[Ffmpeg-devel-irc] ffmpeg-devel.log.20131021

[burek]

[00:26] <cone-729> ffmpeg.git 03Michael Niedermayer 07master:712eff49789a: fate: add synthetic swresample resampling tests
[01:22] <michaelni> mraulet, smarter, it seems icc doesnt like the hevc code, could like a compiler bug though (http://fate.ffmpeg.org/log.cgi?time=20131020212707&log=compile&slot=x86_32-linux-gnu-icc-2011.4.191)
[01:25] <mraulet> and with --disable-asm?
[01:30] <michaelni> i dont know i dont have icc installed here, its thilos fate client
[01:34] <michaelni> just sent thilo a mail, you are in CC
[11:20] <cone-755> ffmpeg.git 03Mickaël Raulet 07master:09ba986cae14: avcodec/hevc: fix duplicate/wrong nal unit check and difference to openhevc_upstream
[12:57] <durandal_1707> geq=p(X+sin(2*PI*Y/50+N)*5\,Y):p(X+sin(2*PI*Y/50+N)*5\,Y):p(X+sin(2*PI*Y/50+N)*5\,Y)
[13:01] <durandal_1707> 2/win3
[13:10] <saste> michaelni, can't see ffprobe coverity issues, did you mean tee?
[13:10] <Daemon404> i saw it earlier
[13:10] <Daemon404> it listed it as ffprobe.c
[13:12] <durandal_1707> there is no way to find out if expression matches previous one?
[13:14] <saste> durandal_1707, st(0, expr0); st(1, expr1); if(eq(ld(0), ld(1), ...)
[13:15] <saste> Daemon404, yep
[13:15] <durandal_1707> saste: i mean inside code
[13:16] <durandal_1707> idea is to not recalculate some values. which may be same all the time
[13:16] Action: durandal_1707 thinking about making ripple filter, which would basicaly do above expression but much faster
[13:17] <saste> durandal_1707, we discussed some way to mark constant expressions
[13:17] <saste> in some cases it can be done, but we need some tweaks in the eval code
[13:17] <durandal_1707> can i just to memcmp?
[13:24] <saste> durandal_1707, AVExpr is a recursive struct IIRC
[13:39] <cone-755> ffmpeg.git 03Stefano Sabatini 07master:ddaf33f5bba5: ffprobe: fix consistency checks in parse_read_intervals()
[13:47] <Compn> or just optimize geq ?
[13:47] <Compn> :P
[13:47] Action: Compn is late
[13:50] <cone-755> ffmpeg.git 03Carl Eugen Hoyos 07master:87eae03aacaf: Support FF_DISABLE_DEPRECATION_WARNINGS with icc on Linux.
[14:34] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:80b8f78048ba: fate: enable hevc tests
[15:46] <cone-755> ffmpeg.git 03Paul B Mahol 07master:f5ed83c3831f: avformat/mxfdec; parse random index pack to get last partition offset and use it when footer partition offset is missing
[15:57] <cone-755> ffmpeg.git 03Paul B Mahol 07master:c65521693914: make it possible to compile ffmpeg without swscale
[16:24] <mateo`> anyone at linuxcon ?
[16:25] <durandal_1707> where is it?
[16:30] <mateo`> edinburgh
[16:51] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:ab2bfb85d49b: avfilter/ff_insert_pad: fix order of operations
[17:29] <saste> wm4, what happened to your libswscale API patch?
[17:39] <durandal_1707> picked invisible color
[17:52] <wm4> saste: waiting for this filter stuff...
[17:52] <saste> wm4, will send an updated patch now
[18:05] <Daemon404> michaelni, ping
[18:05] <michaelni> Daemon404, pong
[18:05] <Daemon404> michaelni, what do you mean by do "i" have padidng input bytes?
[18:06] <Daemon404> crash was produced with ffmpeg cli, not my own code
[18:06] <michaelni> then theres a bug
[18:07] <durandal_1707> in overread
[18:08] <Daemon404> not sure what you mean
[18:08] <Daemon404> @ michaelni 
[18:08] <michaelni> can you share the file / testcase that causes the crash ?
[18:09] <Daemon404> let me see
[18:12] <michaelni> if not some valgrind otput that shows where it was allocated might be interresting too
[19:48] <cone-755> ffmpeg.git 03Diego Biurrun 07master:23e6d3d4e27c: fate: vpx: Remove one redundant $(eval) invocation
[19:48] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:e45460aab0a9: Merge commit '23e6d3d4e27c83e06d00aa072c84d057bd1dad23'
[19:55] <durandal_1707> http://tinypic.com/r/wcbw2a/5
[20:07] <cone-755> ffmpeg.git 03Ben Jackson 07master:5152196b2b47: pthread: Avoid spurious wakeups
[20:07] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:2444fca523d7: Merge commit '5152196b2b47d446d9a509b9bfb318daa26a1f2b'
[20:17] <cone-755> ffmpeg.git 03Luca Barbato 07master:ea71aafd6881: avio: Use AVERROR_PROTOCOL_NOT_FOUND
[20:18] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:2f9422d951dc: Merge remote-tracking branch 'qatar/master'
[20:20] <Daemon404> AVERROR_PROTOCOL_NOT_FOUND... thats a new and oddly specific code
[20:21] <durandal_1707> should it be reverted?
[20:22] <Daemon404> no
[20:22] Action: durandal_1707 will push filter without documentation
[20:27] <wm4> it's not really "new"
[20:29] <wm4> though this use was added only in 2012
[20:30] <wm4> and the error code back in 2010 (odd...)
[21:53] <ubitux> is daniel kang on irc?
[21:57] <gnafu> If I recall correctly, that's Jumpyshoes.
[21:57] <ubitux> indeed, just noticed that from grepping irc logs
[21:58] <ubitux> Jumpyshoes: i would suggest to use a macro for the dup in mmx filter line
[21:58] <ubitux> i didn't do that for the inline asm to avoid a big diff (and because it's not really as "free" as with yasm)
[21:59] <ubitux> but i think it would be relevant here
[21:59] <ubitux> also, the comments are actually useful, you should keep them (and puting this code in a macro will avoid the redundancy of them)
[22:03] <ubitux> durandal11707: have fun fixing the deps
[22:05] <durandal11707> ?
[22:14] <durandal11707> ubitux: ???
[22:14] <ubitux> durandal11707: see fate failure
[22:14] <ubitux> i'm assuming this is because of the assumed swscale dep in ffmpeg for the fate tests
[22:15] <durandal11707> who added such flag?
[22:17] <durandal11707> well previously ffmpeg would not get build at all, so nothing was tested at all
[22:17] <durandal11707> i consider that a huge improvement
[22:20] <cone-755> ffmpeg.git 03Paul B Mahol 07master:c9660999b9a7: avfilter: replaygain scanner
[23:36] <durandal11707> michaelni: i get crash with this: ffmpeg -f lavfi -i smptebars -f lavfi -i smptebars -lavfi "[1]cop[x],[0][1]overlay" -f null -
[23:37] <durandal11707> *ffmpeg -f lavfi -i smptebars -f lavfi -i smptebars -lavfi "[1]copy[x],[0][1]overlay" -f null -
[23:40] <durandal11707> or this: ffmpeg -f lavfi -i color -f lavfi -i color -filter_complex "[1]null[x],[0][1]overlay" -f null -
[23:41] <durandal11707> i have basic ripple filter, anybody wants it?
[23:49] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:7e3e653618a5: avformat/bintext: protect against potential overflow of chars_per_frame
[23:49] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:d47e14b53a39: h263dec: use init_get_bits8() and check its return code
[23:58] <cone-755> ffmpeg.git 03Nicolas George 07master:59f809e9922a: lavf/tee: fix leak of strdup/strtoked buffer.
[23:58] <cone-755> ffmpeg.git 03Nicolas George 07master:5a12142aca22: lavf/tee: fix leak of select option.
[23:59] <cone-755> ffmpeg.git 03Nicolas George 07master:c4e6024adc18: lavf/tee: fix leak of bsfs array.
[23:59] <cone-755> ffmpeg.git 03Michael Niedermayer 07master:23c06984995e: Merge remote-tracking branch 'cigaes/master'
[00:00] --- Tue Oct 22 2013