17:00:26 #startmeeting FESCO (2023-12-07) 17:00:26 Meeting started Thu Dec 7 17:00:26 2023 UTC. 17:00:26 This meeting is logged and archived in a public location. 17:00:26 The chair is zbyszek. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions. 17:00:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 17:00:26 The meeting name has been set to 'fesco_(2023-12-07)' 17:00:26 #meetingname fesco 17:00:26 #chair nirik, decathorpe, zbyszek, sgallagh, mhroncok, dcantrell, mhayden, Conan_Kudo, Pharaoh_Atem, Son_Goku, King_InuYasha, Sir_Gallantmon, Eighth_Doctor, tstellar 17:00:26 The meeting name has been set to 'fesco' 17:00:26 Current chairs: Conan_Kudo Eighth_Doctor King_InuYasha Pharaoh_Atem Sir_Gallantmon Son_Goku dcantrell decathorpe mhayden mhroncok nirik sgallagh tstellar zbyszek 17:00:29 #topic init process 17:00:38 .hello2 17:00:40 mhayden: mhayden 'Major Hayden' 17:00:46 .hello2 17:00:47 dcantrell: dcantrell 'David Cantrell' 17:00:55 morning. sort of here. 17:00:59 .hello ngompa 17:01:00 Son_Goku: ngompa 'Neal Gompa' 17:01:40 .hi 17:01:41 sgallagh: sgallagh 'Stephen Gallagher' 17:02:03 tstellar said that he'll miss the meeting today. 17:02:10 .hello salimma 17:02:11 michel-slm: salimma 'Michel Lind' 17:02:47 We don't have quorum yet, but I think that for the first item, we don't need it. 17:02:53 #topic #3089 retiring redhat-lsb in Fedora 17:02:53 .fesco 3089 17:02:54 zbyszek: Issue #3089: retiring redhat-lsb in Fedora - fesco - Pagure.io - https://pagure.io/fesco/issue/3089 17:03:02 .hi 17:03:03 jonathanspw: jonathanspw 'Jonathan Wright' 17:03:04 Is there anything to do here? Can we close this? 17:03:34 I think so, yeah 17:03:39 I marked it with 'meeting' to follow up, but it's been three weeks with no comments. 17:03:48 were we waiting for something from the maintainer? 17:04:11 No, not really. We made a statement of what should happen, but didn't explicitly request any action. 17:04:12 ping carlwgeorge redhat-lsb 17:04:13 I think we can close it; the decision was made and it's now up to the maintainers to implement it 17:04:43 zbyszek: We explicitly required that packages not claim compatibility. 17:04:56 yeah, at this point either the maintainer does the thing or it gets retired for non-compliance with our guidance 17:05:56 OK, I closed the tic.et 17:06:34 For the rest of the agenda, we don't have quorum. 17:06:46 December is tough like that. 17:06:56 Wait, don't we? 17:07:10 Oh, we have. 17:07:11 I count 6 17:08:07 Sorry. 17:08:12 #topic #3097 Change: DNF Conditional File Lists 17:08:12 .fesco 3097 17:08:13 zbyszek: Issue #3097: Change: DNF Conditional File Lists - fesco - Pagure.io - https://pagure.io/fesco/issue/3097 17:08:42 * Son_Goku sighs 17:09:25 so do we know that FPC is ok with MUST NOT? 17:09:27 What's left to discuss here? 17:10:12 Only Neal was against in the FPC ticket. 17:10:15 I don't object in principle to making Fedora's packages use only file entries from primary.xml (I think we need to fix up the filter for it to add stuff from /usr/libexec though...), but I don't really want to approve this because knowing how the politics of things are with DNF development, if we do this, the likelihood of proper handling for this in DNF drops like a rock 17:11:25 having been involved with DNF upstream for almost the entire time DNF has been the package manager for Fedora, I know that issues like these are largely prioritized based on what Fedora does or does not accept as valid 17:11:29 .hi 17:11:30 decathorpe: decathorpe 'Fabio Valentini' 17:11:32 sorry for being late 17:11:44 @Son_Goku I respect that, but I do have to point out that this looks like a "perfect being the enemy of good enough" situation. 17:12:07 sgallagh: if we were happy to ignore third-party repositories and the experience for users, I'd agree 17:12:26 We've been over this. I think that other people think that this would be beneficial for Fedora users. 17:12:35 Honestly? If this breaks third-party repos... then that's their problem. 17:12:40 no, it isn't 17:12:43 it's our problem too 17:12:49 it's ALWAYS our problem 17:12:57 that's why we deferred the SHA-1 thing twice 17:13:08 which third party repos are we talking about here ... all of them? 17:13:13 yes 17:13:17 has anyone looked how many (if any) of the Fedora Workstation external repositories would be broken by this? 17:13:17 no 17:13:39 but feel free to ignore me and implement it anyway 17:13:39 maybe ask the Change owners to do that and put it in the proposal as well? 17:13:45 I'm the only -1 anyway 17:13:47 seems like the change owners should try and quan.... yes 17:13:49 what gotmax said 17:13:50 I don't know for sure, but I've been using dnf5 exclusively for many months, and I haven't had any issues. 17:13:50 nobody else cares 17:13:59 zbyszek: same here. 17:14:48 I'd prefer to just approve this and move on. 17:14:55 fine do that 17:15:33 decathorpe: you were -1 in the ticket. Change? 17:15:41 Nothing is permanent; if we discover that this has real impact on our users, we can change our minds. 17:15:55 But I think this is likely to be FAR less of an issue than the SHA-1 change. 17:16:05 that's probably true 17:16:15 did they at least fix user requesting files? 17:16:17 sgallagh, nirik were +1 17:16:30 because I know people do `dnf install /usr/include/foo.h` 17:16:36 Son_Goku: that has always worked. 17:16:47 it has not 17:16:58 it was, at least at one point, broken in dnf5 17:17:07 I just added a +1 to the ticket for my vote 17:17:12 and I do not know if it will be broken in dnf4 with this change 17:17:17 it sounds like it will 17:17:36 I'm also +1 17:18:43 added my +1 in the ticket just now 17:18:43 I'm remaining -1 17:18:48 Son_Goku: Sounds like it will, but at least: 17:18:50 totally missed the update on that one :( 17:18:54 > We have a plan to provide a message with a suggestion that downloading filelist might resolve the issue including a description how to do it from command line. 17:19:03 * Son_Goku sighs 17:19:07 > The proposed behavior has already been incorporated into the future successor, DNF5 project 17:19:10 I hate crap UX 17:19:15 but whatever 17:19:26 And the way that dnf5 works is that 'dnf install /usr/include/foo.h' works. 17:19:41 that's at least something 17:19:48 And it's been explicitly mentioned a few times that this is about bringing dnf behaviour closer to dnf5. 17:20:04 let's just approve the ticket and move on 17:20:07 So I really expect that this will not be broken. 17:20:52 OK, so we're at +5, 0, -2 17:21:15 decathorpe, are you still with us? 17:22:01 #agree APPROVED (+5, 0, -2) 17:22:11 #topic #3098 Change: Drop sshd Socket 17:22:11 .fesco 3098 17:22:12 zbyszek: Issue #3098: Change: Drop sshd Socket - fesco - Pagure.io - https://pagure.io/fesco/issue/3098 17:22:41 this one is also not okay, and I don't think they're solving the right problem here 17:23:02 sorry, was pulled away by family 17:23:09 So… I didn't have time to look at the code, but I think we should just make sshd socket-activated. 17:23:18 zbyszek: that's what it currently is 17:23:28 isn't it? 17:23:47 right, they want to move away from that I thought? 17:23:50 iirc the proposal is to drop socket activation because there was no rate limiting, but rate limiting is now added 17:23:58 Right 17:24:00 so... the CP should no longer be necessary? 17:24:11 yeah I don't think the reason for this change applies any more 17:24:13 yeah 17:24:16 It is in inet-style single-connection socket activation, and implement normal single-instance activation-on-first-connection. 17:24:45 yeah, this one needs a rethink/redo at least... 17:25:43 I remain -1 on this change. With the rate limiting available, this seems like an overreaction to a non-problem. 17:25:54 If you'd be OK with that, I'd talk with Dmitry to figure out if it'd be possible to implement proper socket activation. 17:26:08 If they want to improve things, zbyszek's suggestion makes more sense, but even then it seems like a minor gain. 17:26:28 zbyszek: go for it 17:26:38 I remain -1 on this 17:27:00 yeah, I can be -1 for now. At least it needs reworking 17:27:09 #action zbyszek to talk with Change Owner about socket activation with Accept=no 17:27:11 -1 for current proposal as well 17:27:14 OK, let's punt for now. 17:27:22 #topic #3101 Change: Remove OpenSSL Compat 17:27:22 .fesco 3101 17:27:23 zbyszek: Issue #3101: Change: Remove OpenSSL Compat - fesco - Pagure.io - https://pagure.io/fesco/issue/3101 17:27:58 zbyszek: I'm not sure we needed to punt. I think we have -5 (assuming you also disagree with the plan as written) 17:28:07 #undo 17:28:07 Removing item from minutes: 17:28:59 OK, I can vote -1, with the understanding that we can revisit an updated proposal. 17:29:09 That's always true, yes. 17:29:12 #agree REJECTED (0, 0, -5) 17:29:20 #topic #3101 Change: Remove OpenSSL Compat 17:29:20 .fesco 3101 17:29:21 zbyszek: Issue #3101: Change: Remove OpenSSL Compat - fesco - Pagure.io - https://pagure.io/fesco/issue/3101 17:29:28 I just added a -1 vote for me to issue 3098 17:30:10 Folks, please stop voting in the tickets during the meeting. This just creation confusion. 17:30:28 *creation confusion process is creating. 17:30:38 **or something. 17:30:39 I can't tell if you're calling a vote here or not 17:30:41 :D 17:30:47 seems like you're only reading votes in the ticket, apologies 17:31:15 I think the concerns that were raised for Python 3.6 were still not addressed? 17:31:27 * nirik nods 17:31:45 I put a vote of existing votes in the ticket to clarify what the state was. 17:32:18 decathorpe: yeah. So I think it's a valid concern. We want python3.6 to remain rebuildable. 17:32:44 is Python 3.6 the only concern at this point? 17:32:58 if openssl1.1 is explicitly kept around so people can build stuff for RHEL 8 (where both openssl1.1 and python3.6 remain maintained) I don't think dropping it now is a good idea 17:33:19 and openssl1.1 _has_ to be maintained for RHEL 8 anyway 17:33:43 why wouldn't you just... build on rhel8 in that case? 17:34:05 yeah, I don't follow how the rhel8 maintenance requirement needs this is Fedora 17:34:17 are we still supporting back to python 3.6 in fedora? 17:35:24 opensmtpd and gloo also link to openssl1.1. 17:35:32 dcantrell: We still ship it, mainly for running test suites across many Python releases. 17:35:36 dcantrell: https://discussion.fedoraproject.org/t/f40-change-proposal-removing-openssl-1-1-package-system-wide/92899/2 17:35:45 this was Miro's response 17:36:03 ^^ That 17:36:12 OK, I see. that makes sense then 17:36:14 ok. and no answer from change owners. 17:36:24 I say as long as things are linked with openssl1.1, we can't remove it 17:36:40 Well, that can be a catch-22 17:36:52 as long as things are linked to it and there's no commitment or effort to migrate/drop them 17:37:07 sgallagh: which is why should also think long and hard about what packages we bring in to the distribution 17:37:14 if someone wants to make a change to drop it, they're also going to have to figure out the revdeps 17:37:19 Son_Goku: that too 17:37:21 there's also the question of third-party apps 17:37:35 the plan to "just use newer versions of them" is ... interesting 17:37:35 I think we should ask for an explicit answer. Maybe the porting of python3.6 is simple and then I'd be +1 to the change. 17:37:59 in an ideal world, deprecated packages begin a march towards becoming a leaf package that nothing requires and the retirement or removal becomes a trivial thing later 17:38:01 decathorpe: Many third-party apps just bundle their deps anyway, so *shrug* 17:38:23 zbyszek: agreed, we need more info to be able to vote 17:39:03 punt? 17:39:35 and re-ask change owners to answer about python at least and other things too if possible 17:39:38 Punt and ask for more info? it was already delayed for weeks because we had no meetings / quorum 17:39:52 Yes. 17:40:05 give it one more ask and then we vote later 17:40:15 #info We're waiting for a reply from the Change Owner about python3.6. 17:40:28 I asked in the ticket. 17:40:28 #topic #3103 Change: Tuned Replaces Power-profiles-daemon 17:40:29 .fesco 3103 17:40:30 zbyszek: Issue #3103: Change: Tuned Replaces Power-profiles-daemon - fesco - Pagure.io - https://pagure.io/fesco/issue/3103 17:41:30 decathorpe: you wanted to discuss this one 17:42:49 My big concern with this proposal is that it doesn't seem that the Change submitters have been talking to our Edition WGs about it. 17:43:06 Workstation at least is not clearly in favor. 17:43:14 I don't think this is a good idea 17:43:34 they did... 17:43:40 there's a workstation working group ticket... 17:44:03 I get the feeling that this was motivated by the main maintainer archiving the project on GitLab because they got reassigned within Red Hat 17:44:04 My understanding isn't that they aren't in favor, it's that they don't yet know.... 17:44:49 decathorpe: it is entirely motivated by that 17:45:24 ok, then I can also dislike it out of principle :) 17:45:26 though the change from ppd to tuned seems to interest at least some folks for providing better power profile capabilities 17:45:28 nirik: Right, my phrasing may not have been clear. I meant there wasn't a clear consensus in favor. 17:46:47 I think we should just wait for the Workstation WG to make their decision. 17:46:52 I think it's probibly ok to just leave this... and if the workstation working group decides not to take any of it, we just drop it then (as long as the decision is before beta) 17:47:17 I do not think the workstation wg will make any decision 17:47:53 outside of me, no member of the working group feels knowledgeable enough about it 17:48:24 and the maintainer of power-profiles-daemon is not a member of the wg 17:49:35 if they don't want to make any decision, I am not sure where that leaves us... letting the change happen, or requiring it to stay with status quo. 17:49:57 No decision presumably means making no changes 17:49:58 It sounds like it's too early… If the Change Owners make tuned work as a drop-in replacement for p-p-d and it turns out to work great, then it'll be a much easier decision. 17:50:23 But right now we're being asked to approve a transition when we don't really know how the end state will look. 17:50:29 I am concerned about the deps, but I guess thats not a big problem on workstations... 17:50:46 image size go yay big again? 17:50:50 well, changes are often asperational... 17:51:30 Proposal: FESCo does not believe that the Change is adequately designed at this time. Please coordinate with the Workstation WG and upstream and resubmit later. 17:51:42 +1 17:51:52 +1 17:52:30 why can't we just invoke contengency if it's not ready by beta? 17:52:31 sgallagh: the change owners are implementing ppd support in tuned 17:52:52 that is, no change from gnome/kde/budgie/etc. to use tuned 17:53:40 -1. I don't think "inadequatly designed" isn't really the issue. 17:53:52 I'm mostly concerned that they're designing a solution that no one knows if it will be accepted. 17:53:56 in the Change document, it is explicitly mentioned that the power-profiles-daemon API will be provided by tuned 17:54:14 I am not ready to dictate that Fedora adopt it. 17:54:56 Is PPD upstream explicitly abandoned? 17:55:06 sgallagh: it's been revived. 17:55:40 I'm not prepared to override GNOME upstream on this unless we have solid evidence that it's a better option. 17:56:12 it's not part of GNOME either 17:56:25 I don't think there's overriding happening yet... we don't know. no decision has been made. 17:57:28 Proposal: It's too early to decide if/when tuned should replace p-p-d. We reject that part of the proposal at this point, but look forward to seeing the compatibility interface in tuned and an updated proposal to replace p-p-d when it's possible to evaluate an existing solution. 17:58:28 so does that imply then rejecting the change for now? 17:58:36 I agree with the first sentence. The rest ... not so sure 17:58:58 Hmm, let me reword. 17:59:19 Proposal: It's too early to decide if/when tuned should replace p-p-d. The proposal is rejected for now. 18:00:18 +1 18:00:23 I guess... it seems like we could just wait and see instead of rejecting, but if thats the way people want to go... 18:00:38 I'd rather reject than wait and see 18:00:41 +1 18:00:58 I'd also rather reject, because it'll be months before we have enough info. 18:01:35 nirik, please give an explicit vote 18:01:45 sorry, +1 18:01:49 Son_Goku, mhayden? 18:02:13 +1 on the proposal 18:02:28 +1 18:02:51 Oh, sgallagh, I missed you. I can't count today. 18:03:30 +0 18:03:46 #agree It's too early to decide if/when tuned should replace p-p-d. The proposal is rejected for now. (+6, 1, 0) 18:03:56 #topic #3119 Numerous package git repos fail git-fsck, causing issues for mirroring, and need to be fixed 18:03:59 .fesco 3119 18:04:00 zbyszek: Issue #3119: Numerous package git repos fail git-fsck, causing issues for mirroring, and need to be fixed - fesco - Pagure.io - https://pagure.io/fesco/issue/3119 18:04:48 I'm +1, with the caveat that dcantrell raised, that we need a way to store the old state somewhere. 18:05:05 FWIW, I don't think this needs to be fast-tracked. We worked around it for the sixteen packages in CS 10. That said, I'm obviously in favor of fixing it. 18:05:07 looks like a bunch of comments this morning that I have not had time to read or ponder on. ;) 18:05:22 Also, I think we should adjust distgit-branch-unused to handle those rewritten repos properly. 18:06:05 also, I think this should probibly be discussed more widely. 18:06:15 sgallagh: do we need fast-track still? 18:06:16 because its going to surprise some people/maintainers that have checkouts. 18:06:44 is anybody even maintaining those Java packages? 18:06:46 * decathorpe ducks 18:06:47 zbyszek: I just said above: "FWIW, I don't think this needs to be fast-tracked. We worked around it for the sixteen packages in CS 10. That said, I'm obviously in favor of fixing it." 18:07:04 I dropped the tag from the issue. 18:07:48 * nirik needs to go soon, do we have much more today? 18:08:18 Let's continue the discussion on this in the ticket. 18:08:40 I think nirik is correct: we should move this to a wider discussion 18:08:47 fedora-devel and/or Discussion 18:08:56 Yeah. 18:09:14 OK, let's start wrapping this up. 18:09:18 #topic Next week's chair 18:09:24 Volunteers? 18:09:33 * nirik will off on PTO 18:09:43 what meetings do we wish to cancel for the holidays? 18:09:51 I can probably do it. It'll be my last meeting before PTO 18:10:11 Same for me, I'll be there next week and want to take the next week off. 18:10:12 will there be new members next time, or only in two weeks? 18:10:18 I propose cancelling the 20th and 27th meetings 18:10:23 dcantrell: +1 18:10:36 you mean 21st and 28th? 18:10:37 dcantrell: +1 18:10:37 21st and 28th 18:10:40 but +1 18:10:46 yeah, sorry 18:10:55 jan 4th is also... a likely one to cancel IMHO. 18:10:59 +1 18:11:24 I don't think cancelling three meetings in a row is a good idea 18:11:41 I should be around on the 4th 18:11:43 there have been issues with "submitted late" change proposals in previous years 18:11:50 Yeah, we can try to meet on the 4th. 18:12:03 all we can do is try to meet and if there's quorum, meeting happens 18:12:06 Tell you what, someone else pick up next week and I’ll run the 4th, if we have quorum. 18:12:06 when is the submission deadline? 18:12:27 a lack of planning on someone else's part does not constitute an emergency for us 18:12:46 2 Change Checkpoint: Proposal submission deadline (Changes requiring infrastructure changes) Wed 2023-12-20 Wed 2023-12-20 0 18:12:46 3 Change Checkpoint: Proposal submission deadline (Changes requiring mass rebuild) Tue 2023-12-26 Tue 2023-12-26 0 18:12:46 4 Change Checkpoint: Proposal submission deadline (System Wide Changes) Tue 2023-12-26 Tue 2023-12-26 0 18:12:46 5 Change Checkpoint: Proposal submission deadline (Self Contained Changes) Tue 2024-01-16 Tue 2024-01-16 0 18:13:26 the deadline is boxing day? really? 18:13:28 wow 18:13:41 yep. ;( 18:13:49 that conflicts with the Harlem Globetrotters game here :) 18:13:58 only thing that could have been worse is actually christmas itself 18:14:16 OK, is there a volunteer for the 14th? 18:14:36 * nirik heads out, have a great holidays everyone! 18:14:38 I'll do it 18:14:53 #action Son_Goku will chair next meeting, December 14th. 18:14:53 #action sgallagh will chair the meeting on January 4th. 18:14:53 #info Meetings on the December 21st and 28th are cancelled. 18:14:58 that'll be the first with new fesco right? 18:15:01 Thank you. 18:15:08 Yep. 18:15:23 We might have a new time, but we'll figure that out in a ticket. 18:15:28 #topic Open Floor 18:16:08 nothing from me 18:16:14 If nothing, I'll close in a minute. 18:16:22 I got nothing 18:16:29 nada 18:16:38 #endmeeting