:28] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r76e8b336cd 10ffmpeg/libavcodec/proresenc_kostya.c: 
[11:28] <CIA-17> ffmpeg: proresenc_kostya: Mention likely ancestry in the license header.
[11:28] <CIA-17> ffmpeg: As the encoder contained the same bug and has similar structure
[11:28] <CIA-17> ffmpeg: to anatoliys encoder, it is possibly based on it.
[11:28] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:28] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rdd7d08ee89 10ffmpeg/libavcodec/proresenc_anatoliy.c: 
[11:28] <CIA-17> ffmpeg: proresenc_anatoly: fix encoded mantissa.
[11:28] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:54] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * recd7455e96 10ffmpeg/libavcodec/ (aaccoder.c aacenc.c): 
[11:54] <CIA-17> ffmpeg: aacenc: Fix issues with huge values of bit_rate.
[11:54] <CIA-17> ffmpeg: Do not pointlessly call ff_alloc_packet2 multiple times,
[11:54] <CIA-17> ffmpeg: and fix an infinite loop by clamping the maximum
[11:54] <CIA-17> ffmpeg: number of bits to target in the algorithm that does
[11:54] <CIA-17> ffmpeg: not use lambda.
[11:54] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[11:54] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r796097c696 10ffmpeg/tests/ (6 files in 2 dirs): (log message trimmed)
[11:54] <CIA-17> ffmpeg: Add tests for audio encoders: aac, dca, ra144 and nellymoser.
[11:54] <CIA-17> ffmpeg: The PSNR values are of varying usefulness, though at least
[11:54] <CIA-17> ffmpeg: the DTS and AAC ones are useful with the right shift value.
[11:54] <CIA-17> ffmpeg: Note: due to usage of floats some of these may fail on other
[11:55] <CIA-17> ffmpeg: architectures.
[11:55] <CIA-17> ffmpeg: In that case they should be converted into a CMD = stddev
[12:16] <btsimonh> hi, trying to use ffmpeg to capture/forward webcam on openwrt, but not getting any output file.  Anyone any experience of ffmpeg in small environments?
[12:44] <michaelni> btsimonh, what are you doing exactly, i mean command line, error messages, ...
[12:45] <btsimonh> i think it's something to to with the build, but all looks ok.
[12:46] <btsimonh> let me dig out the simplest example....
[12:48] <btsimonh> e.g. ffmpeg -f video4linux2 -i /dev/video0 /tmp/video.avi - trouble now is i've just done a rebuild, and now i don't even get any stdout feedback :(, but also get no error and no output file?
[12:54] <btsimonh> trying to get back to where i was..  just rebooting the router
[13:04] <btsimonh> @michaelni: http://pastebin.com/4FJ0uc7t - output of ffmpeg on openwrt
[13:05] <btsimonh> mjpg-streamer runs, and produces video no problem.  when mjpg-streamer is running, ffmpeg can't get video, when it's not, it 'seems' to be working (but low cpu usage...), but after ctrl-c, no file produced.
[13:09] <btsimonh> (i wanted to use ffpmeg to stream video and audio from this webcam to a distribution website; mjpg-streamer is great, but my upload speed won't support more than one user)
[13:25] <michaelni> btsimonh, i suggest you try latest ffmpeg
[13:26] <michaelni> and if it still fails try to run it under gdb and see where it hangs
[13:26] <btsimonh> what's it on at the moment, verison-wise?
[13:27] <michaelni> latest release in is topic :)
[13:30] <btsimonh> 0.10.2 :) i'll see if i can update the source in my openwrt tree later, a bit scary going that many versions forward :)
[13:31] <ubitux> releases are done more often than previously
[13:31] <ubitux> it's not such a huge "jump"
[13:47] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r10e3dd15a5 10ffmpeg/tests/ (codec-regression.sh ref/acodec/aac ref/acodec/nellymoser): 
[13:47] <CIA-17> ffmpeg: Remove AAC and nellymoser tests that do not work.
[13:47] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[13:47] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r6d3166fcf0 10ffmpeg/tests/ (Makefile fate/aac.mak): 
[13:47] <CIA-17> ffmpeg: Add AAC FATE encode tests.
[13:47] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[13:56] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r087b9c326f 10ffmpeg/tests/lavf-regression.sh: 
[13:56] <CIA-17> ffmpeg: fate: try to force le for the dpx decoder side too
[13:56] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:59] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r2f83681c79 10ffmpeg/ (11 files in 3 dirs): 
[13:59] <CIA-17> ffmpeg: lavfi: port libmpcodecs remove-logo filter
[13:59] <CIA-17> ffmpeg: The code is based on the remove-logo filter in MPlayer/libmpcodecs, by
[13:59] <CIA-17> ffmpeg: Robert Edele, relicensed to LGPL with consent of the author.
[13:59] <CIA-17> ffmpeg: Address trac issue #249.
[13:59] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r6cb01c9dba 10ffmpeg/ (5 files in 3 dirs): 
[13:59] <CIA-17> ffmpeg: lavfi/mp: remove remove_logo wrapper
[13:59] <CIA-17> ffmpeg: removelogo has been ported to libavfilter, and the wrapper was not
[13:59] <CIA-17> ffmpeg: working.
[13:59] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r32b1022f76 10ffmpeg/doc/filters.texi: doc/filters: remove reference to removed delogo wrapper
[13:59] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * rfd00662d5a 10ffmpeg/doc/muxers.texi: 
[13:59] <CIA-17> ffmpeg: doc/muxers: update documentation for the framecrc muxer
[13:59] <CIA-17> ffmpeg: Add mention to the pts and duration log fields, added in
[13:59] <CIA-17> ffmpeg: d2afbd9a56e4e863bad01923006dd9b77f747035.
[13:59] <CIA-17> ffmpeg: Fix trac ticket #1159.
[15:13] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * rcb25a186eb 10ffmpeg/tests/ (codec-regression.sh ref/acodec/dca regression-funcs.sh): 
[15:13] <CIA-17> ffmpeg: Fix DCA regression test to work across architectures.
[15:13] <CIA-17> ffmpeg: Only the decoding step is not bit-exact, so avoid checksum for that.
[15:13] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[15:16] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r82ca1f8004 10ffmpeg/tests/ (lavf-regression.sh ref/lavf-fate/mp3): 
[15:16] <CIA-17> ffmpeg: Test mp3 container muxing (and thus ID3 writing).
[15:16] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[16:35] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r4395c058d2 10ffmpeg/ (libavcodec/dcaenc.c tests/codec-regression.sh): 
[16:35] <CIA-17> ffmpeg: dcaenc: Allow encoding without specifying a channel_layout.
[16:35] <CIA-17> ffmpeg: It will print a warning, making the behaviour consistent
[16:35] <CIA-17> ffmpeg: with the AC3 encoder.
[16:35] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[16:35] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * re80ce4adea 10ffmpeg/tests/ (lavf-regression.sh ref/lavf-fate/mp3): 
[16:35] <CIA-17> ffmpeg: Do not decode the mp3 data in lavf test since that is not bit-exact.
[16:36] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r192efcf768 10ffmpeg/libavcodec/h264_ps.c: 
[20:57] <CIA-17> ffmpeg: h264_ps: check log2_max_frame_num for validity.
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r6560fa390e 10ffmpeg/libavcodec/get_bits.h: 
[20:57] <CIA-17> ffmpeg: get_bits: check the number of bits parameter through av_assert2()
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rd3b0fb1048 10ffmpeg/libavcodec/put_bits.h: 
[20:57] <CIA-17> ffmpeg: put_bits: add av_assert2() to check out of array writes.
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r6c249392fa 10ffmpeg/libavcodec/mpeg4videodec.c: 
[20:57] <CIA-17> ffmpeg: mpeg4videodec: make sure f/b_code are not invalid values.
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r20676711fe 10ffmpeg/libavcodec/vc1.c: 
[20:57] <CIA-17> ffmpeg: vc1: fix incorrect show_bits() usage.
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:57] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r6c3d6a214c 10ffmpeg/libavcodec/alsdec.c: 
[20:57] <CIA-17> ffmpeg: alsdec: Check k used for rice decoder.
[20:57] <CIA-17> ffmpeg: Values that fail this check will cause failure of decode_rice()
[20:57] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:11] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r5012c07336 10ffmpeg/ (4 files in 2 dirs): (log message trimmed)
[22:11] <CIA-17> ffmpeg: Support detecting and demuxing EIA-608 subtitles in mov.
[22:11] <CIA-17> ffmpeg: The format is slightly proprietary.
[22:11] <CIA-17> ffmpeg: DVDs use a format of
[22:11] <CIA-17> ffmpeg: code byte (0x00, 0x01, 0xfe or 0xff), two data bytes
[22:11] <CIA-17> ffmpeg: MOV uses instead
[22:11] <CIA-17> ffmpeg: cdat/cdt2 atom, two data bytes
[23:03] <CIA-17> ffmpeg: 03Joakim Plate 07master * rba24f12982 10ffmpeg/libavformat/utils.c: 
[23:03] <CIA-17> ffmpeg: libavformat: Only require first packet to be known for audio/video streams
[23:03] <CIA-17> ffmpeg: It can take a long time before subtitles or data streams show up,
[23:03] <CIA-17> ffmpeg: so we shouldn't wait for those before assuming we have all info
[23:03] <CIA-17> ffmpeg: for streams.
[23:03] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[23:03] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r4648fe6324 10ffmpeg/avconv.c: avconv: remove now pointless variables.
[23:03] <CIA-17> ffmpeg: 03Anton Khirnov 07master * re06d2dfcde 10ffmpeg/ (Changelog avconv.c configure): 
[23:03] <CIA-17> ffmpeg: avconv: drop support for building without libavfilter.
[23:03] <CIA-17> ffmpeg: Since the mandatory memcpy in vsrc_buffer has been eliminated, there
[23:03] <CIA-17> ffmpeg: shouldn't be any significant reason to build without lavfi anymore.
[23:03] <CIA-17> ffmpeg: This will make upcoming support for complex filtergraphs easier to do.
[23:03] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r951e715ceb 10ffmpeg/avplay.c: 
[23:03] <CIA-17> ffmpeg: avplay: Don't try to scale timestamps if the tb isn't set
[23:03] <CIA-17> ffmpeg: If get_filtered_video_frame failed above, tb might not be
[23:03] <CIA-17> ffmpeg: initialized at all, so don't scale using it.
[23:03] <CIA-17> ffmpeg: This fixes cases where avplay could crash if aborting
[00:00] --- Sun Apr  8 2012