[Ffmpeg-devel-irc] ffmpeg-devel.log.20170425

[burek]

[00:51:45 CEST] <cone-399> ffmpeg 03Paul B Mahol 07release/3.3:da693f8daa62: avcodec/dnxhd_parser: fix parsing interlaced video, simplify code
[01:14:00 CEST] <cone-399> ffmpeg 03Michael Niedermayer 07master:d976d2ec7874: tools/target_dec_fuzzer: Fix build with default FFmpeg build flags
[01:14:01 CEST] <cone-399> ffmpeg 03Michael Niedermayer 07master:5b499bf4a05a: Make tools/target_dec_*_fuzzer buildable with configure and make
[01:14:02 CEST] <cone-399> ffmpeg 03Michael Niedermayer 07master:177608aa7498: fate: Add test for pkt_size of ffprobe
[01:31:58 CEST] <Compn> [15:15] <durandal_1707> at least carl is against
[01:32:07 CEST] <Compn> durandal_1707 : btw, carl is not against merging or working with libav
[01:32:19 CEST] <Compn> i've asked this myself to him directly, face to face
[01:32:26 CEST] <Compn> durandal_1707 : so please stop spreading misinformation
[01:33:54 CEST] <Compn> yes, carl can call people thieves and liars and still work with them. it seems to be a position at odds with itself. 
[01:54:35 CEST] <BBB> Compn: thats not entirely fair, Ive discussed it with him also and what he basically wants is total submission (they are free to contribute to the ffmpeg project as-is)
[01:54:51 CEST] <BBB> Compn: that is not a fair-and-balanced project merge
[01:54:57 CEST] <BBB> Compn: did his opinion change?
[01:59:13 CEST] <Compn> BBB : what he wants was not the question that durandal_1707 was answering, iirc
[01:59:33 CEST] <Compn> i could be misinterpreting the conversation, my internet died half way through
[01:59:48 CEST] <Compn> and if so, i apologize
[01:59:56 CEST] <BBB> its ok, weve got logs ;)
[02:00:05 CEST] <Compn> j-b : how do you handle / judge coc violations btw
[02:00:13 CEST] <BBB> I think j-b said he doesnt
[02:00:17 CEST] <Compn> j-b : i mean... tribunal? trac? process ?
[02:00:21 CEST] <Compn> i mean vlc in general
[02:00:21 CEST] <BBB> (because he doesnt have mailman/irc admin powers)
[02:00:25 CEST] <BBB> oh I see
[02:00:34 CEST] <Compn> not him specifically, but thanks for clarifying :D
[02:02:45 CEST] <Compn> also > waiting 12 hours for irc log
[02:04:17 CEST] <Compn> anyways... ancient history needs to be buried :P
[02:04:19 CEST] <Compn> but whatever
[02:05:36 CEST] <BBB> +1
[02:06:11 CEST] <BBB> I dont think we need a vote to approve of coc rules, the coc are the rules
[02:08:02 CEST] <BBB> but enforcement ws missing, I think some people didnt like those parts (I dont remember)
[02:08:08 CEST] <BBB> and yes that makes enforcement difficult
[02:08:14 CEST] <BBB> I think we need a bofh for enforcement
[02:15:19 CEST] <jamrial> enforcement is hard because voting is a hassle and most people that can vote are not even active enough to know what's being voted
[02:15:58 CEST] <jamrial> or accepted being in the voting comittee because they wanted to vote for technical stuff, not social or netiquette
[02:16:55 CEST] <jamrial> so a "VOTE: ban $name" is probaby not very appealing
[02:17:30 CEST] <Compn> right
[02:17:43 CEST] <Compn> i'd rather moderate ml and just discard any non development mails
[02:17:50 CEST] <Compn> because i'd rather not ban anyone
[02:18:06 CEST] <Compn> but thats my opinion
[02:18:12 CEST] <Compn> what do you think, kierank ?
[02:19:02 CEST] <Compn> i'd also propose a new rule "no opinions about other devs" haha
[02:19:57 CEST] <atomnuker> Compn: do you really think you're being useful at all?
[02:20:04 CEST] <atomnuker> because you're not really
[02:20:37 CEST] <Compn> advancing coc is not helping ?
[02:20:42 CEST] <atomnuker> no, it is not
[02:21:00 CEST] <Compn> why is it not ? if you can explain ?
[02:21:14 CEST] <atomnuker> I gave my explanation on why it sucks a year ago
[02:21:26 CEST] <Compn> the coc itself? yes i agree i dont like the coc
[02:21:27 CEST] <atomnuker> just let people argue, shutting off access to mls won't
[02:21:35 CEST] <atomnuker> or ircs
[02:21:41 CEST] <Compn> but arguing on the list is driving some developers away
[02:21:44 CEST] <atomnuker> if people don't argue then nothing gets solved
[02:24:03 CEST] <atomnuker> no it doesn't, if you don't let people argue on the list they'll just argue privately
[02:25:01 CEST] <atomnuker> I'd agree that's a good idea though
[02:27:02 CEST] <atomnuker> though its a better idea to say "please sort this off privately" rather than "BANHAMMER FOR YOU HAHAHAHA YOU VIOLATED THE COC SECTION 5"
[02:29:33 CEST] <Compn> atomnuker : yes, absolutely arguing is driving off some developers
[02:29:42 CEST] <Compn> i dont mean technical arguing, but insults and the like
[02:30:15 CEST] <Compn> and by "developers" i have specific name
[02:30:37 CEST] <Compn> but since this is private email i'm not going to post ... although i want to
[02:31:22 CEST] <Compn> atomnuker : what is your email, i will forward it
[02:33:21 CEST] <wm4> can I get it too
[02:34:25 CEST] <Compn> wm4 : it was sent to you 20 mar 2017 "IRC" subject
[02:34:31 CEST] <Compn> i was the only one to reply :(
[02:34:49 CEST] <Compn> to everyone at least...
[02:35:13 CEST] <wm4> oh that
[02:35:27 CEST] <wm4> I made a pretty salty reply to the sender
[02:35:54 CEST] <wm4> it was self-pitying bullshit
[02:36:18 CEST] <wm4> not even surw why I got thet mail too
[02:36:25 CEST] <wm4> *that
[02:36:39 CEST] <wm4> atomnuker: an enforced CoC might actually be useful
[02:36:51 CEST] <atomnuker> I guess, having read that
[02:36:54 CEST] <jamrial> did someone left and emailed a few of you privately about it?
[02:36:59 CEST] <wm4> but in ffmpeg it's only randomly used as "weapon" to shut others up (i.e. kindergarten bullshit)
[02:37:12 CEST] <wm4> not left
[02:37:26 CEST] <atomnuker> I gess Compn might be right in saying moderating the ML could help
[02:37:53 CEST] <alevinsn> I don't think its necessary
[02:37:54 CEST] <Compn> moderation of devs on ml. not the entire ml, i hope :P
[02:37:56 CEST] <BBB> that means any email would need to be in moderation, or moderation would be a 24/7 job to instantly jump on a flare
[02:38:00 CEST] <BBB> both are not a good idea
[02:38:10 CEST] <alevinsn> its time consuming
[02:38:16 CEST] <BBB> so in theory its a nice idea but in practice its somewhat cumbersome
[02:38:26 CEST] <Compn> just temporarily , hopefully 
[02:38:34 CEST] <Compn> not indefinitely :P
[02:38:39 CEST] <alevinsn> and I think the issues raised between specific individuals can be resolved
[02:38:44 CEST] <wm4> in general it's indicative of the lack of leadership, direction, and decision power in the ffmpeg project
[02:38:48 CEST] <alevinsn> without needing to take these more drastic measures
[02:38:51 CEST] <Compn> alevinsn : hahahaha, welcome to the internet my friend!
[02:38:56 CEST] <BBB> j-b is the only one who has practically succeeded in getting a flamey community to calm down
[02:39:04 CEST] <wm4> (seems like nobody wants to take i up...)
[02:39:05 CEST] <BBB> maybe we should take his input more seriously :)
[02:39:06 CEST] <wm4> *it
[02:39:28 CEST] <alevinsn> I responded to compn's e-mail
[02:39:45 CEST] <alevinsn> I'd suggest trying something like that first before doing something more drastic
[02:39:55 CEST] <alevinsn> I think it is limited to perhaps two people :-)
[02:40:11 CEST] <Compn> alevinsn : my email is for the rules for everyone
[02:40:15 CEST] <Compn> not just one or two devs this week 
[02:40:18 CEST] <alevinsn> I'm aware
[02:40:20 CEST] <Compn> k
[02:40:24 CEST] Action: Compn hasnt read email
[02:40:36 CEST] <alevinsn> all I'm saying is, I don't think anything good will come out of trying to do what you requested comments for
[02:42:23 CEST] <Compn> i'm more surprised they dont speak german at each other
[02:42:36 CEST] <Compn> wm4 : you are a native german speaker, right ?
[02:42:42 CEST] <wm4> yes
[02:42:57 CEST] <Compn> do they speak german in austria ?
[02:43:02 CEST] Action: Compn is ignorant of such things
[02:43:12 CEST] <wm4> actually I once wrote an email to cehoyos in English, but he absolutely insisted to continue the conversation in German
[02:43:18 CEST] <wm4> it was awkward to say the least
[02:44:04 CEST] <Compn> an attempt to remove the language barrier per hapse
[02:44:19 CEST] <Compn> although i believe both of you speak english well
[02:44:51 CEST] <Compn> i see at least one problem is that neither one of you believe the other one :D
[02:45:20 CEST] <alevinsn> I remember a couple weeks ago, jamrial responded to something I had said on IRC, and I sort of snapped back (although, it wasn't much of one, but I remember being annoyed), and I realized later that I was just tired
[02:45:46 CEST] <alevinsn> sometimes a good night's rest (although not everyone can get one) does wonders for me
[02:45:55 CEST] <alevinsn> but, I just spent 3.5 hours at a spa
[02:45:56 CEST] <alevinsn> :-)
[02:46:07 CEST] <Compn> alevinsn : well the coc was created because the unwritten rule about being nice was being violated. so the project voted to institute the coc
[02:46:23 CEST] <Compn> my email is just wondering how to enforce the coc, when violations occur, as some developers have asked.
[02:46:38 CEST] <Compn> i think either wm4 or kierank asked how to enforce the coc. i could be wrong
[02:47:25 CEST] <wm4> it simply isn't
[02:47:50 CEST] <Compn> which is why i asked the project how to enforce , judge, etc the coc
[02:48:04 CEST] <wm4> anyway, I won't allow cehoyos' defamatory bullshit against Libav (or me) anymore
[02:48:17 CEST] <wm4> if you have a problem with it, then this proiblem will remain unsolved
[02:48:48 CEST] <Compn> my point is both of you should discuss this off list
[02:48:58 CEST] <wm4> no, he is saying this publicly
[02:49:04 CEST] <wm4> on the ffmpeg-devel list
[02:49:11 CEST] <Compn> and he shouldnt
[02:49:15 CEST] <wm4> if we let it stand, outsiders will think we think this is ok
[02:49:20 CEST] <wm4> it's not ok
[02:49:25 CEST] <Compn> which is why i told him to take it off list
[02:49:36 CEST] <Compn> or asked him, depending on what i said 
[02:49:47 CEST] <wm4> no, youz told us to take our "fight" off list
[02:49:57 CEST] <Compn> i'm talking about a couple days ago as well
[02:50:01 CEST] <Compn> or maybe week or two
[02:50:10 CEST] <wm4> there's no penalization against his anti-Libav behavior
[02:50:29 CEST] <alevinsn> a bunch of people reprimanded him for that
[02:50:50 CEST] <wm4> that's how it should be
[02:50:52 CEST] <alevinsn> I don't think what he said with regard to libav this time around was defamatory
[02:51:14 CEST] <Compn> it should be -devel is for development only
[02:51:16 CEST] <Compn> :D
[02:51:24 CEST] <wm4> yeah thisn time it was onbly the usual passive aggressive bullshit
[02:51:35 CEST] <wm4> and we shouldn't allow that either
[02:51:50 CEST] <Compn> your passive agressive replies are tiring as well
[02:51:55 CEST] <Compn> on the list that is
[02:51:59 CEST] <Compn> off the list, feel free...
[02:52:01 CEST] <wm4> what's passive about them
[02:52:07 CEST] <alevinsn> they aren't passive
[02:52:11 CEST] <alevinsn> exactly
[02:52:23 CEST] <alevinsn> but saying "are you joking?"
[02:52:30 CEST] <alevinsn> is sort of code for "are you insane?" :-)
[02:53:02 CEST] <Compn> also i would say that one person on a mailing list does not tain the whole project
[02:53:07 CEST] <Compn> aisde from maybe reiserfs
[02:53:08 CEST] <Compn> :P
[02:53:14 CEST] <wm4> ok with this I'm trolling, just a little bit (I did it because cehoyos asked "are you joking" in this way)
[02:53:37 CEST] <Compn> but i dont view "linux" as being linus torvolds only...
[02:53:51 CEST] <Compn> but maybe i am grasping at straws
[02:54:05 CEST] <wm4> anyway, feel free to deal with him in a "proper" way and you'll see no posts from me
[02:54:43 CEST] <Compn> i would like to encourage you to not engage , and instead ask people to take their opinions off the development mailing list
[02:54:56 CEST] <Compn> as it is a quick way to tell people to fuck off without saying it directly :P
[02:55:13 CEST] <wm4> again, if he posts some bullshit about Libav, and nobody else corrects him, I'll do it
[02:55:18 CEST] <wm4> because I won't let it stand
[02:55:40 CEST] <wm4> ah you mean actually telling them on-list
[02:55:50 CEST] <alevinsn> it is possible to do that without being confrontational I guess
[02:56:06 CEST] <wm4> I guess I should do that if there's a next time
[02:56:15 CEST] <Compn> thanks wm4
[02:56:38 CEST] <Compn> but also with everyone else
[02:56:46 CEST] <Compn> not singling out anyone in particular
[02:56:50 CEST] <Compn> i get weird too
[02:56:56 CEST] <wm4> Compn: your reply wasn't very proper anyway
[02:57:05 CEST] <wm4> and what do you mean this about "shit"
[02:57:33 CEST] <Compn> what did you mean when you said ffmpeg needs to get rid of shit
[02:57:34 CEST] <Compn> define shit
[02:57:51 CEST] <Compn> we cant read mins
[02:57:54 CEST] <Compn> minds*
[02:57:57 CEST] <wm4> read those sentences again
[02:58:28 CEST] <Compn> i'm not going to guess the definition
[02:58:40 CEST] <wm4> the way you put it it sounds like it's emphasizing some unfavorable interpretation of that last part
[02:58:44 CEST] <Compn> thats where the problems occur, with the asumptions
[02:59:03 CEST] <Compn> omg my typing is getting worse 
[02:59:48 CEST] <wm4> anyway I should go to sleep
[03:00:13 CEST] <alevinsn> good night :-)
[03:02:56 CEST] <Compn> alevinsn : do you see pm ?
[03:03:59 CEST] <iive> Compn: when i read the word "tribunal" my first association is with this https://en.wikipedia.org/wiki/Revolutionary_Tribunal
[03:04:24 CEST] <iive> I think it might be good idea to don't go that path. ;)
[03:05:40 CEST] <Compn> iive : well heres what i read , when i read "tribunal" https://en.wikipedia.org/wiki/International_Human_Rights_Tribunal
[03:06:29 CEST] <Compn> wait wrong one
[03:07:04 CEST] <Compn> plenty of other tribunals in wikipedia :P
[03:09:01 CEST] Action: iive shakes head
[03:14:38 CEST] <cone-399> ffmpeg 03Michael Niedermayer 07master:550a9c547ec4: tools/target_dec_fuzzer: Remove FuzzerInterface.h dependancy
[09:27:08 CEST] <durandal_170> so are those regressions going to be fixed or what?
[13:36:29 CEST] <BtbN> How do I properly review a Patch E-Mail with the patch as an attachment?
[13:36:42 CEST] <BtbN> Do I just take the patch, prepend > in front of every line, and paste it into an E-Mail?
[13:37:45 CEST] <nevcairiel> with many email clients you can treat text-based attachments like inline things, may need to override the mimetype somewhere
[13:38:58 CEST] <BtbN> Pasting the patch makes a mess out of the linebreaks
[14:36:02 CEST] <cone-654> ffmpeg 03Martin Vignali 07master:59d219b91ef6: fate/exr : fix pix_fmt
[15:05:48 CEST] <durandal_170> i will write yet another scope video filter for fun
[15:14:06 CEST] <wm4> durandal_170: one filter request I once got is a filter for low-pass filtering the LFE (and only the LFE)
[15:33:49 CEST] <durandal_170> hmm,  i might extend biquads filters to operate only on certain channels
[15:34:18 CEST] <durandal_170> why this havent come to me earlier ...
[16:12:40 CEST] <nevcairiel> wm4: durandal_1707: a per channel lowpass would also be something I might use
[18:21:48 CEST] <cone-979> ffmpeg 03Paul B Mahol 07master:9d08c7bd424c: avfilter/af_biquads: allow filtering only selected channels
[18:24:39 CEST] <durandal_1707> ^ nevcairiel, wm4 : -af lowpass=c=LFE
[18:32:39 CEST] <jamrial> nevcairiel: bad practice in a987c686ba from you lavfilters ffmpeg fork :p
[18:36:00 CEST] <nevcairiel> jamrial: bad practice how so?
[18:36:06 CEST] <jamrial> use av_content_light_metadata_alloc() then add_side_data with the returned size :P
[18:36:30 CEST] <nevcairiel> I give no single f' about ABI
[18:36:38 CEST] <jamrial> heh
[18:38:39 CEST] <nevcairiel> i  copied the same pattern as the mastering thing above, which was originally copied from the ffmatroska demuxer
[18:38:43 CEST] <jamrial> nevcairiel: also, if SEI messages in extradata are desirable (judging by another commit there), should i revert the patch that moved extradata parsing out of hevcdec?
[18:39:06 CEST] <nevcairiel> not sure if its really desirable, but its possible
[18:40:15 CEST] <nevcairiel> the proper solution would be to refactor sei decoding to no longer need a full context
[18:40:19 CEST] <nevcairiel> but thats effort
[18:40:24 CEST] <jamrial> a lot of it
[18:40:47 CEST] <jkqxz> That scaling list bug is mildly amusing.  Seems to be a commit from three years ago which changed the indexing of the 32x32 scaling lists, but missed one case in the reading.
[18:42:56 CEST] <jkqxz> I wonder whether someone actually ran into the specific case which fails (a nondefault list for 32x32 intra and then explicitly copy it for inter, which is the default thing to do anyway), or if that was just from inspecting the code.
[18:44:03 CEST] <jkqxz> (That reordering is also why libav and ffmpeg disagree about the scaling list ordering, which caused confusion (to me at least) with hwaccels.)
[19:03:55 CEST] <cone-979> ffmpeg 03Anton Khirnov 07master:fa1749dd34c5: vp9: split superframes in the filtering stage before actual decoding
[19:03:56 CEST] <cone-979> ffmpeg 03Clément BSsch 07master:d417e95af76c: Merge commit 'fa1749dd34c55fb997c97dfc4da9383c9976ab91'
[19:05:01 CEST] <ubitux> atomnuker: here we go, the next commit is a pain because of the bitstream
[19:05:03 CEST] <ubitux> :(
[19:05:11 CEST] <ubitux> conflicts all over
[19:05:22 CEST] <ubitux> not hard to fix, but a pita anyway
[19:19:38 CEST] <cone-979> ffmpeg 03Anton Khirnov 07master:730c02326094: binkaudio: switch to the new send/receive API
[19:19:39 CEST] <cone-979> ffmpeg 03Clément BSsch 07master:91f8ccdda060: Merge commit '730c02326094bcfb1fa67f10a7e7b22f03f5a88f'
[19:32:52 CEST] <ubitux> what do people prefer: 786032cad8ecabe577d9cff0356da6e9e9488a2d or 86157e6db2c7a9222f77fa7e7f50fb9aebc3aa81 ?
[19:38:14 CEST] <durandal_1707> links?
[19:39:04 CEST] <tmm1> i'm trying to figure out why a53cc isn't working with qsv (https://trac.ffmpeg.org/ticket/5969).. anyone have any hunches as to what might be causing the weird behavior?
[19:39:20 CEST] <jamrial> ubitux: poke BBB, this seems to touch threading
[19:40:05 CEST] <BBB> ?
[19:40:33 CEST] <BBB> I need links or pastebin patches
[19:43:21 CEST] <ubitux> we do http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=786032cad8ecabe577d9cff0356da6e9e9488a2d
[19:43:24 CEST] <ubitux> libav does https://git.libav.org/?p=libav.git;a=commitdiff;h=86157e6db2c7a9222f77fa7e7f50fb9aebc3aa81
[19:43:40 CEST] <ubitux> reducing diff is better, but i'm asking for opinions on this
[19:43:48 CEST] <ubitux> maybe our code is better, maybe not
[19:51:07 CEST] <BBB> poke elenril, poke michaelni, tell them to figure it out
[19:51:38 CEST] <BBB> Im not willing to ivnest significant time in hevc, sorry
[20:54:02 CEST] <wm4> eh, I don't like either of them
[20:54:54 CEST] <wm4> ubitux: I'd go with Libav's... just in the interest of reducing the diff
[00:00:00 CEST] --- Wed Apr 26 2017