[Ffmpeg-devel-irc] ffmpeg-devel.log.20131231

[burek]

[01:33] <Daemon404> 30
[01:34] <BBB> ubitux: right, plus that the function pointers are set such that they should all get specific and special simd optimizations
[01:35] <BBB> so daemon404's password is 30?
[01:35] <Daemon404> lol
[01:35] <Daemon404> i mistyped /g 30 as / 30
[01:35] <BBB> ah yes
[01:36] <BBB> yesterday you forgot the /
[02:30] <cone-510> ffmpeg.git 03Anssi Hannula 07master:4a4437c0fbc8: avformat/mpegts: demux timed ID3 metadata
[02:30] <cone-510> ffmpeg.git 03Anssi Hannula 07master:53765ae33b8b: avformat/id3v2: parse ID3 Private frames as extra metadata
[02:30] <cone-510> ffmpeg.git 03Anssi Hannula 07master:d52882faef36: avformat/http: allow the caller to select a request range
[04:29] <cone-510> ffmpeg.git 03Anssi Hannula 07master:7fdf245ab9cb: avformat/id3v2: allow ID3 parsing without AVFormatContext
[09:33] <ubitux> so, we still don't have any slowmotion filter?
[09:35] <rcombs> other than halving the frame rate?
[09:36] <rcombs> :3
[09:38] <ubitux> :)
[11:37] <ubitux> &
[11:37] <ubitux> wtf?
[12:20] <cone-294> ffmpeg.git 03Reimar Döffinger 07master:cbeaf678889b: Avoid using empty macro arguments.
[12:22] <kierank> ubitux: trying to exploit trac i think
[12:23] <ubitux> that's what i thought but it doesn't look like a backdoor
[12:23] <ubitux> maybe it's just random php files for testing exec
[12:23] <nevcairiel> all the posts include a url too, maybe just weird spam
[12:47] <michaelni> Kurty 	Smartmil8 	smartmil8 at gmail.com
[12:50] <michaelni> 37.45.103.193
[13:07] <kierank> michaelni: did you see that paper that's making the rounds about a new arithmetic coder
[13:07] <kierank> michaelni: might be interesting for ffv1?
[13:09] <michaelni> kierank, no
[13:09] <michaelni> and yes, might be interresting
[13:09] <kierank> michaelni: https://dl.dropboxusercontent.com/u/12405967/poster.pdf
[13:09] <kierank> http://arxiv.org/abs/0902.0271
[13:09] <kierank> he's trying to get people to use it on various mailing lists
[13:10] <kierank> https://github.com/Cyan4973/FiniteStateEntropy
[13:13] <ubitux> http://encode.ru/threads/1821-Asymetric-Numeral-System
[13:13] <ubitux> Kostya was saying yesterday "it's both stupid and ineffective"
[13:15] <ubitux> why isn't he baned?
[13:15] Action: durandal_1707 have no powr
[13:15] <michaelni> his account is already deleted
[13:17] <michaelni> ubitux, do you see any new spam on trac ?
[13:17] <michaelni> or just the fflogger notes ?
[13:17] <ubitux> i guess fflogger was lagging
[13:17] <ubitux> only fflogger
[13:18] <michaelni> find burek in that case and troll him into fixing fflogger
[13:29] <ubitux> [default @ 0x359d8c0] 1 invalid UTF-8 sequence(s) found in string 'Lavf53.32.100ê', replaced with ''
[13:29] <ubitux> ERM.
[13:29] <ubitux> why the fuck do we have special chars in the encoder string?
[13:29] <ubitux> is/was our id3 muxer broken?
[13:44] <nevcairiel> this don moir person sure has crazy ideas, now he wants an interrupt callback for decoding, because he can't wait 5ms for the frame to finish
[13:58] <michaelni> ubitux, is that a regression ?
[13:59] <BBB> nevcairiel: use threading
[13:59] <BBB> nevcairiel: i.e. have one thread for the ui and have things like decoding and demuxing in a special thread
[14:00] <BBB> don't overcomplicate stuff in ffmpeg because application developers are incompetent
[14:04] <durandal_1707> add that solution to FAQ
[14:04] <durandal_1707> we want to help users...
[14:17] <cone-811> ffmpeg.git 03Michael Niedermayer 07master:4bfcedb33f7b: avformat: dont run update_initial_duration() twice
[14:32] <nevcairiel> BBB: apparently he wants to stop decoding so he can respond faster to seek requests
[14:32] <nevcairiel> its crazy if you ask me
[14:50] <ubitux> michaelni: it looks like a regression in the demuxer
[14:50] <ubitux> probably the recent id3 patch
[14:50] <ubitux> oh wait maybe not
[14:51] <ubitux> it was present already
[14:51] <ubitux> i still don't know if it's a muxing or demuxing issue
[14:51] <ubitux> it was muxed with Lavf53.32.100
[14:51] <ubitux> there might have been a bug at that point
[14:52] <ubitux> if i remux the mp3 it doesn't happen
[15:12] <j-b> http://buildbot.videolan.org/builders/ios-device/builds/37/steps/App%20compilation/logs/stdio
[15:12] <j-b> FFmpeg fails on ARM asm, while libav does not, just FYI
[15:13] <ubitux> probably related to the recent patch in swscale
[15:14] <j-b> ubitux: and postproc probably affected too
[15:15] <ubitux> there is arm in postproc?
[15:16] <ubitux> our arm instances aren't in the red
[15:17] <nevcairiel> yeah, the reason libav doesnt fail is that it doesnt have those swscale optims
[15:17] <ubitux> it's probably 1c67ad9d93031839e6e5032190cd95f6dfda8f59
[15:17] <ubitux> committed a few days ago
[15:18] <nevcairiel> probably
[15:18] <nevcairiel> thats the files it fails on at least
[15:35] <michaelni> there was a bug fixed in that arm code (af228a9f9f00746615baa4fb72b71a81930c5916) is the build failure prior or after that ?
[15:38] <j-b> michaelni: I'll check.
[15:39] <michaelni> thx, if its after it ill mail the author of that code, he fixed the last issue quickly last time
[15:39] <j-b> michaelni: we use 74a9c92840d3494393938874c1cdd5bdb95cc918
[15:40] <j-b> so, yes, it's including this commit
[15:45] <michaelni> mail sent
[15:50] <j-b> michaelni: thx
[15:50] <cone-811> ffmpeg.git 03Michael Niedermayer 07master:97fca5492a18: avformat/version: bump micro for addition of internal field to AVStream
[15:50] <cone-811> ffmpeg.git 03Michael Niedermayer 07master:21a2fb7e0579: avcodec/hevc: clear HEVClc when its deallocated in hevc_decode_free()
[23:26] <JEEB> happy new year, folks :)
[23:31] <BBB> nevcairiel: if he takes on writing patches and the maintenance burden of keeping it working in the source tree along with fate tests etc., then it's totally fine
[23:31] <BBB> nevcairiel: but he sounds like he can't code, so then yes, it's not worth it
[23:54] Action: Daemon404 reads backlog
[23:55] <Daemon404> person has never heard of threads?
[00:00] --- Wed Jan  1 2014