[Ffmpeg-devel-irc] ffmpeg-devel.log.20151106

[burek]

[00:35:17 CET] <cone-639> ffmpeg 03Michael Niedermayer 07master:cea9eb9520fa: avcodec/dnxhddec: Make mb_scan_index a fixed length array
[01:19:03 CET] <cone-639> ffmpeg 03Michael Niedermayer 07master:973c3dba27d0: avcodec/mpeg12dec: Do not call show_bits() with invalid bits
[02:57:09 CET] <cone-639> ffmpeg 03Michael Niedermayer 07master:9b6fac11da47: avformat/xmv: factor return check out of if/else
[02:57:10 CET] <cone-639> ffmpeg 03Michael Niedermayer 07master:79c4a338e4b2: avformat/xmv: Discard remainder of packet on error
[03:49:11 CET] <cone-639> ffmpeg 03Sebastian Dröge 07master:01509cdf9287: mpegtsenc: Add support for muxing Opus in MPEG-TS
[03:49:12 CET] <cone-639> ffmpeg 03Sebastian Dröge 07master:7d6a4797f17f: mpegtsenc: Implement writing of Opus trim_start/trim_end control values
[03:53:34 CET] <cone-639> ffmpeg 03Ganesh Ajjanagadde 07master:c8780822bacc: swresample/resample: speed up build_filter for Blackman-Nuttall filter
[11:48:42 CET] <canaar> there's this snippet in ffserver.c : https://paste.hamaralinux.org/?e717ecf091d6dfe6#/j0EV6iL4fW10tjfdX7O+oZnLmg+rOW+4zf3TJEvkqY=
[11:49:04 CET] <canaar> Why does it return 0 when its a fail case? I am talking about the label fail1
[11:53:48 CET] <flux> canaar, have you tried asking the mailing list? certainly it doesn't seem to get answered that well on the irc channel ;-).
[11:54:10 CET] <canaar> :D 
[11:54:15 CET] <flux> canaar, in addition, the code seems to be 13 years old, chances are nobody remembers the reason (and doesn't want to bother finding out)
[11:54:27 CET] <canaar> Exactly
[11:54:34 CET] <canaar> I'll try the mailing list
[12:01:28 CET] <kierank> durandal_1707: fuzzing vp9 sooo slow
[12:02:37 CET] <durandal_1707> kierank: buy supercomputer
[12:02:45 CET] <kierank> I might try doing distributed fuzzing
[12:03:11 CET] <flux> no need to buy anything, just click cpu from amazon ;)
[12:03:19 CET] <flux> (..or some other more affordable cloud vendor)
[12:10:48 CET] <durandal_1707> ms made 5 codecs from single one
[12:21:10 CET] <durandal_1707> michaelni: ok to push xma?
[12:36:05 CET] <michaelni> durandal_1707, sure
[13:51:11 CET] <durandal_1707> supporting multichannel xma will be one big mess
[13:51:22 CET] <wm4> what did they (MS) do?
[13:54:24 CET] <cone-272> ffmpeg 03Paul B Mahol 07master:b456ece55731: XMA1 and XMA2 stereo decoders
[13:58:27 CET] <durandal_1707> wm4: packet have only one or two channels, for others you need another packet
[13:59:45 CET] <durandal_1707> and packet may have part of bitstream of previous packet
[14:38:40 CET] <Compn> approved one mail out of moderation queue
[15:38:50 CET] <cone-272> ffmpeg 03Ganesh Ajjanagadde 07master:6f1ddc726f07: avdevice/dshow_enummediatypes: check return of av_malloc
[15:44:08 CET] <Daemon404> that is a weird request for hls
[15:44:15 CET] <Daemon404> whats the use case?
[15:44:24 CET] <Daemon404> (kierank maybe knows?)
[15:46:24 CET] <rcombs> gotta be ad insertion
[16:00:18 CET] <JEEB> yeah
[16:15:41 CET] <nevcairiel> wow the dcadec guy is alive again
[16:19:43 CET] <wm4> I've also just seen that
[16:19:48 CET] <wm4> that's a surprise
[16:34:24 CET] <durandal_1707> who's gonna assimilate dcadec?
[16:51:51 CET] <jamrial> oh nice, dcadec dev is alive
[16:52:39 CET] Action: rcombs flips coin labeled "stable release" and "assimilation into lavc"
[16:52:44 CET] <rcombs> shit, landed on its side
[16:54:53 CET] <jamrial> he will probably do the stable release he had it planned since june
[16:54:59 CET] <jamrial> doubt he's interested in porting it to lavc
[16:56:26 CET] <jamrial> for that we might need a gsoc code slave if nobody volunteers :p
[16:57:17 CET] <nevcairiel> well it would really need someone that already knows the avcodec infrastructure to make a clean integration
[16:57:32 CET] <nevcairiel> if thats even something people want
[17:10:52 CET] <durandal_1707> sh4rm4^bnc: you still have issues with ffm?
[17:13:20 CET] <wm4> the vp9 parser sure gets a lot of patches (even I sent one once)
[17:22:03 CET] <jamrial> could be people finding issues with youtube-dl when it tries to remux webm dash vp9 streams using ffmpeg
[17:42:33 CET] <BBB> wm4: I admittedly spent several months making the vp9 decoder good, and I probably spend about 1 minute writing the vp9 parser
[17:42:39 CET] <BBB> wm4: so it does suffer from a lack of love and attention
[17:43:52 CET] <wm4> I have no idea what the new patch does, but at least it removes the big macro
[17:44:06 CET] <BBB> you dont like the macro? :(
[17:44:15 CET] <wm4> oh, the new code was copy pasted from vpx?
[17:44:18 CET] <wm4> lol no
[17:44:33 CET] <BBB> ...
[17:44:43 CET] <BBB> why dont you review it then
[17:44:44 CET] <wm4> multiline macros are evil
[17:48:29 CET] <BBB> I think all you need to do to fix the bug is change the return *out_size in the macro with return s->n_frames > 0 ? *out_size : size /* i.e. include idx tail */;
[17:48:45 CET] <BBB> that way it consumes the entire packet in the case of a single-frame superframe
[17:48:56 CET] <BBB> I also think we should discourage that for vp10&
[17:57:42 CET] <BBB> there, much better
[18:09:52 CET] <fritsch> philipl: are you at the GsoC finals currently?
[18:10:02 CET] <fritsch> philipl: if yes - you could meet fernetmenta in person
[18:10:09 CET] <fritsch> he is there with our gsoc student
[20:42:31 CET] <sh4rm4^bnc> durandal_1707, i'd still like to find a way to play the copy of my ffserver's ffm file
[20:43:22 CET] <durandal_1707> how it was created?
[20:44:12 CET] <sh4rm4^bnc> ffserver -d -f /etc/ffserver.conf
[20:44:23 CET] <sh4rm4^bnc> it's the ffm file from /tmp
[20:46:30 CET] <durandal_1707> what ffmpeg version created it?
[20:47:17 CET] <sh4rm4^bnc> ffserver version 2.6.3 Copyright (c) 2000-2015 the FFmpeg developers
[20:49:03 CET] <durandal_1707> have you tried to decode it with exactly same ffmpeg?
[20:49:12 CET] <sh4rm4^bnc> yes
[20:49:44 CET] <sh4rm4^bnc> it appears as if the blocks of the file are out of order
[20:50:15 CET] <sh4rm4^bnc> [ffm @ 0xf183a0] invalid stream index 203
[20:50:15 CET] <sh4rm4^bnc> [ffm @ 0xf183a0] invalid stream index 10
[20:50:15 CET] <sh4rm4^bnc> [ffm @ 0xf183a0] invalid stream index 239
[20:53:08 CET] <durandal_1707> what video codec?
[20:54:30 CET] <durandal_1707> perhaps you could force raw video demuxer and hope it will find all frames
[20:55:10 CET] <durandal_1707> h264 for example
[22:16:59 CET] <cone-479> ffmpeg 03Michael Niedermayer 07master:daefd8ab2f2a: avcodec/dirac_parser: Fix undefined memcpy() use
[23:12:19 CET] <cone-479> ffmpeg 03Ronald S. Bultje 07master:20db54906f55: vp9_parser: allow superframes with a single frame.
[00:00:00 CET] --- Sat Nov  7 2015