15:00:13 #startmeeting FESCO (2021-02-03) 15:00:13 Meeting started Wed Feb 3 15:00:13 2021 UTC. 15:00:13 This meeting is logged and archived in a public location. 15:00:13 The chair is zbyszek. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:13 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:00:13 The meeting name has been set to 'fesco_(2021-02-03)' 15:00:13 #meetingname fesco 15:00:13 The meeting name has been set to 'fesco' 15:00:13 #chair nirik, ignatenkobrain, decathorpe, zbyszek, sgallagh, mhroncok, dcantrell, cverna, Conan_Kudo, Pharaoh_Atem, Son_Goku, King_InuYasha, Sir_Gallantmon, Eighth_Doctor 15:00:13 Current chairs: Conan_Kudo Eighth_Doctor King_InuYasha Pharaoh_Atem Sir_Gallantmon Son_Goku cverna dcantrell decathorpe ignatenkobrain mhroncok nirik sgallagh zbyszek 15:00:16 #topic init process 15:00:18 .hello2 15:00:19 zbyszek: zbyszek 'Zbigniew Jędrzejewski-Szmek' 15:00:23 .hello kevin 15:00:25 nirik: kevin 'Kevin Fenzi' 15:00:31 .hello2 15:00:31 bcotton: bcotton 'Ben Cotton' 15:00:43 .hello2 15:00:43 decathorpe: decathorpe 'Fabio Valentini' 15:00:55 nirik: no "morning"? 15:01:00 heh. morning. ;) 15:01:47 sgallagh said he's lurking, and mhroncok will be late 15:03:55 .hello ngompa 15:03:56 King_InuYasha: ngompa 'Neal Gompa' 15:03:57 * King_InuYasha waves 15:04:16 So we're at 4.5, not really quorum. 15:04:26 But let's chat 15:04:33 #topic #2558 F34 Change: Remove Guile Support From Toolchain 15:04:33 .fesco 2558 15:04:34 zbyszek: Issue #2558: F34 Change: Remove Guile Support From Toolchain - fesco - Pagure.io - https://pagure.io/fesco/issue/2558 15:05:11 King_InuYasha you're the only one who's really against this... 15:05:23 I'm not sure how I feel about this. He's my favorite Street Fighter character... 15:05:27 hello, sorry for being late 15:05:41 Based on both debian's popcon and codesearch, this is completely unused. 15:06:06 I think it should be fine to remove unused features, esp. when a significantly better alternative exitss. 15:06:17 for make, no alternative exists 15:06:56 and the arguments used in the Change are specious at best and disingenuous at worst 15:07:29 would it help to make a 'make-with-guile' package people who need/want this could use instead? 15:07:34 Yeah, but make already has it's own language that is turing-complete. 15:07:34 yes 15:07:36 if this is going to be removed, I guess it should be done upstream first? 15:07:52 nirik: I would have been fine with this if we had a make-guile package being introduced like debian 15:08:22 straight up removing something without upstream also deciding to remove it on the basis of "maintenance" when they are also the upstream developers is not a valid justification to me 15:09:02 decathorpe: I don't think so. It's pretty normal to not enable some upstream features in Fedora packages. 15:09:04 * nirik has never encountered anyone or anything that used this, but I suppose someone must have for them to have added it. 15:09:22 I have encountered a few Makefiles using it before 15:09:25 I can't share them though 15:10:00 it's rare for "public" projects to use it because of macOS compatibility 15:10:13 I don't like telling maintainers to make an effort to create a package which has no users. 15:10:17 Guile support was added in Make 4.0, and macOS stuck to Make 3.x 15:10:22 so, either we ask change owners if they would be willing to make a make-guile (and get King_InuYasha on board) or we approve it over objections? 15:10:28 pretty much 15:10:37 .hello2 15:10:38 dcantrell: dcantrell 'David Cantrell' 15:10:39 I'm comfortable with being alone in my objection 15:10:50 if nobody else cares about make users, then at least it's known 15:10:50 zbyszek: right, and then the 'less maint' is not as compelling, since they again have to maintain it. 15:11:30 and the buildroot thing is *also* not compelling since we just removed Make from the default buildroot 15:12:31 King_InuYasha: read that as not "default buildroot", just "buildroot"... 15:12:52 OK, should we vote? 15:13:00 if we want a make-minimal without guile for buildroot, fine 15:13:16 if we want a make-guile and make default make not have guile, fine 15:13:37 but I don't want guile support removed from make entirely 15:13:56 not unless make upstream wants to remove it and deal with that wart 15:14:49 King_InuYasha: I suppose a volunteer an package make-guile separately if they are interested, tight? 15:14:54 *right? 15:14:56 mhroncok: probably not 15:15:04 King_InuYasha: why not? 15:15:19 most likely the maintainers would pitch a fit 15:15:27 same reason why we don't have alternative kernels 15:15:31 or alternative bootloaders 15:17:24 IMO, this is a pointless minimization 15:17:27 we have some alternative bootloaders. ;) uboot/zipul/yaboot/syslinux. ;) 15:17:35 nirik: they are not alternative 15:17:43 with the exception of syslinux, I guess 15:17:54 all of those have to be used in conjunction with grub2 15:18:06 we can't have refind, for example 15:18:10 King_InuYasha: sd-boot 15:18:16 anyhow, this seems drifting off topic 15:18:25 zbyszek: doesn't practically work, not signed efi binary 15:18:37 King_InuYasha: my laptop begs to differ ;) 15:18:38 King_InuYasha: you don't need yaboot 15:19:01 dcantrell: you do for ppc systems pre POWER9 15:19:34 anyhow, should we vote? or punt back to change owners to adjust anything? 15:19:41 King_InuYasha: I don't see why this would be a problem: maintainers of make don't want to maintain this, fine. other maintainers might want to do that, fine as well. what am I not getting? 15:19:42 * King_InuYasha shrugs 15:20:00 my vote: +1 15:20:05 ignatenkobrain: +1 in the ticket 15:20:09 King_InuYasha: that's also not true (grub2 for POWER7 and above). but that doesn't matter. yaboot is a leftover now from really old ppc systems 15:20:10 * nirik is +1 15:20:20 -1 15:20:22 I vote 0 15:20:44 +1 (also posted in the ticket) 15:20:59 I will abstain. I think if this is going to be removed, it should happen upstream first. All else are hollow arguments. 15:21:07 I don't feel this benefits Fedora much, but I won't block the maintainers to make a decision that doe snot affect other packages 15:21:33 I wouldn't care so much about this if upstream (which *they are*) were *also* removing Guile support 15:21:47 the fact that this is a Fedora-specific choice is not acceptable to me 15:22:05 upstream first :) 15:22:24 King_InuYasha: and if this is really about ELN / RHEL, this should not be a change in Fedora at all 15:22:27 That gives us +4, 1, -1, 1 abstain 15:22:31 decathorpe: bingo 15:22:38 the particularly galling thing about this is that they are the upstream developers and they didn't even bother doing anything upstream first 15:22:44 clearly they should make guile support dlopenable and just support it if you have guile installed. (not it for implementing that :) 15:23:17 the Python team is an excellent example of how they handle these things 15:23:26 * mhroncok blushes 15:23:46 I've generally been disappointed by the Toolchain team in how they handle Changes and interface with Fedora, and this is yet another example of it 15:24:23 ƒ💛🐍 15:24:42 there's only one other that I've had a worse experience with, and I'm sad neither team has improved 15:25:07 zbyszek: so, not passed... ask them to retry ? 15:25:16 nirik: waiting for sgallagh 15:25:18 I'm going to go with -1 as well. 15:25:30 ok 15:25:34 I'll ask them whether they intend this for ELN instead 15:25:36 zbyszek: my "abstain" was meant to be a 0 vote :) 15:25:55 #agree REJECTED (+4, 2, -2) 15:26:08 Err, that's not really rejected. 15:26:18 heh 15:26:27 not APPROVED is not REJECTED? 15:26:30 it's a tie, though we default to reject in that scenario 15:26:39 Right, I suppose that is true. 15:26:40 > A majority of the committee (that is, five out of nine) is required to pass a proposal in a meeting. 15:26:46 ? 15:26:52 zbyszek: you're correct 15:26:59 Right, default to rejected. Never mind. 15:27:37 #topic Next week's chair 15:27:42 volunteers? 15:27:56 This is another case to support my belief that "0 votes should have a more clear impact", though 15:28:15 Or else we should disallow them at all. 15:28:23 sgallagh: yeah, IMO, that should mean we should send them back for refinement 15:28:27 Sorry, I'll save that for Open Floor 15:28:27 sgallagh: True. They are like -1 effectively. 15:28:36 * mhroncok can chair next week if bcotton is ready to cover 15:29:11 #action mhroncok will chair the next meeting. 15:29:16 I'll cover if bcotton can't. 15:29:37 i can if needed 15:29:48 Or actually, I can cover either way. I expect to be there. 15:29:56 #topic Open Floor 15:30:06 sgallagh, go 15:30:21 Proposal: When voting on a proposal, a 0 vote is equivalent to "I am disqualifying myself from this vote, thereby reducing the number required to reach a majority". 15:30:29 +1 15:30:31 -1 15:30:44 I don't think it should have the consequence of reducing majority 15:30:52 Erego, if two people voted 0, a +4 would be required to pass instead of +5 15:30:55 -1 15:31:05 Question: We need +5 for APPROVE, how many -1s do we need for REJECT? 15:31:26 decathorpe: 0 15:31:36 decathorpe: technically, one in the ticket 15:31:36 decathorpe: it needs to be < +5 and > -1 15:31:50 decathorpe: ... because 5 people voting +0 mean that the proposal cannot pass 15:32:00 it can pass via ticket 15:32:07 I think +5 for approve is appropriate. Either way the 0 or -1 votes are doesn't matter, if it's not +5 or more it's not approve. 15:32:08 * decathorpe is confused why rules that are different depending on where votes are counted 15:32:22 async vs sync 15:32:30 more or less, ^ 15:32:34 tickets get approved if there is +3/+1 and no -s 15:32:54 but we do meetings when there is at least one -1 in he ticket 15:33:00 *in the 15:33:01 decathorpe: yes, I think that is ugly. By not voting int he ticket we are effectively abstaining. 15:33:13 have we actually needed that rule (pass with less than +5) even once in the past year? 15:33:15 Well, if voting 0 doesn't change the required +1 count, what effect does it have? Should we then just disallow it? 15:33:19 decathorpe: yes 15:33:21 decathorpe: Yes 15:33:28 we've used it for a lot of in-ticket voting 15:33:41 where not enough votes occur 15:33:42 would those tickets have reached +5 after a few more days anyway? 15:33:48 a lot of them were +4,0,0 15:33:58 * decathorpe wants receipts :) 15:34:08 decathorpe: Experience has shown us that it's the best way to get more tickets processed without wasting meeting time 15:34:44 * mhroncok is happy with the current setup 15:34:45 not sure how you can disallow voting 0... I mean, if you just never vote at all isn't it the same? (if it doesn't change the number needed to pass something) 15:34:51 well, at the expense of making voting rules really complicated in corner cases 15:34:55 sgallagh: it was a great idea. I was sceptical, but it really helps process non-controversial changes without two-hour meetings. 15:35:05 zbyszek: Thanks :) 15:35:05 mhroncok++ 15:35:05 nirik: Karma for churchyard changed to 10 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:35:31 cookies :) 15:35:36 Right, we cannot disallow 0/not-voting, so we should at least decide what it actually means. 15:35:42 mhroncok++ 15:35:42 Conan_Kudo: Karma for churchyard changed to 11 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:35:52 I would argue that if it's not clear if the result of a vote is approved or rejected, then the rules are ... not good 15:36:07 it is clear to me 15:36:24 to approve on a meeting, you need +5 15:36:32 * nirik nods. 15:36:39 At which point, a 0 vote is exactly equivalent to a -1 vote 15:36:50 there could technically be a problem if one of the fesco members resigns and we don't yet have a new one, but... that's rare 15:36:51 net +5 (sum of votes) or just 5 x +1 votes? 15:37:10 at least 5 +1 votes 15:37:33 net +5 would be functionally equivalent to my proposal above, I think 15:37:49 in the former case, whether people vote 0 or -1 actually makes a difference. not in the second case. 15:37:50 Well, maybe not quite 15:38:06 net +5 means that if you have +5,3,-1, it won't be approved, correct? 15:38:13 yes 15:38:19 that's bad 15:38:20 wait, no, that would be approved. 15:38:28 * nirik doesn 15:38:30 no, because that would have a net +4 15:38:35 t understand the problem we are trying to solve. 15:38:47 how do you count it? 5*+1 + 3*0 + 1*-1 == +4 15:39:09 wait. (5 * +1) + (3*0) + (1 * -1) = +4 so not approved 15:39:16 I think that should pass. 15:39:21 and it does 15:39:31 Right 15:39:34 becasue we don't require net +5, we require at least 5 +1s 15:39:40 right 15:40:01 Or in my proposal above, that would pass because the three 0's would make the passing requirement only +3. 15:40:02 but then there's no difference between 0 and -1? 15:40:17 the idea is to have majority approval, not sliding windows for majorities 15:40:21 decathorpe: That's why I'm asking if we should address that. 15:40:29 net +5 would mean that 3 -1 votes would mean a veto 15:40:31 Specifically: what does it mean if I vote 0? 15:40:41 if you only count +1 votes then there's no difference between 1) not voting, 2) abstaining, 3) voting 0, 4) voting -1 15:40:45 so I've just voted 0 15:40:50 I'm not sure why we even have the 0 vote 15:40:58 Should it mean that I am effectively voting -1 (as now) or that I don't have an opinion and I am accepting the remaining majority opiniojn? 15:40:59 it means that the effect on the result is the same as if I voted -1 15:41:07 King_InuYasha: Because not voting at all == 0 15:41:15 no, that I get 15:41:22 I mean why do we let people declare 0 15:41:32 King_InuYasha: in the ticket, it makes a difference 15:41:39 I think it might be useful for people voting for fesco members later? 15:41:40 on the meeting, not so much 15:41:49 I see 15:42:00 0 -> I didn't bock this, but didn't approve either, -1 -> I didn't want this at all/was against ? 15:42:03 King_InuYasha: because sometimes you really don't have an opinion. Even because of lack of time or because the arguments for and against are too close. 15:42:11 fair 15:42:11 I tend to do the following on the meeting: I disagree with the change fundamentally: -1 15:42:27 a 0 vote makes it clear that you're intentionally not voting. so if a vote is +8,1,-1 it's clear that everyone voted and the decision can be processed immediately 15:42:38 makes sense 15:42:47 my rubric is similar to nirik's actually 15:42:51 so 0 == abstain? 15:42:53 it's just rare that I don't have an opinion :) 15:43:03 heh 15:43:10 decathorpe: yes 15:43:19 King_InuYasha: there's things you don't have an opinion on? :) 15:43:20 * nirik runs 15:43:27 * King_InuYasha shrugs 15:43:28 And that's why I was suggesting that abstaining should have a real meaning. 15:43:30 there are a few :P 15:43:49 My proposal above is roughly equivalent to "My vote goes with the majority" 15:44:11 It's just a mathematical description of how that works. 15:44:35 Anyway, it seems there's no support for changing the rules. If y'all agree, I think we should end the discussion. Maybe we can return to it during some devconf meeting. 15:44:42 I am also not a fan of inventing procedural changes during a meeting and voting about them at that very meeting. happy to discuss, but please don't do proposals like this to be voted about immediatelly 15:44:49 sgallagh: now I understand. though I don't think reducing the number of required votes is good. if 4 people abstain then one +1 vote would be enough? 15:44:51 sure 15:45:05 decathorpe: No, that would still require +3 15:45:27 * decathorpe 's brain checks out 15:45:28 A majority of the non-abstaining voters 15:45:43 3 being more than half of the remaining five votes. 15:46:24 Any other subjects for Open Floor? 15:47:06 Flipping it around, it would be the same as 0 == I vote the same as whatever side has the most votes. So a vote of (3, 4, -2) would be equivalent to (7, -2) 15:47:11 devconf meeting? 15:47:36 nirik: *flock 15:48:22 ah, ok. Wondering if I missed a devconf session being scheduled. ;) 15:48:41 sgallagh: I think it'd be fairer to divide the votes in half, ie. (3,4,2)→(5,-,4) 15:48:43 Is Flock a thing this year? Do we have info on that? 15:48:58 too early to tell I think 15:49:37 zbyszek: That could work 15:49:49 * sgallagh will take that to a ticket 15:49:57 what if I vote -½ and +¼ and abstain with the last ¼? :) 15:50:09 .fire nirik 15:50:09 adamw fires nirik 15:50:12 sgallagh++ 15:50:14 zbyszek: Karma for sgallagh changed to 8 (for the current release cycle): https://badges.fedoraproject.org/tags/cookie/any 15:50:17 .fire nirik 15:50:17 adamw fires nirik 15:50:33 double fired! hurray. my day is off to a great start! 15:50:48 I'll close in a minute 15:51:18 See you next week. 15:51:18 #endmeeting