[Ffmpeg-devel-irc] ffmpeg-devel.log.20130711

[burek]

[00:04] <Daemon404> figures the first avreample failure is on an OS nobody uses or will ever use
[00:04] <Daemon404> :V
[00:04] <Daemon404> wait a pain to fix
[00:04] <Daemon404> what*
[00:16] <durandal_1707> what OS? plan9?
[00:16] <Daemon404> haiku
[00:18] <durandal_1707> that is really nice one
[00:20] <michaelni> Daemon404, wait till the next run of haiku before you investigate
[00:20] <Daemon404> ?
[00:20] <michaelni> i think its missing a file
[00:20] <Daemon404> oh
[00:25] <Guest5607> damned phone calls at the wrong time...
[00:50] <cone-59> ffmpeg.git 03Stefano Sabatini 07master:e7a6962e7650: doc/developer: update sections about external development
[00:50] <cone-59> ffmpeg.git 03Stefano Sabatini 07master:b3ac8a2a606c: doc/formats: add section describing stream specifiers
[00:50] <cone-59> ffmpeg.git 03Stefano Sabatini 07master:19ea08a11a8e: lavf/segment: add initial_offset option
[00:50] <cone-59> ffmpeg.git 03Stefano Sabatini 07master:459dc81a64a3: lavf/segment: reindent after last commit
[00:50] <cone-59> ffmpeg.git 03Stefano Sabatini 07master:71a2e5245ca5: lavf/segment: use AV_OPT_TIME_DURATION for time_delta
[01:06] <durandal_1707> TFTR?
[01:10] <saste> durandal_1707, Too Fast To Read
[01:11] <saste> durandal_1707, "time duration specification" or "duration specification" works for me
[01:12] <saste> users will have to look for the specification in man ffmpeg-utils
[01:13] <durandal_1707> so just add: specify time of ... see time duration specification. ?
[01:13] <saste> or ... the argument is a duration specification.
[01:13] <saste> bonus if you provide a reference to the spec
[01:16] <durandal_1707> i fill just link time to duration specification thing...
[01:20] <durandal_1707> its @anchor{time duration syntax}
[01:20] <durandal_1707> but its in another file
[01:25] <durandal_1707> michaelni: hmm, but atrim filters really trim frame samples
[01:27] <durandal_1707> saste: muxers mentions time duration specification, but there is no such thing
[01:28] <durandal_1707> but there is 'duration syntax'
[02:40] <cone-59> ffmpeg.git 03Michael Niedermayer 07master:94e86ae15a43: mjpegdec: initialize source variables before gbr remap
[02:40] <cone-59> ffmpeg.git 03Michael Niedermayer 07master:15cee5e5628a: mjpegdec: Fix used quant index for gbr
[11:21] <cone-135> ffmpeg.git 03Luca Barbato 07master:3865ba7b21ae: iff: K&R formatting cosmetics
[11:21] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:56090b5b8043: Merge commit '3865ba7b21aef5d60183719e238361ec8797ab5a'
[11:33] <cone-135> ffmpeg.git 03Luca Barbato 07master:7d65e960c72f: iff: Do not read over the source buffer
[11:33] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:2efc2b53be96: Merge commit '7d65e960c72f36b73ae7fe84f8e427d758e61da9'
[11:41] <cone-135> ffmpeg.git 03Luca Barbato 07master:04e9853a210e: wmavoice: return meaningful error codes
[11:41] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:5698f6bd7334: Merge commit '04e9853a210e07b1a150aa2b8e9af0aa3601ff21'
[11:50] <cone-135> ffmpeg.git 03Luca Barbato 07master:f5c48f5adaea: wmavoice: use init_static_data
[11:50] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:45cc62f0a099: Merge commit 'f5c48f5adaead5259660d4eaab2e695d1bac3a08'
[11:58] <cone-135> ffmpeg.git 03Luca Barbato 07master:d14a26edb7c4: wmavoice: conceal clearly corrupted blocks
[11:58] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:531f0169444c: Merge remote-tracking branch 'qatar/master'
[12:13] <cone-135> ffmpeg.git 03Michael Niedermayer 07release/2.0:9711b527394c: avfilter/af_earwax: Fix out of array accesses on odd packets
[12:13] <cone-135> ffmpeg.git 03Piotr Bandurski 07release/2.0:1cda4aa1e0e1: avformat/utils: avformat_find_stream_info fix a crash in case of oom
[12:13] <cone-135> ffmpeg.git 03Paul B Mahol 07release/2.0:56bf38859b93: lavfi/aconvert: unbreak
[12:13] <cone-135> ffmpeg.git 03Piotr Bandurski 07release/2.0:c1c84f0a55f8: avformat/utils: avformat_find_stream_info set value for ret in case of oom
[12:13] <cone-135> ffmpeg.git 03Michael Niedermayer 07release/2.0:ea28e7420517: avcodec/qdm2: store bits in an integer instead of float variable
[12:13] <cone-135> ffmpeg.git 03Michael Niedermayer 07release/2.0:37268dcc863e: avcodec/qdm2: initialize sign_bits
[12:13] <cone-135> ffmpeg.git 03Luca Barbato 07release/2.0:a7315116dd4a: wmavoice: conceal clearly corrupted blocks
[12:13] <cone-135> ffmpeg.git 03Carl Eugen Hoyos 07release/2.0:fc3dec8b626f: lavf/utils.c: Avoid a null pointer dereference on oom after duration_error allocation. (cherry picked from commit c9eb5c9751c88caaed62af5ffe908fe545022e7e)
[12:13] <cone-135> ffmpeg.git 03Carl Eugen Hoyos 07release/2.0:fd2cf9c45d3b: Suggest recompilation with openssl or gnutls if the https protocol is not found.
[12:13] <cone-135> ffmpeg.git 03Michael Niedermayer 07release/2.0:6127f792f900: mjpegdec: initialize source variables before gbr remap
[12:13] <cone-135> ffmpeg.git 03Michael Niedermayer 07release/2.0:7740e36a89cd: mjpegdec: Fix used quant index for gbr
[12:15] <ubitux> is Luca re-doing every j00ru fixes?
[15:14] <durandal_1707> can @ref{} point out of doxy file?
[15:41] <durandal_1707> why people thing place for bug reporting is support forum?
[15:42] <durandal_1707> *think
[15:42] <durandal_1707> ubitux: did you read this http://forum.doom9.org/showthread.php?t=168185 ?
[15:47] <durandal_1707> http://forum.doom9.org/showthread.php?t=143624&page=17
[15:48] <durandal_1707> michaelni: yuv444 -> rgb issues
[16:07] <wm4> how do you write picture attachments to a file, e.g. mp3? -attach Doesn't Work
[16:10] <durandal_1707> iirc you do it as normal video stream
[16:10] <durandal_1707> michaelni: did that swf issue got finally fixed?
[16:16] <michaelni> durandal_1707, why do you never ask a clear question and why do you always ask me ?
[16:17] <durandal_1707> michaelni: there was only one security issue with swfdec for over 6 months
[16:17] <durandal_1707> its about not really fixed security issues
[16:18] <durandal_1707> michaelni: is there some more experienced dev for asking about swscale conversion issues?
[16:22] <michaelni> durandal_1707, we fix all security issues for which sufficient information is provided to reproduce them or that contain a detailed description of the issue
[16:23] <durandal_1707> we?
[16:24] <michaelni> yes, you dont want to fix security issues ?
[16:27] <michaelni> about swf, iam not aware of any open security issues
[16:37] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:7afa68d37df9: mjpegdec: parse app-colr
[16:37] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:96747e6cbe75: mjpegdec: parse app-xfrm
[16:37] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:8024b4880ccc: avcodec/jpeglsdec: support xfrm 1-3
[16:37] <cone-135> ffmpeg.git 03Michael Niedermayer 07master:0ef40446a864: mjpegdec: print length too in APPx debug code
[18:27] <Compn> michaelni : i think durandal is talking about the securina / vlc spat (which is about the swf parser in libavformat)
[18:27] <Compn> which hit slashdot and others
[18:30] <Compn> michaelni : http://seclists.org/fulldisclosure/2012/Dec/96
[18:30] <Compn> :P
[18:33] <Compn> michaelni : heres the .swf file that crashed vlc . http://www39.zippyshare.com/v/91522221/file.html
[18:34] <Compn> from 2012 so it might be fixed since then
[18:34] <Compn> dont know if anyone tested ffmpeg
[18:34] <Compn> j-b : did you test swf with ffmpeg? (i know you want to talk about swf and securnia more :D)
[18:34] <j-b> Compn: fixed since a long time, you know...
[18:35] <ubitux> Compn: j-b linked to the fix in ffmpeg upstream in his blog post
[18:35] <j-b> http://git.videolan.org/?p=vlc/vlc-2.0.git;a=blob;f=contrib/src/ffmpeg/swfdec.patch;hb=HEAD
[18:56] <Compn> j-b : oh ok :)
[18:56] <Compn> i saw that you linked to vlc git , i wasnt paying attention if it were a fix for ffmpeg or vlc's ffmpeg ....
[18:56] <Compn> and isnt vlc using libav ? so confused
[19:07] <j-b> Compn: vlc compiles with both
[19:10] <Compn> very professional answer that didnt quite answer the question ;P
[19:10] <Compn> my question wasnt precise either of course
[19:11] <j-b> Compn: it depends on the maintainer
[19:11] <j-b> Compn: hence I cannot answer the question
[19:11] <wm4> what are you using for development?
[19:12] <j-b> I use FFmpeg, because my distro is Arch
[19:22] <durandal_1707> VideoLAN is paying trips to Paris?
[19:30] <Compn> for Videolan dev days, i think they have invited ffmpeg devels in the past
[19:30] <Compn> dunno if they currently have funds for it
[19:30] <Compn> best ask j-b or in #videolan
[19:40] <j-b> durandal_1707: of course.
[21:53] <cone-135> ffmpeg.git 03Dan Flett 07master:43d36599fe33: lavd/fbdev.c: Support RGB565 input colour space.
[22:23] <durandal_1707> should i write adelay filter?
[22:37] <Compn> durandal_1707 : you mean for offsetting audio ? like mplayer -delay or -af delay ?
[22:37] <saste> durandal_1707, time to move on with the framework?
[22:38] <saste> what's going on on the l. side with regards to API/framwork?
[22:38] <saste> anybody knows? So we don't need to redo from scratch after they NIH
[22:39] <saste> but probably the best thing to do at this point is to work on bindings
[22:40] <durandal_1707> saste: framework what?
[22:41] <saste> durandal_1707, reconfiguration for example
[22:41] <saste> extend input movie, add output movie
[22:41] <durandal_1707> you just call config_props for each link
[22:41] <durandal_1707> i don't touch movie filters as i consider them evil
[22:42] <saste> durandal_1707, why?
[22:42] <saste> why people hate movie
[22:42] <saste> ?
[22:42] <durandal_1707> becaue it does not play movie you want
[22:42] <saste> especially since the alternative (-filter_complex) is a huge application-level hack
[22:43] <saste> durandal_1707, what do you mean??
[22:44] <durandal_1707> it doesn't play well
[22:45] <durandal_1707> it doesn't work with http/ etc...
[22:45] <saste> can't you specify http://...
[22:45] <saste> ?
[22:46] <durandal_1707> iirc it never worked for me
[22:46] <saste> it's the usual pain to escape, but what does it prevent to work as expected?
[22:46] <saste> "it doesn't play well" => too vague
[22:46] <durandal_1707> it doesn't expand ~
[22:47] <durandal_1707> i need to give absolute path
[22:51] <durandal_1707> also isn't (a)movie cloning ffmpeg funcionality and is missing some hacks?
[22:51] <saste> durandal_1707, the plan is having movie substitute ffmpeg
[22:52] <saste> so you can implement ffmpeg functionality in your program, with (relatively) few lines of code
[22:52] <durandal_1707> interesting idea, dunno if its doable in reasonalbe time frame
[22:53] <durandal_1707> there are other filters that ffmpeg could be using, but does not
[22:54] <durandal_1707> and instead do it on its own, like setfps
[22:54] <durandal_1707> and trim, which is used by avconv, but not ffmpeg
[22:55] <durandal_1707> i think with my patch, ffmpeg could use it though
[22:55] <durandal_1707> michaelni: ^
[22:55] <saste> not everything can be done with filters
[22:55] <saste> for example for manipulating packets timestamps you need essentially - a packet filter
[22:56] <durandal_1707> yes, so why not write one?
[22:56] <saste> because you need a whole packet filtering framework
[22:56] <saste> also maybe we should merge it somehow to lavfi
[22:57] <durandal_1707> merge AVPacket with AVFrame
[22:57] <saste> for example we could add data filtering to lavfi
[22:57] <saste> but this requires careful design
[22:57] <saste> and I have no more energy for it
[22:57] <durandal_1707> what? you are retiring?
[22:58] <saste> http://sophia.estec.esa.int/socis2013/?q=node/13
[22:58] <saste> still waiting (after almost two weeks)
[22:58] <durandal_1707> others got approved or?
[22:59] <saste> no, only thinks we know since a week is that the list will be published "soon"
[22:59] <saste> it's a bit frustrating
[23:00] <durandal_1707> well, ESA Summer of Code in Space becomes ESA Autumn of Code in Space
[23:05] <saste> durandal_1707, because in space, no one can hear when you wait
[23:29] <nevcairiel> you just have to wait long enough, then summer is on the southern hemisphere
[23:51] <Daemon404> g 42
[23:54] <durandal_1707> so should i do adelay filter, the only way currently one can do delay, is by splitting channels, and joining them again after processing
[00:00] --- Fri Jul 12 2013