[Ffmpeg-devel-irc] ffmpeg-devel.log.20190114

[burek]

[12:15:20 CET] <durandal_1707> atomnuker: will you please answer?
[13:40:45 CET] <cone-383> ffmpeg 03Jerome Borsboom 07master:fc6e53b0b662: avcodec/vc1: fix B predictor validity for 4-MV MBs
[15:04:21 CET] <kurosu> pross: btw, does you new code kills some other one? eg dsp functions?
[18:33:10 CET] <cone-194> ffmpeg 03Rafaël Carré 07master:75ea329b78f4: api-h264-slice-test: fix arguments and help
[18:33:10 CET] <cone-194> ffmpeg 03Michael Niedermayer 07master:6e23736aefa8: avcodec/gdv: Optimize and factorize scaling loops
[18:33:10 CET] <cone-194> ffmpeg 03Michael Niedermayer 07master:c15972f0af76: avcodec/tests/rangecoder: initialize array to avoid valgrind warning
[18:34:11 CET] <durandal_1707> michaelni: who is sponsoring your commits?
[18:37:35 CET] <kierank> durandal_1707: the CIA
[18:37:55 CET] <durandal_1707> more KGB
[21:10:28 CET] <pross> kurosu: yeah, VP4 and VP6 filters are identical so they ought to share code. (VP5 is identical too, but I won't touch that in this round.)
[21:34:56 CET] <durandal_1707> atomnuker: should i use cosine distance instead of squared euclidean in anlmdn?
[21:40:14 CET] <JEEB> wee http://up-cat.net/p/586a1fe8
[21:40:21 CET] <JEEB> now I just need to make the decoder wrapper
[21:48:05 CET] <pross> åí?
[21:48:11 CET] <JEEB> yup
[21:48:16 CET] <JEEB> MBS broadcasting sample, this one
[21:49:04 CET] <JEEB> I hate myself for not enabling the libclang based plugin in my editor though, since modifying the avcodec codec_desc etc made me re-compile all the things
[21:49:07 CET] <JEEB> and then I got lazy
[21:49:34 CET] <JEEB> and suddenly > I forget to export the profile array > I forgot the = between the array[] and { > I forget the closing ; after }
[21:49:36 CET] <durandal_1707> old news
[21:49:42 CET] <JEEB> which are all dumb as hell mistakes
[21:50:35 CET] <JEEB> well at least I care about this stuff on some level :P
[21:50:56 CET] <JEEB> I'm going to have to start reviewing stuff in other projects soon, but for now I'm trying to actually *do something*
[21:51:04 CET] <pross> no clang plugin for ed
[21:51:22 CET] <durandal_1707> ed?
[21:51:38 CET] <JEEB> yea it's fun when your editor screams at you instead of you waiting for full compilation and finding out you did something really dumb
[22:12:53 CET] <atomnuker1> durandal_1707: if it works better, though for this use case I think SSE is alright
[22:13:20 CET] <BradleyS> every time i see anlmdn i read "animal dn"
[22:30:58 CET] <cone-194> ffmpeg 03Carl Eugen Hoyos 07master:5d958f0955fe: lavf/mov: Do not fail hard for more invalid atoms.
[22:34:51 CET] <atomnuker1> I read **** microsoft developer network
[22:38:50 CET] <BradleyS> :P
[23:08:35 CET] Action: Luigi12 was wondering if you could cut another 3.2.x release.  MythTV is bundling an older release from that branch.  The 3.2.x branch has fixes for CVE-2018-15822 and other security issues committed since 3.2.12 was released.
[23:10:17 CET] <Luigi12> although one of my fellow distro developers just told me that CVE commit is in the tarball even though that's not what git shows...
[23:12:21 CET] <Luigi12> ahh no he's wrong
[23:13:13 CET] <Luigi12> I see michaelni added a bunch of fixes in git on 20181118 but didn't cut a new release
[23:17:25 CET] <durandal_1707> Luigi12: ask michaelni directly, he is paid for that
[23:17:38 CET] <Luigi12> thanks
[23:17:52 CET] <Luigi12> michaelni: just making sure you saw my request for a new 3.2.x release.  Thanks!
[23:19:12 CET] <kierank> durandal_1707: how much is he paid
[23:20:59 CET] <durandal_1707> kierank: that info is top secret, sorry
[23:21:41 CET] <kierank> durandal_1707: very important for the cia to "fix" fuzzed files that take time to decode
[23:21:42 CET] <kierank> I understand
[23:21:44 CET] <kierank> top secret
[00:00:00 CET] --- Tue Jan 15 2019