[Ffmpeg-devel-irc] ffmpeg-devel.log.20141102

burek burek021 at gmail.com
Mon Nov 3 02:05:02 CET 2014


[00:08] <cone-460> ffmpeg.git 03Lukasz Marek 07master:e9899ca3ddc1: ffserver_config: dont pass pointer to enum as pointer to int
[00:45] <cone-460> ffmpeg.git 03Carl Eugen Hoyos 07release/1.2:a8fd50bb4eeb: Fix compilation after aa9d7058.
[00:51] <jamrial> michaelni: "ffmpeg -i input output.webp" still defaults to image2
[00:51] <jamrial> is that intended or you forgot to remove the webp extension from image2 muxer?
[01:01] <cone-460> ffmpeg.git 03Michael Niedermayer 07master:f5abd12c4d6c: avformat/img2enc: remove webp extension as we have a dedicated webp muxer
[01:25] <Compn> thanks jamrial , good spot :)
[01:58] <michaelni> jamrial, thx !
[02:08] <cone-460> ffmpeg.git 03Michael Niedermayer 07master:547fce95858e: avcodec/h264_slice: Clear table pointers to avoid stale pointers
[02:15] <jamrial> no prob :)
[02:42] <cone-460> ffmpeg.git 03Michael Niedermayer 07release/2.4:043f32606046: avcodec/h264_slice: Clear table pointers to avoid stale pointers
[03:08] <cone-460> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n2.4.3': unknown revision or path not in the working tree.
[03:08] <cone-460> Use '--' to separate paths from revisions
[03:08] <cone-460> refs/tags/n2.4.3:HEAD: avcodec/h264_slice: Clear table pointers to avoid stale pointers
[11:13] <JEEB> ugh
[11:13] <JEEB> that ticket
[11:14] <JEEB> someone should just say that x264opts passes options straight to libx264, and that libx264 doesn't do any ref calculation if level is set. only x264cli does :P
[11:16] <Compn> you want me to copy paste that to the ticket JEEB ? (honestly cant tell if serious or not)
[11:17] <JEEB> I tried to write something but I hadn't saved my trac password to my passphrase chain
[11:18] <JEEB> x264opts is meant for straight passing of options to libx264 without lavc's extra intervention, other options can have specific logic within lavc :P
[11:18] <Compn> JEEB : use login bug/bug
[11:18] <Compn> anonymous login :)
[11:18] <J_Darnley> Not anymore
[11:18] <Compn> no ?
[11:18] <Compn> or maybe that was mplayer trac login
[11:18] <J_Darnley> As in now everyone knows it.
[11:19] <J_Darnley> I guess it is still anonymous though.
[11:19] <JEEB> but doesn't work :)
[11:19] <Compn> sorry that was mplayer trac login, i got confused
[11:19] <Compn> too early, damn daylight savings time 
[11:24] <J_Darnley> JEEB: THere is an email about this "bug" on ffmpeg-user if you want to send an email there.
[11:25] <JEEB> I don't want to touch that. I get enough measles and lose SAN points from being on #ffmpeg
[11:26] <J_Darnley> fair enough
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:25a9823eb44a: avcodec/h264_slice: Clear table pointers to avoid stale pointers
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:42669252e8f5: postproc/postprocess: fix quant store for fq mode
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:23b3fe00bb0d: postproc: fix qp count
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:c36f5df34aca: avcodec/diracdec: Use 64bit in calculation of codeblock coordinates
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:341cf9ed8167: avcodec/diracdec: Tighter checks on CODEBLOCKS_X/Y
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:35cf24eee996: avcodec/dirac_arith: fix integer overflow
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:d59804bafbaf: avcodec/dxa: check dimensions
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:6e45a99a7cc4: avcodec/dnxhddec: treat pix_fmt like width/height
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:b92ccfefc3a3: avcodec/utils: Align dimensions by at least their chroma sub-sampling factors.
[11:33] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:96dac432f790: avcodec/svq1dec: zero terminate embedded message before printing
[11:36] <cone-921> ffmpeg.git 03Michael Niedermayer 07release/1.2:7e8b4506e0aa: update for 1.2.10
[11:53] <cone-921> ffmpeg.git 03Michael Niedermayer 07master:70b7cf9c39d0: avformat/webpenc: removed unused variable
[13:54] <Daemon404> nodejs fate eh
[13:56] <JEEB> u wot mate
[13:56] <Daemon404> https://github.com/TimothyGu/fateserver-node
[13:57] <Daemon404> it's certainly more readable
[13:57] <Daemon404> i have wondered for a while now if programmers 20 years from now will look back at js like js people look back at 90s cgi perl
[19:29] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:a52cb42ba662: avformat/matroskadec: use gmtime_r() for thread saftey
[19:29] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:013c3eb05cbb: avfilter/vf_drawtext: use gm_time_r() for thread saftey
[20:18] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:bab09864b423: avcodec/jacosubdec: use time_internal.h, simplify code
[20:18] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:32a2876b12f6: avformat/segment: use time_internal.h, simplify code
[20:18] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:76886589eea7: avformat/wavenc: Use localtime_r() for thread saftey
[20:18] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:5ece4f8b7362: avformat/sbgdec: Use localtime_r() for thread saftey
[20:18] <cone-344> ffmpeg.git 03Michael Niedermayer 07master:63e62cfbe23d: avformat/img2enc: Use localtime_r() for thread saftey
[21:40] <ubitux> from frame.h: "Presentation timestamp in time_base units (time when frame should be shown to user)." on top of pts doxy
[21:40] <ubitux> what time_base is it refering to?
[21:40] <ubitux> AVStream.time_base?
[21:41] <ubitux> if so, we have AVPacket.pts == AVFrame.pts, never a rescaling?
[21:41] <ubitux> (i'm trying to follow the same model for subtitles)
[21:42] <wm4> yes AVStream
[21:42] <wm4> and yes
[21:42] <wm4> the decoder doesn't even know the time_base
[21:42] <wm4> or at least that's how it _should_ work
[21:43] <ubitux> alright, so there is no automatic rescaling system in place in the decoding part?
[21:43] <ubitux> i was wondering if some users would need that
[21:43] <ubitux> (requesting all the frame for every media to be rescaled homogeneously)
[21:44] <ubitux> frame ts*
[21:44] <wm4> the subtitle converter API actually needs the time_base (breaking all expectations), so I set a fake time_base
[21:44] <ubitux> the AVSubtitles are in ms tb or something currently
[21:45] <ubitux> and yeah it sucks, i removed all the rescaling in the decoders, it's insane :p
[21:45] <ubitux> anyway, ok, thanks
[21:57] <ubitux> i wonder if that time_base is not relative to the AVFormatContext.time_base
[21:57] <ubitux> oh well..
[22:05] <wm4> isn't time_base per AVStream?
[22:07] <ubitux> i meant AVCodecContext sorry
[22:07] <ubitux> (the sub decoders rescale using AVCodecContext->time_base)
[00:00] --- Mon Nov  3 2014


More information about the Ffmpeg-devel-irc mailing list