[Ffmpeg-devel-irc] ffmpeg-devel.log.20190118

[burek]

[01:50:09 CET] <rcombs> looking for additional feedback on my "enable TLS peer verification by default" patch
[01:52:40 CET] <rcombs> actually come to think of it, I should probably only make that the default for client mode
[01:52:45 CET] <rcombs> (what's server mode even for)
[12:34:41 CET] <cone-096> ffmpeg 03Carl Eugen Hoyos 07master:399c8e860f48: lavu/frame: Fix typo.
[14:59:28 CET] <j-b> 'morning
[15:38:54 CET] <kierank> https://github.com/google/oss-fuzz/issues/2095
[16:25:36 CET] <Compn> j-b : how many codecs are known to exist?
[16:26:11 CET] <Compn> i wonder if microsoft keeps its riff list up to date (just not publically)
[16:49:48 CET] <kurosu__> (26+10)^4?
[16:52:03 CET] <Compn> way to spoil the fun :P
[16:56:17 CET] <funman> is CARL codec registered yet ?
[16:59:05 CET] <durandal_1707> no, but CEH is
[16:59:54 CET] <kurosu__> Compn, would an appropriate sed/awk/... line on libavformat/riff.c be funnier ?
[17:00:50 CET] <kurosu__> oh, it could be case sensitive, so (2*26+10)^4 + (2*26+10)^2, counting audio codecs ?
[17:14:21 CET] <Compn> kurosu : no, exist, not in ffmpeg
[17:14:50 CET] <j-b> Compn: 10000
[17:14:58 CET] <Compn> in theory i want full exist - ffmpeg = codecs to find and document :P
[17:15:26 CET] <durandal_1707>   /KB Compn 
[17:18:31 CET] <durandal_1707> that .drv is using VFW or?
[17:27:08 CET] <durandal_1707> Compn: have more ARBC files?
[20:13:28 CET] <BBB> michaelni: I think the generic idea of a "decoding deadline" (in seconds or %) or "corruption threshold" (in % or pixels) is useful,but I think putting it in individual decoders may not be a great starting point
[20:13:41 CET] <BBB> michaelni: how about making it more generic?
[20:15:11 CET] <BBB> make it a real interface, and make it configurable
[20:15:35 CET] <BBB> I think it's totally worth it, and the reason people dislike it is not b/c it's a bad idea, but because your current approach is arbitrary
[20:15:47 CET] <BBB> why 5%? why start at that decoder?
[21:03:13 CET] <atomnuker> I think the deadline for decoders should be based on the number of samples/pixels
[21:04:21 CET] <atomnuker> 5% is quite a low amount imo considering a bitstream with an unoptimized coding tool could set it off probably
[21:19:00 CET] <michaelni> BBB, I can make it configurable. Happy to do so if that makes more people happy.
[21:19:48 CET] <BBB> don't make me happy; do it because you agree it's right[TM]
[21:20:17 CET] <BBB> my happiness is not affected by this much
[21:21:38 CET] <michaelni> i agree its slightly better to have it configurable, in reality i doubt that parameter will be used much
[00:00:00 CET] --- Sat Jan 19 2019