[Ffmpeg-devel-irc] ffmpeg-devel.log.20121019

burek burek021 at gmail.com
Sat Oct 20 02:05:02 CEST 2012


[00:04] <durandal_1707> michaelni: why i get stream start is in unreadable format? 2286340.981000
[00:32] <durandal_1707> what to use to convert ascii to integer?
[00:34] <pengvado> strtol
[00:35] <durandal_1707> and how to check for failure?
[00:36] <pengvado> *endptr==ptr implies no conversion happened
[01:48] <cone-541> ffmpeg.git 03Michael Niedermayer 07df727d408c7e: mlp_parser: print error when ff_combine_frame() fails to add the current buffer * 03http://tinyurl.com/8zok49203
[01:48] <cone-541> ffmpeg.git 03Michael Niedermayer 0705b0337025f6: motionpixels/mp_decode_frame_helper: assert that the first pixel doesnt reuse the last. * 03http://tinyurl.com/9etjwdy03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 07aadb7b3ac4fb: lavc/c93: use meaningful error codes * 03http://tinyurl.com/9ezxumo03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 07527224830aad: idcinvideo: if decoding fails return error * 03http://tinyurl.com/9q3wtnr03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 074c6c6a266e9c: jvdec: use more meaningful error code * 03http://tinyurl.com/8twxyjs03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 070bf40e0ef421: lavc/yop: remove redudant YOP in av_log() messages * 03http://tinyurl.com/9nz3q8u03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 07be536f084a98: xxan: return more meaningful error codes * 03http://tinyurl.com/9p8n6zs03
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 07445f36d7c83f: kmvc: use meaningful error codes * 03http://tinyurl.com/9yj2mu203
[02:08] <cone-541> ffmpeg.git 03Paul B Mahol 074ebf30595106: lavc/tta: do not overwrite bits_per_coded_sample * 03http://tinyurl.com/9r9toe503
[02:20] <durandal_1707> some guy on libav-user ml says that libswscale is too slow
[02:21] <Compn> lets rm it and rewrite it hen
[02:21] <Compn> :P
[03:03] <cone-541> ffmpeg.git 03Michael Niedermayer 07c0b17ea106b9: roqaudioenc: Fix crash with very small roq files * 03http://tinyurl.com/975s86m03
[03:51] <cone-541> ffmpeg.git 03Michael Niedermayer 07a06f943f9d65: roqvideodec: replace dead code by assert * 03http://tinyurl.com/9kmsbqy03
[03:51] <cone-541> ffmpeg.git 03Michael Niedermayer 07d50aa006fb34: tiffenc: fix integer overflow * 03http://tinyurl.com/9u432en03
[09:32] <divVerent> ubitux: I have seen the commit
[09:32] <divVerent> what is with all this tinyurl waste BTW?
[09:32] <ubitux> i don't know
[09:32] <ubitux> i don't like it :(
[09:32] <divVerent> do we really need to allocate a new tinyurl for each commit?
[09:32] <divVerent> it just worsens the tinyurl service for everyone else
[09:33] <divVerent> why not http://ffmpeg.org/gd50aa00
[09:33] <divVerent> would be a job for mod_rewrite ;)
[09:33] <ubitux> even the full url with a reduced hash is fine
[09:33] <TimNich>  +1
[09:33] <divVerent> well, I see the point of using short urls
[09:33] <divVerent> but compare
[09:33] <divVerent> http://tinyurl.com/9u432en
[09:33] <divVerent> http://ffmpeg.org/gd50aa00
[09:33] <divVerent> equally long
[09:33] <divVerent> and the latter doesn't eat up tinyurl namespace
[09:34] <divVerent> but only means ffmpeg can no longer have URLs starting with g and having no dot inside ;)
[09:34] <divVerent> could even do http://ffmpeg.org/d50aa00 with some careful regexing
[09:34] <ubitux> ffmpeg.org/h/<hash>
[09:34] <divVerent> or that
[09:34] <divVerent> that's one more character than the tinyurl though ;)
[09:34] <ubitux> i think we have similar redirections with sources.ffmpeg.org
[09:35] <ubitux> source.ffmpeg.org sorry
[09:35] <divVerent> o the tinyurl TOS even allow this?
[09:35] <divVerent> yes, it is
[09:35] <divVerent> well, then that's tinyurl's fault ;)
[09:36] <ubitux> curl -I 'source.ffmpeg.org/?test' [...] Location: http://git.videolan.org/?p=ffmpeg.gittest&test
[09:36] <ubitux> something looks fishy :D
[09:36] <divVerent> hehe
[09:36] <divVerent> that'd work too
[09:36] <divVerent> http://ffmpeg.org/?hash
[09:36] <ubitux> curl -I 'source.ffmpeg.org/?foo=bar' [...] Location: http://git.videolan.org/?p=ffmpeg.gitfoo=bar&foo=bar
[09:37] <ubitux> the dup is weird
[09:37] <divVerent> ubitux: it MAY still make sense
[09:37] <divVerent> there may be a separator that isn't contained
[09:37] <divVerent> but I can't find the one :P
[09:37] <ubitux> curl -sI 'source.ffmpeg.org/?foo=bar' | cat -e [...] Location: http://git.videolan.org/?p=ffmpeg.gitfoo=bar&foo=bar^M$
[09:38] <ubitux> doesn't look like to
[09:38] <divVerent> was just a guess, becuase such duping happens easily in shell scripts splitting strings
[09:38] <divVerent> if s is foo/bar, then ${s%%/*} is foo and ${s#*/} is bar
[09:38] <divVerent> but if s is just foo, both will be foo
[09:39] <divVerent> so MAYBE source.ffmpeg.org uses a matching technique with similar issues
[09:39] <divVerent> sure not a shell script though ;)
[09:40] <ubitux> source.ffmpeg.org/?;a=commit;h=d50aa006
[09:41] <ubitux> i'm able to trick it with this
[09:41] <ubitux> but it's ugly :p
[09:41] <divVerent> but it still dupes, I suppose
[09:41] <divVerent> just that this then at least works
[09:41] <ubitux> yes
[09:42] <ubitux> Location: http://git.videolan.org/?p=ffmpeg.git;a=commit;h=d50aa006&;a=commit;h=d50aa006
[09:42] <ubitux> it's a big ugly :D
[09:42] <divVerent> source.ffmpeg.org/?;h=d50aa006
[09:42] <divVerent> suffices BTW
[09:42] <ubitux> indeed
[09:42] <divVerent> because a=commit is default if h is given
[09:42] <ubitux> i like a=commitdiff though
[09:43] <ubitux> i'd go for source.ffmpeg.org/?;a=commitdiff;h=d50aa006
[09:43] <divVerent> I don't, for a very simple reason
[09:43] <divVerent> because the subject line then is less visible :P
[09:43] <ubitux> how so? oO
[09:44] <ubitux> commitdiff has the commit description
[09:44] <ubitux> in addition
[09:44] <ubitux> ah no sorry it's also in commit as well
[09:50] <divVerent> yes
[09:50] <divVerent> it is there in commitdiff
[09:50] <divVerent> but in commit, it is right above the commit message
[09:51] <divVerent> and in committdiff it's way more above
[10:49] <ubitux> hey, it would be nice to have a shout protocol
[13:36] <cone-191> ffmpeg.git 03Martin Storsjö 07b760ffdd07ea: aviobuf: Remove a senseless ifdef in avio_seek * 03http://tinyurl.com/9ukqgnp03
[13:36] <cone-191> ffmpeg.git 03Martin Storsjö 0753e8cd68b722: configure: Split out msvc as a separate target OS * 03http://tinyurl.com/8sttjkg03
[13:36] <cone-191> ffmpeg.git 03Martin Storsjö 07fc085c5b33a9: gxf: Add a local copy of the relevant parts of the frame rate table * 03http://tinyurl.com/8rrhfe903
[13:36] <cone-191> ffmpeg.git 03Martin Storsjö 07eaa9b2e66c04: avcodec: Rename avpriv_frame_rate_tab to ff_mpeg12_frame_rate_tab * 03http://tinyurl.com/9g8tx6h03
[13:36] <cone-191> ffmpeg.git 03Martin Storsjö 07d66c52c2b369: Add support for building shared libraries with MSVC * 03http://tinyurl.com/8ehw33803
[13:36] <cone-191> ffmpeg.git 03Mans Rullgard 07c0329748b04e: fate: add a dependency helper macro * 03http://tinyurl.com/8bw2wlb03
[13:36] <cone-191> ffmpeg.git 03Michael Niedermayer 07b0554fec0470: Merge commit 'c0329748b04e1f175dad8c9c2ebf22a5e2dc5b72' * 03http://tinyurl.com/9kxlvbp03
[13:54] <nevcairiel> guess now i have to figure out what needs to be fixed for shared msvc ffmpeg builds
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 07eccd0671cc89: fate: dependencies for demux tests * 03http://tinyurl.com/9u2k9wf03
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 07b55dda4a7205: fate-als: add dependencies * 03http://tinyurl.com/9ogn8x503
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 072e05143a8df9: fate-indeo: add dependencies * 03http://tinyurl.com/9f2fhbg03
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 0708c6a12a562d: fate-mpc: add dependencies * 03http://tinyurl.com/8pr5yor03
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 0722f5149ec461: fate-twinvq: add dependencies * 03http://tinyurl.com/97d4waa03
[13:54] <cone-191> ffmpeg.git 03Mans Rullgard 0755351de9f43d: fate: add dependencies for misc microsoft codecs * 03http://tinyurl.com/8hmbtak03
[13:54] <cone-191> ffmpeg.git 03Martin Storsjö 0722310eef9f78: changelog: Mention the MSVC DLL support * 03http://tinyurl.com/9h8hzyf03
[13:54] <cone-191> ffmpeg.git 03Martin Storsjö 070af1fe845a9d: avformat: Fix references to the removed function av_write_header in comments * 03http://tinyurl.com/9btmu3d03
[13:54] <cone-191> ffmpeg.git 03Michael Niedermayer 0793f244e3abc3: Merge commit '0af1fe845a9d7112da0a58d33a4fc81fe7c47e95' * 03http://tinyurl.com/95tlxyk03
[14:10] <cone-191> ffmpeg.git 03Diego Biurrun 077b2121e7e2e1: riff: Move functions around to be covered by appropriate #ifdefs * 03http://tinyurl.com/8jr27k203
[14:10] <cone-191> ffmpeg.git 03Diego Biurrun 07e8fe208be810: fate: dependencies for screen codec tests * 03http://tinyurl.com/9x3kjv303
[14:10] <cone-191> ffmpeg.git 03Mans Rullgard 077a12d97eb1aa: aac: fix build with hardcoded tables * 03http://tinyurl.com/8dcsffq03
[14:10] <cone-191> ffmpeg.git 03Martin Storsjö 070de9380be54e: rtp: Update the check for distinguishing between RTP and RTCP * 03http://tinyurl.com/8ey2ojh03
[14:10] <cone-191> ffmpeg.git 03Martin Storsjö 071c37744963ac: rtsp: Vertically align a constant definition * 03http://tinyurl.com/8bf6l4k03
[14:10] <cone-191> ffmpeg.git 03Martin Storsjö 073f055f8f5f71: rtsp: Allow setting the reordering buffer size via an AVOption * 03http://tinyurl.com/8qtg8s203
[14:10] <cone-191> ffmpeg.git 03Mans Rullgard 071cd432e167b1: configure: fix libcdio check * 03http://tinyurl.com/8jjwqzr03
[14:10] <cone-191> ffmpeg.git 03Michael Niedermayer 0781ff0c24ef05: Merge commit '1cd432e167b1a80853760c89a33606e2b5f229c2' * 03Error03
[14:20] <cone-191> ffmpeg.git 03Kostya Shishkov 078774d5835806: bmv: get a new frame on every decode_frame(), so we can use direct rendering * 03http://tinyurl.com/9hmms5e03
[14:20] <cone-191> ffmpeg.git 03Kostya Shishkov 07169514c440a7: indeo3: do not try to output more lines than we can fit * 03http://tinyurl.com/9vrhule03
[14:20] <cone-191> ffmpeg.git 03Martin Storsjö 0791485e744f8c: fate: Add proper dependencies in lossless-video.mak * 03http://tinyurl.com/9ntr68f03
[14:20] <cone-191> ffmpeg.git 03Martin Storsjö 0761cc99748c11: fate: Add proper dependencies in qt.mak * 03http://tinyurl.com/8htepdb03
[14:20] <cone-191> ffmpeg.git 03Michael Niedermayer 07c5fd9d3c35cb: Merge remote-tracking branch 'qatar/master' * 03http://tinyurl.com/9gkanw703
[15:02] <cone-191> ffmpeg.git 03jamal 079434ead2f34b: fate: Handle lavf-fate tests in a makefile * 03http://tinyurl.com/9vmkneu03
[15:13] <gnafu> Don't we really only need tiny URLs if we plan to tweet about a commit?
[15:13] Action: gnafu scuttles.
[15:36] Action: cbsrobot advises gnafu to fix his tinyurl before fixing others: http://tinyurl.com/gnafu
[15:39] <gnafu> Huh, there's a Gnafu on Twitter.
[15:54] <durandal_1707> troller too?
[16:11] <cone-191> ffmpeg.git 03Petter Ericson 077abf394814d8: mov.c: Check for stsd + m1s tag indicating MOV-wrapped MPEG-PS, and force continued probing if found. * 03http://tinyurl.com/9erz4sh03
[16:26] <gnafu> durandal_1707: Huh, I guess I never quite considered myself a troller.  I didn't think I was good enough.  I do look up to kshishkov and av500, though ;D.
[16:27] <gnafu> And I dunno about that Gnafu, as it seems most of their tweets are in French (and I didn't bother trying to translate).
[16:27] <durandal_1707> gnafu: sorry, I was just trolling
[16:27] <gnafu> :-D
[17:01] <durandal_1707> Compn: ping
[17:07] <j-b> http://technet.microsoft.com/en-us/security/msvr/msvr12-017 
[17:10] <durandal_1707> they are still secret, so?
[17:13] <durandal_1707> now that there is riff tags it is time add exif tags support
[17:22] <ubitux> the hard part with supporting old weird subtitles format is to actually find some samples :P
[17:22] <nevcairiel> how about a tool that can create them? :)
[17:23] <ubitux> might do the trick
[17:23] <ubitux> i have one sample for subviewer1, one for mpsub (though, i should be able to generate some more with mplayer)
[17:23] <mateo`> hi, dumb question about the sony D10 spec, is 8 channels at 24bit allowed ?
[17:23] <ubitux> now about vplayer, aqt or pjs..
[17:28] <ubitux> j-b: would vlc be interested in using lavf/lavc for subtitles at some point?
[17:30] <durandal_1707> j-b: would vlc be interested in QDMC ?
[17:31] <j-b> ubitux: subtitles being?
[17:32] <j-b> durandal_1707: I think this is used in RTSP too right?
[17:33] <ubitux> j-b: right now we have subrip, realtext, sami, subviewer2, jacosub, webvtt (basic markup for this one though), and a few others; basically the more complex ones
[17:33] <ubitux> i plan to add the old ones supported by mplayer pretty soon
[17:35] <ubitux> we're going to change the api anyway, but i was wondering if at some point the vlc project would be interested in moving to lavf/lavc for this
[17:37] <j-b> ubitux: no, I mean: do you mean decoding or rendering?
[17:37] <ubitux> demux and decode
[17:37] <ubitux> right now you would get an ASS string which you could directly send to libass
[17:38] <ubitux> it will change
[17:38] <ubitux> here is how it would look like: http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2012-September/130474.html
[17:39] <ubitux> basically you would be in the video player path in the diagram
[17:39] <ubitux> (btw if you have any comment/suggestion on this)
[17:39] <ubitux> (work is already started)
[17:41] <j-b> decoding, maybe, re-rendering, no.
[17:42] <ubitux> you would be responsible for the rendering
[17:42] <ubitux> what do you mean by "re" rendering?
[17:43] <ubitux> the "libass to render" is in your scope; if your player has libass you can decide to encode the avsubtitle into ass markup; otherwise you can encode it in "raw text" so it can be printf'ed or displayed with a basic osd
[17:44] <ubitux> (that's what the rendering filter in lavfi would do)
[17:44] <nevcairiel> i wish someone would finish the work on libass to make it free of fontconfig on windows
[17:47] <j-b> zgreg did
[17:49] <Plorkyeran> a patch to vlc is not exactly what I would call non-fontconfig support in libass
[17:50] <j-b> He released 0.10.1 and updated his work on harfbuzz in order to go to this point
[17:50] <nevcairiel> there is a branch in the google code git repository, but it has some open TODOs
[18:11] <j-b> nevcairiel: but it is on the way
[18:20] <cone-191> ffmpeg.git 03Ash Hughes 072470851f1228: lavc: enable recursively using avcodec_open2/close. * 03http://tinyurl.com/9w5p8t903
[18:54] <ubitux> -    h->cbp= cbp;
[18:54] <ubitux> +    h->cbp                              = cbp; s->current_picture.f.mb_type[mb_xy] = mb_type;
[18:54] <ubitux> am i the only one to find this ridiculous?
[18:54] <ubitux> (\n after ';', irssi ate it)
[18:56] <ubitux> same for the switch cases.. :/
[18:56] Action: ubitux really doesn't get it
[18:57] <Compn> durandal_1707 : pong
[18:57] <Compn> ubitux : you saw the sub samples in the repo, right ?
[18:58] <durandal_1707> Compn: does playing QDMC works for you?
[18:59] <Compn> in mplayer ?
[18:59] <Compn> i dont have latest svn mplayer, a lot of things are broken due to planar output atm
[18:59] <durandal_1707> but svn mplayer have fixed that
[19:00] <Compn> http://samples.mplayerhq.hu/A-codecs/QDMC/rumcoke.mov works with lavf demuxer and binary quicktime codec
[19:00] <durandal_1707> and it does not need to be latest, because QDMC works only via binary codec
[19:00] <ubitux> Compn: the mpsub in a lost directory on mphq
[19:01] <Compn> for me , using r34835
[19:01] <ubitux> (not the samples directory iirc)
[19:01] <durandal_1707> Compn: good
[19:02] <Compn> ubitux : whut? theres http://samples.ffmpeg.org/sub/
[19:02] <cone-191> ffmpeg.git 03Michael Niedermayer 07d86ef54476b3: avcodec_align_dimensions2: add missing breaks * 03http://tinyurl.com/999d77v03
[19:02] <cone-191> ffmpeg.git 03Michael Niedermayer 070fa26bd4703c: utvideoenc: fix theoretical integer overflow * 03http://tinyurl.com/8l96hz603
[19:02] <Compn> ubitux : http://samples.ffmpeg.org/sub/manyfmts/ contains the mpsub samples generated by some random subtitle program
[19:05] <ubitux> wow. :)
[19:05] <ubitux> thx
[19:06] <Compn> theres also this 
[19:06] <Compn> http://www1.mplayerhq.hu/DOCS/tech/mpsub.sub
[19:12] <Compn> durandal_1707 : anything else you want me to test with quicktime stuff ?
[19:13] <durandal_1707> no, i just need brainz to start REing
[19:16] <Compn> ah
[19:16] <Compn> wonder if kostya did any work on it
[19:17] <Compn> qdm2 has lots of authors :)
[19:17] <Compn> decoder i mean
[19:20] <Compn> durandal_1707 : you've been busy RE'ing lots of stuff :)
[19:20] <durandal_1707> Compn: that was old times...
[19:20] <Compn> ben larsson wrote this up : http://multimedia.cx/mirror/qdmc2.pdf
[19:21] <Compn> has some info on qdmc i think
[19:22] <Compn> oh , no it doesnt :D
[19:22] Action: Compn says stuff before reading stuff
[19:34] <Compn> qdm2.c is hard to follow :)
[19:56] <rrivera> Hey guys, oss colleague of mine is sitting down with the Visual C++ guys re: why oss devs might not be using Visual Studio/VC++ compiler. Looking to collect your feedback and pass it on as I've hacked on ffmpeg/win32 in the past.
[19:56] <rrivera> I understand C99 is important, anything else?
[19:56] <JEEB> C99 features are the biggest reason unless the project is C++
[19:57] <JEEB> ffmpeg now has MSVC support because there's a clang-based converter now
[19:57] <JEEB> which converts /some/ C99 features into C89
[19:57] <rrivera> Ah yeah, saw that. That's kinda lame tho isn't it, lol
[19:57] <JEEB> and makes ffmpeg compile'able in the first place
[19:57] <JEEB> rrivera, well how else would you support it then o_O Because it's pretty sure that MSVC ain't going to support C99/11 any time soon
[19:58] <JEEB> you take the most realistic path that's possible
[19:58] <rrivera> I hear you, yup.
[19:59] <Compn> rrivera : patches are welcome to make ffmpeg work easier in vc++ compilers and apps
[19:59] <JEEB> and then there's plenty of windows hate in various OSS circles, certain people not being ready to make limitations or extra changes regarding windows or MSVC specifically unless pushed a lot
[19:59] <Compn> as long as they dont clutter the code in huge ways...
[20:00] <JEEB> those are the two biggest reasons I know why OSS projects don't use the MS compilers
[20:00] <Compn> actually we support a lot of mingw/msvc now JEEB
[20:00] <JEEB> yes I know that
[20:00] <JEEB> I just noted
[20:00] <Compn> especially because chrome uses msvc to compile ffmpeg ...
[20:00] <Compn> ah
[20:00] <cone-191> ffmpeg.git 03Michael Niedermayer 0731fdf3065dac: vf_idet: reorder operations to avoid division by 0 * 03http://tinyurl.com/9f6zo4903
[20:00] <cone-191> ffmpeg.git 03Michael Niedermayer 07042a738b4599: vf_mp: check list in querry_format() * 03http://tinyurl.com/8jtsj5903
[20:00] <cone-191> ffmpeg.git 03Michael Niedermayer 07aaf78e4d14b4: vf_mp: fix null ptr deref in case of ENOMEM * 03http://tinyurl.com/9mg26dh03
[20:00] <JEEB> I'm just talking about overall (F)OSS traits that lead to MSVC not being used
[20:00] <Compn> ah
[20:01] <rrivera> Yep gotcha. I'll put another tick in the C99 list haha
[20:01] <Compn> rrivera : most of the devels use linux. so they stick to linux tools.
[20:01] <Compn> some use msvc and its debugging tho
[20:02] <JEEB> the first being the fact that stuff uses C99/11 features (and possibly GNU extensions), second being linux/*nix developers and possible wish to not "filth" the project with MSVC-specifics (given the thing can even be built with MSVC to begin with)
[20:02] <Compn> and one of ffmpeg's core goals is to be cross-os-compatable
[20:03] <Compn> so vc++ stuff probably doesnt work well on mac for instance...
[20:03] <Compn> or on ppc, arm, and other arch
[20:04] <JEEB> they did re-add ARM again lately, but yeah
[20:04] <JEEB> it's very specific, and if you're cross-platform then...
[20:04] <Compn> yeah i dont think visualstudio was cross-plat back in the year 2000/2001 when ffmpeg was started
[20:05] <rrivera> I would think Windows is a big consumer of ffmpeg
[20:05] <Compn> it is
[20:05] <rrivera> Interesting that it's a back burner OS for ffmpeg-dev
[20:05] <Compn> vlc distributes millions of copies every year for windows 
[20:06] <kierank> 19:00:17 <Compn> especially because chrome uses msvc to compile ffmpeg ... --> iirc not yet
[20:06] <kierank> not sure about that though
[20:06] <Compn> there was some talk of a script that converts c99 to what vc++ stuff needs
[20:06] <Compn> a pre-compiler script
[20:07] <Daemon404> you mustnt be very observant >.>
[20:07] <Daemon404> chrome does use msvc now (or will extremely soon-- theres stuff sitting in gerrit)
[20:07] <Daemon404> and ffmpeg has built with msvc for a while now
[20:07] <JEEB> <rrivera> Interesting that it's a back burner OS for ffmpeg-dev <- people've used it with mingw toolchains for years, then linking to their MSVC-built software
[20:08] <Compn> yeah, people have been using ffmpeg on windows successfully 
[20:08] <Compn> there is a ffmpeg-msvc port , why arent more vc++ people helping to develop that ?
[20:09] <Daemon404> Compn, ffmpeg git master builds with msvc
[20:09] <Daemon404> stop spreading misinfortmation
[20:09] <Compn> Daemon404 : tell it to kierank
[20:09] <kierank> i mean i don't think chrome use ffmpeg with msvc yet
[20:09] <kierank> iirc
[20:10] <Compn> rrivera : what problems are there wtih ffmpeg ? submitting some bug reports is the quickest way to get ffmpeg to fix things ....
[20:10] <Compn> http://trac.ffmpeg.org/trac/ffmpeg
[20:10] <rrivera> Compn: Hmm, what?
[20:11] <Compn> Daemon404 : http://trac.ffmpeg.org/trac/ffmpeg/wiki/KnownCompilationIssues and http://trac.ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide both mention msvc being uncompilable
[20:11] <Daemon404> i read it on teh internet
[20:11] <Compn> Daemon404 : if the wiki needs to be updated, yell at it for spreading misinfo
[20:11] <Daemon404> on a wiki even!
[20:11] <Daemon404> must be true!
[20:11] <Compn> our wiki! 
[20:11] <rrivera> lol
[20:11] <Daemon404> never mind http://ffmpeg.org/platform.html
[20:11] <Daemon404> that OFFICIAL docs
[20:12] <Compn> oh yeah
[20:12] <Compn> rrivera : so what problems are there with ffmpeg and visual studio ?
[20:12] <Compn> i ask because i dont know them
[20:12] <rrivera> Compn: Not sure, I typically cross-compile. I'm just asking for feedback to pass onto Visual C++ team. :)
[20:13] <Compn> heh
[20:13] <Compn> which team / company are you talking to ? :)
[20:13] Action: Compn updates wiki
[20:13] <rrivera> I'm just passing it onto my colleague Garrett (http://twitter.com/fearthecowboy/status/259340782337671168)
[20:14] <Compn> well the answer is ---- ffmpeg supports vc++ :)
[20:17] <Compn> there.
[20:18] <Compn> rrivera : send him that info on the c99 to c89 script. thats probably useful for a lot of projects...
[20:21] Action: durandal_1707 compiling mplayer, result: libavcodec/x86/cabac.h:170:9: error: ran out of registers during register allocation
[20:22] <Compn> durandal_1707 : ehe, michael vs gcc bug :)
[20:22] <Compn> it should have been fixed
[20:22] <Compn> re run the last command with -fomit-frame-pointer 
[20:22] <durandal_1707> i'm using clang
[20:22] <Compn> ah
[20:22] <Compn> well -fomit-frame-pointers then (i cant remember pointers / pointer)
[20:22] <Compn> if thats possible with clang...
[20:23] <durandal_1707> why mplayer recompiles ffmpeg - how to forece it to use shared lib?
[20:23] <Compn> ./configure --help | grep -i ffmpeg |grep -i external
[20:23] <Compn> or something
[20:23] <Compn> or rm -rf mplayer/ffmpeg
[20:23] <Compn> maybe
[20:23] <Compn> no idea
[20:24] <Daemon404> the better question is
[20:24] <Daemon404> why does mplayer use a shitty fucked up custom build system for ffmpeg
[20:24] <Daemon404> at least mplayer2 got rid of that iirc
[20:25] <durandal_1707> Daemon404: last time i checked it still use some internal stuff
[20:25] <Compn> have to ask dondiego 
[20:26] <Compn> mplayer uses internal stuff, for some filters and other things
[20:26] <Daemon404> durandal_1707, i thought mplayer2 supported system ffmpeg now
[20:26] <Compn> mplayer2 got rid of those filters i think :D
[20:26] <Compn> or maybe just for mplayer2 shared ffmpeg
[20:26] <Daemon404> i bet theyre nto even useful filters
[20:26] <durandal_1707> Daemon404: mplayer2 and mpv support system libav*
[20:27] <Daemon404> yeah
[20:27] <Daemon404> i havent used 'normal' mplayer in a long time.
[20:28] <durandal_1707> wtf, libpostproc can not compile with latest lavu
[20:28] <durandal_1707> according to mplayer output
[20:29] <beastd> durandal_1707: huh
[20:29] <durandal_1707> Checking for libpostproc ... Package 'libpostproc' requires 'libavutil = 51.41.100' but version of libavutil is 51.65.100
[20:30] Action: Compn runs away from version hell
[20:31] <beastd> durandal_1707: Some pkg_config problem it seems
[20:31] <durandal_1707> yea
[20:31] <durandal_1707> i forgot to remove it when removing lib
[20:39] <durandal_1707> Compn: looks like with latest mplayer i cant play QDMC
[20:41] <durandal_1707> forget that it works
[20:41] <durandal_1707> something was corrupted in binary blobs
[20:53] <Compn> durandal_1707 : yeah , theres some versions of the binary which dont work (qt 6.5 i think does not, but qt 6.3 does)
[21:58] <ubitux> <@Compn> http://www1.mplayerhq.hu/DOCS/tech/mpsub.sub // i was talking about this one :)
[22:55] <cone-191> ffmpeg.git 03Michael Niedermayer 073d48dd01fd3d: avidec: remove unneeded null check * 03http://tinyurl.com/9nyphoz03
[22:56] <cone-191> ffmpeg.git 03Michael Niedermayer 07d30351363f6f: url_alloc_for_protocol: fix use of uninitialized variable * 03http://tinyurl.com/9cb38se03
[22:56] <cone-191> ffmpeg.git 03Michael Niedermayer 07a96577df3846: avio: fix sizeof argument * 03http://tinyurl.com/9hfsfp803
[23:36] Action: durandal_1707 Yes!!
[23:37] <ubitux> @_@
[23:40] <cone-191> ffmpeg.git 03Michael Niedermayer 07395caf3de84a: hls: fix integer overflow * 03http://tinyurl.com/8snqvdk03
[23:40] <cone-191> ffmpeg.git 03Michael Niedermayer 07ba39303050c1: gxfenc: fix null ptr dereference * 03http://tinyurl.com/9osqqhj03
[23:40] <cone-191> ffmpeg.git 03Michael Niedermayer 078fb8d539a4c5: hlsproto: fix integer overflow * 03http://tinyurl.com/96eg79o03
[00:00] --- Sat Oct 20 2012


More information about the Ffmpeg-devel-irc mailing list