[Ffmpeg-devel-irc] ffmpeg-devel.log.20140918

[burek]

[00:46] <cone-913> ffmpeg.git 03Vitor Sessak 07master:55d11d277bf5: swscale/x86: do not expect registers to be preserved across inline ASM blocks
[00:50] <compn> Ffmpeg made Debian? Yay
[00:53] <wm4> yes
[00:53] <nevcairiel> well, experimental
[00:54] <nevcairiel> until its at least promoted to unstable, i won't celebrate
[00:55] <wm4> true
[00:55] <wm4> well I have no clue about debian politics
[00:55] <wm4> aren't they promoted semi-automatically?
[00:55] <nevcairiel> i doubt it has any chances to still get into testing (ie. next stable), so it will be a while until it is wide spread
[00:59] <cone-913> ffmpeg.git 03Andreas Cadhalpun 07master:739f179dd6a2: vf_deshake: rename Transform.vector to Transform.vec to avoid compiler confusion
[01:00] <J_Darnley> 2.4 will gestate for 9 months till it gets birthed into unstable...
[01:00] <J_Darnley> then it will mature over a few years to make stabele
[01:00] <J_Darnley> *stable
[01:11] <cone-913> ffmpeg.git 03Katerina Barone-Adesi 07release/2.4:b45ab61b24a8: apetag: Fix APE tag size check
[01:11] <cone-913> ffmpeg.git 03Michael Niedermayer 07release/2.4:ab02548c8a71: Merge commit 'b45ab61b24a8f2aeafdd4451491b1b30b7875ee5' into release/2.4
[01:35] <iive> nevcairiel: going through the ftp masters is the hardest part.
[01:41] <cone-913> ffmpeg.git 03Michael Niedermayer 07release/2.4:d694ab846cf1: avcodec/x86/vp9lpf: Always include x86util.asm
[01:41] <cone-913> ffmpeg.git 03Andreas Cadhalpun 07release/2.4:dc4e34a2f059: vf_deshake: rename Transform.vector to Transform.vec to avoid compiler confusion
[02:10] <cone-913> ffmpeg.git 03Jörg Krause 07master:3f2c70355ab7: configure: Use the right variables in check_host_cpp
[02:10] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:343643fdca4f: Merge commit '3f2c70355ab722bc9f741bd3ed8224c7cfb62379'
[02:33] <cone-913> ffmpeg.git 03Martin Storsjö 07master:2bb2c2bd75e5: rtpenc_chain: Pass the initial time_base hint on to the chained muxer
[02:33] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:986f63d17fa1: Merge commit '2bb2c2bd75e5f4b28a945511cda77e0a1a44c758'
[03:15] <kepstin> ooh, nice, the fate tests on my k6 finished.
[03:16] <kepstin> assuming I'm interpreting the results correctly, they passed too.
[03:19] <kepstin> took a little over 10 hours total for one run.
[03:20] <kepstin> I suppose I should set it up to run automatically and submit the results, now :)
[03:43] <pross> kepstin: k6-2 or iii?
[03:48] <kepstin> pross: no
[03:48] <kepstin> "AMD-K6tm w/ multimedia extensions" (i.e with mmx only)
[04:17] <jamrial> kepstin: I'd say disable hevc, vp9 and similar codecs that are not relevant for pre sse processors if you really intend to run fate on that k6
[04:17] <jamrial> it will shave an hour or two of runtime
[04:52] <cone-913> ffmpeg.git 03James Almer 07master:77f9a81ccaad: x86/me_cmp: combine sad functions into a single macro
[05:05] <kepstin> yeah, I suppose without sse, there's no assembly in any of those that would run on this chip
[07:28] <ubitux> michaelni: how did you spot that issue in vp9lpf?
[07:47] <wudi> lpf?
[07:47] <wudi> loop filter?
[07:52] <ubitux> wudi: yes
[08:01] <jamrial> nevcairiel: interesting commit http://git.1f0.de/gitweb?p=ffmpeg.git;a=commitdiff;h=4cc6c06fb7c298dbf081ccb43a272f8606504d0e :p
[08:07] <sunye_> hi , everyone, I am Sun Ye ^ ^
[11:44] <michaelni> ubitux: for f in libavcodec/x86/*.o ; do echo $f ; readelf -S $f | grep GNU-stack || break ; done
[11:44] <ubitux> noted
[11:44] <ubitux> thank you
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:074eef56ef4a: fate: add fate-filter-pp1
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:9f9ebe631dc1: libpostproc/postprocess: avoid some if()
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:44dabf1f422b: postproc: simplify forwarding return codes
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:921caf620386: postprocess: make some variables in pp_get_mode_by_name_and_quality() const
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:9e8be462521e: postproc/postprocess: use av_strtok()
[14:55] <cone-41> ffmpeg.git 03Michael Niedermayer 07master:27f936eca8a1: Merge branch 'postprocwork'
[14:55] <michaelni> j-b, please add  *.mak to the files which may contain tabs in the ffmpeg git hook. We have tabs in *.mak files due to Makefile syntax and it requires each change to such files to be done as a merge currently, this is ugly
[14:56] <michaelni> or tests/fate/*.mak if *.mak would only cover the "root" directory
[15:00] <ubitux> heh
[15:00] <ubitux> second time this happens :)
[15:44] <Daemon404> cd ..
[15:44] <Daemon404> woops
[16:05] <kierank> wow that guy's first name is guerilla
[16:05] <kierank> oh gorilla
[20:25] <ubitux> do we enable PIC by default with shared builds?
[20:27] <ubitux> (it seems to be the case after a simple test)
[20:29] <wm4> I suppose most systems need PIC for shared libs
[20:51] <wm4> how can I make ffprobe output as many CodecContext fields as possible?
[20:52] <wm4> and formatcontext
[21:01] <ubitux> what does that mean?
[21:01] <wm4> there are lots of AVCodecContext fields that might be interesting somehow, and ffprobe should print them
[21:02] <nevcairiel> invent some sort of reflection based code to automatically print the entire context
[21:02] <wm4> also, oh god, mp3dec updates AVCodecContext->skip_samples on seeking
[21:03] <nevcairiel> should send side data
[21:03] <wm4> maybe utils.c actually does
[21:03] <wm4> yep it does
[21:03] <nevcairiel> thats ok then
[21:03] <nevcairiel> not the cleanest solution, but it works
[23:43] <weirddan455> hey is there anyone here who can approve me for the mailing list?  i tried submitting a patch and it says i need approval
[23:44] <J_Darnley> Someone will get to it.
[23:45] <J_Darnley> If you're not subscribed emails get sent for moderation, or something.
[23:51] <llogan> weirddan455: approved
[23:53] <llogan> if you don't want to wait for the lazy moderator in a weird timezone then you can subscribe (but it could be considered a high volume list).
[23:53] <weirddan455> llogan, ok thank you :)
[23:59] <jamrial> weirddan455: you will not get the replies delivered to your email unless you're CCed (or you subscribe to the list)
[00:00] --- Fri Sep 19 2014