wiki:

FFmeeting

/

2015-09

This FFmeeting was hosted on irc://irc.freenode.net/ffmpeg-meeting on 2015-09-12, at 15 UTC.

Pre-scheduled topic

  1. ABI compatibility policy
  2. general policy decision process
  3. VDD15
  4. Outreachy funding for the next round (winter 2015)
  5. use of Github/Gitorious? for pull requests
  6. Any other business

Full meeting log

Sep 12 17:00:59 <saste>	allright time to go
Sep 12 17:01:18 <saste>	i summoned this meeting to discuss some relevant topics
Sep 12 17:01:27 *	Shiz (~shiz at hydrogenium.shiz.me) has joined #ffmpeg-meeting
Sep 12 17:01:49 <saste>	this is also meant as a sort of preparatory meeting for the real-life meeting which will be held in paris the next weekend
Sep 12 17:02:07 <saste>	you can see in the topics the pastebin containing the topics of the day to discuss
Sep 12 17:02:22 <saste>	http://pastebin.com/e6Q0pY6Z
Sep 12 17:02:29 <saste>	first topic is
Sep 12 17:02:41 <saste>	ABI compatibility policy
Sep 12 17:03:25 <saste>	please note that i just compiled the list of topics but i'm not very involved with ffmpeg development, so don't expect me to chat a lot about the merit of each topic
Sep 12 17:03:32 <BBB>	so are we going to just do a vote on that? or do you want to re-discuss it also?
Sep 12 17:03:50 <atomnuker>	wasn't wm4 the person who proposed it in the first place? where is he?
Sep 12 17:03:54 <saste>	I don't know, maybe someone can spend a few words describing the proposed options
Sep 12 17:04:02 <BBB>	(there was a long … “discussion” :) … on the mailinglist already on the abi compat subject, and it’s fair to say that we disagreed)
Sep 12 17:04:13 <nevcairiel>	well the options are pretty simple, do or don't
Sep 12 17:04:23 <saste>	then we can delay the voting to any other means, not necessarily we have to decide/vote right now
Sep 12 17:04:39 <Cigaes>	nevcairiel: indeed, that is the root alternative; there are a few subquestions after this is decided.
Sep 12 17:04:50 <ubitux>	atomnuker: wm4 kind of ragequited irc because it wasn't going fast enough apparently; try to /invite him
Sep 12 17:05:06 <nevcairiel>	do we spent effort to maintain the ABI compat, which in itself is not and cannot really be fully tested due to mis-matching behavior, or well, do we simply not
Sep 12 17:05:13 <atomnuker>	I'll ping him on ffmpeg-devel, that should get his attention
Sep 12 17:05:50 <ubitux>	we probably need to agree about how we "advertise" the policy
Sep 12 17:06:02 <ubitux>	be it a news, or a dedicated pages to "current goals"
Sep 12 17:06:18 <ubitux>	just to ease taking decisions and keeping up with them
Sep 12 17:06:23 <BBB>	so the discussion was more about how it was advertised or “how the patch was concealing its purpose”, right?
Sep 12 17:07:12 *	Easyfab (~chatzilla at ip-62-241-112-16.evc.net) has joined #ffmpeg-meeting
Sep 12 17:07:21 *	Easyfab (~chatzilla at ip-62-241-112-16.evc.net) has left #ffmpeg-meeting
Sep 12 17:07:23 <Cigaes>	BBB: from my PoV, right; I have mixed opinions on the decision itself.
Sep 12 17:07:31 <BBB>	who was actually advocating _for_ keeping the abi compat options (as opposed to the discussion around it being concealed)?
Sep 12 17:07:47 <nevcairiel>	I don't think anyone was directly for keeping it
Sep 12 17:07:48 *	Easyfab (~chatzilla at ip-62-241-112-16.evc.net) has joined #ffmpeg-meeting
Sep 12 17:08:18 <nevcairiel>	The discussion mostly went in circles around the policy issue
Sep 12 17:08:24 *	llogan2 (lou at 2600:3c01::f03c:91ff:fe70:2ed2) has joined #ffmpeg-meeting
Sep 12 17:08:50 <BBB>	right, I’m re-reading it now
Sep 12 17:09:14 <BBB>	carl eugen and nicolas george didn’t like us removing it without an explicit, dedicated discussion that allows us to decide whether we want to change policy on abi compat or not
Sep 12 17:09:17 <BBB>	so …
Sep 12 17:09:36 <BBB>	cehoyos isn’t here, is he?
Sep 12 17:09:45 <nevcairiel>	dont think so
Sep 12 17:09:49 <nevcairiel>	unless he uses a new name
Sep 12 17:10:08 *	jamrial (~jamrial at 181.22.42.38) has joined #ffmpeg-meeting
Sep 12 17:10:13 <BBB>	does nicolas do irc?
Sep 12 17:10:25 <nevcairiel>	thats Cigaes i thought
Sep 12 17:10:51 <BBB>	ah :) ok sorry didn’t know
Sep 12 17:11:01 <BBB>	all makes sense now
Sep 12 17:11:13 <saste>	so we're basically about topic #2, the policy decision process
Sep 12 17:11:23 <ubitux>	it's kind of related
Sep 12 17:11:28 <BBB>	one sort of morphed into the other… we can do #2 before we do #1
Sep 12 17:11:32 <BBB>	that might make more sense
Sep 12 17:11:37 <saste>	do you have technical arguments to discuss about the ABI policy to adopt
Sep 12 17:11:52 <saste>	or things which were not discussed on the ML
Sep 12 17:11:56 <nevcairiel>	unless someone wants to speak for keeping it?
Sep 12 17:12:46 <iive>	just a note. I think the best place to notify about ABI policy is somewhere in the changelog entries of 3.0 release.
Sep 12 17:13:02 <ubitux>	yeah, could be just that
Sep 12 17:13:06 <nevcairiel>	I don
Sep 12 17:13:18 <nevcairiel>	I don't think a  news entry or something is needed, changelog is mandatory of course
Sep 12 17:13:28 <nevcairiel>	the feature seems hardly used
Sep 12 17:13:34 <nevcairiel>	(if at all)
Sep 12 17:13:42 <ubitux>	we had a section for important behaviour changes in 2.6 
Sep 12 17:13:45 <ubitux>	iirc
Sep 12 17:13:50 <BBB>	APIChanges also exists
Sep 12 17:13:54 <jamrial>	what would happen with the sonames if we drop the ABI compatibility policy?
Sep 12 17:14:10 <saste>	nevcairiel, I don't think anybody is using it, but if they do they are probably very advanced FFmpeg users, and they probably are used to read the logs
Sep 12 17:14:18 <saste>	so a news entry could be useless in that case
Sep 12 17:14:29 *	yayoi (~sndh at 204.28.118.156) has joined #ffmpeg-meeting
Sep 12 17:15:04 <BBB>	jamrial: how did we fix that in the past?
Sep 12 17:15:08 <nevcairiel>	jamrial: we could bump them by 100 like nicolas suggested for all I care, just to make sure its clearly disjunct from libav
Sep 12 17:15:10 <BBB>	jamrial: is that why we have a 100 delta?
Sep 12 17:15:40 <Cigaes>	Distributing two incompatible libraries with the same SONAME is irresponsible.
Sep 12 17:15:45 <jamrial>	that's a configure option that i'm not sure anybody ever used, but yeah, can work as a solution
Sep 12 17:15:55 <Cigaes>	Bumping the major version once and for all is a simple solution.
Sep 12 17:16:12 <BBB>	so can we just bump to a different major than libav?
Sep 12 17:16:14 <Cigaes>	If we decide to drop ABI compatibility with libav, I would like to bring once again the suggestion of merging the libraries.
Sep 12 17:16:21 <BBB>	like, they use libavcodec.57, we use libavcodec.58?
Sep 12 17:16:30 <BBB>	(or the inverse, I don’t really care)
Sep 12 17:16:32 *	J_Darnley has quit (Ping timeout: 244 seconds)
Sep 12 17:16:33 <nevcairiel>	strictly speaking we already distribute incompatible libraries with the same soname
Sep 12 17:16:33 <Cigaes>	BBB: no, because they will eventually use 58.
Sep 12 17:16:47 <nevcairiel>	because noone uses the flag to enable the compat mode
Sep 12 17:17:03 <Cigaes>	nevcairiel: yes, but the ABI is mostly compatible even without it.
Sep 12 17:17:05 *	J_Darnley (~J_Darnley at d51A44418.access.telenet.be) has joined #ffmpeg-meeting
Sep 12 17:17:12 <nevcairiel>	"mostly" is irrelevant :)
Sep 12 17:17:21 <BBB>	we could just say “it’s not our problem” since ffmpeg isn’t the fork that caused there to be two identical sonames
Sep 12 17:17:24 <atomnuker>	Cigaes: IIRC there were plans to talk about that at VDD
Sep 12 17:17:30 <michaelni>	will we bump to 200 if theres another fork that doest bump ?
Sep 12 17:17:40 <jamrial>	nevcairiel: true, and in general people that use one project don't use the other
Sep 12 17:17:48 <ubitux>	note: we still need to keep the .100 micro as a mean to differentiate libraries, otherwise checks are going to be a pain for people trying to support post & past 3.0 + libav 
Sep 12 17:17:53 <BBB>	I think sonames conflict is something to be discussed with libav
Sep 12 17:18:05 <BBB>	so the most efficient way to deal with it may be at vdd, not here
Sep 12 17:18:13 <iive>	we can also rename the libraries
Sep 12 17:18:13 <BBB>	since “they” are not here to agree on a solution with us
Sep 12 17:18:19 <nevcairiel>	its not our responsibility to keep compatible with every single fork out there, if we decide now to bump to get the soname conflict out of the way, then it should be a one-time thing
Sep 12 17:18:45 <iive>	we can use the debian naming scheme by default.
Sep 12 17:19:05 <Cigaes>	nevcairiel: agreed.
Sep 12 17:20:33 <BBB>	saste: I propose we move to #2 and then revisit #1 afterwards if needed
Sep 12 17:20:41 <nevcairiel>	(a proper fork should rename their SONAME entirely, anyway)
Sep 12 17:20:47 <Cigaes>	BBB: seconded.
Sep 12 17:20:56 <jamrial>	i personally would prefer if we don't bump. as i said most distros care only about one of the two projects at a time, and those that use both afaik don't ship binaries
Sep 12 17:20:58 <atomnuker>	nevcairiel: I agree as well
Sep 12 17:21:03 <saste>	BBB: I agree
Sep 12 17:21:16 <Cigaes>	jamrial: distros are not the only distribution channel.
Sep 12 17:21:20 <saste>	if there are no objections we are moving to point #2
Sep 12 17:21:32 <nevcairiel>	jamrial: or they ship renamed binaries, ie using the --build-suffix option
Sep 12 17:22:56 <BBB>	ok, so decision making process it is then… I guess michaelni should be given some time to give opinion as “old boss” here?
Sep 12 17:23:07 <BBB>	you’re very quiet michaelni 
Sep 12 17:23:15 <jamrial>	Cigaes: true, but where else does it really matter? afaik all these abi considerations were put in place specifically because of distros
Sep 12 17:23:26 <michaelni>	BBB, ive nothing to say :)
Sep 12 17:23:48 <Cigaes>	jamrial: not only. If someone does "./configure --enable-shared && make install", it should not break their system either.*
Sep 12 17:24:10 <saste>	it looks to me the most controversial point is having/not having vetoes
Sep 12 17:24:47 <BBB>	well there’s also the more general “when consensus cannot be reached, now what?”
Sep 12 17:24:55 <saste>	with one formal or de-facto leader it was relatively easy to set controversies allowing the leader to decide
Sep 12 17:24:56 <BBB>	vetoes are just one part of that question
Sep 12 17:25:36 <llogan2>	how do other big projects deal with the situation?
Sep 12 17:25:52 <saste>	anybody is for or against a committe or something?
Sep 12 17:25:58 <saste>	no more than three people
Sep 12 17:26:00 <BBB>	some have a bofh
Sep 12 17:26:16 <BBB>	others have a committee (although that didn’t go well for xfree86)
Sep 12 17:26:18 <Cigaes>	saste: indeed. That is the reason I believe a leader is needed (or a leading committee). But it does not need to be the person who does all the heavy work.
Sep 12 17:26:19 <nevcairiel>	llogan2:  They usually tend to have a subsystem maintainer or an overall leader to decide, from what I hear
Sep 12 17:26:20 <saste>	that's a form of leadership of course, regarding the overall project design
Sep 12 17:26:52 <BBB>	committee is good, but membership needs to be rotational, in the sense that you’re not a member for life, and it’s not up to you to relinquish your membership
Sep 12 17:26:56 <BBB>	(that leads to xfree86 situations)
Sep 12 17:27:00 <jamrial>	we could also have different people for different parts of the project. maintainers if you will
Sep 12 17:27:16 <BBB>	jamrial: well, we already have that, this is more about global decisions
Sep 12 17:27:20 <saste>	jamrial, the problem is about deciding overall design, like the ABI things
Sep 12 17:27:21 <BBB>	or are you advocating global maintainers?
Sep 12 17:27:23 *	reynaldo (~rverdejo at c-67-169-80-122.hsd1.ca.comcast.net) has joined #ffmpeg-meeting
Sep 12 17:27:28 <BBB>	(like a bofh :) )
Sep 12 17:27:36 <saste>	local maintainer always worked pretty well with FFmpeg, I think
Sep 12 17:27:46 <nevcairiel>	We also have parts of the code base that is very generic code, and not maintained by a single person
Sep 12 17:27:58 <reynaldo>	Sat 8:30AM here, thanks and hello o/ :)
Sep 12 17:28:02 <Cigaes>	BBB: the leader need to be accepted by the other developers, that is the crux of the issue (and what failed in 2011).
Sep 12 17:28:11 <BBB>	right
Sep 12 17:28:15 <atomnuker>	the MAINTAINERS file sometimes list people from libav as well
Sep 12 17:28:48 <saste>	atomnuker, cleaning MAINTAINERS is from hard to impossible, with no explicit reply from old maintainers
Sep 12 17:28:48 *	lglinskih has quit (Quit: This computer has gone to sleep)
Sep 12 17:28:54 <llogan2>	reynaldo: im to your timezone on the left
Sep 12 17:29:21 <ubitux>	what were the last things the project had problems to make up its mind?
Sep 12 17:29:31 <nevcairiel>	ubitux: pkg-config :D
Sep 12 17:29:34 <ubitux>	i remember pkg-config... and this recent abi thing, but what else?
Sep 12 17:29:38 <saste>	or better, the only way would be to ask active maintainers to confirm their will to maintain those parts of the code
Sep 12 17:29:45 <ubitux>	nevcairiel: yeah right, but is there anything else in these last 4 years?
Sep 12 17:29:55 <jamrial>	BBB: maybe something like said committe, but one for different part of the code
Sep 12 17:30:05 <llogan2>	saste: a grat purge of MAINTAINERS?
Sep 12 17:30:05 <Cigaes>	ubitux: subtitles character encoding API.
Sep 12 17:30:17 <reynaldo>	llogan2: !
Sep 12 17:30:19 <ubitux>	actually, same question about similar problems libav had without leader to take a decision
Sep 12 17:30:37 <BBB>	libav tried to find common agreement
Sep 12 17:30:38 <nevcairiel>	libav didnt really have those discussions
Sep 12 17:30:51 <iive>	they probably had the on the phone
Sep 12 17:30:52 <BBB>	and if that didn’t happen, the developer typiclly fell off the boat
Sep 12 17:30:56 <nevcairiel>	when someone objects, its usually on a technical level
Sep 12 17:31:06 <nevcairiel>	or yeah, the developer went away
Sep 12 17:31:11 <nevcairiel>	like mru
Sep 12 17:31:20 <BBB>	or me :)
Sep 12 17:31:34 <Cigaes>	nevcairiel: you know for a fact that mru went away because of disagreements?
Sep 12 17:31:50 <ubitux>	loosing a developer everytime there is a disagreement is kind of an expensive cost
Sep 12 17:32:04 <nevcairiel>	i couldnt attest 100% to it, but i think diego's refactoring of some build system things pissed him off eventually because he disagreed
Sep 12 17:32:58 <saste>	in case we want a committe, we need some metrics to decide if a developer can have voting rights
Sep 12 17:33:07 <Cigaes>	Since mru was one of the most prominent devlopers on the side of the fork at the time, I take it as a sign that leaderless does not work.
Sep 12 17:33:11 <llogan2>	BBB: ...rotational makes sense to me.
Sep 12 17:33:19 <Cigaes>	saste: that is the big problem indeed.
Sep 12 17:33:55 <BBB>	I think it’s a sign that leaderless without a disagreement resolution mechanism does not work
Sep 12 17:34:03 <nevcairiel>	Cigaes: these days their active core is so small that everyone just does their own thing and the others dont really mind
Sep 12 17:34:16 <Cigaes>	llogan2/BBB: as long as the leader/committee is willing and the other developers are satisfied with him, there is no need to force a rotation.
Sep 12 17:34:28 <BBB>	that’s true
Sep 12 17:34:32 *	J_Darnley has quit (Ping timeout: 250 seconds)
Sep 12 17:34:39 <BBB>	a rotation can mean you stay if others are ok with it
Sep 12 17:34:42 <llogan2>	but what if others want to participate?
Sep 12 17:34:56 *	J_Darnley (~J_Darnley at d51a44418.access.telenet.be) has joined #ffmpeg-meeting
Sep 12 17:35:12 <nevcairiel>	then the other developers are clearly not satisfied anymore, and the condition doesnt apply .)
Sep 12 17:35:39 <Cigaes>	llogan2: then said others are not satisfied, and indeed, they should have a chance.
Sep 12 17:36:06 <reynaldo>	so we decide on a leader + voters(developers) setup already. Missed a few minutes, sorry. just trying to understand where we are at
Sep 12 17:36:17 <reynaldo>	the first one was a question ^ ?
Sep 12 17:36:44 <BBB>	reynaldo: no, nothing was decided yet
Sep 12 17:36:51 <reynaldo>	BBB ok, thanks
Sep 12 17:37:14 <atomnuker>	I think there should be someone like a leader who only steps in when two developers are at an absolute impasse, and have argued for at least a few days
Sep 12 17:37:30 <nevcairiel>	you mean a judge
Sep 12 17:37:42 <atomnuker>	yes, or a voting process like reynaldo said
Sep 12 17:38:05 <reynaldo>	I'd have both
Sep 12 17:38:07 <nevcairiel>	or mediator or arbitrator if those are better words
Sep 12 17:38:20 <Cigaes>	atomnuker: who steps in when developers call upon him/her.
Sep 12 17:38:25 <llogan2>	media-tor
Sep 12 17:38:28 <reynaldo>	Cigaes: the voters
Sep 12 17:38:37 <atomnuker>	but the key point is not to intervene too early or too late
Sep 12 17:39:14 <reynaldo>	and just in case my position is not clear enough, I think our community *needs* this kind of strict setup, we are wild as that
Sep 12 17:39:19 <Cigaes>	reynaldo: that was not what I meant. I mean: A proposes a patch, B disagrees stubbornly, A calls upon Leader.
Sep 12 17:39:32 <reynaldo>	Cigaes: Im thinking beyond patch disagreements
Sep 12 17:39:41 <saste>	I think a committe of ideally three developers would be fine, so that who settle the controversy doesn't attract all the hate
Sep 12 17:39:48 <reynaldo>	Im thinking on the same structure handling every disagreement
Sep 12 17:39:52 <saste>	that what happens when you have a single one developer
Sep 12 17:39:59 <Cigaes>	Of course, for important questions, the decision must be from all developers, not just leader/committee.
Sep 12 17:40:03 <saste>	but then the more you have to decide, the slower the process is
Sep 12 17:40:07 <BBB>	I am happy with saste’s 3-developer committee thing
Sep 12 17:40:07 <Cigaes>	If only: nominating the leader/committee.
Sep 12 17:40:49 <BBB>	also, they don’t have to be the activest of developers; rather, they need to be respected across sections of the community
Sep 12 17:40:50 <reynaldo>	saste, BBB that's 3 devs & no single leader ?
Sep 12 17:40:52 <Cigaes>	We can vote on the number of members in the leading committee the first time we vote.
Sep 12 17:40:57 <llogan2>	saste: sounds good to me
Sep 12 17:41:15 <BBB>	reynaldo: yeah. although leader seemed to work fine also
Sep 12 17:41:52 <llogan2>	Benevolent Triumvorate for Life
Sep 12 17:42:19 <saste>	BBB: and they need to be able to judge about the technical merits of the decisions to take
Sep 12 17:42:20 <iive>	yes, fixed structure tend to accumulate power with time.
Sep 12 17:42:34 <saste>	then we would need a rotation mechanism
Sep 12 17:42:49 <BBB>	right, that’s why I mentioned rotation mechanism
Sep 12 17:42:58 <ubitux>	what if the leader & 1 person agree against 10 persons?
Sep 12 17:43:13 <nevcairiel>	then you didnt pick a good leader
Sep 12 17:43:14 <iive>	there is another solution. If there is dead-lock, we can pick the solution randomly
Sep 12 17:43:16 <ubitux>	10 persons being more or less active developers
Sep 12 17:43:23 <reynaldo>	saste, BBB so if desicion X is so complex that it warrants one member of our commitee to vote "whatever", then a decision might actually never be reached?
Sep 12 17:43:28 <Cigaes>	ubitux: then why did they chose that leader?
Sep 12 17:43:29 <atomnuker>	ubitux: down with the leader
Sep 12 17:43:40 <reynaldo>	this is why I think we should have 1 leader + small commitee ^
Sep 12 17:43:52 <Cigaes>	reynaldo: if leaders are unsure, they can ask for counsel.
Sep 12 17:43:53 <saste>	i'm not even sure about the leader
Sep 12 17:43:54 <reynaldo>	so leader & commitee can oversee each other and step in if needed
Sep 12 17:44:00 <saste>	probably a committe is better
Sep 12 17:44:09 <saste>	so we avoid conflict between leader and committe
Sep 12 17:44:13 <llogan2>	i odn't think anyone wants to be "the leader"
Sep 12 17:44:45 <Cigaes>	saste: I insist: we can decide formally on leader/committee when we have a voting process. We can even change each time the committee is rotated.
Sep 12 17:44:49 <jamrial>	ubitux: have a voting first, if those 10 people are the majority, leader/committee doesn't intervene?
Sep 12 17:45:10 <ubitux>	ok ok
Sep 12 17:45:23 <iive>	Is there another solution that doesn't involve Politics?
Sep 12 17:45:35 <llogan2>	armwrestling
Sep 12 17:45:36 <reynaldo>	jamrial: not only that Im afraid, in my experience there are matters that might only be discussed among the commitee and/or leader
Sep 12 17:45:45 <reynaldo>	this is rather important guys ^^
Sep 12 17:45:53 <iive>	llogan2: I was thinking of lottery, but that's find too :)
Sep 12 17:45:55 <reynaldo>	and something that needs to be considered
Sep 12 17:46:09 <atomnuker>	iive: what we currently do: call a meeting once conflicts happen
Sep 12 17:46:30 <saste>	atomnuker, the truth is that the meeting doesn't resolve things, most of the time
Sep 12 17:46:43 <saste>	unless you give voting power to the attendees
Sep 12 17:46:52 <saste>	in that case you need to define who are the attendees
Sep 12 17:47:00 <saste>	this applies in case there are controversies
Sep 12 17:47:15 <saste>	note that in most cases decisions are resolved with no conflict
Sep 12 17:47:16 <atomnuker>	saste: they should mostly resolve themselves given time
Sep 12 17:47:26 <saste>	that's more than 99% of the issues
Sep 12 17:47:45 <saste>	the decision process is about that 1% of issues which can't be settled with the "normal" means
Sep 12 17:47:58 <Cigaes>	saste: hear, hear. Let us focus on that please.
Sep 12 17:48:03 <BBB>	++
Sep 12 17:48:58 <saste>	can we settle reasonable criteria for selecting voters?
Sep 12 17:49:15 <reynaldo>	fwiw I think the voters right should be granted after X commits and lost after a _large_ period of inactivity
Sep 12 17:49:29 <reynaldo>	plain criteria, easy to follow
Sep 12 17:49:35 <llogan2>	why inactivity?
Sep 12 17:49:39 <saste>	that's the whole issue, and one of the reasons of the fork, since we couldn't agree about the validity of a votation
Sep 12 17:50:11 <reynaldo>	llogan2: thats my take, lets hear others and then discuss on their details
Sep 12 17:50:17 <BBB>	llogan2: my term “mplungarians” is only partially derogatory
Sep 12 17:50:18 <Cigaes>	reynaldo: for the initial voting rights, that is of course a must. For later evolutions, co-optation by vote is possible too.
Sep 12 17:50:20 <nevcairiel>	llogan2: iif someone wasnt involved with the development for a sufficiently long time, he may not really know the current state of the project sufficiently
Sep 12 17:50:31 <reynaldo>	guys, can you propose your formulas and then we can discuss
Sep 12 17:50:34 <reynaldo>	lets focuss please
Sep 12 17:50:45 <reynaldo>	otherwise this will just drag on and on
Sep 12 17:50:58 <BBB>	I support reynaldo’s notation
Sep 12 17:51:36 <reynaldo>	thanks. do we have any other alternative? guys? anything else to suggest ?
Sep 12 17:51:45 <BBB>	as for large period, I’d say 1 or 2 years with no commits is sufficiently large, but I’m open to other ideas
Sep 12 17:51:50 <reynaldo>	if not, we can discuss on the details on my proposal and settle on something
Sep 12 17:52:02 <BBB>	but yeah we can discuss details on ML
Sep 12 17:52:05 <Cigaes>	BBB: discussing who loses voting rights can wait.
Sep 12 17:52:08 <BBB>	and then let’s move on to #3
Sep 12 17:52:16 <saste>	allright
Sep 12 17:52:18 <llogan2>	criteria: someone who is active, someone who wants to be a voter, and for the first "triumvorate" someone who has been around for a "while"
Sep 12 17:52:19 <reynaldo>	no, Id rather do it here than to the mailing list
Sep 12 17:52:31 <reynaldo>	this is the single most important desicion we should make
Sep 12 17:52:37 <BBB>	ok
Sep 12 17:52:57 <saste>	llogan2, sounds good
Sep 12 17:53:20 <nevcairiel>	llogan2: define active
Sep 12 17:53:32 <reynaldo>	so, inactivity: 2 years / number of commits to reach voting rights: 50 in one year
Sep 12 17:53:41 <reynaldo>	sounds like something you'd be able to agree on ^ ?
Sep 12 17:53:54 <BBB>	ok
Sep 12 17:54:05 <llogan2>	nevcairiel: at least some sort of activity within the last 6 months? shows that they have an interest in the project.
Sep 12 17:54:14 <Cigaes>	I think we can not decide on a criterion without having a few stats.
Sep 12 17:54:19 <reynaldo>	sorry, mean to say "50" not, "50" in one year
Sep 12 17:54:31 <ubitux>	anyone to share a git command to raise those names?
Sep 12 17:54:53 <ubitux>	(so we can evaluate how much people are involved and the concerned ppl know about that)
Sep 12 17:55:05 <reynaldo>	my bet its it will come out to ~20ppl or maybe less
Sep 12 17:55:08 <reynaldo>	a maneagable set
Sep 12 17:55:15 <llogan2>	my beef with inactive is that they are just not interested in the project, nor would they be informed about the detailes of the decision to be made
Sep 12 17:55:21 <nevcairiel>	ubitux: git shortlog -s -n
Sep 12 17:55:36 <reynaldo>	llogan2: not every single time, sometimes life just takes over but you remain lurkin around
Sep 12 17:55:41 <reynaldo>	just not actively contributing
Sep 12 17:56:10 <ubitux>	nevcairiel: now we need the 2-year parameters
Sep 12 17:56:18 <ubitux>	nevcairiel: do we invite libav ppl to vote too? :)
Sep 12 17:56:22 <reynaldo>	happens when ppl change jobs, have kids, get sick, you name it
Sep 12 17:56:22 <jamrial>	ubitux: git shortlog -ns --no-merges n2.5..n2.8 maybe
Sep 12 17:56:31 <reynaldo>	libav has nothing to do with this unless they are "dual"
Sep 12 17:56:48 <ubitux>	jamrial: release scoped then, not time scoped?
Sep 12 17:56:48 <llogan2>	reynaldo: but that means they are now too bust to deal with FFdecisions
Sep 12 17:56:48 <Cigaes>	reynaldo: in that case, maybe it is normal they do not vote.
Sep 12 17:56:48 <saste>	reynaldo, commit criteria are also unjust towards people doing other kind of services (helping users or with the administration), but we need to converge towards some (somehow arbitrary) criteria
Sep 12 17:56:53 <llogan2>	*busy
Sep 12 17:57:01 <jamrial>	use commits instead of tags then
Sep 12 17:57:08 <nevcairiel>	ubitux: find a rev in time thats 2 years ago then, dont think it gets any easier
Sep 12 17:57:21 <Cigaes>	OTHO, someone reviewing patches on the ML without producing them should have vote right too.
Sep 12 17:57:30 <reynaldo>	saste: I understand that, and as I plan to give the voters every single right I think they should be the ones choosing how to deal with that and other issues
Sep 12 17:57:39 <ubitux>	(nevcairiel: pretty sure git has a time parser for that kind of stuff)
Sep 12 17:58:22 <reynaldo>	so, pending confirmation, we are at 50 commits and activity in the last 2 years, maybe we can settle in 1 year of inactivy llogan2 ?
Sep 12 17:58:29 <Cigaes>	I propose this: we decide on an objective criterion for the initial set, then the selected people coopt worthy people who were left over, if any.
Sep 12 17:58:44 <reynaldo>	we need to discuss "reentry" how can you get back to voting after a period of inactivty
Sep 12 17:58:56 <BBB>	reynaldo: same as initial entry?
Sep 12 17:59:00 <saste>	anybody against the reynaldo commit criteria about what an active developer is?
Sep 12 17:59:05 <reynaldo>	Cigaes: yeah, they can decide that and pretty much anything for that matter
Sep 12 17:59:09 <BBB>	saste: sgtm
Sep 12 17:59:11 <reynaldo>	BBB: good
Sep 12 17:59:14 <saste>	are we also going to count merged commits?
Sep 12 17:59:15 <reynaldo>	maybe half that ?
Sep 12 17:59:19 <reynaldo>	BBB ^
Sep 12 17:59:37 <Cigaes>	saste: if we consider reuniting with libav, then libav guys must have voting right.
Sep 12 17:59:44 <iive>	well, whoever wants to vote, could do 50 K&R formatting commits and be done :D
Sep 12 17:59:50 <saste>	Cigaes, I'm fine with that
Sep 12 17:59:54 <reynaldo>	I dont want to miss the vote from someone with community experience out of waiting for him to get back at full speed
Sep 12 18:00:10 <durandal_1707>	Lol
Sep 12 18:00:11 <reynaldo>	Cigaes: reuninting with libav is not something that has to be decided now
Sep 12 18:00:22 <reynaldo>	thats why we have voters
Sep 12 18:00:29 *	lglinskih (~lglinskih at verda.kir.elagin.me) has joined #ffmpeg-meeting
Sep 12 18:01:05 <llogan2>	i dont think anyone who has been inactive for x years is going to give a shit about spending time making votes
Sep 12 18:01:19 <saste>	llogan2, indeed
Sep 12 18:01:25 <reynaldo>	lets settle on a 1y inactivity ?
Sep 12 18:01:30 <reynaldo>	saste: ^
Sep 12 18:01:33 <llogan2>	"hey, babtiste, you're a voter"
Sep 12 18:01:48 <reynaldo>	like, a kid, a sickness, etc
Sep 12 18:02:00 <ubitux>	nevcairiel: git shortlog -ns --since '{2 years ago}'
Sep 12 18:02:19 <jamrial>	nice
Sep 12 18:02:26 <llogan2>	reynaldo: perhaps i do not understand your critera.
Sep 12 18:02:27 <saste>	so we have this criterion: a developer is considered active if in the last year he has at least 50 committs (merged committs are good as well)
Sep 12 18:02:53 <ubitux>	should we target a number of commits or just the most actives ?
Sep 12 18:03:13 <reynaldo>	llogan2: 50 commits in a 1 year window to become a voter, 1 year of inactivty to loose your right, 25 commits in one year to regain it
Sep 12 18:03:14 <jamrial>	saste: we have developers that don't have nearly as much as that but are active in the ml
Sep 12 18:03:26 <ubitux>	like, git shortlog -ns --since '{2 years ago}' | head -n20 maybe
Sep 12 18:03:40 <ubitux>	this looks like a good list to me ^
Sep 12 18:03:48 <saste>	jamrial, I think we can agree that any criteria will be flawed in some way
Sep 12 18:03:53 <ubitux>	or maybe 25
Sep 12 18:04:00 <reynaldo>	jamrial: activity on the mailing list, incredible as it sounds, has been a problem some times :)
Sep 12 18:04:04 <Cigaes>	reynaldo: do you insist to have the same criterion for the initial list and later proceedings?
Sep 12 18:04:09 <reynaldo>	I think we can let the voters decide on such issues
Sep 12 18:04:09 <llogan2>	reynaldo: yes, i was wrong about what you wanted. i thought you required that the voter must be inactive for x years to be "impartial". 
Sep 12 18:04:17 <saste>	we just need an objective metrics, then the new voters can fine tune the system
Sep 12 18:04:21 <ubitux>	30 looks fine too, it reaches the 50 commits one
Sep 12 18:04:25 <reynaldo>	Cigaes: I'd let the rules to be stabilized by the voters
Sep 12 18:04:29 <reynaldo>	so lets make them initial
Sep 12 18:04:30 <jamrial>	saste: true, which is why we should think how to include said people
Sep 12 18:04:34 <reynaldo>	and see where it leads us
Sep 12 18:04:36 <Cigaes>	reynaldo: then let us only discuss the initial criterion.
Sep 12 18:04:52 <ubitux>	saste: git shortlog -ns --since '{2 years ago}' | head -n30 sounds like a fine list to me currently
Sep 12 18:04:56 <reynaldo>	yes, I had the impression that's what we were dealing with
Sep 12 18:05:35 <jamrial>	example, rcombs has been very active this year, but he has 48 commits
Sep 12 18:05:36 <Cigaes>	reynaldo: you wrote about "loose your right" and "regain it"
Sep 12 18:05:48 <Cigaes>	That is not relevant for the immediate vote.
Sep 12 18:06:02 <reynaldo>	oh, point there
Sep 12 18:06:11 <iive>	i have other proposla. let make a wide commitie from the people who are top committer in the last year. e.g. top 15 commiters ?
Sep 12 18:06:11 <jamrial>	(why am i sixth on that list, wth)
Sep 12 18:06:16 <reynaldo>	well, consider those guideline suggestions :)
Sep 12 18:06:25 <ubitux>	jamrial: redundant merges probably
Sep 12 18:06:47 <ubitux>	i shouldn't be counted though
Sep 12 18:06:57 <nevcairiel>	pretty hard to filter out the dupes
Sep 12 18:07:07 <nevcairiel>	unless you directly subtract the same list from libav
Sep 12 18:07:09 <BBB>	I have a secret way
Sep 12 18:07:14 <BBB>	right
Sep 12 18:07:14 <reynaldo>	iive: top commiters can change by the minute, you need something more stable
Sep 12 18:07:17 <ubitux>	i actually don't see dups
Sep 12 18:07:41 <nevcairiel>	send a patch to both libav and ffmpeg, and you get twice the commit count, is all i mean :D
Sep 12 18:07:52 <iive>	reynaldo: not with 15 people.
Sep 12 18:08:01 <Cigaes>	reynaldo: "top committers at $date".
Sep 12 18:08:33 <reynaldo>	I'd rather go with the voting right after X commit formula I proposeed, maybe we can assemble both lists and compare?
Sep 12 18:08:36 <reynaldo>	we are humas after all
Sep 12 18:08:42 <reynaldo>	one will make more sense, Im sure
Sep 12 18:08:43 <reynaldo>	;)
Sep 12 18:08:51 <iive>	1. giving and taking voting rights gives bad vibe...
Sep 12 18:08:52 <Cigaes>	reynaldo: for a punctual choice, they are equivalent.
Sep 12 18:09:02 <reynaldo>	I understand at least BBB and saste already back that formula up ^
Sep 12 18:09:08 <iive>	2. having a lot of people arguing is bad.
Sep 12 18:10:06 <iive>	I think that Prakinson gives 12 as maximum of the people who can work productively in a group.
Sep 12 18:10:25 <jamrial>	i back it as long as there's some way to include people that don't fit into it but are still noticeably active in the project
Sep 12 18:10:34 <ubitux>	iive: most people won't vote
Sep 12 18:10:52 <BBB>	jamrial: you mean as a later extension to the formula?
Sep 12 18:10:53 <ubitux>	when there is an issue like pkg-config, at most 10 ppl will get involved
Sep 12 18:10:55 <reynaldo>	iive: have you heard of any community following this approach succesfully? I for ine dislike it for being too dynamic, you need stability in power to be able to make desicions that might spawn across several voting sessions
Sep 12 18:10:57 <iive>	most people don't want to get involved in politics
Sep 12 18:10:57 <ubitux>	most of the others don't care
Sep 12 18:11:00 <jamrial>	rcombs and kierank would be left out as is, and i consider them active
Sep 12 18:11:01 <reynaldo>	long term planing and what not
Sep 12 18:11:16 <BBB>	jamrial: llogan would be out also (he’s 40, not 50)
Sep 12 18:11:21 <ubitux>	jamrial: what if we reduce to 1 year?
Sep 12 18:11:34 <durandal_1707>	let just vote for pkgconfig thing
Sep 12 18:11:38 <jamrial>	and so would saste
Sep 12 18:12:05 <reynaldo>	the voters commitee can garnt voting rights to ppl not initially selected, thats their prerogative
Sep 12 18:12:05 <Cigaes>	Me too.
Sep 12 18:12:12 <jamrial>	ubitux: i'm checking one year right now. rcomns, kierank, llogan and saste don't have 50 commits
Sep 12 18:12:17 <saste>	jamrial, personally, I don't care as long as we have a bootstrap criteria
Sep 12 18:12:20 <iive>	yes, commit number is bad measure too. because one might have 100 1liners and another can have 2x 1000lines
Sep 12 18:12:35 <llogan2>	this is hard.
Sep 12 18:12:36 <Cigaes>	saste: idem for me.
Sep 12 18:12:41 <saste>	also, I can send some typo-fixing patches if that matters ;-)
Sep 12 18:12:50 <reynaldo>	ok. I have a proposal
Sep 12 18:13:04 <reynaldo>	the one I descrived + include some ppl that should be left out in the initial
Sep 12 18:13:08 <reynaldo>	list
Sep 12 18:13:12 <reynaldo>	talking about saste, llogan, etc
Sep 12 18:13:43 <jamrial>	so i agree to use the 50 votes formula, but we should also use a different criteria for people active in the ml and/or irc (reviewing patches, being part of discussions, etc) that don't meet the commit criteria
Sep 12 18:13:44 <ubitux>	iive: but 100 one liners means the developer is kind of involved in many parts of the project, while a 2x 1000 lines is most likely someone very much scope to a very small area where he doesn't have much opinion on the overall project policy as long as it doesn't involve his pet
Sep 12 18:14:03 <ubitux>	(sorry long run sentence)
Sep 12 18:14:53 <saste>	llogan2, in your case we should probably count also ffmpeg-web committs
Sep 12 18:15:01 <BBB>	so reynaldo’s proposal and have these people include additionals by majority vote?
Sep 12 18:15:05 <BBB>	saste: indeed
Sep 12 18:15:14 <reynaldo>	BBB: yes, I like that
Sep 12 18:15:20 <saste>	anyway, I'm really fine with any objective criterion, then it can be fine-tuned later
Sep 12 18:15:26 <reynaldo>	great
Sep 12 18:15:29 <nevcairiel>	(if someone cares, this is a list from the last year with commits merged from libav excluded, since those really can't be counted "active in ffmpeg" right now http://pastebin.com/kCQc55jb)
Sep 12 18:15:46 <reynaldo>	do we need to vote this or can we call it settled and move on?
Sep 12 18:15:59 <Cigaes>	reynaldo: if "include some ppl" is done by vote by the initial selection, then it is mostly what I argue for too, so +1.
Sep 12 18:16:16 <ubitux>	nevcairiel: so 20 first without libav looks sane?
Sep 12 18:16:28 <reynaldo>	Cigaes: thanks
Sep 12 18:16:42 <BBB>	so can we take that as decision then?
Sep 12 18:16:46 <reynaldo>	looks sane to me, missing some indeed but that can be sort out latter
Sep 12 18:16:46 <nevcairiel>	ubitux: sure, even includes saste again =p
Sep 12 18:16:47 <BBB>	who objects?
Sep 12 18:16:48 <saste>	what if we include libav
Sep 12 18:16:59 <reynaldo>	what list is that one btw?
Sep 12 18:17:17 <saste>	I don't want to make arbitrary distinctions, a committ is a committ
Sep 12 18:17:20 <reynaldo>	the one comming from my proposal ?
Sep 12 18:17:22 <Cigaes>	saste: if we merge, we have to redo the whole decision process with them anyway.
Sep 12 18:17:25 <reynaldo>	saste: +1
Sep 12 18:17:32 <saste>	then if they don't care they won't vote
Sep 12 18:17:37 <Cigaes>	But it will help to have a united proposition.
Sep 12 18:17:46 <nevcairiel>	saste: people contributing to libav dont seem to have an interest in voting on ffmpeg decisions
Sep 12 18:18:01 <nevcairiel>	I would prefer getting people actually itnerested in ffmpeg
Sep 12 18:18:21 <nevcairiel>	the only people dropping out due to that are the main libav contributors anyway
Sep 12 18:18:44 <ubitux>	so 4-5 more people to add?
Sep 12 18:18:47 <reynaldo>	ok, just to confirm: we agreed on 50 commits granting you voting right and 1 year of inactivity making you loose it.
Sep 12 18:19:04 <Cigaes>	reynaldo: stop your sentence before "and".
Sep 12 18:19:23 <reynaldo>	and this WONT be discusses on the mailing list. Most ppl in this meetint care enough about the process to trust their judgement for the initial list
Sep 12 18:19:31 <ubitux>	Cigaes: we really need a time window
Sep 12 18:19:43 <reynaldo>	Cigaes: trying to cut the list a bit shorter, otherwise first council* coordination will be a pain
Sep 12 18:19:57 <reynaldo>	ubitux: +1
Sep 12 18:20:08 <Cigaes>	reynaldo / ubitux: we do not need a criterion to LOSE voting rights for the initial list.
Sep 12 18:20:30 <reynaldo>	we need, otheriwse it will be a huge and not really representative list
Sep 12 18:20:38 <ubitux>	yeah it's not about loosing but accounting in a given time window (--since '{X months ago}' or whatever)
Sep 12 18:20:49 <reynaldo>	ubitux: +1
Sep 12 18:20:55 <reynaldo>	a year is something I feel ok with
Sep 12 18:21:09 <reynaldo>	and matches our idea of regain* window latter on, not that it matters that much, just saying
Sep 12 18:21:14 <reynaldo>	seems to make sense
Sep 12 18:21:15 <Cigaes>	reynaldo: then your sentence was not correctly stated.
Sep 12 18:21:21 <Cigaes>	Let me try to rephrase:
Sep 12 18:21:29 <reynaldo>	Cigaes: might be, I suck at english too early on saturdays
Sep 12 18:21:31 <reynaldo>	:)
Sep 12 18:21:33 <iive>	Cigaes: you want it as "50 commits for 1 year granting you voting right"
Sep 12 18:22:06 <Cigaes>	iive: that was what I was going to write.
Sep 12 18:22:37 <reynaldo>	year being 2015 for the first list ot the last 365 days counting back from today ?
Sep 12 18:22:43 <reynaldo>	ot/or ?
Sep 12 18:22:49 <Cigaes>	But maybe "50 commits in total including X% in the last year" could be slightly better.
Sep 12 18:23:07 <reynaldo>	please dont go beyond rephrasing, we already agreed on a formula
Sep 12 18:23:27 <ubitux>	i propose to use a git command as formula 
Sep 12 18:23:35 <Cigaes>	reynaldo: that would make 2014-09-12T15:00:00Z - 2015-09-12T15:00:00Z (let us take the start of the meeting as the reference point).
Sep 12 18:23:35 <iive>	365(+1) days on the day voting is called.
Sep 12 18:23:50 <Cigaes>	iive: dangerous.
Sep 12 18:24:00 <iive>	politics is dangerous
Sep 12 18:24:01 <Cigaes>	iive: once the criterium is stated, it can be gamed.
Sep 12 18:24:14 <iive>	that's why I'm serious about the lottery.
Sep 12 18:24:47 <saste>	Cigaes, a bad criterion => still better than random choice
Sep 12 18:25:03 <Cigaes>	saste: of course.
Sep 12 18:25:11 <BBB>	so do we agree on this now?
Sep 12 18:25:17 <BBB>	I’d like to move to #3 at some point :)
Sep 12 18:25:21 <Cigaes>	saste: but if only past commits are taken into account, it can not be gamed.
Sep 12 18:25:21 <BBB>	(my family wants lunch)
Sep 12 18:25:30 *	cehoyos (~cehoyos at 80-110-108-0.cgn.dynamic.surfer.at) has joined #ffmpeg-meeting
Sep 12 18:25:42 <Cigaes>	So: 50 comits at the time of the start of the meeting rather than at the time of the vote.
Sep 12 18:25:43 <cehoyos>	Hi!
Sep 12 18:26:20 <reynaldo>	ook, rephrasing: List of initial voters is being decided today, we are giving them rights over whatever they choose to handle and the first group will be the people that have more than 49 commits during the last 365 days
Sep 12 18:26:27 <reynaldo>	please move on ok ?
Sep 12 18:26:52 <saste>	reynaldo, ok
Sep 12 18:26:52 <cehoyos>	Sorrry for the late question:
Sep 12 18:27:05 <cehoyos>	What about merge commits (Just to make sure there are no misunderstandings)?
Sep 12 18:27:10 <Cigaes>	reynaldo: "during the 365 days before the beginning of the meeting"?
Sep 12 18:27:21 <BBB>	cehoyos: no merge commits, but merged commits are fine
Sep 12 18:27:28 <cehoyos>	Ty
Sep 12 18:27:30 <nevcairiel>	BBB: merging is hard work :(
Sep 12 18:27:31 <reynaldo>	Cigaes: yes, yesterday was day 1
Sep 12 18:27:41 <BBB>	nevcairiel: that’s true …
Sep 12 18:27:47 <Cigaes>	cehoyos: only Hendrik and Michael have a lot of those, so it does not matter.
Sep 12 18:27:55 <reynaldo>	Cigaes: UTC/GMT as reference TZ
Sep 12 18:28:00 <BBB>	 219  Hendrik Leppkes
Sep 12 18:28:01 <reynaldo>	Im guessing thats well defined enough
Sep 12 18:28:03 <BBB>	nevcairiel: you’re safe :)
Sep 12 18:28:29 <saste>	who can post the list somewhere, showing the git command?
Sep 12 18:28:31 <nevcairiel>	and I would argue that commits merged from libav shouldn't be counted, as that (1) gives several people an inflated number, and (2) people exclusively commiting for libav wouldnt really have any interesting in ffmpeg decisions at all
Sep 12 18:28:33 <Cigaes>	And now we see Clement trying to get back in the list of voters :-Þ
Sep 12 18:28:45 <reynaldo>	ok, saste, can you take care about publishing the list on the webpage ?
Sep 12 18:28:46 <cehoyos>	Cigaes: As said, I only asked for clarification
Sep 12 18:29:01 <BBB>	 438  Clément Bœsch ?
Sep 12 18:29:09 <nevcairiel>	its a joke BBB :P
Sep 12 18:29:10 <BBB>	maybe my list is b0rk3d
Sep 12 18:29:16 <ubitux>	(Cigaes: you got me, exactly what i just said on #ffmpeg-devel :D)
Sep 12 18:29:19 <reynaldo>	saste: can you ?
Sep 12 18:29:30 <saste>	reynaldo, not today for sure
Sep 12 18:29:35 <reynaldo>	maybe tomorrow ?
Sep 12 18:29:44 <saste>	also, do you have a git command?
Sep 12 18:29:49 <reynaldo>	remeber the choosen timeframe please, just that
Sep 12 18:29:50 <cehoyos>	nevcairiel: That sounds like very important point that needs to be made 100% clear.
Sep 12 18:29:58 <saste>	i'm not yet sure about the merged and merge committs
Sep 12 18:30:09 <saste>	cehoyos, indeed
Sep 12 18:30:15 <nevcairiel>	some things get merged from branches specifically meant for ffmpeg
Sep 12 18:30:20 <nevcairiel>	those shuld be counted for sure
Sep 12 18:30:20 <reynaldo>	saste: I have not but I think ubitux had come up with one already
Sep 12 18:30:29 <Cigaes>	git shortlog -s -n --since=2014-09-12T15:00:00Z --until 2015-09-12T15:00:00Z
Sep 12 18:30:29 <nevcairiel>	but I would exclude libav merges at this time
Sep 12 18:30:38 <reynaldo>	otherwise we can sort it out
Sep 12 18:30:38 <BBB>	how?
Sep 12 18:30:43 <saste>	ok, that list also include merged and merge committs?
Sep 12 18:30:48 <nevcairiel>	saste: yes
Sep 12 18:30:57 <saste>	so I'm fine with that
Sep 12 18:31:00 <reynaldo>	I would exclude libav merges too
Sep 12 18:31:01 <Cigaes>	saste: my command count them.
Sep 12 18:31:05 <reynaldo>	i think we all kinda agree on that
Sep 12 18:31:14 <nevcairiel>	its not trivial to exclude them in one command however
Sep 12 18:31:35 <reynaldo>	Cigaes: would you be willing to work on the command
Sep 12 18:31:37 <BBB>	I just subtract the libav stats from the ffmpeg stats :)
Sep 12 18:31:39 <reynaldo>	we can review it latter on
Sep 12 18:31:44 <BBB>	ok, let’s move on?
Sep 12 18:31:46 <reynaldo>	just to make sure its doing the right thing
Sep 12 18:31:47 <nevcairiel>	you would have to run the same command on their repo and subtract the stats, yes
Sep 12 18:31:48 <reynaldo>	yes
Sep 12 18:31:50 <reynaldo>	lets move on
Sep 12 18:32:04 <ubitux>	(note: it can be a script in tools/ directory to raise the names)
Sep 12 18:32:12 <saste>	I'd prefer to include all merged committs, but if it's just me I'll leave that at you
Sep 12 18:32:21 <Cigaes>	saste: +1
Sep 12 18:32:22 <reynaldo>	ubitux: sounds like a plan
Sep 12 18:32:32 <Cigaes>	Excluding them is sending the wrong message.
Sep 12 18:33:01 <Cigaes>	We are talking about Vittorio, Anton and Luca, and that is all unless I am mistaken.
Sep 12 18:33:02 <cehoyos>	Cigaes: What message is not including them and what message would be including them?
Sep 12 18:33:08 <cehoyos>	Martin
Sep 12 18:33:15 <nevcairiel>	diego probably too still
Sep 12 18:33:19 <cehoyos>	No
Sep 12 18:33:23 <nevcairiel>	although he vanished
Sep 12 18:33:48 <Cigaes>	Their contribution is technically good, their advice matter. And they probably would not want to vote anyway.
Sep 12 18:33:56 <Cigaes>	Diego is at 43.
Sep 12 18:34:14 <reynaldo>	please stop thinking on libav right now
Sep 12 18:34:17 <BBB>	I’m going to grab lunch, I support whatever you guys decide from this point onwards :)
Sep 12 18:34:19 <reynaldo>	it only complicates things
Sep 12 18:34:31 <reynaldo>	BBB just a minute please
Sep 12 18:34:32 <saste>	nevcairiel, if they don't want to vote, they won't (as it's likely), but especially considering that we are considering to reunite the two project I won't make distinction with their committs
Sep 12 18:34:42 <ubitux>	yeah let's move on, it's details
Sep 12 18:34:44 <cehoyos>	reynaldo: sorry, but either a decision is made or no decision is made
Sep 12 18:34:50 <Cigaes>	saste: strong +1
Sep 12 18:34:53 <ubitux>	there are more important things to worry about wrt libav
Sep 12 18:34:58 <reynaldo>	Im not sure everyone is considering reuniting tbh
Sep 12 18:35:04 <reynaldo>	Id leave that to the voting comitee
Sep 12 18:35:05 <Cigaes>	The voting process must be chosen too.
Sep 12 18:35:19 <llogan2>	saste: fine with me
Sep 12 18:35:41 <reynaldo>	ok. So, can we settle on the initial criteria please ?
Sep 12 18:35:51 <saste>	can we count about who wants to consider merged commits and not?
Sep 12 18:36:01 <reynaldo>	I understand the only thing pending is whether to count commits originating in libav
Sep 12 18:36:04 <reynaldo>	?
Sep 12 18:36:07 <saste>	reynaldo, yes
Sep 12 18:36:39 <saste>	so we have in favour: cigaes, saste, llogan2
Sep 12 18:36:43 <Cigaes>	Can we postpone that to after we decide if we want to try and reunite?
Sep 12 18:36:52 <reynaldo>	ok, please say A if you want commit originating in libav to be counting in the initial formula, b otherwise. please dont write anything else
Sep 12 18:37:00 <reynaldo>	counted/
Sep 12 18:37:10 <nevcairiel>	strictly speaking its not only about those 4, but also those that post patches to both projects and get twice the count, although it doesnt seem to push anyone over the 50 today as it is
Sep 12 18:37:10 <saste>	A
Sep 12 18:37:13 <reynaldo>	-------------------------------
Sep 12 18:37:14 <reynaldo>	B
Sep 12 18:37:15 <Cigaes>	Because basically, I would say: if we want to reunite, then we MUST include libav developers.
Sep 12 18:37:46 <saste>	Cigaes, that's my point too, also it won't make any difference in practice since they won't probably vote anyway
Sep 12 18:38:06 <reynaldo>	thats a HUGE conditional for an initial voting group with unlimited powers
Sep 12 18:38:23 <Cigaes>	reynaldo: do you believe they will abuse it?
Sep 12 18:38:36 <ubitux>	B (because i think it's not clear right now if they want to use this vote for toxicity purpose or not, and i believe including them should be postponed - the metrics is about evaluating the personal involvement in ffmpeg)
Sep 12 18:38:36 <reynaldo>	I think its not worst the risk, even if nil
Sep 12 18:38:38 *	rcombs (rcombs at irc.rcombs.me) has joined #ffmpeg-meeting
Sep 12 18:38:42 <reynaldo>	worth/
Sep 12 18:38:43 <Cigaes>	We can not trust the initial voters and distrust them at the same time.
Sep 12 18:38:43 <saste>	saste, so please let's vote on it, since we can't apparently settle
Sep 12 18:39:00 <reynaldo>	ok, we are B=2 A=1 cn you guys keep voting please
Sep 12 18:39:04 <Cigaes>	A
Sep 12 18:39:04 <nevcairiel>	B
Sep 12 18:39:17 <reynaldo>	B=4 A=1
Sep 12 18:39:17 <iive>	 
Sep 12 18:39:26 <nevcairiel>	reynaldo can't count :)
Sep 12 18:39:27 <Cigaes>	reynaldo: Ahem...
Sep 12 18:39:34 <reynaldo>	oh, 3,2 sorry
Sep 12 18:39:34 <ubitux>	haha
Sep 12 18:39:36 <reynaldo>	:)
Sep 12 18:39:40 <Loriker>	A
Sep 12 18:39:54 <durandal_1707>	B
Sep 12 18:40:11 <reynaldo>	4,3
Sep 12 18:40:34 <saste>	still two minutes then we go on, please vote if you didn't and care about it
Sep 12 18:41:17 <ubitux>	is llogan2 vote accounted?
Sep 12 18:41:34 <atomnuker>	A
Sep 12 18:41:53 <saste>	still one minute
Sep 12 18:42:59 <cehoyos>	B
Sep 12 18:43:15 <reynaldo>	ok, 2 minutes expired
Sep 12 18:43:18 <reynaldo>	initial group decided
Sep 12 18:43:19 <saste>	any more votes?
Sep 12 18:43:27 <reynaldo>	the window expired already
Sep 12 18:43:30 <saste>	allright
Sep 12 18:43:52 <reynaldo>	formula as descrived previously, not counting commits originating in libav
Sep 12 18:44:08 <saste>	A=4 B=5
Sep 12 18:44:09 <reynaldo>	I think we can leave everything else to the voting commitee once its published by monday
Sep 12 18:44:17 <saste>	so let's move on
Sep 12 18:44:20 <saste>	VDD 2015
Sep 12 18:44:22 <Cigaes>	reynaldo: no, the voting process must be decided.
Sep 12 18:44:36 <reynaldo>	I trust the voting comitee can do that
Sep 12 18:44:43 <saste>	some of us will be there
Sep 12 18:45:03 <Cigaes>	I propose: public ballots, on the mailing-list, 1 week deadline, and same kind of ballot than Debian.
Sep 12 18:45:19 <atomnuker>	Cigaes: yep, sounds good
Sep 12 18:45:35 <reynaldo>	votes dont need to be public for that matter, but again, this can be decided by the voting comittee itself
Sep 12 18:45:39 <saste>	I think we will have some real-life meeting where to discuss things 
Sep 12 18:45:50 <saste>	Cigaes, cehoyos, will you be at VDD?
Sep 12 18:45:53 <Cigaes>	reynaldo: we do not need it, but it is simpler, no need for infrastructure, authentication, etc.
Sep 12 18:45:59 <reynaldo>	saste: when is vdd taking place ? next month?
Sep 12 18:46:01 <Cigaes>	saste: I will be there.
Sep 12 18:46:08 <Cigaes>	Next week.
Sep 12 18:46:09 <saste>	reynaldo, next weekend, in paris
Sep 12 18:46:53 <cehoyos>	saste: Yes
Sep 12 18:46:55 <reynaldo>	I wont, would have loved to but just moved :/
Sep 12 18:47:04 <reynaldo>	need to be here for my family
Sep 12 18:47:11 <saste>	cehoyos, good
Sep 12 18:47:18 <cehoyos>	You sure?
Sep 12 18:47:44 <saste>	BBB, yayoi, ubitux, nevcairiel should be there as well
Sep 12 18:47:51 <saste>	maybe llogan2?
Sep 12 18:47:51 <nevcairiel>	I am not
Sep 12 18:47:58 <saste>	nevcairiel, too bad
Sep 12 18:48:17 <yayoi>	i am really broke..i would love to meet your guys though.. 
Sep 12 18:48:24 <saste>	anyway, if there is nothing to discuss here then we can move to the next point
Sep 12 18:48:35 <reynaldo>	yayoi: where are you based?
Sep 12 18:48:41 <yayoi>	san francisco...
Sep 12 18:48:42 *	Sulik (4cfe4741 at gateway/web/freenode/ip.76.254.71.65) has joined #ffmpeg-meeting
Sep 12 18:48:44 <saste>	yayoi, I believed you was going to attend, i was confused
Sep 12 18:48:57 <reynaldo>	yayoi: im in san jose, you can sure meet me ;)
Sep 12 18:49:04 <yayoi>	let's do that :)
Sep 12 18:49:10 <reynaldo>	sure
Sep 12 18:49:17 <llogan2>	saste: unfortunately, i won't be there.
Sep 12 18:49:21 <saste>	yayoi, Videolan foundation is going to refund travel and pay for hosting
Sep 12 18:49:31 <saste>	the only thing is that you need to register in time
Sep 12 18:49:37 <reynaldo>	yeah, thats an option ^
Sep 12 18:49:46 <yayoi>	well i asked VVD to fund me at registration.. for my air..but no answer.. so..
Sep 12 18:49:53 <yayoi>	sorry VDD
Sep 12 18:50:02 <saste>	yayoi, mmh OK :-(
Sep 12 18:50:04 <yayoi>	oh i see
Sep 12 18:50:10 *	am_ (ca4fcb63 at gateway/web/freenode/ip.202.79.203.99) has joined #ffmpeg-meeting
Sep 12 18:50:11 <saste>	ok, let's move on to the next topic
Sep 12 18:50:12 <yayoi>	it must be too late then
Sep 12 18:50:49 <Cigaes>	saste: if "next" is "4 outreachy", could we expedite 5 first?
Sep 12 18:51:02 <saste>	Cigaes, allright, anyone against it?
Sep 12 18:51:08 <nevcairiel>	I assume we delegate #1 to the commitee now?
Sep 12 18:51:21 <saste>	nevcariel: yes
Sep 12 18:51:27 <reynaldo>	short one: sponsoring structure, amounts, perks, etc. I begun drafting this last month, will send to the voters comitee for review once its public on monday
Sep 12 18:51:29 <saste>	next topic: use of Github/Gitorious for pull requests
Sep 12 18:51:37 <reynaldo>	not really an issue, just thought about pointing it out ^
Sep 12 18:51:56 <saste>	well, gitorious is dead, so this is only about github
Sep 12 18:52:11 <reynaldo>	saste: I'd say no but isnt this something that should be handled by the voters comitee ?
Sep 12 18:52:15 <saste>	I think this could be decided by the voting committee as well
Sep 12 18:52:17 <saste>	reynaldo, indeed
Sep 12 18:52:19 <reynaldo>	yes
Sep 12 18:52:32 <Cigaes>	For myself, I would very much prefer that all patches arrive on the mailing list as such.
Sep 12 18:52:38 <saste>	i'm also against github pull requests, if they are not backed by mailing-list patches
Sep 12 18:52:54 <saste>	anyone in favour of github pull requests?
Sep 12 18:53:10 <cehoyos>	Sorry, but afaict nobody from inside FFmpeg supports github pull requests, so there will be no voting necessary.
Sep 12 18:53:11 <reynaldo>	yeah, I think pretty much everyone out of highschool by now will agree on not using these web abominations as sources for changesets
Sep 12 18:53:22 <ubitux>	i like the ml exclusivity as well, in particular for archival purposes
Sep 12 18:53:23 <atomnuker>	I agree about mailing list only patches
Sep 12 18:53:27 <reynaldo>	+1
Sep 12 18:53:34 <saste>	well, at least here we seem to agree
Sep 12 18:53:37 *	Sulik has quit (Quit: Page closed)
Sep 12 18:54:00 <saste>	if we have no more comments we can proceed to the next point
Sep 12 18:54:03 <cehoyos>	The only question is if the current maintainer is for some funny reason forbidden to merge pull requests from github  that he likes
Sep 12 18:54:10 <jamrial>	someone mentioned adding a line to readme.md (which is shown on github) to let people know we don't accept pull requests
Sep 12 18:54:24 <reynaldo>	cehoyos: I wouldnt forgive that, no
Sep 12 18:54:34 <reynaldo>	just say that it sends the wrong message
Sep 12 18:54:37 <cehoyos>	Well, that is the only question...
Sep 12 18:54:45 <reynaldo>	but maybe a change is just too good to let it slip, who knows
Sep 12 18:55:03 <reynaldo>	jamrial: yes, that'd be a good idea
Sep 12 18:55:05 <reynaldo>	wana do it?
Sep 12 18:55:08 <cehoyos>	Sorry, I thought you were joking: Yes, this does happen
Sep 12 18:55:08 <llogan2>	i wish there was a way to disable it in github, but i don't think there is
Sep 12 18:55:23 <saste>	maintainers are free to handle their pull requests as they will, I think
Sep 12 18:55:26 <jamrial>	and make it clear that git send-email is encouraged but not required, since some people seem to have problems getting it to run
Sep 12 18:55:33 <reynaldo>	saste: +1
Sep 12 18:55:43 <Cigaes>	reynaldo: if a merge request is too good to let it pass, send it to the mailing-list.
Sep 12 18:55:44 <saste>	but we shouldn't advertise that method, since it doesn't work for most changes
Sep 12 18:55:46 <ubitux>	saste: depends if we allow merge commits outside the libav scope
Sep 12 18:55:58 <reynaldo>	but as a guideline I would take it out of the "recommended ways of contributing to FFmpeg" and all official documentation
Sep 12 18:56:02 <Cigaes>	jamrial: git send-email or git format-patch correctly attached.
Sep 12 18:56:06 <jamrial>	people will be discouraged if they need to tinker with git, but will not if they know they can attach a patch to an email
Sep 12 18:56:09 <jamrial>	yeah
Sep 12 18:56:25 <ubitux>	i really don't like the idea of loosing discussion about a patchset
Sep 12 18:56:33 <ubitux>	even if the maintainer didn't care about it at that time
Sep 12 18:56:35 <iive>	+1 git format-patch
Sep 12 18:56:38 <nevcairiel>	i really dont like format-patch, it just makes me copy-paste the patch into the mail myself to review it, but oh well
Sep 12 18:56:58 <llogan2>	nevcairiel: what client?
Sep 12 18:57:06 <Cigaes>	nevcairiel: I do not understand, format-patch is just the same thing as send-email done manually.
Sep 12 18:57:18 <jamrial>	nevcairiel: afaik most clients let you "quote" the text you select
Sep 12 18:57:42 <reynaldo>	yeah, thats a MUA issue, not a development issue
Sep 12 18:57:47 <nevcairiel>	no, send-email includes the patch in the body, while people using format-patch will attach it as an attachment, which my client offers me as a file, and not inline content i can directly quote and comment
Sep 12 18:57:50 <reynaldo>	lets not deviate though
Sep 12 18:58:13 <reynaldo>	do you guys have anything to discuss right now that the voting comitee should not be handling?
Sep 12 18:58:22 <Cigaes>	nevcairiel: this can be fixed on client side, let us discuss it later.
Sep 12 18:58:31 <saste>	next point?
Sep 12 18:58:44 <reynaldo>	a brief on publishing the list of voters
Sep 12 18:58:49 <Cigaes>	reynaldo: let us not vote when we can agree.
Sep 12 18:59:07 <reynaldo>	this will be done by monday by saste, cigaes will send the script to /tools/ and we can take it from there
Sep 12 18:59:14 <reynaldo>	was this what we agreed on ?
Sep 12 18:59:17 <reynaldo>	just confirming
Sep 12 18:59:46 <Cigaes>	reynaldo: ??? what am I supposed to do?
Sep 12 19:00:06 <saste>	reynaldo, i will publish the list to ffmpeg-devel
Sep 12 19:00:13 <reynaldo>	maybe Im wrong, give me just one sec to take a look at my backlog
Sep 12 19:00:16 <reynaldo>	saste: wprks
Sep 12 19:00:19 <reynaldo>	works/
Sep 12 19:00:37 <saste>	allright, should we move to the next point?
Sep 12 19:00:45 <saste>	which is: Outreachy funding for the next round (winter 2015)
Sep 12 19:00:45 <llogan2>	sure
Sep 12 19:00:54 <reynaldo>	just need to figure out who is providing the script saste
Sep 12 19:01:00 <reynaldo>	so you can have a list by monday
Sep 12 19:01:04 <reynaldo>	anyone volunteers ?
Sep 12 19:01:11 <reynaldo>	ubitux ^ ?
Sep 12 19:01:23 <Cigaes>	reynaldo: someone who voted B :-Þ
Sep 12 19:01:28 <ubitux>	not really :(
Sep 12 19:01:42 <reynaldo>	ok. I will
Sep 12 19:01:48 <llogan2>	yayoi 'n lglinskih: ping. we are now attempting to talk about Outreachy funding.
Sep 12 19:01:59 <yayoi>	yes
Sep 12 19:02:05 <yayoi>	i have a lot of questions actually
Sep 12 19:02:07 <Cigaes>	For Outreachy: I will probably have no time to mentor during this period, I will withdraw from the discussion mostly.
Sep 12 19:02:34 <yayoi>	yeah outreachy is asking for 5 hours a week commitment for mentoring...
Sep 12 19:02:38 <saste>	yes, the first question is if we have any volunteering mentor
Sep 12 19:02:45 <saste>	then we can seek for the funding
Sep 12 19:02:50 <saste>	or use part of our money
Sep 12 19:02:51 <yayoi>	make sense
Sep 12 19:03:43 <michaelni>	saste, i might volunteer to mentor 1 applicant maybe, depends on applicants ad exact projects ...
Sep 12 19:03:46 <saste>	which is about 8.5 K$ and 9.125 K€
Sep 12 19:03:56 <reynaldo>	saste: maybe we can start workingon an ideas page and proceed from there
Sep 12 19:04:02 <saste>	reynaldo, ok
Sep 12 19:04:16 <saste>	how much time do we have to apply?
Sep 12 19:04:28 <reynaldo>	not sure
Sep 12 19:04:37 <reynaldo>	but thats usually a bit flexible for us
Sep 12 19:04:42 <llogan2>	reynaldo: do you think samsung would be interested in funding or partial funding again?
Sep 12 19:04:43 <yayoi>	i am not sure for participants org
Sep 12 19:04:45 <reynaldo>	we are in good terms with the org
Sep 12 19:05:00 <reynaldo>	llogan2: yes, I think and hope so. ust have to confirm
Sep 12 19:05:04 <reynaldo>	just/
Sep 12 19:05:06 <yayoi>	it was not very clear as far as reading their website..
Sep 12 19:05:33 <reynaldo>	their website is never too clear tbh
Sep 12 19:05:37 <yayoi>	i see
Sep 12 19:05:39 <llogan2>	i'll try to do a better job at mentioning the funding organizations.
Sep 12 19:05:49 <yayoi>	but you can start accepting applicants.. end of the sep? 
Sep 12 19:06:04 <yayoi>	well i can ask them
Sep 12 19:06:17 <reynaldo>	llogan2: we can work together on that. I'd like to write a news entry about last one, thanking samsung for funding us twice at th every least
Sep 12 19:06:26 <yayoi>	well if they sponsor one intern, their deadline is Nov2.
Sep 12 19:06:31 <yayoi>	i mean for us
Sep 12 19:06:46 <llogan2>	reynaldo: ok. and maybe a summary of what was achieved.
Sep 12 19:07:05 <saste>	BTW, at the moment only me and michaelni are deciding how to use the project fund
Sep 12 19:07:08 <reynaldo>	yeah, that was the idea
Sep 12 19:07:18 <reynaldo>	saste: and I think you guys rock at that
Sep 12 19:07:22 <saste>	this should probably change once we have a voting committee
Sep 12 19:07:32 <reynaldo>	yes, it might
Sep 12 19:07:37 <reynaldo>	but Im happy with you guys doing it
Sep 12 19:07:43 <llogan2>	saste: can we afford a student if we don't get funding?
Sep 12 19:07:47 <reynaldo>	havent seen ppl arguing about it at least
Sep 12 19:08:13 <saste>	llogan2, we have the money, so we can
Sep 12 19:08:30 <saste>	llogan2, the question is how much money to use from our general fund
Sep 12 19:08:35 <reynaldo>	guys do we have any other pressing issue to discuss? I'd like to leave now
Sep 12 19:08:41 <reynaldo>	family weekend and stuff
Sep 12 19:08:44 <saste>	at the moment this is the situation:
Sep 12 19:08:58 <saste>	FFmpeg SPI general fund: 8446.80 $
Sep 12 19:08:59 <michaelni>	saste, its the people on the ML deciding on funds really in theory IIRC what we agreed on
Sep 12 19:09:12 <saste>	FFmpeg SPI OPW fund: 23.40
Sep 12 19:09:31 <saste>	FFmpeg ffis fund: 9125 € 
Sep 12 19:09:50 <reynaldo>	ok, leaving. Glad we made some desicions & happy to help. see you guys around o.
Sep 12 19:09:53 <reynaldo>	o/
Sep 12 19:09:55 <saste>	michaelni, yes, in practice in case there is no consensus we need to both agree
Sep 12 19:10:13 <saste>	reynaldo, see you, thanks
Sep 12 19:10:21 <llogan2>	yayoi: btw, i will try to take a look at your email template sometime soon, just so you know it isn't being ignored
Sep 12 19:10:28 <michaelni>	reynaldo, have fun!
Sep 12 19:10:34 <yayoi>	sure
Sep 12 19:10:35 <reynaldo>	saste: will ping you when the script is ready
Sep 12 19:10:38 <reynaldo>	thank guys
Sep 12 19:10:39 <reynaldo>	bye bye
Sep 12 19:10:56 <yayoi>	bye see you in san jose :)
Sep 12 19:11:09 <saste>	should we move on to the next topic?
Sep 12 19:11:22 <saste>	yayoi, anything else to say/comment about?
Sep 12 19:11:30 <yayoi>	not at this moment
Sep 12 19:11:45 <yayoi>	i like to know how many will be a mentor and fund needs to be raiseed or not
Sep 12 19:12:00 <llogan2>	we could always use funds.
Sep 12 19:12:06 <saste>	yayoi, at least we have a mentor, michaelni
Sep 12 19:12:07 <yayoi>	true
Sep 12 19:12:16 <yayoi>	but my question is mostly logistics
Sep 12 19:12:26 <saste>	michaelni, do you agree to use our money if we don't find a sponsor or a partial sponsor?
Sep 12 19:12:35 <yayoi>	like bank account and etc...
Sep 12 19:12:44 <yayoi>	also contact list would be nice..
Sep 12 19:12:59 <michaelni>	saste, no objection from me but i would prefer to find a sponsor
Sep 12 19:13:09 <saste>	michaelni, same for me
Sep 12 19:13:14 <llogan2>	yayoi: a sponsor contact list? I don't think we have one.
Sep 12 19:13:18 <yayoi>	oh
Sep 12 19:13:28 <yayoi>	okay how you get donation usually?
Sep 12 19:13:32 <saste>	so let's try to find a sponsor to cover at least part of the expense
Sep 12 19:13:53 <saste>	yayoi, http://ffmpeg.org/donations.html
Sep 12 19:14:05 <saste>	or we contact sponsors directly, like in the case of samsung
Sep 12 19:14:10 <yayoi>	right
Sep 12 19:14:18 <yayoi>	but don't they give you their contact information though?
Sep 12 19:14:32 <yayoi>	like at the end of the year,,, i assume they want their tax document etc??
Sep 12 19:14:35 <saste>	yayoi, ask reynaldo, he was dealing with that
Sep 12 19:14:41 <yayoi>	oh okay
Sep 12 19:14:46 <yayoi>	i will 
Sep 12 19:15:09 *	am_ has quit (Ping timeout: 246 seconds)
Sep 12 19:15:14 <saste>	next point is: any other business
Sep 12 19:15:24 <llogan2>	yayoi: we will also make a news entry, tweets, etc.
Sep 12 19:15:35 <yayoi>	nice
Sep 12 19:15:41 <llogan2>	begg...asking for monies
Sep 12 19:15:42 <saste>	about the use of the money: it's something we should handle once we have a decision system
Sep 12 19:15:42 <yayoi>	are you going to update the website?
Sep 12 19:16:04 <saste>	possibly if we go with a committee that should be decided by the committee
Sep 12 19:16:07 <llogan2>	yeah, unless i forget or get lazy
Sep 12 19:16:20 <yayoi>	haha
Sep 12 19:16:33 *	reynaldo has quit (Ping timeout: 246 seconds)
Sep 12 19:17:14 <llogan2>	actually, i just thought of a local organization i could ask. i know the director personally.
Sep 12 19:17:23 <yayoi>	nice
Sep 12 19:17:45 <durandal_1707>	can we talk about lavfi limitations?
Sep 12 19:18:12 <yayoi>	well please let me know if you are happy about the letter
Sep 12 19:18:27 <saste>	so let's move to the next topic?
Sep 12 19:18:28 <yayoi>	then i will start sending the email to the company i could think of potentially give the money to the project
Sep 12 19:18:33 <llogan2>	will do. may take me a few days to get to
Sep 12 19:18:43 <llogan2>	saste: ok.
Sep 12 19:19:02 <saste>	last point is about any other business
Sep 12 19:19:07 <jamrial>	durandal_1707: i don't think that requires to be discussed here. it can be done in the devel channel just fine
Sep 12 19:19:24 <saste>	durandal_1707, I'm fine if you want to discuss it here, but agree with jamrial
Sep 12 19:19:54 <saste>	especially since we're over three hours with the meeting
Sep 12 19:20:20 <saste>	no it's over two hours in truth
Sep 12 19:20:23 <durandal_1707>	on ml list my questsions get unanswered...
Sep 12 19:20:39 <saste>	durandal_1707, allright go on
Sep 12 19:21:42 <durandal_1707>	well, one of it is limited query_formats function
Sep 12 19:22:07 <jamrial>	i remember ubitux wanted to discuss how to handle lavfi regarding subtitles, but it required a decission about abi compatibility
Sep 12 19:22:16 <llogan2>	durandal_1707: sometimes i find your replies hard to read but i blame gmail for breaking quoting (and i guess my client). but trimming the relevent section could help.
Sep 12 19:22:52 <cehoyos>	Sorry that I have been late: What was the outcome of the compatibility discussion? We still need to change bitrate, probesize and analyzeudration to 64bitm either with HAVE_INCOMPATIBLE_LIBAV_ABI or without---
Sep 12 19:23:07 <durandal_1707>	I can just list bunch of them but seems Cigaes and ubitux are gone
Sep 12 19:23:19 <ubitux>	no i'm here
Sep 12 19:23:28 <saste>	cehoyos, that will be decided by the voting committee
Sep 12 19:23:36 <ubitux>	the problem with subtitles in lavfi is that i need to redesign the api/abi of subtitles
Sep 12 19:23:42 <cehoyos>	In two months? For the next bump?
Sep 12 19:23:58 <ubitux>	if we redefine our politics wrt to abi compat with libav etc, it might remotivate me to get done with my stuff
Sep 12 19:24:19 <jamrial>	so the abi discussion will be left for the voting committee in the end?
Sep 12 19:24:27 <saste>	jamrial, I think so
Sep 12 19:24:38 <iive>	who here is to keep the ABI?
Sep 12 19:24:55 <saste>	since it wasn't possible to find an agreement here, and it looked like it was an important decision to take without a formal decision system
Sep 12 19:25:12 <cehoyos>	I will soon send a patch adding HAVE_INCOMPATIBLE_LIBAV_ABI to the definitions of probesize and analyzeduration, please review, because I will push soon!
Sep 12 19:25:18 <jamrial>	afaik nobody here wants to keep it. the only thing that was talked about was how to drop it (soname bump or not, etc)
Sep 12 19:25:20 <michaelni>	the existing format negotiation and merging resulted out of the wish for a strict limit on how far to be away from the global optimal solution for arbitraray filter graphs we are with the amount of inserted converts. But this was never fully implemented so the current system kind of sucks
Sep 12 19:25:29 <cehoyos>	The bump was already days ago=-(
Sep 12 19:25:48 <cehoyos>	jamrial: No, the thing missing is a commit messageiirc
Sep 12 19:25:59 <iive>	let's keep the voting for matters that cannot be resolved otherwise
Sep 12 19:25:59 <cehoyos>	We just bumped, so we don't have to bump again.
Sep 12 19:26:12 <michaelni>	durandal_1707, so yes something should be done wih query_formats, iam not sure what is best though
Sep 12 19:26:14 <jamrial>	cehoyos: please don't push abi patches without review, and especially not before we reach a consensus about the whole dela
Sep 12 19:26:15 <jamrial>	*deal
Sep 12 19:26:30 <cehoyos>	The consensus was to go to 64 bit!
Sep 12 19:26:44 <cehoyos>	Please read the mailing list archives about it!
Sep 12 19:26:56 <durandal_1707>	I need to decide out format depending on input format
Sep 12 19:27:00 <jamrial>	i'm talking about abi compatibility, not that especific change
Sep 12 19:27:24 <durandal_1707>	And eagain sucks for that
Sep 12 19:27:35 <cehoyos>	I don't care about the compatibility: Everybody wants to drop it, nobody has written a commit message.
Sep 12 19:27:40 <ubitux>	(i need that for a local patch as well)
Sep 12 19:27:52 <cehoyos>	(I will not do it because I don't care if it gets dropped or not.)
Sep 12 19:29:23 <jamrial>	iive: i don't want to keep it, but i also don't want the soname to get a bump to 1xx like ganesh or Cigaes suggested
Sep 12 19:30:01 <cehoyos>	Aren't threse two completely unrelated issues?
Sep 12 19:30:53 <jamrial>	no
Sep 12 19:31:05 <jamrial>	anyway, back on track. saste?
Sep 12 19:31:38 <iive>	is ganesh here?
Sep 12 19:32:15 <iive>	Cigaes: would you repeat why you want to (keep) minor +100 ?
Sep 12 19:32:17 <jamrial>	don't think so
Sep 12 19:32:33 <jamrial>	iive: not minor
Sep 12 19:32:43 <iive>	patch
Sep 12 19:33:06 <ubitux>	you mean micro?
Sep 12 19:33:12 <jamrial>	ganesh patch was to bump major to 1xx. i don't want that
Sep 12 19:33:15 <michaelni>	durandal_1707, i agree but i dont know atm how to best solve it
Sep 12 19:33:15 <ubitux>	it's important to keep micro 100+
Sep 12 19:33:19 <iive>	micro
Sep 12 19:33:19 <jamrial>	micro should remain 100 as minimu,m yes
Sep 12 19:33:50 <iive>	why?
Sep 12 19:34:45 <jamrial>	<ubitux> note: we still need to keep the .100 micro as a mean to differentiate libraries, otherwise checks are going to be a pain for people trying to support post & past 3.0 + libav
Sep 12 19:34:56 <durandal_1707>	michaelni: for simple filters use current solution, for others build relation map from input to output formats
Sep 12 19:35:26 <michaelni>	durandal_1707, yes thats what iam thinking of too
Sep 12 19:35:33 <michaelni>	but iam unsure about details
Sep 12 19:35:53 <iive>	ok. then micro+100 remains.
Sep 12 19:36:11 <michaelni>	the system must work with arbitrary filter graphs, with loops and stuff
Sep 12 19:36:24 <iive>	is there anything else to decide about ABI compat?
Sep 12 19:37:56 <jamrial>	what to do with major. ganesh, Cigaes and others want to bump to 1xx. I personally don't
Sep 12 19:38:05 <ubitux>	what's the purpose?
Sep 12 19:39:02 <iive>	since micro is already bumped, what would major bump be about?
Sep 12 19:39:24 *	J_Darnley has quit (Ping timeout: 264 seconds)
Sep 12 19:39:29 *	jlfhdvfjabsdfg (~J_Darnley at d51A44418.access.telenet.be) has joined #ffmpeg-meeting
Sep 12 19:39:31 *	jlfhdvfjabsdfg is now known as J_Darnley
Sep 12 19:40:02 <iive>	I think their proposal is major to be bumped instead. However app script compatibility would be problem.
Sep 12 19:41:05 <jamrial>	the idea of bumping major to something different than libav is to stop distributing two incompatible libraries with the same soname, basically
Sep 12 19:41:32 <iive>	but they would still be with same soname
Sep 12 19:41:33 <jamrial>	but as nevcairiel pointed out earlier today, we have been technically doing that all this time anyway
Sep 12 19:42:33 *	kurosu (Kurosu at 2a01:e35:8ae7:63a0:199a:f324:6481:cf2d) has left #ffmpeg-meeting ("Leaving")
Sep 12 19:43:27 <iive>	I think debian uses suffix, so the libs are libavcodec_ffmpeg.so 
Sep 12 19:44:10 <jamrial>	yeah, that was mentioned as an argument in favor of keeping major versions as they are right now
Sep 12 19:44:56 <iive>	so is there anything else to decide about ABI compat?
Sep 12 19:46:27 <saste>	anything else?
Sep 12 19:46:34 <saste>	or should we close the meeting?
Sep 12 19:46:49 <saste>	technical discussion can go on #ffmpeg-devel and in the ML
Sep 12 19:47:31 <jamrial>	yeah, we can close the meeting i guess
Sep 12 19:47:44 <saste>	allright
Sep 12 19:47:50 <saste>	so the meeting is closed
Sep 12 19:48:03 <saste>	i'm going to publish the log on ffmpeg-devel soon
Sep 12 19:48:18 <saste>	thanks all for your time and contribution!
Sep 12 19:48:36 <jamrial>	alright. thanks everyone as well
Sep 12 19:48:51 *	jamrial (~jamrial at 181.22.42.38) has left #ffmpeg-meeting
Sep 12 19:49:01 *	atomnuker (~atomnuker at 2001:19f0:6800:89a5:5400:ff:fe0e:35da) has left #ffmpeg-meeting
Sep 12 19:49:21 *	yayoi has quit (Quit: yayoi)
Sep 12 19:49:29 *	nevcairiel (quassel at WoWUIDev/WoWAce/Ace3/nevcairiel) has left #ffmpeg-meeting ("http://quassel-irc.org - Chat comfortably. Anywhere.")
Sep 12 19:49:57 *	ubitux (~ubitux at bre75-1-78-192-242-8.fbxo.proxad.net) has left #ffmpeg-meeting ("WeeChat 1.1.1")
}}
Last modified 17 months ago Last modified on Nov 1, 2015, 8:53:51 PM