[Ffmpeg-devel-irc] ffmpeg-devel.log.20120718

[burek]

[01:53] <CIA-41> ffmpeg: 03Samuel Pitoiset 07master * r775c4d3625 10ffmpeg/ (5 files in 2 dirs): 
[01:53] <CIA-41> ffmpeg: rtmp: Rename rtmphttp to ffrtmphttp
[01:53] <CIA-41> ffmpeg: The prefix makes it easier to distinguish the proper end-user
[01:53] <CIA-41> ffmpeg: protocols from the internal ones.
[01:53] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:53] <CIA-41> ffmpeg: 03Kostya Shishkov 07master * r0b40153d20 10ffmpeg/libavcodec/tscc2.c: 
[01:53] <CIA-41> ffmpeg: tscc2: DCT output should not be clipped
[01:53] <CIA-41> ffmpeg: This fixes decoding some TSCC2 files with large quantisers.
[01:53] <CIA-41> ffmpeg: 03Anton Khirnov 07master * r5417efbbf3 10ffmpeg/avconv.c: 
[01:53] <CIA-41> ffmpeg: avconv: print an error message when demuxing fails.
[01:53] <CIA-41> ffmpeg: Also exit in such a case if -xerror is used.
[01:53] <CIA-41> ffmpeg: Fixes bug 329.
[01:53] <CIA-41> ffmpeg: 03Samuel Pitoiset 07master * r6aedabc9b6 10ffmpeg/ (9 files in 3 dirs): 
[01:53] <CIA-41> ffmpeg: RTMPS protocol support
[01:53] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:53] <CIA-41> ffmpeg: 03Samuel Pitoiset 07master * r86991ce2dd 10ffmpeg/ (9 files in 3 dirs): 
[01:53] <CIA-41> ffmpeg: RTMPTS protocol support
[01:53] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[01:53] <CIA-41> ffmpeg: 03Michael Bradshaw 07master * rb43a7bb4f9 10ffmpeg/libavcodec/ (Makefile libopenjpeg.c): 
[01:53] <CIA-41> ffmpeg: libopenjpeg: rename decoder source file.
[01:54] <CIA-41> ffmpeg: libopenjpeg: introduce encoding support
[01:54] <CIA-41> ffmpeg: Based on FFmpeg version from
[01:54] <CIA-41> ffmpeg: commit 713a7854e06964abc8f7d015b94acbed27769d24
[01:54] <CIA-41> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[01:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r56ae5926f7 10ffmpeg/: (log message trimmed)
[01:54] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[01:54] <CIA-41> ffmpeg: * qatar/master:
[01:54] <CIA-41> ffmpeg:  libopenjpeg: introduce encoding support
[01:54] <CIA-41> ffmpeg:  libopenjpeg: rename decoder source file.
[01:54] <CIA-41> (2 lines omitted)
[01:54] <durandal_1707> michaelni: i hope that openjpeg code did not overwrote blame history
[02:04] <Hawk-TAO> hi
[02:04] <Hawk-TAO> anyone online?
[02:05] <Hawk-TAO> i have a question
[02:06] <Hawk-TAO> what is the C API that reads the video from harddisk in ffmpeg?
[02:07] <durandal_1707> libavformat
[02:07] <durandal_1707> but actually it is file protocol
[02:10] <Hawk-TAO> can u show me the function?
[02:11] <durandal_1707> what function?
[02:11] <Hawk-TAO> read the video from harddisk,like fread().
[02:13] <durandal_1707> as i said really low level stuff is handled with URLProtocol API
[02:14] <durandal_1707> Hawk-TAO: read this http://ffmpeg.org/doxygen/trunk/index.html
[02:14] <Hawk-TAO> thx
[03:29] <TAO> hi
[03:29] <TAO> no one speaking?
[03:30] <TAO> who is from China in this chatgroup?
[03:31] <durandal_1707> only you
[03:31] <TAO> oh
[03:32] <TAO> it is a bad news
[03:34] <TAO> do you wanna visit China?
[03:40] <TAO> is the manager of this community whose name contains "@" ?
[03:41] <durandal_1707> TAO: @ means channel operator
[03:46] <TAO> hmm
[03:46] <TAO> thx
[03:49] <TAO> and what is meaning of "+"?
[03:50] <durandal_1707> + means master of universe
[04:01] <TAO> university?
[04:01] <TAO> hmm,good
[04:03] <Compn> @ means ffmpeg committer
[04:03] <Compn> + also means ffmpeg committer or other helpful person
[04:14] <TAO> hmm
[04:46] <highgod> Hi,I add opencl to libavutil in ffmpeg.But I don't know how to modify the makefile to make the ffmpeg to link opencl.lib.Can anyone know how to modify?Thanks
[04:52] <TAO> hi,highgod,you write the opencl code to ffmpeg?
[04:52] <highgod> yes
[04:52] <TAO> which part?
[04:53] <highgod> scale,and colorspace conversion
[04:53] <highgod> But I didn't pass the first step.haha
[04:53] <TAO> oh 
[04:53] <highgod> I don't know how to modify the make file to make ffmpeg link opencl.lib
[04:54] <highgod> Can you give me some advice?
[04:55] <TAO> wait
[04:58] <TAO> add the lib in OBJS
[05:24] <TAO> highgod
[05:24] <TAO> it can be compiled now?
[10:17] <Spideru> morning
[12:59] <saste> http://sophia.estec.esa.int/socis2012/?q=node/13
[12:59] <saste> we are in the list :-)
[13:00] <Spideru> good :)
[13:01] <saste> michaelni: maybe you want to mentor a task related to encoder reconfiguration (adaptive streaming), it may be somewhat related to ESA interest
[13:53] <michaelni> saste, dunno how encoder reconfig is related to ESA :)
[13:54] <michaelni> also i dont think it needs any API changes
[13:54] <michaelni> just change parameters and if encoder supports it it reconfigs
[13:55] <michaelni> also libx264 with ffmpeg already supports reconfigs
[13:59] <michaelni> only thing missing is maybe a CODEC_CAP_RECONFIG(_ASPECT/INTERLACING/DIMENSIONS)
[14:03] <michaelni> A problem internal to GDB has been detected,
[14:03] <michaelni> further debugging may prove unreliable.
[14:03] <michaelni> Quit this debugging session? (y or n) 
[14:03] <michaelni> :/
[14:26] <burek> you've found a bug in a bug de-bugger :)
[14:33] <Compn> good question, how do you debug gdb ?
[14:33] <Compn> gdb gdb ? :P
[14:44] <Zor> that actually works yes
[14:54] <Compn> better to report bug to gdb , let them sort it :P
[15:05] <iive> have you tried the latest git gdb? can you provide backtrace ? ;)
[15:06] <Compn> ehe
[15:06] <Compn> sounds familiar...
[15:06] <iive> :)
[16:27] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rba0e69994d 10ffmpeg/libswscale/input.c: 
[16:27] <CIA-41> ffmpeg: sws/input: switch to av_assert
[16:27] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:28] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r60f3291086 10ffmpeg/libavcodec/mlpdec.c: 
[16:28] <CIA-41> ffmpeg: mlpdec: switch to av_assert
[16:28] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rdece4f4693 10ffmpeg/ffmpeg.c: 
[16:54] <CIA-41> ffmpeg: ffmpeg: revert 3ba90d9 (cosmetics to reduce difference to qatar by about 90 lines)
[16:54] <CIA-41> ffmpeg: This commit caused Ticket1490 (Infinity loop / lock in export image)
[16:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:17] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r712d82130e 10ffmpeg/libavfilter/avfiltergraph.h: lavfi/avfiltergraph: apply misc fixes to fix avfilter_graph_request_oldest() @return doxy
[17:17] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r0cc9077a87 10ffmpeg/configure: 
[17:17] <CIA-41> ffmpeg: configure: fix help message for --disable-avfilter
[17:17] <CIA-41> ffmpeg: Increase consistency with the other similar messages.
[17:17] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r6a314d59d5 10ffmpeg/configure: 
[17:17] <CIA-41> ffmpeg: configure: simplify dependencies for x11_grab_device
[17:17] <CIA-41> ffmpeg: XShmCreateImage dependency is implied by x11grab.
[17:42] <iive> cosmetics causing bugs?!!
[17:44] <michaelni> it looked like cosmetics but theres a difference in the initial iteration vs no iteration
[17:49] <iive> :)
[17:56] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * ref1c0b5295 10ffmpeg/libavfilter/avfiltergraph.c: 
[17:56] <CIA-41> ffmpeg: avfiltergraph: add some debug info about selected pixel formats
[17:56] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:56] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r31a192f387 10ffmpeg/libavcodec/imgconvert.c: 
[17:56] <CIA-41> ffmpeg: imgconvert: favor pixel formats without resolution loss
[17:56] <CIA-41> ffmpeg: Fixes Ticket1517
[17:56] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:11] <burek> ubitux, on this link http://ffmpeg.org/ffplay.html the page title says FFmpeg documentation, instead of FFplay documentation
[18:12] <burek> also, meta description and meta keywords are not correctly set
[18:12] <burek> <meta name="description" content=": ">
[18:12] <burek> it should be something like <meta name="description" content="FFplay documentation">
[18:13] <burek> also "ff" through out the doc might be maybe changed to "FF"?
[18:13] <burek> nah.. I'm going to find some git tutorial
[18:13] <burek> this will take too much time this way
[18:15] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r739ec710c4 10ffmpeg/libavformat/utils.c: 
[18:15] <CIA-41> ffmpeg: lavf/utils: suggest what to do in case of missing codec information
[18:15] <CIA-41> ffmpeg: In particular, fix trac issue #218.
[18:15] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r5784a46986 10ffmpeg/libavformat/utils.c: 
[18:15] <CIA-41> ffmpeg: lavf/utils: show the number of stream in case of missing parameter
[18:15] <CIA-41> ffmpeg: Increase feedback provided in avformat_find_stream_info().
[18:15] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * ra293549b32 10ffmpeg/libavformat/utils.c: 
[18:15] <CIA-41> ffmpeg: lavf/utils: extend has_codec_parameters() to make it show what info is missing
[18:15] <CIA-41> ffmpeg: Improve feedback.
[18:37] <Compn> j-b : is vlc able to use ffmpeg VDA now ?
[18:37] <j-b> yes
[18:37] <Compn> ah
[18:37] <Compn> i think the decoder is more for mplayer use then :)
[18:38] <Compn> or mplayer just too lazy to do it how vlc does it 
[18:38] <Compn> re: vda patch on ffmpeg-devel
[18:47] <nevcairiel> I dont think it should be applied because someone doesnt know how to implement hwaccel support
[18:47] <burek> does ffmpeg "resets" timestamps after -i (when using something like: ffmpeg -ss 10 -i input -ss 12 ...)
[18:47] <burek> the idea is to tell ffmpeg to skip frames quickly until some point and then to fine seek into the final point
[18:48] <burek> (see #ffmpeg)
[19:15] <burek> ubitux, is it enough to just change the file, do a "git diff" and then post it on the devel ML? is that all?
[19:17] <nevcairiel> its preferred to actually create a commit and use git format-patch
[19:17] <durandal_1707> burek: better commit change to own branch with nice commit message and than just use format-patch and git send-email
[19:19] <burek> ok :) thanks :)
[19:39] <ubitux> < burek> ubitux, on this link http://ffmpeg.org/ffplay.html the page title says FFmpeg documentation, instead of FFplay documentation // the texi has no "FFmpeg documentation" string
[19:39] <ubitux> it's just a one doc page of the FFmpeg project
[19:40] <ubitux> dunno were to set this
[19:40] <ubitux> same of the meta
[19:40] <ubitux> about changing ff to FF i'm not sure
[19:40] <ubitux> i'd refer to ffplay all the time
[19:40] <ubitux> you have ffmpeg,ffplay,ffprobe,ffserver tools, within FFmpeg project
[19:40] <ubitux> there is no much meaning in "FFplay"
[19:41] <burek> ubitux, read the 1st line of ffplay's doc
[19:41] <ubitux> burek: you should check your irc log where i explained a bit that stuff
[19:41] <burek> after the index
[19:41] <burek> :)
[19:41] <burek> yes im reading it
[19:41] <burek> :)
[19:41] <ubitux> burek: yes i saw it in the html output
[19:42] <ubitux> this is most likely in the config script
[19:42] <ubitux> see doc/t2h.init
[19:43] <ubitux> that script/config "needs" to be changed, and then fixed on the server
[19:54] <durandal_1707> Daemon404: when you gonna fix zerocodec?
[19:54] <Daemon404> durandal_1707, i actually do have a job, so hopefully tonight
[19:55] <Daemon404> you are free to fix it yourself in the meanwhile
[19:56] <durandal_1707> Daemon404: i have no idea what is wrong :(
[19:56] <Daemon404> durandal_1707, decode_close is not necessarily called when decode_frame returns error, but is guaranteed to be called when decode finishes
[19:56] <Daemon404> so if a frame fails, it could lose prev/curr buffers
[19:57] <Daemon404> it actually always loses teh curr buffer if decode_frame fails, even if decode_close is called
[19:57] <Daemon404> ~memleaks~
[19:57] <nevcairiel> there must be something else wrong
[19:57] <durandal_1707> but there are many assumptions in code that docode_close is called....
[19:58] <nevcairiel> the decoder never worked for me properly, and somehow i doubt the file was only broken
[19:58] <nevcairiel> although its possible
[19:58] <Daemon404> yes
[19:58] <Daemon404> that is busted too
[19:58] <Daemon404> i made a bad assumption about how get_buffer worked
[19:59] <Daemon404> since ffplay etc use a custom get_buffer
[19:59] <Daemon404> i never noticed it
[19:59] <Daemon404> it's a simple fix
[19:59] <nevcairiel> i always wanted to mention it, but i forgot until i saw the comments about it yesterday :d
[19:59] <Freakshow> guys... how could I go about getting some eyes on this ticket?
[19:59] <Freakshow> https://ffmpeg.org/trac/ffmpeg/ticket/1446
[19:59] <Daemon404> basically i assumed that get_bugger would always obliterate the old data[N] buffers
[20:00] <Daemon404> but teh default get_buffer fails if data[N] != NULL
[20:00] <Daemon404> i should start testing in things that are !ffmpeg/avconv
[20:01] <Daemon404> brb meeting...
[21:28] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rc66978e29a 10ffmpeg/libswscale/swscale.h: 
[21:28] <CIA-41> ffmpeg: sws: more complete doxy for sws_setColorspaceDetails()
[21:28] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:44] <CIA-41> ffmpeg: 03Diego Biurrun 07master * r7a249625fd 10ffmpeg/libavformat/os_support.c: os_support: K&R formatting cosmetics
[22:44] <CIA-41> ffmpeg: 03Diego Biurrun 07master * r47876d661a 10ffmpeg/configure: configure: cosmetics: drop some unnecessary backslashes
[22:44] <CIA-41> ffmpeg: 03Diego Biurrun 07master * r87246953d8 10ffmpeg/configure: configure: mark libfdk-aac as nonfree
[22:44] <CIA-41> ffmpeg: 03Kostya Shishkov 07master * r3c6c19184c 10ffmpeg/libavcodec/tscc2.c: tscc2: fix typo in DCT
[22:44] <CIA-41> ffmpeg: 03Kostya Shishkov 07master * r4cfb0d871d 10ffmpeg/libavcodec/tscc2.c: 
[22:44] <CIA-41> ffmpeg: tscc2: do not add/subtract 128 bias during DCT
[22:44] <CIA-41> ffmpeg: It turns out that the reference decoder subtracts 128 from DC during block
[22:44] <CIA-41> ffmpeg: decode but adds it back during reordering block with zigzag pattern.
[22:44] <CIA-41> ffmpeg: Transforming block with incorrect DC caused heavy visual artifacts for
[22:44] <CIA-41> ffmpeg: many quantisers.
[22:44] <CIA-41> ffmpeg: 03Mans Rullgard 07master * rab9f987661 10ffmpeg/ (4 files in 4 dirs): 
[22:44] <CIA-41> ffmpeg: build: add CONFIG_VP3DSP, reduce repetition in OBJS lists
[22:44] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[22:44] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r28f9ab7029 10ffmpeg/libavcodec/ (19 files in 4 dirs): 
[22:44] <CIA-41> ffmpeg: vp3: move idct and loop filter pointers to new vp3dsp context
[22:44] <CIA-41> ffmpeg: This moves all VP3-specific function pointers from dsputil to a
[22:44] <CIA-41> ffmpeg: new vp3dsp context. There is no reason to ever use the VP3 IDCT
[22:45] <CIA-41> ffmpeg: where an MPEG2 IDCT is expected or vice versa.
[22:45] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[22:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r204c4e953d 10ffmpeg/: (log message trimmed)
[00:00] --- Thu Jul 19 2012