[Ffmpeg-devel-irc] ffmpeg-devel.log.20131105

[burek]

[00:00] <iive> llogan: ?
[00:06] <llogan> iive: ffmpeg.org is having issues
[00:08] <cone-910> ffmpeg.git 03Stefano Sabatini 07master:988e2e7c1c18: lavfi/ladspa: cast return value of av_x_if_null to char*, fix warnings
[00:08] <cone-910> ffmpeg.git 03Stefano Sabatini 07master:246e323de8a1: lsws/utils: fix typo
[00:14] <ubitux> BBB: ping
[00:35] <Compn> ffmpeg.org still down ?
[00:46] <ubitux> Compn: seems so
[01:00] <Compn> maybe one should mail root at mphq 
[01:02] <llogan> how would they get it if it is also the mail server (assuming more than httpd is down)?
[01:03] <michaelni> it seems more is down, i didnt had any luck with reaching other boxes on the subnet
[01:17] <Compn> llogan : i had that wonder too
[01:21] <Compnn> michaelni : any chance for fixing up the bayer rgb patch i pinged about ?
[01:21] <lukaszmluki> Hi, it is just me or ffmpeg.org/trac/ffmpeg/ is not working?
[01:23] <michaelni> Compnn, i guess i should ...
[01:28] <llogan> lukaszmluki: it is not just you
[01:39] <BBB> ubitux: pong sortof
[01:40] <BBB> (will be on and off in bits)
[01:43] <cone-910> ffmpeg.git 03Michael Niedermayer 07master:04047b6166b3: avformat/http: print http headers at AV_LOG_DEBUG level
[01:47] <michaelni> btw, until the server is online again (which i hope will be very soon, but i have no idea what is wrong), dont hesitate to CC patches to me, so i can push them if they are trivial & ok
[01:51] <Compnn> is anyone going to md5 the entire tree to make sure it wasnt backdoored when we werent looking? :P
[01:52] <llogan> i don't think the problem is with the server itself
[01:52] <Compnn> did they launch the university into orbit ?
[01:52] <Compnn> is hungary still a country?
[01:54] <wm4> Compnn: git is already sha-hashed by design, literally every ffmpeg developer would notice if someone tempered with the tree
[01:55] <wm4> the other stuff is a different question of course
[02:02] <llogan> Compnn: did pross make the initial bayer patch?
[02:28] <Compnn> uhm
[02:30] <Compnn> On Sun, 16 Dec 2012 12:48:21 +1100, Peter Ross wrote:
[02:30] <Compnn> >This is an rebase and cleanup of some earlier patches that make
[02:30] <Compnn> i dont know
[02:30] <Compnn> http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/2008-March/056719.html
[02:31] <Compnn> [MPlayer-dev-eng] [PATCH] add 8 bit Bayer format and filter
[02:31] <Compnn> Guennadi Liakhovetski g.liakhovetski at pengutronix.de
[02:32] <Compnn> i guess it was written as a filter back in 2008
[02:32] <Compnn> for mplayer
[02:34] <llogan> oh. thanks.
[02:41] <Compnn> 8bit bayer should be listed in v4l docs
[02:41] <Compnn> i have no idea what r3d uses with bayer
[02:46] <Paranoialmaniac> nevcairiel: the isom and mov spec doesn't forbid that audio has stss atom. stss atom just marks random access points. well, more restrictions depend on derived specifications though. 
[07:26] <burek> ffmpeg.org seems to be down
[07:27] <Compnn> yes.
[07:28] <ubitux> still down? meeh
[07:28] <ubitux> :(
[09:31] <ubitux> yay it's back
[09:35] <ubitux> smtpd still down?
[09:35] <ubitux> or maybe under heavy load :p
[10:31] <cone-557> ffmpeg.git 03Diego Biurrun 07master:e2b5b097898c: x86: rv40dsp: Use PAVGB instruction macro where appropriate
[10:31] <cone-557> ffmpeg.git 03Michael Niedermayer 07master:934e489ee8f4: Merge commit 'e2b5b097898c9155f4bdff4d83cdc54d5eef6930'
[11:09] <ubitux> yay mail reup
[11:19] <michaelni> mailman had a stale lock ...
[11:20] <cone-557> ffmpeg.git 03Diego Biurrun 07master:70cbf33405f5: vdpau: Add missing #includes to fix standalone header compilation
[11:20] <cone-557> ffmpeg.git 03Michael Niedermayer 07master:e1459f4b0bb2: Merge commit '70cbf33405f50dfaf77e85f382a188acf17dc71a'
[12:14] <cone-557> ffmpeg.git 03Diego Biurrun 07master:48715285f69d: graph2dot: Fix use of deprecated API
[12:14] <cone-557> ffmpeg.git 03Michael Niedermayer 07master:de970d49fede: Merge commit '48715285f69d70879353737d7bed049f89f58096'
[12:17] <cone-557> ffmpeg.git 03Diego Biurrun 07master:82c86d0b1839: graph2dot: Add missing #include for av_get_channel_layout_string()
[12:17] <cone-557> ffmpeg.git 03Michael Niedermayer 07master:6a7de2361b94: Merge commit '82c86d0b1839b01b3cd0dd848fb6416ff21506dd'
[12:25] <cone-557> ffmpeg.git 03Diego Biurrun 07master:7d1a68e95baf: avfilter/buffer: Suppress deprecation warning for avfilter_unref_buffer
[12:25] <cone-557> ffmpeg.git 03Michael Niedermayer 07master:be74cb5159ef: Merge remote-tracking branch 'qatar/master'
[13:11] <sspiff> I'm trying to figure out how the demuxing/decoding of dvb subtitles works, but I can't seem to figure it out.
[13:13] <sspiff> I'm looking at libavformat/mpegts.c, which is used for TS demuxing, but the only subtitle specific bit I can find is in ff_parse_mpeg2_descriptor, and it doesn't seem to pass through the subtitle case for my test streams.
[13:14] <kierank> dvb subtitling is done through a descriptor
[13:15] <kierank> and there's a table somewhere in the demux
[13:15] <sspiff> kierank: StreamType DESC_types[] { ... } is what you're talking about?
[13:16] <kierank> probably
[13:16] <kierank> can't remember the name
[13:17] <sspiff> looks like it, so these entries contain an AVCodecID, which I presume is used to identify the codec to pass the packets to?
[13:18] <kierank> yes
[13:21] <ubitux> saste: static const in the ScaleAlgo patch maybe
[13:22] <sspiff> kierank: this is used in ff_parse_mpeg2_descriptor, which I was looking at (and only there according to cscope), and it doesn't enter the switch-case clause that handles the subtitles
[13:25] <cone-557> ffmpeg.git 03Paul B Mahol 07master:7722ae35eb63: avcodec/c93: stop using deprecated avcodec_set_dimensions
[13:28] <cone-557> ffmpeg.git 03Paul B Mahol 07master:1a244c22ea83: avcodec/brender_pix: stop using deprecated avcodec_set_dimensions
[13:39] <cone-557> ffmpeg.git 03Paul B Mahol 07master:22c00962c283: avcodec/flashsv: use init_get_bits8()
[13:39] <cone-557> ffmpeg.git 03Paul B Mahol 07master:6e094f702639: avcodec/flashsv: stop using deprecated avcodec_set_dimensions
[13:39] <cone-557> ffmpeg.git 03Paul B Mahol 07master:290be34a745f: avcodec/brender_pix: stop calling av_image_check_size() twice
[13:41] <sspiff> never mind, I was wrong - turns out #define DEBUG 1 didn't turn on the av_dlog prints
[13:41] <durandal_1707> where you put it ?
[13:43] <ubitux> durandal_1707: ...
[13:43] <ubitux> -    init_get_bits(&gb, avpkt->data, buf_size * 8);                                        
[13:43] <ubitux> +    if ((ret = init_get_bits(&gb, avpkt->data, buf_size)) < 0) 
[13:43] <ubitux> where is the 8?
[13:43] <durandal_1707> awwwwwww
[13:45] <cone-557> ffmpeg.git 03Paul B Mahol 07master:ad1dfe04a1bf: avcodec/tscc2: use init_get_bits8()
[13:45] <cone-557> ffmpeg.git 03Paul B Mahol 07master:c3311515d513: avcodec/flashsv: fix typo
[13:46] <ubitux> thx
[13:48] <sspiff> durandal_1707: just under the header comment, before the includes
[13:48] <durandal_1707> sspiff: and what file is that?
[13:49] <durandal_1707> looks like atrac+ decoder is lost because of Don
[13:51] <durandal_1707> by insisting in putting stuff that only one decoder uses in separate .c file
[14:03] <cone-557> ffmpeg.git 03Stefano Sabatini 07master:2a6c95d85f98: lsws/utils: introduce scale_algorithms array
[14:05] <Compnn> don moir ?
[14:07] <durandal_1707> Compnn: Don Corleone
[14:10] <sspiff> durandal_1707: mpegts.c
[14:11] <Compnn> durandal_1707 : i dont see him in the list?
[14:11] <Compnn> maybe i should go back further
[14:12] <durandal_1707> sspiff: and no av_dlog appear?
[14:12] <sspiff> durandal_1707: none whatsoever
[14:12] <sspiff> replacing them by av_log(..., 0, ...) makes it work
[14:13] <sspiff> so there's something strange going on with the DEBUG
[14:13] <durandal_1707> perhaps you do not use -v debug?
[14:13] <Compnn> durandal_1707: i dont see don careone name anywhere in ffmpeg or libav list ?
[14:13] <Compnn> r carleone
[14:14] <durandal_1707> Compnn: watch film and you will see it
[14:15] <sspiff> durandal_1707: I was unaware :(
[14:15] <Compnn> oh lol
[14:20] <durandal_1707> 'spaghetti factor'
[14:21] <cone-557> ffmpeg.git 03Paul B Mahol 07master:b79f543c5869: avcodec/mss1: use init_get_bits8()
[14:21] <cone-557> ffmpeg.git 03Paul B Mahol 07master:82f5c4d737ca: avcodec/flashsv: use av_freep() for tmpblock
[14:21] <cone-557> ffmpeg.git 03Paul B Mahol 07master:e3f384b42835: avcodec/mss2: use init_get_bits8()
[14:38] <cone-557> ffmpeg.git 03Lenny Wang 07master:668255479085: avutil/opencl: compile kernels separately
[14:38] <cone-557> ffmpeg.git 03Lenny Wang 07master:89a3be8921e2: avfilter/opencl: compile kernels separately
[15:55] <sspiff> ubitux: dvb-to-dvd subtitle transcoding was also broken, right?
[15:59] <ubitux> sspiff: problably, i admit i didn't try
[15:59] <sspiff> ubitux: it works here
[16:00] <sspiff> (using my system ffmpeg, not built from source)
[16:00] <ubitux> ah, surprising, ok
[16:00] <ubitux> did you find the root cause of the problem?
[16:00] <sspiff> transcoding to theora/dvdsub mkv's works for me
[16:00] <sspiff> no
[16:00] <sspiff> still looking
[16:01] <sspiff> but I've been looking in the decoding/demuxing, which is probably fine since dvdsub transcoding works?
[16:01] <ubitux> honestly i don't know
[16:01] <sspiff> neither do I :)
[16:01] <sspiff> I'll continue looking
[16:02] <ubitux> dvb -> dvb remux only or transcoding are broken
[16:02] <ubitux> dvd -> dvb works
[16:02] <sspiff> does ffmpeg decode/re-encode for dvb->dvb?
[16:02] <ubitux> yes
[16:02] <sspiff> and dvb -> dvd works
[16:02] <ubitux> you can can't convert without transcode
[16:03] <sspiff> well you don't really convert anything if you stick with dvb subtitles, unless you want to change the clocks etc?
[16:03] <kierank> you can lower the colour depth iirc
[16:03] <sspiff> ah I see
[16:46] <ubitux> arg multistream sami
[16:46] <ubitux> why the hell do ppl still use sami?
[17:39] <michaelni> mateo`_, as you are movenc maintainer you should have op here
[19:37] <cone-834> ffmpeg.git 03Clément BSsch 07master:87434cf37340: avcodec/vp9: add ff_vp9_idct_idct_{4x4,8x8}_ssse3().
[19:56] <ubitux> no one else wants to work on vp9?
[20:30] <llogan> anyone find out what the deal was with the server connection?
[20:53] <Compnn> llogan : why not ask arpi ?
[21:06] <michaelni> llogan, Compnn long power break
[22:06] <cone-834> ffmpeg.git 03Diego Biurrun 07master:92352a94c165: configure: Move -Wno-maybe-uninitialized check into gcc section
[22:06] <cone-834> ffmpeg.git 03Michael Niedermayer 07master:6cfaf9cde875: Merge remote-tracking branch 'qatar/master'
[23:39] <BBB> durandal maybe?
[23:42] <BBB> ubitux: you don't like vp9 anymore? ;( :P
[23:57] <cone-834> ffmpeg.git 03Michael Niedermayer 07master:66872161cbb7: avutil/common: error out with clear message if __STDC_CONSTANT_MACROS is not defined with c++
[00:00] --- Wed Nov  6 2013