[Ffmpeg-devel-irc] ffmpeg-devel.log.20150210

[burek]

[00:06] <wm4> michaelni: I don't get that failure anyway, don't know why
[00:07] <wm4> though I guess my patch breaks the assumptions the patch checks
[00:43] <cone-22> ffmpeg.git 03James Almer 07master:06fe6dfe12d7: x86/hevc_sao: make sao_band_filter work on x86_32
[03:12] <pross> \
[03:34] <cone-22> ffmpeg.git 03Andreas Cadhalpun 07master:078be09dd713: examples/demuxing_decoding: abort decoding when width, height or pix_fmt change
[10:39] <cone-815> ffmpeg.git 03Paul B Mahol 07master:a5398aa56cfb: avcodec/hevc_parser: check for av_mallocz() failure
[10:39] <cone-815> ffmpeg.git 03Paul B Mahol 07master:fd6768a4b8e2: avcodec/huffman: check if map was allocated too
[10:39] <cone-815> ffmpeg.git 03Paul B Mahol 07master:e4eba9e21d71: avcodec/mss4: use av_malloc_array()
[11:30] <cone-815> ffmpeg.git 03Mark Reid 07master:9b22e460adb2: test/fate: added mxf_opatom regression tests
[12:06] <wm4> av_assert0(((size_t)-1) > 0); // C guarantees this but if false on a platform we care about revert at least b284e1ffe343d6697fb950d1ee517bafda8a9844
[12:07] <wm4> that commit does
[12:07] <wm4> -    if (size <= 0 || nmemb >= INT_MAX / size)
[12:07] <wm4> +    if (!size || nmemb >= INT_MAX / size)
[12:07] <wm4> with size being size_t
[12:07] <wm4> wtf.
[12:29] <nevcairiel> size_t is unsigned, so a < 0 check is pointless
[12:38] <wm4> nevcairiel: yeah, but a non-conformant C compiler may treat it as signed!!!!111111
[12:39] <durandal_1707> it is just an assert
[14:02] <ubitux> durandal_1707: did you plan to do a apng muxer?
[14:03] <ubitux> ah my bad, Benoit was the was doing the demuxer, i got mistaken
[14:14] <wm4> why would you write an apng muxer
[14:16] <wm4> isn't it already a dead format?
[14:16] <kierank> wm4: ffeatures
[14:20] <ubitux> wm4: because it's better than gif?
[14:20] <ubitux> and supported by browser
[14:20] <ubitux> (+s?)
[14:21] <wm4> there's a browser yet other than firefox which supports apng?
[14:21] <ubitux> no idea
[14:21] <ubitux> well if ffmpeg supports it, maybe chromium will ;)
[14:22] <wm4> I very much doubt they use ffmpeg to load png images
[14:22] <wm4> not even the firefox debian package supports apng, because it's built with official libpng
[14:22] <ubitux> they could :)
[14:22] <wm4> no
[14:22] <ubitux> why not?
[14:22] <wm4> images need incremental loading
[14:22] <ubitux> ok
[14:24] <durandal_1707> apng muxer should be trivial, but encoder is different story...
[15:22] <rcombs> wm4: Safari supports it
[15:22] <cone-815> ffmpeg.git 03Paul B Mahol 07master:392d975efe94: avcodec: fix clobbered ff_get_buffer()
[15:44] <cone-815> ffmpeg.git 03wm4 07master:3e18dc235e9c: avutil: move internal function out of public header
[15:57] <cone-815> ffmpeg.git 03wm4 07master:92e3a6fdac73: avutil: check pixdescs in a different place
[16:13] <cone-815> ffmpeg.git 03wm4 07master:c247fc63e4dc: pixdesc: clarify AV_PIX_FMT_FLAG_ALPHA doxygen
[17:01] <cone-815> ffmpeg.git 03Kevin Wheatley 07master:7fc33dca395c: libavformat: DNxHD in .mov, switch unspecified color_range to mpeg
[17:34] <cone-815> ffmpeg.git 03Supraja Meedinti 07master:8d5a85011952: libavutil: camellia: remove unwanted memory loads
[20:43] <cone-716> ffmpeg.git 03Paul B Mahol 07master:8b77c4dd424d: avfilter: Add repeatfields filter (Port of mp=softpulldown)
[21:56] <gerion> t
[21:56] <gerion> sorry, mistyped
[22:42] <llogan> michaelni: what unit is -frame_drop_threshold? a number of frames? time?
[22:44] <michaelni> i think it was in relation to the frame rate  but can be fractional values like 0.123
[22:53] <wm4> llogan: heh, I was wondering this too
[22:53] <cone-716> ffmpeg.git 03Clément BSsch 07master:d47eeff2741a: tests: drop bc dependency
[23:14] <cone-716> ffmpeg.git 03Michael Niedermayer 07master:133d81535d39: avutil/error: list most common error code in error_entries when strerror_r() is unavailable
[23:53] <cone-716> ffmpeg.git 03Marton Balint 07master:68850090b6d2: ffplay: update frame timer based on last updated clock time when toggling pause
[23:53] <cone-716> ffmpeg.git 03Marton Balint 07master:d6910c4b5ea9: ffplay: make eof part of videostate and signal it when opening a stream
[23:53] <cone-716> ffmpeg.git 03Marton Balint 07master:ef1ccea94551: ffplay: factorize thread starting and stopping code into decoder
[23:53] <cone-716> ffmpeg.git 03Michael Niedermayer 07master:838c755f62a1: Merge remote-tracking branch 'cus/stable'
[23:56] <Timothy_Gu> how do you produce a sweep audio signal with aevalsrc again?
[23:56] <Timothy_Gu> I did it once then I forgot
[23:57] <Timothy_Gu> oh ok i got it
[23:57] <Timothy_Gu> used n instead of t
[00:00] --- Wed Feb 11 2015