[Ffmpeg-devel-irc] ffmpeg-devel.log.20111214

burek burek021 at gmail.com
Thu Dec 15 02:05:04 CET 2011


[00:43] <ubitux> make: *** [fate-lavfi-crop] Error 1
[00:43] <ubitux> mmh
[00:43] <ubitux> things start to get ugly
[00:49] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * reda3758c87 10ffmpeg/libavcodec/ac3dec.c: (log message trimmed)
[00:49] <CIA-49> ffmpeg: ac3dec: update checked AV_EF flags.
[00:49] <CIA-49> ffmpeg: A user who wishes to use default error concealment cannot set the
[00:49] <CIA-49> ffmpeg: AV_EF_CRCCHECK flag because not every CRC in every format is a
[00:49] <CIA-49> ffmpeg: reliable indicator of bitstream damage. In some formats crcrs
[00:49] <CIA-49> ffmpeg: can be nonsensical in absence of any damage. We thus add the
[00:49] <CIA-49> ffmpeg: AV_EF_CAREFUL flag in addition to the AV_EF_CRCCHECK flag to
[04:01] <CIA-49> ffmpeg: 03Derek Buitenhuis 07master * r64a0b826e1 10ffmpeg/tests/ (fate2.mak ref/fate/v410dec ref/fate/v410enc): 
[04:01] <CIA-49> ffmpeg: fate: Add FATE tests for v410 encoder and decoder
[04:01] <CIA-49> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis at gmail.com>
[04:01] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:01] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r635a301683 10ffmpeg/MAINTAINERS: 
[04:01] <CIA-49> ffmpeg: MAINTAINERS: add 0.9
[04:01] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:02] <CIA-49> ffmpeg: 03Derek Buitenhuis 07master * r309ffa4fe3 10ffmpeg/ (11 files in 4 dirs): 
[04:02] <CIA-49> ffmpeg: v410 encoder and decoder
[04:02] <CIA-49> ffmpeg: v410 is a packed 10-bit 4:4:4 YCbCr format used in
[04:02] <CIA-49> ffmpeg: QuickTime.
[04:02] <CIA-49> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis at gmail.com>
[04:02] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:35] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r0a11e7ffba 10ffmpeg/tests/fate2.mak: 
[05:35] <CIA-49> ffmpeg: fate: fix v410 on big endian
[05:35] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:37] <Daemon404> oh. duh.
[05:38] <durandal_1707> :)
[05:40] <GordonFreeman> hi
[05:40] <GordonFreeman> i made a simple PRNG :)
[05:40] <GordonFreeman> 64 bit randomz2 vs /dev/urandom dieharder full test http://pastebin.com/Zs9CcBxe
[05:40] <GordonFreeman> /dev/urandom has a weak line, and mine passes all
[07:30] <CIA-49> ffmpeg: 03Clément BSsch 07master * rc608776ce9 10ffmpeg/Changelog: cosmetic: consistent spacing over the Changelog.
[09:50] <qrtt11> hello
[09:52] <qrtt1> I have an iOS app which use the ffmpeg library, and iOS should compile in static link. According to the LGPLv2, do I release my source code about related my UI ?
[10:09] <JEEB> qrtt1, basically the LGPL IIRC was about allowing the user to switch libraries if one wants to do that. Which would be either sources for the complete app (ffmpeg + non-ffmpeg parts), or object files / whatever they're called @ iOS and the source code for the ffmpeg part (so one can rebuild the ffmpeg part and link it and get a working app)
[10:10] <JEEB> but before that, IIRC the Apple store had problems with (L)GPL software
[10:10] <JEEB> so you might want to check up on that
[10:11] <av500> Apple imposes additional restrictions on top of GPL/LGPL
[10:11] <av500> like only to install on 5 machines
[10:11] <av500> that was found a problem for e.g. VLC
[10:14] <qrtt1> Apple doesnt allow LGPL library ?
[10:14] <av500> I did not say that
[10:14] <qrtt1> av500: sorry, I misunderstnad it
[10:15] <av500> apple store has some restrictions that are not compatible with GPL
[10:15] <qrtt1> I found some information. I can open the bussiness logic in obj-files format 
[10:16] <qrtt1> av500: thanks. 
[10:20] <qrtt1> I have the modified ffmpeg on the public github repo https://github.com/qrtt1/ffmpeg_icy . In my product information, just to contain it. does it comply distubtion ?
[10:29] <ohsix> you can't comply with the gpl and apples terms if you distribute it through them
[10:33] <qrtt1> ohsix: Do you mean I should distribute the information (ffmpeg source code and non-ffmpeg objs) in my app, if it is possible ?
[10:34] <qrtt1> ohsix: does the 'gpl' typo ? I use ffmpeg in lgpl mode.
[11:11] <ubitux> michaelni: "poll_frame() is a bit more tricky, if the sources filters poll returns 0, it should return 0 too"  why? if i have a thumbnail available, why should i wait for the sources?
[11:12] <ubitux> (http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2011-December/117367.html)
[14:58] <michaelni> ubitux, if you have a frame available you should return 1 but if you have one available i would have thought it was pushed out already
[14:59] <ubitux> oh so it's a "can you compute one?" state
[14:59] <ubitux> or well, queue the frame
[15:00] <ubitux> lavfi is really confusing, i feel lost most of the time, even thought it's pretty well documented
[15:17] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r9c18a056b1 10ffmpeg/libavformat/riff.c: Detect 'yuv2' as rawvideo also in avi.
[15:17] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * ra8e93ce4cb 10ffmpeg/libavformat/wav.c: 
[15:17] <CIA-49> ffmpeg: Do not reject LIST tags in wav with size 4.
[15:17] <CIA-49> ffmpeg: Fixes ticket #745.
[15:17] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r7862bd3d4a 10ffmpeg/libavformat/wav.c: Cosmetics: Improve 'too short LIST' error message in wav.c.
[15:19] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07release/0.9 * r238c7ffd7d 10ffmpeg/libavformat/wav.c: 
[15:19] <CIA-49> ffmpeg: Do not reject LIST tags in wav with size 4.
[15:19] <CIA-49> ffmpeg: Fixes ticket #745.
[15:19] <CIA-49> ffmpeg: (cherry picked from commit a8e93ce4cb74dc26153bdccfbe8a487674e583a6)
[16:21] <CIA-49> ffmpeg: 03Matthieu Bouron 07master * r366605ff5d 10ffmpeg/libavcodec/mpeg12enc.c: 
[16:21] <CIA-49> ffmpeg: mpeg12enc: fix drop_frame_timecode value
[16:21] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:35] <evil_andy> so I think kierank said yesterday that I should talk to bcoudurier about adding private data stream support to mpegTS.  I was messing around with it yesterday and managed to get it to recognize the stream (via the program element descriptor in the PMT)
[16:36] <evil_andy> as for how to make that play nice with the rest of FFmpeg, I'm not sure. I used ffplay to see if it was working, but had to add quite a bit of code just to get it to read the packets (because it's a codec for AVMEDIA_TYPE_DATA, not video, audio, or subtitles)
[16:39] <jermy> I think we read timecode chunks by calling av_read_frame as normal on a suitable context for the data stream. Presumably ffplay doesn't care about non-AV streams
[16:41] <jermy> I suspect making it play nice with libavformat is quite different (and probably easier) than making it play nice with the ffmpeg/ffplay binaries
[16:43] <evil_andy> Hmm... Ok. so, av_read_frame does the 'right' thing (that is it grabs a packet of the data stream and stores it)
[16:44] <evil_andy> the issue with ffplay was that it sets all the streams to AVDISCARD_ALL before enabling the ones you specify (and it only supports audio, video, and subtitles)
[16:45] <evil_andy> so (for testing purposes) I just added the necessary code to add a data codec into the mix.
[16:46] <MuscaluDiana> who is a GCI mentor?
[16:47] <MuscaluDiana> i have claimed http://www.google-melange.com/gci/task/view/google/gci2011/7172288 . can a mentor accept the claim?
[16:48] <evil_andy> heh, I didn't even know that existed. I found a bug too :( but I already fixed it lol
[16:50] <michaelni> MuscaluDiana, accepted, good luck :)
[16:50] <evil_andy> granted. it's taken me a year of 'free time' to finally sit down to work on this
[16:50] <MuscaluDiana> michaelni: what files do i need to report? valgrind,gdb,bt reports and the corrupted file?
[16:51] <michaelni> yes, that sounds good
[16:51] <MuscaluDiana> michaelni: do i need to report the bug on bugzilla like the libav bugs?
[16:52] <Compnn> we use trac, http://www.ffmpeg.org/trac/ffmpeg
[16:52] <Compnn> you can report to trac or ffmpeg-user mailing list
[16:52] <MuscaluDiana> ok, perfect thx Compnn
[16:52] <Compn> np
[16:53] <Compn> evil_andy : sounds like a bug that really irritated you if you remember it after a year :)
[16:54] <evil_andy> Compn, well, the bug that really bothered me for a year got fixed, and a new one was introduced
[16:54] <michaelni> evil_andy, what new bug?
[16:55] <evil_andy> michaelni, looking back through my notes, 1 sec
[16:56] <michaelni> also if you are interrested in mpeg, theres a unfinished mpeg-ts muxer bugfix patch laying around
[17:01] <evil_andy> Hmm... actually, looking at it again, it's just the logic in mpegts_set_stream_info doesn't follow my own... the first thing it does is set the codec_type to AVMEDIA_TYPE_DATA, then it tries to find out what kind of stream it is based on stream type, then if it didn't find it, sets the codec_type back to whatever it was coming in (type unknown).  I guess it looked weird, since AVMEDIA_TYPE_DATA was actually correct for what I was 
[17:01] <evil_andy> working wit
[17:01] <evil_andy> h
[17:02] <evil_andy> Then I added a REGD_types entry for KLVA, and that gets picked up correctly in the new ff_parse_mpeg2_descriptors function
[17:05] <evil_andy> also, if I specify a StreamType as having CODEC_ID_NONE, will that actually still do anything? or will it automatically drop those packets because there's no associated codec?
[17:07] <ubitux> MuscaluDiana: you should read http://ffmpeg.org/bugreports.html too
[17:07] <michaelni> it should not drop the packets
[17:07] <MuscaluDiana> okei ubitux
[17:07] <ubitux> thank you in advance :)
[17:08] <ubitux> michaelni: is it possible to duplicate the task?
[17:08] <ubitux> the "find & report a new bug" can be done multiple times :p
[17:08] <michaelni> ubitux, i did open multiple :)
[17:09] <ubitux> ok, ok :)
[17:34] <CIA-49> ffmpeg: 03Panagiotis H.M. Issaris 07master * r9b01a8ad5e 10ffmpeg/libavformat/utils.c: 
[17:34] <CIA-49> ffmpeg: Prevent segfault if iformat was not yet initialized.
[17:34] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:23] <ubitux> any idea why vf_yadif is checking avfilter_poll_frame() == 1 in the poll_frame() callback (val==1) and not just "val"?
[18:36] <Compn> ubitux : dunno, thats black magic by michaelni ;D
[18:36] <Compn> only he knows the magic :P
[18:37] Action: Compn finally comes up with solution to isom/fourcc list that maybe bcoudurier could agree on
[18:37] <Compn> if we cant combine the lists, what about just checking the other list for a matching code
[18:37] <Compn> in case the first list fails to have it
[18:38] <michaelni> sounds good to me, 
[18:38] <michaelni> i think this was suggested in the past already though
[18:39] <Compn> such a simple idea, should have been suggested already
[18:40] <Compn> probably rejected because then we'll just add everything to RIFF and ignore isom :D
[18:44] <Compn> do i bikeshed too much ?
[18:46] <michaelni> ubitux, when there are more the 1 frames available from the previous filter then we dont need to request any at that point, we know that we can return at least one
[19:09] <ubitux> michaelni: mmh ok, thx
[19:53] <ubitux> michaelni: should i split off the map channel fate tests to ease your merge with libav?
[20:04] <michaelni> i didnt notice any extra work from the map channels test ...
[20:04] <ubitux> well, Diego is splitting the fate tests into dedicated files
[20:04] <ubitux> which is btw a good idea
[20:05] <ubitux> but the patches will likely have conflict because of the mapchannel tests being in the middle of the current tests
[20:05] <ubitux> i can take them out easily in a dedicated file
[20:06] <ubitux> (actually i asked when i already started and i'm already done)
[20:06] <michaelni> ok thx
[20:07] <ubitux> you will just have conflict on the bunch of includes
[20:08] <ubitux> but it's imo much less problematic than dealing with the chunks in fate2.mak
[20:08] <CIA-49> ffmpeg: 03Clément BSsch 07master * r54c7e81a53 10ffmpeg/tests/ (Makefile fate/mapchan.mak fate2.mak): 
[20:08] <CIA-49> ffmpeg: fate: split off map_channel tests.
[20:08] <CIA-49> ffmpeg: Also introduce a fate-mapchan rule for all of them.
[21:04] <OanaStratulat> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7172288 please review
[21:12] <michaelni> DianaMuscalu, good work, i can confirm it segfaults :)
[21:12] <DianaMuscalu> ok, can i take another one?
[21:13] <michaelni> of course
[21:13] <DianaMuscalu> michaelni: the files reported are good? or i should submit more reports?
[21:14] <michaelni> DianaMuscalu, please test with ffmpeg_g instead of ffmpeg in the future as ffmpeg_g contains debug symbols
[21:14] <j-b> 'morning
[21:15] <michaelni> night ;)
[21:15] <DianaMuscalu> michaelni: okei, are the same arguments -i filename -f nulll -?
[21:15] <michaelni> -i filename -f null is fine
[21:16] <DianaMuscalu> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7179243 please accept claim
[21:17] <michaelni> done
[21:18] <DianaMuscalu> michaelni: thx, i will file another bug in ~ an hour
[21:33] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r76b9a0961c 10ffmpeg/libavformat/riff.c: 
[21:33] <CIA-49> ffmpeg: riff: Fix freeing of random value.
[21:33] <CIA-49> ffmpeg: Fixes Ticket752
[21:33] <CIA-49> ffmpeg: Bug-found-by: Diana Elena Muscalu
[22:07] <DianaMuscalu> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7179243 please review. 10x
[22:09] <ubitux> DianaMuscalu: just curious, how do you find them?
[22:09] <DianaMuscalu> ubitux: i have tested the ffmpeg_g with some videos of mine zzufed with different seeds from 1 to 2000
[22:09] <DianaMuscalu> ubitux: and different ratio
[22:10] <ubitux> ok :)
[22:11] <michaelni> DianaMuscalu, reveiw done, good work! 
[22:11] <DianaMuscalu> michaelni: thx
[22:12] <ubitux> the "find a bug & report it" task was a good idea :)
[22:13] <Daemon404> no human indent machine tasks?
[22:13] <ubitux> please no :(
[22:13] <DianaMuscalu> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7179242 please accept claim
[22:13] <Daemon404> lol
[22:14] <michaelni> DianaMuscalu, done :)
[22:16] <DianaMuscalu> michaelni: thx
[22:17] <pasteeater> anyone know how to remove a user in trac? we have a spambot.
[22:17] <michaelni> pasteeater, name ?
[22:17] <pasteeater> AlysiaSpike
[22:19] <michaelni> deleted AlysiaSpike
[22:19] <pasteeater> michaelni: thanks
[22:31] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * re1ba29c764 10ffmpeg/libavcodec/4xm.c: 
[22:31] <CIA-49> ffmpeg: 4xm: check if there are bits left in decode_i_block()
[22:31] <CIA-49> ffmpeg: Fixed Ticket753
[22:31] <CIA-49> ffmpeg: Bug Found by: Diana Elena Muscalu
[22:31] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:38] <CIA-49> ffmpeg: 03Marton Balint 07master * r9fb2b41258 10ffmpeg/ffplay.c: 
[22:38] <CIA-49> ffmpeg: ffplay: force setting video mode on fullscreen toggle
[22:38] <CIA-49> ffmpeg: Fixes ticket #736. Based on the patch by glangshaw.
[22:38] <CIA-49> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[22:38] <CIA-49> ffmpeg: 03Panagiotis H.M. Issaris 07master * r91a3ea671a 10ffmpeg/ (doc/ffplay.texi ffplay.c): 
[22:38] <CIA-49> ffmpeg: ffplay: add 10 minute seek support to ffplay
[22:38] <CIA-49> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[22:38] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * ra5a1e3cb8a 10ffmpeg/ffplay.c: 
[22:38] <CIA-49> ffmpeg: ffplay: Fix got_frame type.
[22:38] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:38] <CIA-49> ffmpeg: Signed-off-by: Marton Balint <cus at passwd.hu>
[22:38] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rdc5f001878 10ffmpeg/: (log message trimmed)
[22:38] <CIA-49> ffmpeg: Merge remote-tracking branch 'cus/stable'
[22:38] <CIA-49> ffmpeg: * cus/stable:
[22:38] <CIA-49> ffmpeg:  ffplay: clear pkt_temp when pkt is freed.
[22:38] <CIA-49> ffmpeg:  ffplay: Fix got_frame type.
[23:14] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rf129f593f7 10ffmpeg/libavcodec/mjpegdec.c: 
[23:14] <CIA-49> ffmpeg: ljpeg: Check that lowres is 0 as lowres is not possible with ljpeg.
[23:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:20] <CIA-49> ffmpeg: 03Jean First 07master * ra1be5bc79d 10ffmpeg/libavcodec/dnxhdenc.h: 
[23:20] <CIA-49> ffmpeg: dnxhdenc: remove unused variable from context
[23:20] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:38] <Diana_Muscalu> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7179242 please review 
[23:43] <michaelni> Diana_Muscalu, done
[23:49] <Diana_Muscalu> michaelni: http://google-melange.appspot.com/gci/task/view/google/gci2011/7172289 please accept the claim. 10x
[23:50] <michaelni> Diana_Muscalu, done :)
[23:51] <Diana_Muscalu> michaelni: 10x 
[00:00] --- Thu Dec 15 2011


More information about the Ffmpeg-devel-irc mailing list