[Ffmpeg-devel-irc] ffmpeg-devel.log.20140118

burek burek021 at gmail.com
Sun Jan 19 02:05:02 CET 2014


[00:06] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:7d91fb305e54: ffmpeg: check ost->finished in reap_filters()
[00:06] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:4ff77d445121: ffmpeg: check ost->finished in check_output_constraints()
[00:06] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:706741e8d797: ffmpeg: properly close down muxers on av_interleaved_write_frame() failure
[00:07] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:5c459504f697: ffmpeg/flush_encoders: dont mux packets once ost->finished has been set
[00:07] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:d73f8976691f: ffmpeg: update statistics only when a packet is actually muxed
[00:07] <cone-6> ffmpeg.git 03Michael Niedermayer 07master:740e6042a016: ffmpeg: change ost->finished to an enum
[01:18] <BBB> michaelni: can you merge this patch from x264 in our x86inc.asm (libavutil/x86/)? it'll fix the assembly speed issue
[01:19] <BBB> michaelni: https://github.com/DarkShikari/x264-devel/commit/8a27cd68d3d6137b8a38ccef1770a1b3ef78bdfb
[01:29] <cone-6> ffmpeg.git 03Loren Merritt 07master:4d55fe720438: x86inc: speed up compilation with yasm
[01:38] <BBB> \o/ ty
[03:25] <jnvsor> Alphamerge should make black into transparent and white into full opaque yes?
[07:00] <cone-109> ffmpeg.git 03Michael Niedermayer 07master:0117ba559470: avcodec/mpegvideo_enc: fix custom matrixes with jpeg
[13:11] <Rodeo> :-)
[13:11] <Rodeo> yeah, software developers absolutely need strong muscles
[13:59] <wm4> maybe we should auto-delete tickets that don't follow the cehoyos template?
[14:09] <BBB> it's kind of an exploit that all tickets are logged in this irc channel
[14:09] <BBB> maybe they should be logged only after cehoyos moves them from unconfirmed to reproduced
[14:23] <Compn> lol
[14:24] <Compn> michaelni : how you lock trac from new users ?
[14:33] <jnvsor> fflogger is repeating himself :P
[14:36] <wm4> spaaaaam
[15:16] <iive> are 3321/2/3 just spam?
[15:18] <michaelni> Compn, llogan, others, Theres a bad content page where you can list things that should be identified as spam, see for example https://trac-hacks.org/wiki/BadContent, the spam problem will not get solved by "ohh spam go crying to admin", you all can do something about it  by creating and maintaining that page
[15:37] <Compn> lol
[15:38] <Compn> i dont know much about trac so i should probably read something 
[15:39] <wm4> I think my suggestion is reasonable
[15:39] <Compn> whas was ?
[15:39] <wm4> cehoyos marks "bad" tickets as invalid anyway
[15:40] <wm4> wm4> maybe we should auto-delete tickets that don't follow the cehoyos template?
[15:40] <Compn> :P
[15:41] <Compn> michaelni : i dont think i have permissions to create that page is what i mean :P
[15:42] <Compn> ooh nevermind
[15:43] <Compn> i see button now!
[15:43] <Compn> hooray
[15:43] <Compn> wm4 : any suggestions for me being blind ? :P
[15:43] <wm4> hm?
[15:44] <Compn> nothing :)
[15:45] <michaelni> Compn, just created it
[15:45] <Compn> ok, is the badcontent plugin enabled in trac? 
[15:45] <Compn> or it automatically follows those rules 
[15:50] <michaelni> also you can use some regular expressions on that page, but not sure where/if they are documented
[15:51] <michaelni> i think the badcontent is either part of trac itself or the spam filter plugin which we have
[15:51] <michaelni> also we will see if it works, if it doesnt then ill investigate why
[15:55] <Compn> should i try to create a ticket with some badcontent rules or will my superuser status get past the spam filter ?
[15:55] <Compn> to see if my ticket gets blocked i mean
[16:00] <Compn> nevermind i'm too lazy
[16:00] Action: Compn goes looking for more to do
[16:06] <michaelni> Compn, dont submit spam, you would get marked as spammer (and we also submit that to external services like akismet utomatically)
[16:06] <michaelni> or maybe just when its manually marked as spam dunno
[16:15] <Compn> ah
[16:15] <Compn> ok
[16:23] <Daemon404> ... lol
[16:26] <jnvsor> Can't even spell muscles it seems
[16:26] <michaelni> Compn, see, they help us testing :)
[16:29] <Compn> ehe
[16:29] <Compn> dang mussels
[16:30] Action: michaelni wonders if the irc log will get marked as spam
[16:31] <ubitux> michaelni: i'm more concerned about the mail trac notification TBH
[16:31] <ubitux> it's making the smtpd actually mail spam to a bunch of developers
[16:36] Action: michaelni droped the karma of authenticated users from 10 to 5
[16:41] <Compn> block all tickets that dont contain 'ffmpeg' or 'video' :P
[16:45] <michaelni> that would block valid things, theres ffplay, there are other applications that just use libavcodec
[16:46] <ubitux> there is still the annoying but probably more efficient captcha thing
[16:47] <jnvsor> If you think they're bots you could add a field hidden with css and dump anything that fills it in
[16:48] <ubitux> yes i think i suggested that a while ago
[16:48] <ubitux> but it's probably trac targetted bot here
[16:57] <michaelni> jnvsor, done
[17:15] <cone-851> ffmpeg.git 03James Almer 07master:26800e386454: vp9/x86: rename ff_avg[48]_sse to ff_avg[48]_mmxext
[17:25] <ubitux> 17:20:55 <+JEEB> I see that MCW is using some OSS faces for x265 marketing https://www.youtube.com/watch?v=yYy5nD25PPc
[17:25] <ubitux> hello j-b & Skyler :))
[17:27] <JEEB> :)
[17:38] <Compn> its that guy from thescreensavers
[17:39] <Compn> good to see hes still got work
[17:40] <Compn> why am i watching this video haha
[19:38] <cone-851> ffmpeg.git 03Martin Storsjö 07master:3dd04cbcf76b: swresample: Add arm&x86 clobber tests
[19:39] <BBB> ubitux: how's lf8/4?
[19:40] <ubitux> BBB: i won't have much time this week end, maybe a few hours tomorrow
[19:40] <ubitux> i'm working on something else
[19:41] <ubitux> 'trying to rush it so i can go back to asm :)
[19:42] <BBB> awh poor you
[21:21] <cone-851> ffmpeg.git 03Reinhard Tartler 07master:210881110d3e: Prepare for 10_alpha2 Release
[21:21] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:06637c92952c: Merge commit '210881110d3e22cca36edeaca118b5e22c998266'
[21:36] <cone-851> ffmpeg.git 03Anton Khirnov 07master:d1b3fabe6945: h264: reset first_field if frame_start() fails for missing refs
[21:36] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:357a733f91ac: Merge commit 'd1b3fabe6945e511bb20fc9ca52b47eb952526ee'
[21:46] <cone-851> ffmpeg.git 03Anton Khirnov 07master:7f0e81db3c4e: h264: limit allowed pred modes in ff_h264_check_intra_pred_mode() to 3
[21:46] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:f2cb3b36aca2: Merge commit '7f0e81db3c4ee6f8ce15058bafa72ce928a89f3f'
[21:53] <Daemon404> 40
[21:55] Action: ubitux wonder if all Daemon404's irc channel switches failure aren't just for showing off the number of channels
[21:56] <Daemon404> if anything i should be ashamed
[21:56] <ubitux> :D
[22:00] <BBB> more simd on ML
[22:02] <ubitux> yay :)
[22:02] <ubitux> BBB: not hitting the macro limit this time?
[22:03] <BBB> michaelni merged the fix from pengvado from x264
[22:03] <BBB> "fix" =~ workaround really
[22:03] <BBB> but anyway
[22:03] <ubitux> yes but you're adding more macros :)
[22:03] <ubitux> so maybe it's going to rise again
[22:03] <BBB> in the realm of 10^5, a handful here or there don't matter
[22:03] <ubitux> :)
[22:04] <BBB> I also added just one new macro
[22:04] <ubitux> it's nice
[22:05] <ubitux> anyway i'm almost done with my rush, so i'll probably work on the lpf8/4 tomorrow
[22:05] <BBB> \o/
[22:05] <BBB> I'll try to do the 4x4 iadst, maybe iwht and then start intra_pred
[22:05] <BBB> for intra pred, I'd like to measure for a few whether the vp8/h264 or this new approach can be determined to be "faster"
[22:05] <BBB> but we'll see
[22:07] <cone-851> ffmpeg.git 03Anton Khirnov 07master:00dbff4c3e04: h264: do not call field_end if we do not have a current picture
[22:07] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:329610303aac: Merge commit '00dbff4c3e048b4abd01bf805725aabff0fa5ee1'
[22:47] <cone-851> ffmpeg.git 03Anton Khirnov 07master:98cc9efc5f6b: h264: prevent two matching fields from being both a short/long ref combination
[22:47] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:59fb3c4be826: Merge commit '98cc9efc5f6b4dc847ed3b9b60b1c3bcd14e6791'
[22:57] <cone-851> ffmpeg.git 03Anton Khirnov 07master:0652e024c680: h264: reset ref count if decoding the slice header fails
[22:57] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:87e46dd5fb8b: Merge commit '0652e024c680420d298cdf3719d0a0c030173fe3'
[23:07] <cone-851> ffmpeg.git 03Anton Khirnov 07master:9ecabd7892ff: h264: reset num_reorder_frames if it is invalid
[23:07] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:2510e820c43c: Merge commit '9ecabd7892ff073ae60ded3fc0a1290f5914ed5c'
[23:20] <cone-851> ffmpeg.git 03Anton Khirnov 07master:8b2e5e42bb9d: h264: check that an IDR NAL only contains I slices
[23:20] <cone-851> ffmpeg.git 03Michael Niedermayer 07master:a52fbe511948: Merge remote-tracking branch 'qatar/master'
[00:00] --- Sun Jan 19 2014


More information about the Ffmpeg-devel-irc mailing list