[Ffmpeg-devel-irc] ffmpeg-devel.log.20170125

burek burek021 at gmail.com
Thu Jan 26 03:05:03 EET 2017


[00:00:11 CET] <nevcairiel> will turn the box back on when i get home next week
[00:00:53 CET] <Zeranoe> I'm curious about Windows testing, so if anyone knows of a test that has been run on a Windows box with >4 cores, let me know.
[00:01:36 CET] <Zeranoe> I don't really see any reason it wouldn't work, but who knows.
[00:01:46 CET] <Zeranoe> Maybe it gets slower
[00:02:22 CET] <nevcairiel> by 6 core/12 thread cpu runs avcodec fine on windows
[00:03:04 CET] <nevcairiel> using all 12 threads fully
[00:08:10 CET] <Zeranoe> nevcairiel: Wonderful, that's a result
[00:08:32 CET] <Zeranoe> nevcairiel: You wouldn't happen to be testing with Zeranoe's builds? ;)
[00:08:42 CET] <nevcairiel> no, i make my own
[00:09:29 CET] <Zeranoe> nevcairiel: w32threading?
[00:09:34 CET] <nevcairiel> yes
[00:10:24 CET] <Zeranoe> nevcairiel: Do you need something that isn't in my builds, or is there something wrong?
[00:11:02 CET] <nevcairiel> there is way too much shit in there, but i wouldnt use anyones build no matter what, why wouldnt i just make my own
[00:11:35 CET] <Zeranoe> To each there own I suppose
[00:11:43 CET] <Zeranoe> their*
[00:12:11 CET] <nevcairiel> also i'm a developer, hard to develop if you cant  build the code .)
[00:12:48 CET] <Zeranoe> It does make it difficult
[00:14:54 CET] <Zeranoe> nevcairiel: Do you use MinGW-w64?
[00:15:14 CET] <nevcairiel> both that and msvc
[00:16:27 CET] <Zeranoe> I think I can make the conclusion that my builds would preform the same. Only difference is that I cross compile, which I would argue is better anyway.
[00:17:09 CET] <nevcairiel> how is cross-compiling better then native compiling? :)
[00:17:25 CET] <Zeranoe> nevcairiel: Speed, for one
[00:17:41 CET] <Zeranoe> nevcairiel: You're adding an additional abstraction layer by cross compiling a cross compiler
[00:18:04 CET] <nevcairiel> well thats a one-time thing
[00:18:12 CET] <Zeranoe> I've been scarred
[00:19:01 CET] <Zeranoe> If I build native at all it's with Cygwin, which "attempts" to be Linux on Windows
[00:19:18 CET] <nevcairiel> and thats why cygwin is evil
[00:19:32 CET] <nevcairiel> if i want to build for windows, then dont pretend i'm on linux :D
[00:19:53 CET] <Zeranoe> That's exactly what  MSYS is
[00:20:01 CET] <Zeranoe> It's just that it isn't emulating the kernel
[00:21:20 CET] <nevcairiel> cygwin is mostly designed to exist inside cygwin, while msys is designed to integrate/interact with windows
[00:21:31 CET] <Zeranoe> I rarely even use Cygwin though, pretty much only for MSVC, which at that point you're throwing the build out the window anyway
[00:22:19 CET] <Zeranoe> Well if you compile with Cygwin's GCC yeah, but Cygwin comes with cross compilers, which I've used with success.
[00:23:11 CET] <nevcairiel> but the point is cross-compiling from cygwin is mostly like cross-compiling from an actual linux box, its extremly self-contained and closed-off
[00:23:39 CET] <nevcairiel> while msys you can just call from windows command prompts and whatnot
[00:23:40 CET] <Zeranoe> Right, which is perfect, and how it should be IMHO
[00:23:56 CET] <Zeranoe> Cygwin works great from cmd :)
[00:24:11 CET] <nevcairiel> if all you do is open a cygwin instance
[00:24:20 CET] <nevcairiel> but not call individual small unix tools individually
[00:24:51 CET] <Zeranoe> Why not? cygwin.bat (ships with Cygwin) runs bash -l
[00:25:49 CET] <Zeranoe> You can use mintty (how I'm running WeeChat right now actually), but you can use Cygwin/Bash from cmd just fine.
[00:41:46 CET] <RiCON> Zeranoe: there's no point in --enable-decklink if you don't --disable-w32threads
[00:41:54 CET] <RiCON> ffmpeg silently disables decklink
[00:44:17 CET] <cone-757> ffmpeg 03Carl Eugen Hoyos 07master:75bd4ea02400: lavf/rtmpproto: Make bytes_read variables 64bit.
[00:44:29 CET] <Zeranoe> RiCON: I noticed that too
[00:44:45 CET] <Zeranoe> RiCON: The <pthread.h> is breaking it
[00:46:03 CET] <Zeranoe> RiCON: I think DeckLink would benifit from becoming thread library independent
[00:52:34 CET] <Zeranoe> I find it surprising it's in C++ too
[00:56:26 CET] <cone-757> ffmpeg 03Carl Eugen Hoyos 07master:9d5141d1fb4c: lavd/v4l2: Avoid setting frame_size to a negative value.
[00:56:27 CET] <cone-757> ffmpeg 03Carl Eugen Hoyos 07master:0b607228bf72: Cosmetics: Reindent after last commit.
[00:57:01 CET] <cone-757> ffmpeg 03Michael Niedermayer 07master:2080bc337179: avcodec/utils: correct align value for interplay
[05:48:17 CET] <cone-984> ffmpeg 03compn 07master:5316ed899f35: isom: map xalg and avlg to h264, fixes ticket #6099
[08:01:53 CET] <neeraj> Hi 
[08:02:04 CET] <neeraj> I am a GSOC'17 aspirant
[08:02:27 CET] <neeraj> I want to work on this : https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2017#FITSimagedecoderandencoder project
[08:02:47 CET] <neeraj> Currently, there is no qualification task
[08:03:09 CET] <neeraj> can someone please guide me to proceed further
[08:04:26 CET] <crelloc> email the mentor
[08:06:36 CET] <neeraj> alright! thanks !
[11:49:58 CET] <cone-481> ffmpeg 03Carl Eugen Hoyos 07master:a135b017dec5: lavf/mov: Unscramble dref debug output.
[12:00:59 CET] <cone-481> ffmpeg 03Paul B Mahol 07master:b953aec3c408: avformat: add Sample Dump eXchange demuxer
[12:01:00 CET] <cone-481> ffmpeg 03Paul B Mahol 07master:45ff6ef50e2a: avformat: add Scenarist Closed Captions demuxer
[12:01:01 CET] <cone-481> ffmpeg 03Paul B Mahol 07master:b4a13d442a51: avcodecc/ccaption_dec: remove extra word from long codec description
[12:53:42 CET] <ubitux> durandal_1707: not nice to ignore the fate test request :(
[12:54:55 CET] <durandal_1707> ubitux: i will add it, just need to pick scc file
[12:55:09 CET] <ubitux> alright, thx
[15:32:28 CET] <av500> make: *** No rule to make target 'ffplay'.  Stop.
[15:32:36 CET] <av500> ffplay has been discontinued?
[15:33:44 CET] <BtbN> I don't think there have ever been individual targets for the single programs.
[15:34:09 CET] <BtbN> If ffplay is enabled, a simple make will build it.
[15:36:21 CET] <ubitux> av500: you need sdl3
[15:36:23 CET] <ubitux> sdl2*
[15:40:34 CET] <av500> ubitux: what travesty is this
[15:40:55 CET] <ubitux> what's wrong with it?
[15:41:43 CET] <av500> you mean sdl2?
[15:42:06 CET] <ubitux> i guess, assuming that's what you're talking about
[15:42:49 CET] <av500> ah, sorry missed the sdl2* above :)
[15:43:10 CET] <ubitux> yeah i'm not good with number, above 1
[15:43:25 CET] <ubitux> s/,/s/
[15:43:57 CET] <av500> thx
[16:19:54 CET] <raphael29_> hi. I took a look at FFmpeg gsoc ideas list for this summer and I noticed this interesting idea https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2017?version=9#FITSimagedecoderandencoder
[16:40:00 CET] <Compn> raphael29_ : are you interested in it ?
[16:40:27 CET] <Compn> do you have expericence with coding , and are you able to communicate , share and be counted during a project / deadline ?
[16:41:10 CET] <wm4> "be counted" what
[16:43:11 CET] <raphael29_> Compn: I am a 3rd year CS student and former (2016) GSoC student, so my answear is yes
[16:44:06 CET] <Compn> wm4 : "show up and be counted" e.g. easy to find/reach/contact
[16:44:34 CET] <Compn> raphael29_ : ah, what project did you work on in 2016? 
[16:44:47 CET] <raphael29_> Compn: KStars on Windows within KDE
[16:44:51 CET] <Compn> neato
[16:45:00 CET] <raphael29_> https://community.kde.org/GSoC/2016/StatusReports/RaphaelCojocaru
[16:45:18 CET] <Compn> yeah stick around and talk to the mentor of that gsoc project if you have questions
[16:46:40 CET] <raphael29_> yeah, sure
[17:22:39 CET] <durandal_1707> raphael29_: what you want/need to know?
[17:30:56 CET] <raphael29_> durandal_1707: hi
[17:31:21 CET] <durandal_1707> raphael29_: hi
[17:31:49 CET] <Compn> durandal_1707 is a mentor on that gsoc idea :)
[17:32:00 CET] <raphael29_> as far as I saw, the only prerequisities for this project are C and Git; so this is great
[17:32:02 CET] <raphael29_> yeah, I know
[17:32:11 CET] <raphael29_> I read in the idea description
[17:33:44 CET] <raphael29_> I also have a basic idea of what FITS is
[17:33:59 CET] <raphael29_> I used it in KStars (the astronomy project I worked on)
[17:34:15 CET] <raphael29_> so this would be another good point for this project
[17:35:10 CET] <durandal_1707> raphael29_: you would need first to complete qualification task
[17:35:47 CET] <raphael29_> for now, the qualification task is missing for this idea
[17:36:40 CET] <raphael29_> durandal_1707: what qualification task means? what if I complete it and what if I cannot complete it?
[17:36:59 CET] <raphael29_> and also, what if more than one student complete that task?
[17:37:49 CET] <durandal_1707> thats unlikely
[17:38:16 CET] <durandal_1707> but if it happens another task is given
[17:38:36 CET] <raphael29_> oh, so that task aims to select the most suitable student for that idea
[17:38:41 CET] <wm4> there's someone else on the ML who is interested in this
[17:38:44 CET] <raphael29_> as in interview
[17:43:15 CET] <durandal_1707> wm4: i answered him privately already
[17:53:27 CET] <raphael29_> durandal_1707: so all I have to do now is waiting for the qualification task
[17:55:25 CET] <durandal_1707> raphael29_: well i could give you one already
[17:56:20 CET] <durandal_1707> note that first we must be accepted as project for gsoc
[17:57:09 CET] <Compn> but if you start working now, before we even are accepted, it will give you extra bonus points :)
[17:57:24 CET] <Compn> note that we might not be accepted, google has thousands of projects to choose from
[18:03:48 CET] <raphael29_> yes, I know how this works
[18:04:08 CET] <raphael29_> durandal_1707: I would like to hear the task;
[18:05:44 CET] <durandal_1707> raphael29_: writing xpm decoder or encoder or fixing one of nontrivial bugs on trac
[18:06:17 CET] <durandal_1707> or add dwa decoding to exr
[18:06:31 CET] <durandal_1707> pick what you want
[18:12:57 CET] <raphael29_> durandal_1707: ok, thanks
[20:48:57 CET] <durandal_1707> michaelni: have you received scc file?
[20:53:19 CET] <michaelni> durandal_1707, yes, file uploaded
[22:17:01 CET] <Zeranoe> Is this no longer the case? https://ffmpeg.org/pipermail/ffmpeg-devel/2015-July/176489.html
[22:45:02 CET] <durandal_1707> Zeranoe: Carl is now our dictator
[22:46:01 CET] <BBB> Zeranoe: I believe that officially speaking, we now have a rule by a group of core devs
[22:58:37 CET] <cone-947> ffmpeg 03Sasi Inguva 07master:f227fc4c2a9f: ffmpeg_opt.c: Introduce a -vstats_version option and document the existing -vstats format.
[22:58:38 CET] <cone-947> ffmpeg 03Frank Liberato 07master:95bde49982a8: avformat/flacdec: Check avio_read result when reading flac block header.
[00:00:00 CET] --- Thu Jan 26 2017


More information about the Ffmpeg-devel-irc mailing list