Changes between Initial Version and Version 1 of FFmeeting/2016-05

May 30, 2016, 3:11:37 AM (12 months ago)



  • FFmeeting/2016-05

    v1 v1  
     1This FFmeeting was hosted on ​irc:// on 2015-05-28, at 17 UTC.  
     3== Full meeting log == 
     6**** BEGIN LOGGING AT Sat May 28 18:51:04 2016 
     8May 28 18:51:04 *       Now talking on #ffmpeg-meeting2016 
     9May 28 18:51:21 *       durandal_1707 ( has joined #ffmpeg-meeting2016 
     10May 28 18:51:34 *       jamrial (~jamrial@ has joined #ffmpeg-meeting2016 
     11May 28 18:51:49 *       durandal_170 gives channel operator status to durandal_1707 
     12May 28 18:52:22 *       BBB (~rbultje@ has joined #ffmpeg-meeting2016 
     13May 28 18:52:48 *       mateo` ( has joined #ffmpeg-meeting2016 
     14May 28 18:53:35 *       cehoyos ( has joined #ffmpeg-meeting2016 
     15May 28 18:53:36 *       ubitux ( has joined #ffmpeg-meeting2016 
     16May 28 18:54:55 *       c_14 (~c_14@unaffiliated/c-14/x-8913907) has joined #ffmpeg-meeting2016 
     17May 28 18:55:10 *       saste (~saste___@ has joined #ffmpeg-meeting2016 
     18May 28 18:57:41 *       kurosu_ (020d4d7f@gateway/web/freenode/ip. has joined #ffmpeg-meeting2016 
     19May 28 18:58:06 *       iive (~iive@unaffiliated/iive) has joined #ffmpeg-meeting2016 
     20May 28 18:59:53 *       nevcairiel (nev@WoWUIDev/WoWAce/Ace3/nevcairiel) has joined #ffmpeg-meeting2016 
     21May 28 19:00:24 <durandal_170>  we will start at at 17:15 UTC 
     22May 28 19:03:11 *       Timothy_Gu (~timothy_g@wikipedia/timothy-gu) has joined #ffmpeg-meeting2016 
     23May 28 19:03:45 *       kurosu (020d4d7f@gateway/web/freenode/ip. has joined #ffmpeg-meeting2016 
     24May 28 19:03:48 *       j-b (~jb@videolan/developer/j-b) has joined #ffmpeg-meeting2016 
     25May 28 19:04:09 *       kurosu_ has quit () 
     26May 28 19:04:43 <durandal_170>  so lets start 
     27May 28 19:04:46 <durandal_170>  first topic: Code of Conduct and policy around it 
     28May 28 19:04:52 <BBB>   yes 
     29May 28 19:05:29 <jamrial>       the "first version" was voted and commited, so that's done 
     30May 28 19:06:01 <durandal_170>  michaelni commited first version, should it be extended and other stuff added to it? 
     31May 28 19:06:09 <Timothy_Gu>    I still maintain that some solid repercussions should be specified in the CoC. The ML root idea is good, but it should be written as a reference for future. 
     32May 28 19:06:27 <BBB>   the VLC one is useful for reference 
     33May 28 19:06:56 <Timothy_Gu>    We don't have to go as solid as "this --> 1 day ban; that --> 2-day ban" but it should be clear what could happen 
     34May 28 19:06:56 <jamrial>       atomnuker was against adding that, afaik 
     35May 28 19:07:19 <jamrial>       i also think it should be added 
     36May 28 19:07:32 <jamrial>       BBB where's vlc's? 
     37May 28 19:07:54 <BBB> 
     38May 28 19:09:30 <Timothy_Gu>    This clause doesn't go against the assumption of good faith, as atomnuker seems to suggest. It's more for the clarity of the entire community, to show that our community is a mature one governed by a set of clear rules. 
     39May 28 19:09:51 <kurosu>        I'm also for adding some - in spite of the issues that were raised 
     40May 28 19:10:29 <durandal_170>  like one VLC have or more rigid? 
     41May 28 19:11:27 <atomnuker>     I dislike the notion that very well defined rules and (especially ZERO TOLERANCE POLICIES) are needed to maintain order or say a community is "mature" 
     42May 28 19:11:41 *       DSM_ (~textual@ has joined #ffmpeg-meeting2016 
     43May 28 19:12:10 <nevcairiel>    we would all be happy if its never invoked, but alas we're here because precedence exists that we need it 
     44May 28 19:12:15 <atomnuker>     I think that defining some rules is okay, as long as they're not set in stone and are debateable 
     45May 28 19:12:25 <iive>  nevcairiel: and what is that precedence? 
     46May 28 19:12:29 <BBB>   atomnuker: “NB: Before applying any of those following disciplinary policies, the VideoLAN team will try to discuss the problem with the offender in order to solve it in a more peaceful way.” 
     47May 28 19:12:39 <BBB>   atomnuker: from the VLC CoC “disciplinary actions" 
     48May 28 19:12:49 <atomnuker>     yeah, that's okay 
     49May 28 19:12:50 <kurosu>        VLC's, in my interpretation, is for issues caused by situations going out of control, so it looks fine 
     50May 28 19:13:23 <BBB>   has anyone here read the book “animal farm”? it’s so appropriate in this situation :-p </side-note> 
     51May 28 19:14:47 <durandal_170>  so should we vote for adding basic clear rules? 
     52May 28 19:14:55 <jamrial>       vlc's seems a tad explicit, so lets not copy paste it 
     53May 28 19:15:02 <BBB>   jamrial: agreed 
     54May 28 19:15:08 <saste> what's the process to approve the rules? we need to go through vote on MLs or here on IRC? 
     55May 28 19:15:29 <jamrial>       i'd say ML since some people in the voting comitee are not here, i think 
     56May 28 19:15:40 <BBB>   who will write the changes? 
     57May 28 19:15:42 <kurosu>        I'd suggest ML as IRC could be considered not to reach a quorum (whatever the spelling) 
     58May 28 19:16:03 <atomnuker>     yep, better to define them now and just put them on the ML 
     59May 28 19:16:14 <iive>  I still would like explanation, why do we need that 
     60May 28 19:17:06 <durandal_170>  without it there are no consequences to bad habits 
     61May 28 19:17:57 <iive>  like? 
     62May 28 19:17:59 <jamrial>       iive: we now have a CoC, so there needs to be a list of what happens if you don't follow it 
     63May 28 19:18:48 <BBB>   so action points: who will write the changes to the CoC for disciplinary actions on CoC violations? 
     64May 28 19:20:07 <durandal_170>  I can 
     65May 28 19:20:25 <durandal_170>  ok, so lets add something like: first warning then temporal ban then perma ban? 
     66May 28 19:20:27 <iive>  I think that most of you do not understand what CoC is for. atomnuker had explained it quite well. 
     67May 28 19:20:37 <iive>  but i do suspect that you do not care. 
     68May 28 19:20:45 *       kurosu_ (020d4d7f@gateway/web/freenode/ip. has joined #ffmpeg-meeting2016 
     69May 28 19:20:55 <iive>  You do not want CoC, you just want laws and enforcement. 
     70May 28 19:21:22 <BBB>   iive: “NB: Before applying any of those following disciplinary policies, the VideoLAN team will try to discuss the problem with the offender in order to solve it in a more peaceful way.” 
     71May 28 19:21:28 <BBB>   iive: from the videolan CoC page 
     72May 28 19:22:49 <iive>  BBB: 1. You are proposing to literally replace the current CoC with VideoLans. 2. Why don't we do that now, instead of complicating our lifes with none-sense? 
     73May 28 19:23:00 <Timothy_Gu>    1. No. 
     74May 28 19:23:07 *       Compn ( has joined #ffmpeg-meeting2016 
     75May 28 19:23:16 <Timothy_Gu>    2. In case you didn't notice, atomnuker approved this clause. 
     76May 28 19:24:08 <saste> about the consequences can you (durandal?) be more specific? I think it was about banning from the ML for some time 
     77May 28 19:24:31 <saste> having some simple rules should be better than nothing (as it happens now) 
     78May 28 19:24:44 <durandal_170>  first warning, then 7 days ban than permanent ban 
     79May 28 19:25:14 <saste> and it will be operated by the ML admins? 
     80May 28 19:25:44 <atomnuker>     what about time(x) = 4^(times_warned)? 
     81May 28 19:25:59 <jamrial>       mind, seven day ban should not be for silly stuff like top posting unless the requests are constantly ignored 
     82May 28 19:26:14 *       iive facepalms 
     83May 28 19:26:16 <durandal_170>  it is hard to enforce it on internet for real, but yes it would be ban on IRC and ML 
     84May 28 19:26:26 <Timothy_Gu>    Git push access? 
     85May 28 19:26:32 <durandal_170>  and no its not about top-posting or bottom-posting 
     86May 28 19:26:53 <durandal_170>  Timothy_Gu: yes 
     87May 28 19:27:02 <kurosu_>       Yes 3 strikes for a permanent ban seems heavy handed depending on the topic 
     88May 28 19:27:14 <atomnuker>     yeah, I think the time should be a function of time, not pernament 
     89May 28 19:27:24 <jamrial>       three strikes sounds extreme 
     90May 28 19:27:40 <atomnuker>     64 days for 3 offences seems fine with 4^times_warned 
     91May 28 19:27:46 <Timothy_Gu>    I'd be happy with "If the Contributor repeatedly, intentionally, and severly violates this Code of Conduct, possible repercussions include ..." 
     92May 28 19:28:04 <durandal_170>  agree 
     93May 28 19:28:06 <Timothy_Gu>    I don't want to get too formulaic about such things. 
     94May 28 19:28:14 <atomnuker>     yeah, that too 
     95May 28 19:29:06 <durandal_170>  so lets add that to current CoC and put it for vote on ML? 
     96May 28 19:29:12 <atomnuker>     "possible reprocussions include temporary or pernament ban on ML/IRC" 
     97May 28 19:29:40 <jamrial>       durandal_1707: send a draft first and get comments to improve it 
     98May 28 19:30:09 <durandal_170>  will do, if noone beats me 
     99May 28 19:30:26 <kurosu_>       Yeah this would be more appropriate else time will run short for this meeting 
     100May 28 19:31:08 <durandal_170>  ok, next topic? 
     101May 28 19:31:32 <Timothy_Gu>    "technical development issues" 
     102May 28 19:31:51 <Timothy_Gu>    so AVClass et al. 
     103May 28 19:32:27 <durandal_170>  isn't this topic about recent git server outage? 
     104May 28 19:32:58 <jamrial>       since cehoyos is here, we could maybe talk about his behavior and why the CoC and repercussions for violating it was introduced to begin with 
     105May 28 19:33:39 <durandal_170>  i have nothing new to add, everything have already been said on ML 
     106May 28 19:34:05 <jamrial>       where? 
     107May 28 19:34:28 <iive>  nothing's been said on ML 
     108May 28 19:34:35 <iive>  nobody wants to start flamewar 
     109May 28 19:35:11 <durandal_170>  i can talk about that later, now should we vote for AVClass thing or? 
     110May 28 19:35:26 <atomnuker>     on that topic, there's a good reason why we have stable releases and a git master - git master can be broken at any point and then be fixed so that we don't break the stable version 
     111May 28 19:35:41 <BBB>   is this about michaelni’s idea of adding an AVClass to AVCodecParameters? 
     112May 28 19:35:47 <jamrial>       durandal_1707: some context about the avclass issue first would be good 
     113May 28 19:36:35 <durandal_170>  i'm not really into it, michaelni could better describe it why (if) it's needed 
     114May 28 19:37:10 <durandal_170>  iirc its about exporting codec info from lavf, right? 
     115May 28 19:38:00 <michaelni>     AVClass & AVOption should be added to all public "Context" structs for API consistency and to make it easier for apps to support multiple ffmpeg versios and distros  
     116May 28 19:38:08 <michaelni>     that is IMHO 
     117May 28 19:38:17 <durandal_170>  basically few devs are against it iirc 
     118May 28 19:38:47 <michaelni>     nevcairiel, seemed to be unhappy about adding it to AVCodecParameters 
     119May 28 19:39:38 <durandal_170>  I would prefer it there is better solution 
     120May 28 19:41:21 <kurosu_>       Side note: the name is confusing because I would have assumed from the name it was ok 
     121May 28 19:41:39 <nevcairiel>    I see it as a necessity for contexts that have private options in internal data somewhere, not as a really good API for generic access, which loses strict typing that structs give you, you can't typo a struct member name because the compiler yells at you .. you can in avoption, etc. 
     122May 28 19:41:42 <kurosu_>       (to put such an information there) 
     123May 28 19:42:04 <michaelni>     maybe the problems AVOptions have can be fixed 
     124May 28 19:43:06 <Timothy_Gu>    Are members of AVCodecParameters supposed to be changeable by a consumer? 
     125May 28 19:43:13 <Timothy_Gu>    consumer = user app 
     126May 28 19:43:32 <michaelni>     a user app using only libavcodec or only libavformat has to change them 
     127May 28 19:43:42 <nevcairiel>    if you just use lavf and lavc, there is no reason to change them really, but you could fill it manually if you use another demuxer 
     128May 28 19:44:01 <nevcairiel>    or another mxuer for 
     129May 28 19:44:33 <nevcairiel>    but you dont need AVOptions to fill the struct 
     130May 28 19:44:45 <nevcairiel>    for me, the downsides outweigh the potential benefits, thats all 
     131May 28 19:45:56 <Timothy_Gu>    Do we care about ABI with Libav? No, right? 
     132May 28 19:46:03 <BBB>   no 
     133May 28 19:46:06 *       kurosu has quit (Ping timeout: 250 seconds) 
     134May 28 19:46:06 <nevcairiel>    we got rid of that last meeting 
     135May 28 19:46:55 <michaelni>     what downsides does having support for AVClass/AVOption have ? its completely optional, a user app can just directly access the struct, it just will have to update more often and if it wants tp support stable distros it would need alot of #if which with AVOptions should be alot cleaner 
     136May 28 19:48:25 <michaelni>     for example if we change sample_rate to a AVRational. With AVOptions an app can use the same code to set sample rate from an AVRational for new and old code 
     137May 28 19:48:42 <michaelni>     without AVOption a #if VERSION.. is needed 
     138May 28 19:48:58 <iive>  so, basically avclass would allow to use options, instead of get/set_ functions? 
     139May 28 19:49:14 <nevcairiel>    I would absolutely  favor the explicit #if myself, that way I know wtf is going on :) 
     140May 28 19:49:45 <michaelni>     me too until there are 20 such #ifs then i dont know anything anymore 
     141May 28 19:50:21 <michaelni>     AVOption gives an app this choice it doesnt need to be used 
     142May 28 19:50:31 *       kurosu (020d4d7f@gateway/web/freenode/ip. has joined #ffmpeg-meeting2016 
     143May 28 19:51:27 <michaelni>     about type checking named constants could be added like #define SAMPLE_RATE_KEY "sample_rate" 
     144May 28 19:51:34 <BBB>   so … we’re currently just rehashing the ML discussion 
     145May 28 19:51:37 <BBB>   this isn’t reall helping 
     146May 28 19:51:50 <BBB>   there’s disagreement, that much is clear 
     147May 28 19:52:01 <jamrial>       iive: afaik get/set were added because of libav's abi compatability, since new fields added by us would not have a known offset 
     148May 28 19:52:52 <iive>  jamrial: yes, but same thing applies if we want to add or remove a member. 
     149May 28 19:53:37 <Timothy_Gu>    So is this debate between AVClass/AVOptions vs. explicit getters and setters? 
     150May 28 19:53:45 <BBB>   no 
     151May 28 19:53:47 <nevcairiel>    no, dont let iive derail it 
     152May 28 19:53:49 <saste> also, with AVOption there's the possibility to some (very limited) introspection, like listing options and exporting them to a GUI/UI 
     153May 28 19:54:13 <michaelni>     saste, yes, also theres max/min and defaults 
     154May 28 19:54:36 <BBB>   my issue with avoption is that the introspection isn’t very useful, because you just list all options even though only a very small subset are useful or even set per use case 
     155May 28 19:54:53 <Timothy_Gu>    Plus, AVCodecParameters isn't really intended to be such a end user-accessible structure though. 
     156May 28 19:54:57 <BBB>   imagine AVCodecContext’s introspection - priv options made it a little better, but it’s still criminal 
     157May 28 19:55:18 <kurosu_>       Agreed with BBB: if a technical agreement can't be reached here or the ml, a way forward must be found  
     158May 28 19:55:24 <michaelni>     BBB this is a implementation problem, we could do codec specific AVOption lists 
     159May 28 19:55:46 <kurosu_>       No point in an infinite loop only broke by who gets fed up first 
     160May 28 19:55:51 <nevcairiel>    introspection and options m ight be useful on the actual AVCodecContext, especially with private codec options, but that argument falls off quickly as you get to the other structs 
     161May 28 19:55:58 <BBB>   kurosu_: so true :) 
     162May 28 19:56:29 <michaelni>     kurosu_, yes we need to know AVOption/AVClass in AVCodecParameters yes or no for the release 
     163May 28 19:56:29 <BBB>   so, I don’t think we have consensus that AVClass should be added to AVCodecParameters at this point. so how are we going to continue this discussion? 
     164May 28 19:56:50 *       DSM_ has quit (Read error: Connection reset by peer) 
     165May 28 19:58:12 <durandal_170>  hmm, I'm out of ideas. 
     166May 28 19:58:26 *       baptiste ( has joined #ffmpeg-meeting2016 
     167May 28 19:58:38 <baptiste>      heya 
     168May 28 19:59:06 <durandal_170>  hi 
     169May 28 19:59:08 <kurosu_>       I'm ok of if there's a vote, possibly restricted to people with commits in affected parts of the code in the last x units of time 
     170May 28 19:59:13 <nevcairiel>    if everyone else is in favor of adding it there, then just do so, its not like it actively hurts me or my interests, but I just don't like the concept 
     171May 28 19:59:32 <ubitux>        well the question asking if there is a downside to adding it hasn't been answered (unless i missed it); if we add it, does it promote bad usage? does it cause a maintainance burden?  
     172May 28 19:59:44 <atomnuker>     is there even anyone who agrees with michaelni to add AVClass? 
     173May 28 19:59:50 *       jamrial_ (~jamrial@ has joined #ffmpeg-meeting2016 
     174May 28 20:00:01 <atomnuker>     most of the people on the ML disagreed 
     175May 28 20:00:03 <nevcairiel>    if you want it to be complete, you need to maintain the struct twice, once in the header, and once in its AVOption declaration 
     176May 28 20:00:16 <nevcairiel>    so there is a bit of maintenance 
     177May 28 20:00:28 <michaelni>     iam happy to maintain the extra code 
     178May 28 20:00:35 <jamrial_>      offtopic, but fuck win10 wireless 
     179May 28 20:00:39 <durandal_170>  we could drop header support ;) 
     180May 28 20:00:58 <iive>  i thought that's the end goal. 
     181May 28 20:01:27 <nevcairiel>    if it is, i'm outta here =p 
     182May 28 20:02:03 *       jamrial has quit (Ping timeout: 260 seconds) 
     183May 28 20:02:10 <durandal_170>  ok, we can put this on vote for ML if michaelni really wants this 
     184May 28 20:02:15 <nevcairiel>    making AVOptions the only way to access things was never the goal, and if anyone should ever want to bring that up, we would have quite the discussion =p 
     185May 28 20:02:16 *       jamrial_ is now known as jamrial 
     186May 28 20:02:45 <ubitux>        what about serializations of parameters and such?  
     187May 28 20:03:05 <michaelni>     nevcairiel, iam a C and asm guy really, i wont come up with such idea ... 
     188May 28 20:03:28 <michaelni>     ubitux, yes, that too is a potential usecase 
     189May 28 20:03:30 <nevcairiel>    didnt think you would 
     190May 28 20:03:34 <ubitux>        i guess this enters again in the scope of api users not supposed to though (i feel like it could help wrt ffserver but i'm going to get killed) 
     191May 28 20:03:44 <nevcairiel>    the danger is real 
     192May 28 20:04:21 <ubitux>        but otoh adding the AVClass (without avoptions yet) before the release sounds like a safe bet 
     193May 28 20:04:31 <jamrial>       let ffserver die after avstream->avctx is gone 
     194May 28 20:04:32 <ubitux>        like, we could postpone such discussion 
     195May 28 20:04:36 <Timothy_Gu>    We should rewrite FFmpeg in JavaScript. Built-in introspection. 
     196May 28 20:04:51 <nevcairiel>    if the AVClass exists you might as well add the options, otherwise its just useless 
     197May 28 20:05:20 <ubitux>        i'm with michaelni on the risky aspect of not having that pointer before the release, but i don't know yet if it will be helpful in the future 
     198May 28 20:05:34 <ubitux>        nevcairiel: it's useless but it prevents the abi break later 
     199May 28 20:05:56 <nevcairiel>    well we should just decide now then instead of going with a schrödingers avclass :) 
     200May 28 20:06:01 <jamrial>       wm4 was trying to get libav to first implement the updated parser stuff before making a release. do we care? 
     201May 28 20:06:03 <ubitux>        (of adding it if we decide to use avoption or any related "technology") 
     202May 28 20:06:30 <nevcairiel>    jamrial: they decided not to though, so not sure what there is to care 
     203May 28 20:06:39 <kurosu_>       I think only 2 persons called for a vote to settle this, so besides disagreeing with the concept of vote on this, does this mean I ot enough info is available to make an educated vote?te 
     204May 28 20:06:40 <jamrial>       ah, i see 
     205May 28 20:07:07 <kurosu_>       (sorry on phone :( ) 
     206May 28 20:07:11 <nevcairiel>    well apparently noone else really has any opinions on the matter 
     207May 28 20:07:35 <kurosu_>       I consider I don't get to vote on that topic 
     208May 28 20:07:35 <BBB>   I’m not convinced we need AVOptions at this point 
     209May 28 20:07:44 <BBB>   if we don’t need AVOptions, I don’t think we need AVClass 
     210May 28 20:07:52 <BBB>   (in AVCodecParameters) 
     211May 28 20:08:15 <durandal_170>  yea, it opens cans of worms 
     212May 28 20:08:17 <jamrial>       we also don't need log context in AVCodecParameters 
     213May 28 20:08:18 <nevcairiel>    fwiw, i don't think its one of the structs that will change dramatically and all the time 
     214May 28 20:08:28 <BBB>   avcodecParameters logging is useless 
     215May 28 20:08:39 <BBB>   since the class is not codec-specific 
     216May 28 20:08:45 <BBB>   so it would just log [parameters] bla bla 
     217May 28 20:08:49 <BBB>   instead of [h264] bla bla 
     218May 28 20:08:51 <BBB>   or [mov] bla bla 
     219May 28 20:08:54 <nevcairiel>    and yes, everytime AVCodecParameters is handled you generally have some parent context it belongs to 
     220May 28 20:08:58 <nevcairiel>    be it a  muxer or demuxer 
     221May 28 20:09:00 <nevcairiel>    or something 
     222May 28 20:09:06 <BBB>   so we should actively discourage using AVCodecParameters being used as the class for logging 
     223May 28 20:09:16 <nevcairiel>    if you would try, it would likely crash 
     224May 28 20:09:18 <ubitux>        BBB: i think logging has a mechanism to print the tree? 
     225May 28 20:09:31 <ubitux>        sth like [h264 @ ...] [parameters @ ...] bla 
     226May 28 20:09:37 <nevcairiel>    never seen that 
     227May 28 20:09:49 <BBB>   I’ve never seen that either :-p certainly news to me 
     228May 28 20:09:53 <ubitux>        i remember sth like that with swr but maybe i'm confused 
     229May 28 20:09:53 <michaelni>     ubitux, yes it can print the parent too 
     230May 28 20:09:54 <BBB>   (that’s not to say you’re not right) 
     231May 28 20:11:02 <durandal_170>  ok, can we get back to this later? 
     232May 28 20:11:16 <nevcairiel>    in any case, i don't really care that much if it gets in or not, my opinion is that its not really useful or needed 
     233May 28 20:12:09 <jamrial>       on the ml if possible. a thread stating the potential benefits, drawbacks, concerns about misuse/confusion, etc, for discussion. then a vote in that same thread 
     234May 28 20:14:09 <durandal_170>  sounds good to me 
     235May 28 20:14:46 *       DSM_ (~textual@ has joined #ffmpeg-meeting2016 
     236May 28 20:14:46 <kurosu>        btw, is a vote a codified way to move forward in the face of such a polarizing issue (or more)? 
     237May 28 20:14:46 <michaelni>     if a vote is done it should be started by someone else than me & nevcairiel probabl, not sure if  a vote makes sense for technical issue it would give a yes vs no result though  
     238May 28 20:15:15 <michaelni>     and either yes or no is better than no result 
     239May 28 20:16:04 <BBB>   anyone remotely interested in managing the vote probably has some stake in it either way 
     240May 28 20:16:11 <BBB>   I think it’s fine for you to set up the vote 
     241May 28 20:16:21 <BBB>   (since you want it so you have incentive to do effort for it) 
     242May 28 20:17:32 <michaelni>     I dont understand the drawbacks and concerns about misuse/confusion so iam not able to write a neutral mail 
     243May 28 20:18:26 <durandal_170>  then setup wiki and write drawbacks/etc there? 
     244May 28 20:18:44 <jamrial>       lets not make things complicated 
     245May 28 20:20:31 <jamrial>       may i suggest another technical topic? 
     246May 28 20:20:39 <durandal_170>  yes 
     247May 28 20:21:19 <jamrial>       liabvutil is currently the only non modular library. literally everything is compiled and installed no matter your configure options 
     248May 28 20:21:33 <BBB>   I’ve complained about that 10000x 
     249May 28 20:21:42 <jamrial>       yes, that's why i bring it up :p 
     250May 28 20:21:42 <BBB>   but I think it lacks somebody actually fixing it :) 
     251May 28 20:21:53 <nevcairiel>    i dont think thats necessarily a bad thing 
     252May 28 20:22:09 <nevcairiel>    the modules in avcodec etc dont directly impact public api, they just disable some encoder or something 
     253May 28 20:22:18 <jamrial>       i could give it a try if i have the time, but i want to know how to handle it 
     254May 28 20:22:18 <nevcairiel>    but in avutil the modules practically map to public API functions 
     255May 28 20:22:22 <jamrial>       nevcairiel: that's the thing 
     256May 28 20:22:24 <nevcairiel>    which should still exist even if disabled 
     257May 28 20:22:37 <jamrial>       what would be best? instlal all headers and return NULL/ENOSYS, or not install them? 
     258May 28 20:22:43 <jamrial>       for the disabled modules, that is 
     259May 28 20:22:53 <BBB>   I would just not install them 
     260May 28 20:23:00 <BBB>   the use I have for it is in static libs 
     261May 28 20:23:01 <nevcairiel>    well even if not installed, the ABI should be the same, imho 
     262May 28 20:23:04 <BBB>   so I don’t care for placeholders 
     263May 28 20:24:21 <jamrial>       well, opencl and lzo are headers that only get installed if those components are enabled 
     264May 28 20:24:39 <jamrial>       so there's a precedent of sorts 
     265May 28 20:25:39 <BBB>   I think it’s new territory 
     266May 28 20:25:53 <BBB>   so there’s a reasonable defense for either strategy 
     267May 28 20:26:06 <iive>  what would you like to disable from libavutil? 
     268May 28 20:26:17 <nevcairiel>    shouldnt a decent linker get rid of your unused stuff in static libs anyway? 
     269May 28 20:26:21 <ubitux>        if we end up moving various dsp utils inside lavu (because it needs to be shared between codecs and filters), it makes perfect sense to make lavu modular 
     270May 28 20:26:27 <jamrial>       but at the same time, ubitux's pixelutils gets installed and just returns NULL if it's not enabled 
     271May 28 20:26:42 <ubitux>        yeah, this was a first attempt at it 
     272May 28 20:26:52 <jamrial>       iive: all the crypto stuff and such, for example 
     273May 28 20:26:53 <ubitux>        but pretty annoying to do tbh 
     274May 28 20:27:19 <jamrial>       you may only need md5, crc and aes, but not sha or camelia 
     275May 28 20:28:06 <ubitux>        does anyone mind moving to infrastructure issues after that? 
     276May 28 20:28:27 <ubitux>        i'm not sure i can stay for very long and i'm slightly concerned about that issue 
     277May 28 20:28:35 <jamrial>       as a topic? sure 
     278May 28 20:28:41 <ubitux>        yes 
     279May 28 20:28:44 *       Illya (sid133335@gateway/web/ has joined #ffmpeg-meeting2016 
     280May 28 20:29:06 <Illya> Is this only for contributors or can I sit in and listen?  
     281May 28 20:29:46 <jamrial>       you're welcome to listen and participate 
     282May 28 20:29:55 <durandal_170>  ubitux: so what you want to ask? 
     283May 28 20:30:01 <BBB>   infrastructure, let’s go 
     284May 28 20:30:19 <ubitux>        alright well 
     285May 28 20:30:24 <ubitux>        first there is the dedicated server 
     286May 28 20:30:32 <ubitux>        which i ordered a while ago as a "quick solution" 
     287May 28 20:31:03 *       rcombs ( has joined #ffmpeg-meeting2016 
     288May 28 20:31:06 <ubitux>        it's not like i do really mind the money aspect, but there was many offers 
     289May 28 20:31:19 <kurosu_>       Err what's the result on the previous topic? Send patches and it'll get reviewed but the end goal is ok? (I don't mind) 
     290May 28 20:31:23 <ubitux>        and i'm still handling the server 
     291May 28 20:31:45 <durandal_170>  ubitux: i thought that server is no longer used... 
     292May 28 20:31:47 <BBB>   kurosu_: sounds like it yes 
     293May 28 20:31:52 <ubitux>        durandal_170: ah? 
     294May 28 20:31:57 <iive>  kurosu_: i think nobody objected. Try to make some example patches and we might find issues then :) 
     295May 28 20:32:10 <ubitux>        so we don't use ffbox0 anymore? 
     296May 28 20:32:15 <kurosu_>       Ok just wanted to make the result of the meeting on that topic clear 
     297May 28 20:32:37 <durandal_170>  ubitux: i dunno, I never touched those stuff 
     298May 28 20:32:54 <ubitux>        that's the other issue; we don't really have someone to handle the sysadmin stuff 
     299May 28 20:33:01 <jamrial>       i think michaelni handled the server migration last time 
     300May 28 20:33:14 <ubitux>        does anyone have a sysadmin in his relationships that would be interested in that? 
     301May 28 20:33:15 <michaelni>     "<ubitux> it's not like i do really mind the money aspect, but there was many offers" <-- everyone disappeared except the one from bulgaria 
     302May 28 20:33:28 <jamrial>       ubitux: wasn't it llogan? 
     303May 28 20:34:09 <michaelni>     we have a virtual box in bulgaria that ffbox0 could be moved to if teres a volunteer 
     304May 28 20:34:14 <ubitux>        well, i saw michael handling mailing issue recently, and we have regularly unsolved issue wrt the infrastructure 
     305May 28 20:34:42 <michaelni>     trac is currently on a virtual box there too 
     306May 28 20:35:00 <jamrial>       kierank offered a server as well i remember 
     307May 28 20:35:02 <ubitux>        like, i still receive many unwanted dvdnav moderation mails, and no one seems to know how to stop them 
     308May 28 20:35:22 <iive>  jamrial: trac was on kierank's server, until it died. 
     309May 28 20:35:30 <ubitux>        i feel like michael has to handle all the dirty stuff everytime there is a problem 
     310May 28 20:35:32 <iive>  the server 
     311May 28 20:35:35 *       Shiz ( has joined #ffmpeg-meeting2016 
     312May 28 20:35:36 <ubitux>        or even when we have to upload a sample etc 
     313May 28 20:35:58 <ubitux>        i feel like we really need someone to officially handle the sysadmin stuff 
     314May 28 20:36:07 <michaelni>     jamrial, kieranks server is now limited to 5mbit/sec 
     315May 28 20:36:15 <ubitux>        so if anyone knows someone to help with that i think that would be a good idea 
     316May 28 20:36:15 *       kierank (sid5955@gateway/web/ has joined #ffmpeg-meeting2016 
     317May 28 20:36:50 <BBB>   speaking of the devil :D 
     318May 28 20:37:41 <michaelni>     it would be nice to have some help with sysadmin stuff 
     319May 28 20:37:48 <durandal_170>  i'm really not into such stuff, so I can't do that 
     320May 28 20:38:15 <BBB>   maybe lou is interested in that kind of stuff? 
     321May 28 20:38:43 <durandal_170>  the only experience I have is some GUI web setup of mail server and http server and database ... 
     322May 28 20:39:05 <BBB>   you shouldn’t feel obliged to do stuff you don’t want to do ;) 
     323May 28 20:39:17 <michaelni>     about uploading samples anyone who needs/wants access to that its easy to give 
     324May 28 20:39:22 *       DSM_ has quit (Max SendQ exceeded) 
     325May 28 20:39:43 <jamrial>       michaelni: at least regarding samples write access could be given to people that often makes new tests 
     326May 28 20:40:21 <jamrial>       or maybe to anyone that can git push 
     327May 28 20:40:38 <iive>  can we hear what services are running on what server atm? 
     328May 28 20:40:41 <durandal_170>  michaelni: iirc you gave me access but I don't remember what password was 
     329May 28 20:40:43 <michaelni>     i need to setup account on the box so "anyone that can git push" is a bit hard 
     330May 28 20:40:50 <jamrial>       ah ok 
     331May 28 20:41:49 <ubitux>        pretty sure no one in the community is available for sysadmin stuff, we probably need someone who isn't a developer but more a sysadmin 
     332May 28 20:42:00 <ubitux>        and i think they're not that rare 
     333May 28 20:42:01 <michaelni>     having someone with mail / postfix experience and more free time than tim would be usefull 
     334May 28 20:42:21 <ubitux>        we can probably ask around our close relationships 
     335May 28 20:42:35 <michaelni>     that would be a good idea i think 
     336May 28 20:42:36 <ubitux>        i mean, we need someone to trust 
     337May 28 20:42:41 <michaelni>     yes 
     338May 28 20:43:26 <ubitux>        so if you're working in a company and knows well a bored sysadmin who like opensource 
     339May 28 20:43:31 <ubitux>        maybe hint him? 
     340May 28 20:43:35 <BBB>   don’t we have various people in the project that don’t code? 
     341May 28 20:43:47 <ubitux>        ppl that don't code seems as busy as us 
     342May 28 20:43:56 <BBB>   iive doesn’t seem busy 
     343May 28 20:43:58 <ubitux>        and don't seem to particularly have sysadmin stuff 
     344May 28 20:44:04 <ubitux>        skills* 
     345May 28 20:44:23 <ubitux>        we called many times for sysadmin support 
     346May 28 20:44:31 <BBB>   iive: can you sysadmin? 
     347May 28 20:44:32 <ubitux>        i think we need to look a bit outside this community 
     348May 28 20:45:11 <iive>  BBB: no.  
     349May 28 20:45:56 <iive>  just to be clear. we are talking about moving the maillist to the bulgarian box? 
     350May 28 20:46:15 <michaelni>     iive, i think we dont talk about anything specific 
     351May 28 20:46:41 <michaelni>     also raz has already copied ffbox to bulgaria but its a few month old 
     352May 28 20:46:47 <BBB>   I’m gonna run out for lunch, I’ll be back in a bit 
     353May 28 20:47:38 <kurosu_>       I think we're discussing about asking people worthy of trust to do it. Why can't we do the same as for asking infrastructure ? 
     354May 28 20:47:55 <ubitux>        just an overview about what kind of stuff we need: security updates, managing various accesses, improvements in various setups such as mail or bug trackers, help with synchronizing samples, ... 
     355May 28 20:48:12 <ubitux>        and i'm probably forgetting many things as i'm not even helping a bit 
     356May 28 20:48:20 <ubitux>        (except by paying the server) 
     357May 28 20:48:28 <jamrial>       i also need to leave, but wont be back for a few hours so i'll miss the rest of the meeting 
     358May 28 20:48:47 <ubitux>        i'm going afk pretty soon but i haven't much more to say 
     359May 28 20:49:01 <ubitux>        that was just a desperate call so everyone is aware of the issue 
     360May 28 20:49:10 *       jamrial has quit () 
     361May 28 20:49:47 <iive>  ubitux: what's the issue again? 
     362May 28 20:50:25 <michaelni>     iive theres no real issue i think, it just would be nice to have some help with sysadmin 
     363May 28 20:51:00 <durandal_170>  michaelni: is ubitux paid server used for web page? 
     364May 28 20:51:10 <michaelni>     yes 
     365May 28 20:51:33 <durandal_170>  hmm, why it was never transfered? 
     366May 28 20:52:40 <kierank>       VLC have offered to sysadmin for years 
     367May 28 20:53:03 <Compn> we have to vote on vlc 
     368May 28 20:53:09 <Compn> before we were worried about takeover on vlc side 
     369May 28 20:53:34 <Compn> but now i think its ok for vlc to host.. 
     370May 28 20:54:27 <michaelni>     we have a dedicated server from ubitux and one in bulgaria and one from kieran 
     371May 28 20:54:50 <Compn> ok, just need admins ? 
     372May 28 20:54:55 <Compn> i think i offered to admin before 
     373May 28 20:55:01 <Compn> i put it on the table, my offer to admin again  
     374May 28 20:55:14 <ubitux>        we need someone to refer to for every sysadmin stuff 
     375May 28 20:55:28 <michaelni>     Compn, you know postfix and stuff ? 
     376May 28 20:55:40 <michaelni>     ubitux, probably 
     377May 28 20:55:40 <ubitux>        like if there is an issue, we're always looking for someone for the skills and motivation 
     378May 28 20:56:17 <michaelni>     theres also a dedicated server from baptiste, i almost forgot 
     379May 28 20:56:30 <michaelni>     fate is on that one 
     380May 28 20:56:56 <michaelni>     or i think its a dedicated server at least  
     381May 28 20:57:02 <Compn> michaelni : no, but i'm fast learner :P 
     382May 28 20:57:32 <ubitux>        mails are not a simple thing 
     383May 28 20:57:38 <ubitux>        and it's kind of sensitive 
     384May 28 20:57:48 <michaelni>     ubitux, yes :/  
     385May 28 20:58:44 <michaelni>     we probably should config postfix or spamassasin to check DMARK/DKIM/SPF or part of that on incoming mai (not really important but i thn it doest curretly) 
     386May 28 20:59:48 <michaelni>     also gmail will soon switch DMARK to reject like yahoo, we have 2 methods basically in place to deal with that a custom solution from tim ad malman itself 
     387May 28 21:00:31 <michaelni>     tims solution would then traslate all gmail like yahoo to 
     388May 28 21:00:56 <michaelni>     mailman would put the original gmail into CC and put in from 
     389May 28 21:01:16 <michaelni>     mailman is fully automatic, tims solution needs manual listig of affected servers 
     390May 28 21:01:27 <michaelni>     or we hack mailman to do something else 
     391May 28 21:01:50 <kurosu_>       That's a new item of work, but maybe not the biggest for the volunteer(s)? 
     392May 28 21:02:13 <kurosu_>       Just to make sure people what the request is 
     393May 28 21:02:21 <michaelni>     kurosu_, absolutely no big issue 
     394May 28 21:02:35 <kurosu_>       *people understand what 
     395May 28 21:04:03 <michaelni>     Compn, what server admin stuff yu know ? 
     396May 28 21:05:03 <michaelni>     i mean any expereicene with apache config stuff, git? svn ? ftpd ? 
     397May 28 21:05:19 <michaelni>     (that was a random list)  
     398May 28 21:05:58 <kurosu_>       Just list what you and the others mainly do? 
     399May 28 21:06:00 <Compn> just running home papache 
     400May 28 21:06:09 <Compn> stuff like that 
     401May 28 21:06:16 <Compn> er apache on my own home box, temporarily 
     402May 28 21:06:17 <Compn> nothing major 
     403May 28 21:06:41 <Compn> and on some other site 
     404May 28 21:07:22 <Compn> i just figured it would be easier to ask me to restart some service 
     405May 28 21:07:23 <Compn> than michael 
     406May 28 21:07:28 <Compn> but if not, nevermind 
     407May 28 21:08:00 <durandal_170>  i would like to propose next topic: FFmpeg funding/donations 
     408May 28 21:08:15 <michaelni>     kurosu_, apt-get update ;) occasionally creating an account, looking at why someones mail isnt where it shoud be, then seting up stuff that is needed like new mailman  
     409May 28 21:09:25 <michaelni>     kurosu_, rarely updating dns zone, uploading samples, rarely fixing permissions on samples, once every few month building new doxygen for a major release 
     410May 28 21:10:09 *       durandal1170 ( has joined #ffmpeg-meeting2016 
     411May 28 21:10:18 <michaelni>     its not that much work when everythig works 
     412May 28 21:10:40 <Shiz>  considering the quality of this sysadmin conversation it seems delegating it to VLC would be a wise decision 
     413May 28 21:10:45 <Compn> i'd rather ask michael what he wants  
     414May 28 21:11:07 <kurosu_>       Michaelni, I meant when asking volunteers :-) I'm personally not 
     415May 28 21:12:27 *       durandal_170 has quit (Ping timeout: 240 seconds) 
     416May 28 21:12:27 <michaelni>     yes, i know, but i dont know who to ask, so be listing it a but others could ask people 
     417May 28 21:13:07 <saste> michaelni, what's the implications of asking vlc to do sysadmin stuff? or in other words, why you never considered that option? 
     418May 28 21:13:28 <kierank>       Paranoia of course 
     419May 28 21:13:33 <saste> i'm fine with either, but i'm not the one helping so my opinion doesn't matter, was just trying to understand 
     420May 28 21:13:42 *       durandal_1707 gives channel operator status to durandal1170 
     421May 28 21:15:24 <iive>  or we could move the whole project to github :P 
     422May 28 21:16:26 <michaelni>     saste, we use dedicated servers and multiple virtual machines on the new bulgarian box, vlcs offer was to integrate some of our services into the existing vlc server IIRC 
     423May 28 21:17:42 <michaelni>     its much more restrictive IIUC 
     424May 28 21:18:02 <Illya> So the different components are: website, mailing list, email, git, fate samples ftpd/rsync (have I missed any?). Which services would be integrated into vlc? 
     425May 28 21:18:17 <michaelni>     theres also trac 
     426May 28 21:19:34 <Compn> [15:25] <kierank> Paranoia of course 
     427May 28 21:19:36 <Compn> sounds like insult 
     428May 28 21:19:49 <cehoyos>       That's not correct. 
     429May 28 21:21:03 <Compn> code of conduct  
     430May 28 21:21:04 <Compn> :P 
     431May 28 21:21:35 *       Shiz ( has left #ffmpeg-meeting2016 ("-") 
     432May 28 21:21:52 *       kurosu has quit (Ping timeout: 250 seconds) 
     433May 28 21:22:30 <Illya> iive: makes a good suggestion, GitHub would release at least two services (git and trac). For trac to GitHub you could look at something like: It also might make the project more accessible to new contributors  
     434May 28 21:22:43 <michaelni>     Compn, btw you should have the power to upload fate-samples and also normal samples 
     435May 28 21:23:06 <michaelni>     you are in samples group 
     436May 28 21:23:06 <Illya> And GitHub is fairly reputable, if you're paranoid about that 
     437May 28 21:23:08 <kierank>       michaelni: you are mistaken 
     438May 28 21:23:45 <kierank>       But don't let that get in the way of VLC paranoia 
     439May 28 21:24:32 <kurosu_>       Please. 
     440May 28 21:24:47 <durandal1170>  well if nobody comes then VLC will be used 
     441May 28 21:26:28 <durandal1170>  ok, can we get to next topic? 
     442May 28 21:26:38 <michaelni>     durandal1170, ok with me 
     443May 28 21:26:48 <durandal1170>  hopefully the last topic :) 
     444May 28 21:27:04 <michaelni>     "<durandal1170> ok, can we get to next topic?" 
     445May 28 21:27:33 <durandal1170>  its about FFmpeg funding/donations 
     446May 28 21:28:07 <michaelni>     i want more funding/donations for FFmpeg, sadly thats probably all i can do here 
     447May 28 21:28:21 <michaelni>     that is "wanting" 
     448May 28 21:28:47 <durandal1170>  i want possibility to fund devs to work on specific part of FFmpeg 
     449May 28 21:28:53 <michaelni>     +1 
     450May 28 21:28:59 <iive>  what happened with FFmtech? 
     451May 28 21:29:15 <saste> this was discussed again and again, and we decided that there is nothing preventing us to do that 
     452May 28 21:29:38 <saste> at the moment we have a total of ~15K USD in the SPI and funds 
     453May 28 21:30:19 <saste> OTOH we never asked to use money for that, since it was always only about refunding people for stuff and travel 
     454May 28 21:30:35 <saste> also I'm not sure what was the outcome of the last outreachy 
     455May 28 21:30:58 <michaelni>     saste, can we fund someone maybe to make kierans fuzzing GSoC project a reality ? i mean if people agree to that  
     456May 28 21:31:02 <kierank>       There were no good students 
     457May 28 21:31:59 <durandal1170>  so just need to pick some part of codebase that need refactoring/cleaning up/improving? 
     458May 28 21:32:08 <durandal1170>  like rm demuxer 
     459May 28 21:32:45 <durandal1170>  or swscale 
     460May 28 21:33:47 <michaelni>     iam happy to help cleanup swscale but iam missing some motivation (not money, maybe rather other people joining in and working, i dont know) 
     461May 28 21:35:28 *       saste has quit (Ping timeout: 252 seconds) 
     462May 28 21:36:25 <durandal1170>  guess only left is to propose such thing like, sponsoring dev to for X 
     463May 28 21:36:28 <kierank>       I would be able to fund a non crazy avfilter api 
     464May 28 21:37:52 <durandal1170>  hmm, nicolas is working on some lavfi stuf, unrelated to that, and he is very busy 
     465May 28 21:38:48 <durandal1170>  the main problem I see is what will Libav do if we develop such new API 
     466May 28 21:39:27 <durandal1170>  i'm really interested in better lavfi API for various reasons 
     467May 28 21:40:18 <durandal1170>  kierank: i guess end goal is to feed filters AVFrames directly? 
     468May 28 21:40:29 <nevcairiel>    doesnt that happen today? 
     469May 28 21:40:45 <durandal1170>  you have filtergraph... 
     470May 28 21:40:52 <kierank>       Main goal is to not have weird buffering  
     471May 28 21:41:02 <nevcairiel>    an API that links filters together for a filtering chain is quite useful 
     472May 28 21:41:04 <kierank>       i.e an API suitable for live as well as file 
     473May 28 21:41:42 <nevcairiel>    things just get oddly complicated if you handle multiple un-synced inputs and shit like that 
     474May 28 21:45:51 <durandal1170>  kierank: do you still have candidate for new lavfi API?  
     475May 28 21:47:32 <BBB>   are there any other subjects that come after this one? 
     476May 28 21:48:05 *       saste (~saste___@ has joined #ffmpeg-meeting2016 
     477May 28 21:48:37 <durandal1170>  BBB: any stuff you want to chat about... 
     478May 28 21:49:17 <saste> i was offline for a few minutes, did I miss any reply? 
     479May 28 21:49:20 <BBB>   no I’m just surprised it’s still ongoing after 3 hrs 
     480May 28 21:50:17 <kurosu_>       Yeah not efficient but that's not like people were not doing anything else 
     481May 28 21:51:01 <BBB>   I agre with the several people above that we should re-consider VLC as sysadmin for our stuff 
     482May 28 21:51:11 <BBB>   doing stuff oruselves really doesn’t make sense if we’re strained in manpower 
     483May 28 21:51:31 <BBB>   is there anything concrete we’re going to do w.r.t. derek and carl? 
     484May 28 21:51:57 <iive>  i think that last time when this was discussed, it was mentioned that they cannot admin all our services 
     485May 28 21:52:53 <saste> <saste> michaelni, before proposing that I'd check with SPI if that would be an issue 
     486May 28 21:52:54 <saste> <saste> what would be possible (indeed we already did it in the past) would be to sponsor an outreachy slot as we did in a past edition 
     487May 28 21:52:54 <saste> <saste> IIRC FFmpeg is not involved with the last round of outreachy, right? 
     488May 28 21:52:54 <saste> <saste> also, related to this, at the moment it is me and michaelni who approve the funding requests 
     489May 28 21:52:54 <saste> <saste> also I should report about the donations status (every six months) and I think I missed that in the last year 
     490May 28 21:52:57 <saste> <saste> when we decided the funding procedure it was agreed that it was going to be a temporary solution, but then we ended up with that setup since then (it was 2012) 
     491May 28 21:53:00 <saste> <saste> do you have any specific proposal / comment about funding and donations? 
     492May 28 21:53:02 <saste> <saste> then I think we should move on 
     493May 28 21:53:04 <saste> <saste> I can re-ask the SPI guys again to know if it is acceptable to use the fund for sponsoring development 
     494May 28 21:53:07 <saste> <saste> in the past years people also suggested fancier stuff like crowdfunding but nothing came out of that 
     495May 28 21:53:14 <saste> ^^ this is what I wrote before realizing i was offline, sorry for the spam 
     496May 28 21:54:30 *       durandal_170 ( has joined #ffmpeg-meeting2016 
     497May 28 21:54:54 <kurosu_>       Yeah the funding topic didn't seem to reach a conclusion/decision of what to do 
     498May 28 21:55:25 <saste> what we have now is bad, but better than nothing 
     499May 28 21:55:53 <kurosu_>       Well you've proposed something to do, seems fine 
     500May 28 21:55:56 <saste> at least it helps with collecting money for sponsoring a bit of travelling 
     501May 28 21:56:21 *       durandal1170 has quit (Ping timeout: 264 seconds) 
     502May 28 21:57:15 <kurosu_>       So, topic closed and "action points" noted? 
     503May 28 21:57:48 <saste> I think so 
     504May 28 21:58:05 <durandal_170>  saste: I want to make it possible to fund work on FFmpeg part 
     505May 28 21:58:24 <durandal_170>  notably rm demuxer improvements 
     506May 28 21:59:08 <kurosu_>       Kierank mentioned btw he was willing to fund some work on libavfilter API that would suit his needs, the details of which he'll give to whomever is integrated  
     507May 28 21:59:15 <kurosu_>       *interested 
     508May 28 21:59:25 <saste> before asking that on the ML, let me check again with SPI, then we can use the usual procedure with the fund request on the ML 
     509May 28 21:59:48 <saste> ping me in a few days if you don't hear nothing from me 
     510May 28 21:59:52 <kierank>       So Derek and carl? 
     511May 28 22:00:25 <durandal_170>  I guess can't be done much, Derek left and carl is still here 
     512May 28 22:00:49 <michaelni>     saste, if you get no reply from SPI then put them in a position in which they must reply 
     513May 28 22:01:01 <saste> michaelni, sure 
     514May 28 22:02:07 <durandal_170>  for sure I'm ignoring carl diffs 
     515May 28 22:02:24 <kurosu_>       Although the CoC and the repercussions are neither final nor voted, I'm for this to be acted upon 
     516May 28 22:02:34 <durandal_170>  he already posted at least 1 proper patch so he is able to follow that.... 
     517May 28 22:02:43 <iive>  durandal_170: it's ok if you ignore them all together. 
     518May 28 22:02:50 <kurosu_>       It shouldn't even been allowed to reach this point without intervention 
     519May 28 22:02:52 <iive>  durandal_170: what is not OK is to reject them  
     520May 28 22:03:20 <durandal_170>  iive: be assured if they are important that they will not be rejected 
     521May 28 22:03:34 <iive>  with stuff like "Sorry, life sucks. This patch is unacceptable." 
     522May 28 22:03:50 <kurosu_>       But I'd personally be only for the equivalent of a warning if it would come to that 
     523May 28 22:04:31 <durandal_170>  iive: i replied how bug should be correctly handled after that.. 
     524May 28 22:04:54 <iive>  durandal_170:  that's not the issue. You went on war 
     525May 28 22:05:21 <iive>  "I will comment only this time and if I don't get positive result I will simply block every your patch." 
     526May 28 22:05:29 <BBB>   ... 
     527May 28 22:05:41 <kurosu_>       Everyone is to blame with this mess 
     528May 28 22:05:52 <kurosu_>       People should have been told to cool off 
     529May 28 22:05:59 <iive>  absolutely correct. 
     530May 28 22:06:04 <BBB>   I think that’s the point of the VLC tempbans on the ML 
     531May 28 22:06:17 <BBB>   a 1-day ban is essentially just to cool down 
     532May 28 22:06:23 <durandal_170>  iive: i wrote that, but I looked at later diffs... 
     533May 28 22:06:50 <iive>  durandal_170: but the "life sucks" is action on our promise. 
     534May 28 22:06:55 <kurosu_>       Exactly, although an informal warning would have been a prelude 
     535May 28 22:07:26 <iive>  durandal_170: you know that a polite request could be a lot more effective than threats? 
     536May 28 22:07:31 <kurosu_>       And we are precisely starting such a discussion here 
     537May 28 22:07:33 <BBB>   so conclusion is that the proposed disciplinary actions on the CoC will be written such that they would have prevented this but otherwise no further action? 
     538May 28 22:08:15 <durandal_170>  BBB: further action on what? 
     539May 28 22:08:56 <durandal_170>  iive: i asked politely multiple times 
     540May 28 22:10:30 <BBB>   further action on derek vs carl 
     541May 28 22:11:51 <durandal_170>  anyone could propose what to do 
     542May 28 22:12:08 <iive>  durandal_170: you can give me links to the emails where you do that later. But in future, don't go on war. Ask other developers for help. 
     543May 28 22:12:08 <kurosu_>       The CoC doesn't contain repercussions yet 
     544May 28 22:12:12 <durandal_170>  thing is CoC come after this incident 
     545May 28 22:12:44 <BBB>   I don’t even think the CoC strictly existed when the incident arose 
     546May 28 22:12:54 <nevcairiel>    the CoC is just a formality, if someone wants to argue the behavior was OK because there was no CoC in place yet, i dont want to work with them in the first place 
     547May 28 22:13:37 <BBB>   right 
     548May 28 22:13:40 <durandal_170>  it certainly was not OK... but how can we proceed? 
     549May 28 22:13:54 <kurosu_>       I'm not - I just want CoC to contain it then it to be used for the decision to be made 
     550May 28 22:14:36 <BBB>   going by the VLC CoC ( 
     551May 28 22:14:47 <BBB>   mailing list violations 
     552May 28 22:15:06 <BBB>   first bulletpoint for non-netiquette violations: “24-hour ban from the mailing list in question.” 
     553May 28 22:15:19 <BBB>   and also “As one cannot develop without the mailing lists, a ban from a *-devel mailing list will result in a ban from commit access for one day on the related project.” 
     554May 28 22:15:23 <nevcairiel>    i dont think its as simple as a "mailing list violation" 
     555May 28 22:15:33 <BBB>   no it clearly wasn't 
     556May 28 22:16:07 <BBB>   oh, there’s an escalation section before that 
     557May 28 22:16:09 <BBB>   “The first violation will always result in a simple warning, except if it is a grave or deliberate violation.” 
     558May 28 22:16:20 <BBB>   (I’m assuming this one is considered grave) 
     559May 28 22:16:23 <BBB>   and then: “The following violations will result in some of the disciplinary actions listed in the paragraphs below.” 
     560May 28 22:16:43 <BBB>   so, I think we should formally ban carl from the ML for 24 hrs and ban his commit access for the same period 
     561May 28 22:16:47 <iive>  what exactly is grave situation? 
     562May 28 22:17:04 <BBB>   iive: this is how lawyers get rich 
     563May 28 22:17:18 <iive>  <BBB> (I’m assuming this one is considered grave) 
     564May 28 22:17:25 <kurosu_>       Something that is seen worthy of repercussions by several people 
     565May 28 22:17:32 <iive>  i'd like to know why what you consider grave 
     566May 28 22:17:55 <BBB>   iive: that’s not necessary for this decision 
     567May 28 22:18:06 <BBB>   iive: we merely need to know whether this particular instance was grave or not 
     568May 28 22:18:21 <iive>  and do YOU think it was? 
     569May 28 22:18:23 <BBB>   we don’t need to put a line in the sand on what is grave and what is not, we can do that as situations arise 
     570May 28 22:18:51 <kurosu_>       Well Carl hasn't been treated kindly but the situation has actually been rotting for years now 
     571May 28 22:19:04 <durandal_170>  we can put vote on ML... 
     572May 28 22:19:54 <kurosu_>       CoC would avoid abuse in one way or another, so that this situation doesn't actually rot 
     573May 28 22:23:46 <BBB>   durandal_170: vote on … CoC? or my proposal? 
     574May 28 22:24:02 <kurosu_>       Durandal_170: I'm ok for a vote on Carl, after the CoC contains repercussions (is it a warning? A 1 day ban? Etc) 
     575May 28 22:24:39 <iive>  it's warning 
     576May 28 22:24:56 <durandal_170>  on ML? irc meeting don't have enough members 
     577May 28 22:25:15 <kurosu_>       Durandal_170: ML 
     578May 28 22:25:45 <iive>  you see, this is perfect example why CoC should not punishment clauses 
     579May 28 22:26:03 <iive>  we are not discussing things on principle, we are looking for a way to punish Carl. 
     580May 28 22:26:18 <iive>  and we should be looking for a way to avoid getting there. 
     581May 28 22:26:54 <nevcairiel>    its too late now, and we need to handle the situation at hand 
     582May 28 22:27:03 <kurosu_>       I have the polar opinion, I consider that harassment towards Carl should be equally punished if found a harassment 
     583May 28 22:27:34 <iive>  kurosu_: It should stop. 
     584May 28 22:28:24 <kurosu_>       nevcairiel, do you want a vote here and now, to what effect? 
     585May 28 22:29:00 <nevcairiel>    yes everyone should be happy and get along, but in real-life people don't, and we need to be prepared to handle that if needed, instead of burying our  heads in the sand and losing people over that 
     586May 28 22:30:43 <BBB>   *clap*clap*clap* ++ 
     587May 28 22:31:09 <BBB>   I agree that a vote on the ML would be better to give people that fell asleep here the chance to participate also 
     588May 28 22:34:59 <kurosu_>       It's late here. I'm ok for a vote also, just not sure what kind of offense it would be 
     589May 28 22:35:09 <kurosu_>       That could be part of the vote 
     590May 28 22:35:23 <kurosu_>       BTW public vote? 
     591May 28 22:36:08 <kurosu_>       It seems it has always been so 
     592May 28 22:36:37 <iive>  it's hard to fight slender campaign 
     593May 28 22:37:05 <iive>  as it might not involve strong offensive words. 
     594May 28 22:38:08 <kurosu_>       Well people can make an opinion for themselves 
     595May 28 22:39:05 <iive>  ffmpeg was broken apart once from such campaign. 
     596May 28 22:39:38 <saste> I have to leave now, please send the chat log to the ML when the meeting closes 
     597May 28 22:39:53 <saste> goodbye! ;-) 
     598May 28 22:39:59 <iive>  have fun :) 
     599May 28 22:40:12 <michaelni>     saste, have fun ! 
     600May 28 22:42:27 <kurosu_>       Same here. Debate is done for me, I'll then act upon whatever is decided for vote afterwards 
     601May 28 22:42:41 <kurosu_>       Good night  
     602May 28 22:42:52 *       kurosu_ has quit () 
     603May 28 22:43:29 *       saste (~saste___@ has left #ffmpeg-meeting2016 ("Leaving") 
     604May 28 22:50:10 <atomnuker>     you know, I think we should define a multimedia-based drinking game and put it up on our website 
     605May 28 22:50:35 <nevcairiel>    identify the codec by its artifacts 
     606May 28 22:50:44 <atomnuker>     take a shot for every AVI extension/hack your file has (double for B-frames) 
     607May 28 22:50:59 <atomnuker>     take a shot if your VFR file doesn't play right 
     608May 28 22:51:06 <iive>  atomnuker: drinking 10l cola? 
     609May 28 22:51:15 <nevcairiel>    poor anime people atomnuker 
     610May 28 22:51:26 <atomnuker>     never understood that 10l cola thing 
     611May 28 22:52:01 <atomnuker>     oh oh take a shot if your file has an incorrect colorspace (e.g. 601 flagged 709) 
     612May 28 22:52:23 <iive>  it's from mplayer, a figurative punishment for breaking the build. 
     613May 28 22:52:45 <atomnuker>     ah, that's before my time 
     614May 28 22:52:58 <iive>  oh, and it is the cola brand you don't like :D otherwise it won't be punishment. 
     615May 28 22:53:13 <atomnuker>     doctor pepper is poison 
     616May 28 22:53:37 <atomnuker>     anamorphic video -> take a shot 
     617May 28 22:54:00 <atomnuker>     weird crop/whatever flag in your mkv -> take a shot 
     618May 28 22:54:30 <mateo`>        atomnuker: anamorphic video not flagged as such 
     619May 28 22:54:38 <atomnuker>     I wonder what would be horrible enough to make you want to finish your bottle 
     620May 28 22:55:01 <nevcairiel>    anamorphic isnt that bad, its just terrible if no AR is indicated =p 
     621May 28 22:58:03 <atomnuker>     a shot if you see a jpeg2000 image/video anywhere outside professional stuff 
     622May 28 22:59:21 <nevcairiel>    i have a few DCP trailers of movies, not sure if that counts 
     623May 28 22:59:45 <atomnuker>     a shot for direct (e.g. sped up) 24(/1.001) to 25fps conversion 
     624May 28 23:00:06 <nevcairiel>    thats unfair for people living in PAL land =p 
     625May 28 23:00:13 <nevcairiel>    thats like, all our content! 
     626May 28 23:00:13 <nevcairiel>    :D 
     627May 28 23:00:39 <iive>  do you want to make all developers a chronic alcoholics? :P 
     628May 28 23:01:00 <atomnuker>     those are all small shots, they don't add up 
     629May 28 23:01:23 <atomnuker>     and I can't think of anything horrific to make me drink an entire bottle 
     630May 28 23:01:53 <iive>  you drink it, one shot at a time :) 
     631May 28 23:02:16 <atomnuker>     you can't get drunk like that, gotta drink a whole bottle at a time 
     632May 28 23:02:58 <atomnuker>     well, depends on the alcohol too I guess 
     633May 28 23:05:02 <atomnuker>     take a shot for some fuzzed wav file being recognized as aac/h264/mp3 
     634May 28 23:05:27 <atomnuker>     drink the entire bottle if it gets recognized as... theora 
     635May 28 23:05:35 <atomnuker>     that's sufficiently rare 
     636May 28 23:05:38 <atomnuker>     or VC1 
     637May 28 23:05:41 <nevcairiel>    not sure that can even happen 
     638May 28 23:06:14 <atomnuker>     cosmic rays is the sole cause of all miracles that *may* happen 
     639May 28 23:06:29 <atomnuker>     I believe in them 
     640May 28 23:07:08 <atomnuker>     at this moment billions of neutrinos occupy my laptop in a 1x1x1m cube 
     641May 28 23:07:45 <atomnuker>     all it would take is a couple of thousand to react and flip some bits 
     642May 28 23:08:35 <iive>  i thought neutrinos don't react... that's why they are so hard to detect. 
     643May 28 23:11:48 <atomnuker>     if they don't react how could you even detect them? 
     644May 28 23:12:55 <atomnuker>     they do, but because they're so low energy it takes quite a lot of matter for them to make a difference and generate a photon 
     645May 28 23:20:55 <iive>  well, I wasn't precise...  
     646May 28 23:23:12 <Timothy_Gu>    So is the meeting done? 
     647May 28 23:23:52 <iive>  i guess 
     648May 28 23:26:06 *       Compn waves at cehoyos 
     649May 28 23:26:07 <Compn> ehe 
     650May 28 23:26:17 <Compn> how goes it carl? :) 
     651May 28 23:39:29 <baptiste>      what's with carl ? 
     652May 28 23:48:46 *       BBB has quit (Quit: BBB) 
     653May 28 23:52:10 *       durandal_1707 has quit (Read error: Connection reset by peer) 
     654May 28 23:55:37 *       durandal_1707 ( has joined #ffmpeg-meeting2016 
     655May 28 23:55:57 *       durandal_170 has quit (Quit: leaving) 
     656May 28 23:57:10 *       durandal_1707 ( has left #ffmpeg-meeting2016 
     657May 28 23:58:51 *       BBB ( has joined #ffmpeg-meeting2016 
     658May 29 00:24:48 *       kierank (sid5955@gateway/web/ has left #ffmpeg-meeting2016 
     659May 29 01:39:00 *       iive has quit (Quit: They came for me...) 
     660May 29 04:35:49 *       Timothy_Gu (~timothy_g@wikipedia/timothy-gu) has left #ffmpeg-meeting2016 
     661May 29 04:41:13 *       BBB has quit (Quit: BBB) 
     662May 29 06:21:12 *       arthcp (75c6c0e7@gateway/web/cgi-irc/ has joined #ffmpeg-meeting2016 
     663May 29 06:21:40 *       arthcp (75c6c0e7@gateway/web/cgi-irc/ has left #ffmpeg-meeting2016 
     664May 29 09:21:40 *       nevcairiel (nev@WoWUIDev/WoWAce/Ace3/nevcairiel) has left #ffmpeg-meeting2016 (" - Chat comfortably. Anywhere.") 
     665May 29 10:22:39 *       Illya (sid133335@gateway/web/ has left #ffmpeg-meeting2016 
     666May 29 12:38:34 *       BBB ( has joined #ffmpeg-meeting2016 
     667May 29 13:43:45 *       c_14 (~c_14@unaffiliated/c-14/x-8913907) has left #ffmpeg-meeting2016 ("WeeChat 1.4") 
     668May 29 14:19:35 *       BBB ( has left #ffmpeg-meeting2016 
     669May 29 17:00:46 *       omerjerk (2d791d07@gateway/web/freenode/ip. has joined #ffmpeg-meeting2016 
     670May 29 17:22:36 *       omerjerk has quit (Ping timeout: 250 seconds) 
     671May 29 22:43:06 *       cehoyos has quit (Quit: Konversation terminated!)