fesco
LOGS
17:00:28 <tstellar> #startmeeting FESCO (2023-07-27)
17:00:28 <zodbot> Meeting started Thu Jul 27 17:00:28 2023 UTC.
17:00:28 <zodbot> This meeting is logged and archived in a public location.
17:00:28 <zodbot> The chair is tstellar. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
17:00:28 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:00:28 <zodbot> The meeting name has been set to 'fesco_(2023-07-27)'
17:00:34 <tstellar> #meetingname fesco
17:00:34 <zodbot> The meeting name has been set to 'fesco'
17:00:42 <tstellar> #chair nirik, decathorpe, zbyszek, sgallagh, mhroncok, dcantrell, mhayden, Conan_Kudo, Pharaoh_Atem, Son_Goku, King_InuYasha, Sir_Gallantmon, Eighth_Doctor, tstellar
17:00:42 <zodbot> Current chairs: Conan_Kudo Eighth_Doctor King_InuYasha Pharaoh_Atem Sir_Gallantmon Son_Goku dcantrell decathorpe mhayden mhroncok nirik sgallagh tstellar zbyszek
17:00:46 <dcantrell> .hello2
17:00:47 <zodbot> dcantrell: dcantrell 'David Cantrell' <dcantrell@redhat.com>
17:00:52 <Eighth_Doctor> .hello ngompa
17:00:53 <zodbot> Eighth_Doctor: ngompa 'Neal Gompa' <ngompa13@gmail.com>
17:01:13 <tstellar> #topic init process
17:01:22 <tstellar> .hello tstellar
17:01:23 <zodbot> tstellar: tstellar 'Tom Stellard' <tstellar@redhat.com>
17:01:35 <sgallagh> .hi
17:01:36 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
17:01:41 <mhroncok> .hello churchyard
17:01:42 <zodbot> mhroncok: churchyard 'Miro Hrončok' <mhroncok@redhat.com>
17:02:02 * mhroncok will just fetch some snack
17:02:06 <nirik> morning
17:02:31 <dcantrell> hi, everyone
17:02:45 <dcantrell> I got out of the pool on vacation to join the meeting  :)
17:02:58 <nirik> dedication!
17:03:13 <tstellar> Looks like we have a quorum.
17:04:10 <decathorpe> .hi
17:04:11 <zodbot> decathorpe: decathorpe 'Fabio Valentini' <decathorpe@gmail.com>
17:04:21 <tstellar> OK, let's begin.
17:04:33 <tstellar> #topic #3039 RFC: Roadmap for DNF5 in Fedora 39 / invoking the Contingency Mechanism
17:04:43 <tstellar> .fesco 3039
17:04:44 <zodbot> tstellar: Issue #3039: RFC: Roadmap for DNF5 in Fedora 39 / invoking the Contingency Mechanism - fesco - Pagure.io - https://pagure.io/fesco/issue/3039
17:04:50 <jmracek> Hello
17:05:10 <sbueno> hi everyone
17:05:12 <nirik> so, I'd like to see if we can convince them to land for f40 and still have a way to revert in eln/cs10...
17:05:50 <nirik> awesome. Thanks for coming sbueno and jmracek!
17:05:51 <dcantrell> jmracek, do you have an update since last week's meeting?
17:06:16 * mhroncok is back
17:06:17 <jmracek> The patch for ELN is on the way - https://github.com/rpm-software-management/dnf5/pull/780
17:07:03 <jmracek> Yes, update. We would like to postpone replace DNF by DNF5
17:07:10 <mhroncok> ah, upstream spec files, yay
17:07:23 <jmracek> Yes, upstream spec
17:07:48 <mhroncok> what does this mean for Fedora 40?
17:07:57 <dcantrell> so invoke the contingency plan?  yeah, what does this mean for F40?
17:08:09 <decathorpe> I see two options here: 1) postpone to F41, and revert the change in rawhide until after the F40 branching. or 2) postpone to F40 and revert only after F39 is branched off.
17:08:09 <decathorpe> I think keeping DNF5 in Rawhide for so long but not shipping it would be detrimental.
17:08:53 <dcantrell> of those options, I like #1 more because I am concerned that we won't get enough testing coverage even for F40
17:08:55 <decathorpe> dcantrell looks like you haven't seen this yet: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org/thread/EYE2JY537OM7GFW46EK7YIBLHJ52USAZ/
17:09:09 <jmracek> To not interfer with RHEL branching we would like to postpone it to Fedora 41. We will use given time to add missing features and improve it
17:09:35 <decathorpe> I keep repeating myself, but RHEL schedule should not have impact on Fedora schedule
17:09:45 <nirik> well, but I don't know that it would really interfere would it? it might take some tweaking to make sure it doesn't branch over, but it should be possible right?
17:09:58 <dcantrell> decathorpe: thanks, I hadn't seen that.  haven't looked at anything this week because of PTO
17:10:04 <decathorpe> yes. spec file conditionals exist.
17:10:05 <jmracek> Revert in rawhide will create problems for rawhide users - missing dnf symlink
17:10:32 <sbueno> i'm curious as well re: reverting in rawhide. how would keeping it in rawhide be detrimental?
17:10:41 <decathorpe> by "revert" I mean: have dnf provide dnf again
17:10:51 <sbueno> though to be clear (and thank you for linking my email decathorpe) i do think f41 is the safest option. all rhel schedules aside
17:11:19 <nirik> Right, so after f39 branches off, revert back to dnf4 there. Keep dnf5 in f40... when eln/centos10 branches off, make sure it reverts there too back to dnf4...
17:11:21 <dcantrell> dnf and dnf5 can coexist together, yes?  if so, I think having dnf provide 'dnf' while dnf5 lives alongside is fine
17:11:25 <mhroncok> if f41 is safes, let's revert in rawhide
17:11:35 <tstellar> It's not really up to FESCO, when the change proposal gets implemented.  I think we should just vote whether or not to revert the change in rahwide.
17:11:39 <mhroncok> rawhide isn't a rolling testbed for rhel, it's future Fedora
17:12:03 <decathorpe> I don't think there's ever been a feature that was enabled in rawhide for three release cycles but never shipped?
17:12:12 <dcantrell> and based on the current feedback, we have seen it's not ready for fedora
17:12:47 <decathorpe> if both F39 and F40 are going to ship with DNF 4 as default, we need test coverage for those code paths, including rawhide
17:12:59 <nirik> There's more work to do for sure, but it's already a big win for some things. ;)
17:13:41 <decathorpe> It wouldn't be fun to revert to DNF 4 after branching off from Rawhide and find out that everything explodes
17:13:54 <nirik> yeah, and ultimately, I think the choice is up to dnf team. :) What they are willing to do. The one case I don't think we want to allow is keep dnf5 in rawhide but not target f40
17:14:03 <tstellar> nirik: +1
17:14:05 <jmracek> Keepng DNF5 in rawhide for longer period will help with stabilazing DNF5 for Fedora 41
17:14:26 <decathorpe> jmracek: that might be true, but it will be detrimental to F39 and F40
17:14:30 <nirik> but will not help in stablizing f40
17:14:54 <decathorpe> (hence why I outlined my two options above the way I did)
17:15:19 <jmracek> Yes, but dnf4 is already stable enoght
17:15:28 <mhroncok> :D
17:15:30 <dcantrell> how will dnf5 existing now be detrimental if the default is dnf4?
17:15:35 <tstellar> I'm in favor of option 1.
17:16:06 <decathorpe> jmracek: it might be, but can you prove that dnf 4 as default will work just fine on Fedora 40 if it gets no testing in rawhide until then?
17:16:28 <mhroncok> we have been on this topic for a while
17:17:01 <mhroncok> is there anybody at FESCo who thinks keeping rawhide on dnf5 while having both f39 and f40 on dnf4 is an acceptable plan?
17:17:48 <mhroncok> because that is what the change authors want -- and we should determine whether that's possible (the discussion seems that it won't be, but maybe not everybody has spoken)
17:18:12 <jmracek> We have CI that ensurse that it works. If those test are anought for RHEL gating, why they cannot prove the same thing for Fedora?
17:18:18 <nirik> I'm not in favor of that plan, but we could vote on it if we want to make sure everyone is on the same page
17:18:25 <Eighth_Doctor> I don't think holding something in Rawhide to not be released is a good idea
17:18:57 <tstellar> I think we should vote on decathorpe's option 1.
17:19:04 <Eighth_Doctor> when we killed off Alphas and such, the point of Rawhide was to become semi-stable, and at least some degree useful
17:19:14 <dcantrell> I was misunderstanding.  My assumption was we'd revert *EVEYRTHING* to dnf4 by default, even rawhide.  but merely leave dnf5 in the compose alongside, which does nothing except make it more accessible to those who want to test/early adopt
17:19:15 <decathorpe> ok ... so since upstream developers seem to favor shipping with Fedora 41 now, should I make my Option 1 (revert in rawhide, switch back to dnf 5 after F40 is branched off) a formal proposal?
17:19:20 <jmracek> I was honest and provide all information.
17:19:55 <tstellar> decathorpe: Yes, I think so.  We are right at the 15 minute mark now too.
17:19:56 <nirik> well, would dnf team folks like to discuss again out of band with their team based on this constraint?
17:20:13 <nirik> or do you all prefer option 1 or option 2 if you can't get the path you wanted before?
17:20:34 <tstellar> I think either way we need to vote so we can invoke the contigency in some form.
17:20:44 <sbueno> jmracek what do you think? this is probably a call better made by you
17:20:45 <dcantrell> tstellar: +1
17:20:59 <nirik> (reminder: I see two options here: 1) postpone to F41, and revert the change in rawhide until after the F40 branching. or 2) postpone to F40 and revert only after F39 is branched off.)
17:21:53 <dcantrell> what about 3) invoke contingency plan, have dnf team resubmit change proposal later
17:21:54 <jmracek> Thank nirik -
17:21:57 <tstellar> We don't technically need to postpone, we can just revert in rawhide and then ask the change owners to get back to us whether they want to do it for f40 or f41.
17:22:05 <tstellar> dcantrell: Right.
17:22:21 <nirik> yeah, true... that could be an option
17:23:37 <nirik> would also give more time to consider and see where they are at the f40/f41 submission times...
17:23:37 <jmracek> I am slightly in favor of option #1
17:23:48 <tstellar> I would be in favor on just voting on #3.
17:23:50 <mhroncok> let's vote on option 1
17:23:56 <mhroncok> +1
17:24:00 <tstellar> mhroncok: That's fine with me too.
17:24:08 <tstellar> We just need to vote on something :)
17:24:12 <decathorpe> Proposal: The switch to DNF5 will be postponed (likely to Fedora 41, to be determined by the Change Owners) and reverted in Rawhide for the time being (i.e. dnf 4 will provide /usr/bin/dnf again).
17:24:30 <Eighth_Doctor> Fabio Valentini: +1
17:24:30 <mhroncok> Fabio Valentini: +1
17:24:35 <dcantrell> +1
17:24:37 <nirik> +1
17:24:42 <tstellar> +1
17:24:48 <decathorpe> +1
17:24:57 <tstellar> Wait do we need to use the \# proposal command.
17:25:09 <mhroncok> no
17:25:13 <tstellar> ok
17:25:21 <nirik> there's not one. ;)
17:25:32 <decathorpe> the text will be in the approve thing :)
17:26:10 <tstellar> So 'agree <proposal> APPROVED(+6,0,0)' ?
17:26:23 <mhroncok> that works
17:26:39 <decathorpe> though usually we do 'agree APPROVED (......) <proposal>' I think
17:26:51 <sgallagh> +1
17:26:56 <sgallagh> Sorry, multitasking
17:27:04 <nirik> I'd like to also thank the dnf5 team. It can't be easy to retarget something you were working on so long. My thanks for being honest about the status and realistic about timeframes and needed work. Kudos!
17:27:21 <tstellar> #agree APPROVED(+7,0,-0) The switch to DNF5 will be postponed (likely to Fedora 41, to be determined by the Change Owners) and  reverted in Rawhide for the time being (i.e. dnf 4 will provide /usr/bin/dnf again).
17:27:33 <decathorpe> Yes, I'd like to second that. Thank you for being transparent about the state of things!
17:27:41 <tstellar> Yes, thank you.
17:27:51 <tstellar> Ok next topic.
17:27:56 <sbueno> thank you. it definitely was not easy, and the support is much-appreciated. we really do want to make sure we deliver a high quality experience to everyone :)
17:28:01 <tstellar> #topic
17:28:09 <tstellar> Oops
17:28:10 <tstellar> #undo
17:28:10 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x7f1e01a94a90>
17:28:16 <decathorpe> Making things crunch for F39 doesn't sound like it would've been worth it
17:28:30 <tstellar> #topic #3035 Change: Build JDKs once, repack everywhere
17:28:38 <tstellar> .fesco 3035
17:28:39 <zodbot> tstellar: Issue #3035: Change: Build JDKs once, repack everywhere - fesco - Pagure.io - https://pagure.io/fesco/issue/3035
17:29:38 <tstellar> This ticket stands at (+2,0,-2).
17:30:06 * nirik needs more coffee. I'm still a weak +1 (per ticket vote)
17:30:38 <tstellar> Or maybe +3 sgallagh voted minutes before the tally was reset.
17:30:53 <dcantrell> I haven't voted in the ticket, but generally speaking I am not in favor of it.  So I'm going to say -1 for me.  I'll at it to the ticket
17:31:00 <decathorpe> Is there something to discuss or do we just need to make a final vote? Other than "bodhi changes would require a rewrite of the entire app" I see no new information.
17:31:02 <sgallagh> I remain +1
17:31:43 <mhroncok> I wasn't around when this was proposed again -- is there a working solution that does not involve relengs tagging updates manually
17:31:44 <mhroncok> ?
17:32:00 <tstellar> Does anyone want to discuss this further?
17:32:03 <decathorpe> mhroncok: as far as I know, no
17:32:42 <mhroncok> wait, wasn't that the requirement when we first refused this?
17:32:43 <nirik> yes...
17:32:58 <nirik> they can use sidetags, there's no more releng involvement needed.
17:33:03 <tstellar> hm, I thought the new process did no involve releng.
17:33:53 <nirik> make sidetag for oldest stable, build, certify/test that, then make sidetags for other releases, tag that into it, build there, untag it and make bodhi updates from sidetags
17:34:38 <nirik> that assumes the 'portable' one is only released on oldest stable
17:34:57 <nirik> (and other releases get a readme pointing to it)
17:34:58 <mhroncok> and the "actually built" package will only be part of the "oldest stable" bodhi update?
17:35:07 <mhroncok> ack
17:35:09 <nirik> right.
17:35:10 <mhroncok> +1
17:35:53 <nirik> it's really not great... but under the constraints it seems like it would work. Of course the contingency of 'maintain less' seems attractive too. ;)
17:38:20 <tstellar> By my count, we are at (+4,0,-3) does anyone else want to vote: Eighth_Doctor, mhayden ?
17:39:06 <Eighth_Doctor> I'm extremely weak +1 on the provision we're not mixing distros for JDK builds
17:39:26 <Eighth_Doctor> I really don't like this, but we're not left with much
17:40:17 <mhroncok> Conan Kudo: from the summary: "(build JDK) ...in oldest live Fedora repack in all live Fedoras. If jdk is buitl in epel, it will be built in oldest possible epel and repacked in newer live epels. "
17:40:24 <mhroncok> aka not mixing distros
17:41:28 <Eighth_Doctor> then we should be... "okay"
17:42:53 <tstellar> alright we can close the voting then.
17:43:10 <tstellar> #agree APPROVED(+5,0,-3)
17:43:44 <mhroncok> what is the count then?
17:43:46 <mhroncok> ack
17:44:14 <tstellar> Someone can double check my counting, but I think that's correct.
17:45:22 <tstellar> #topic Next week's chair
17:45:30 <tstellar> Any volunteers?
17:45:41 <dcantrell> should we skip next week since it occurs during Flock?
17:46:02 <nirik> yes. +1 skip for flock. ;)
17:46:09 <tstellar> Makes sense +1
17:46:19 <dcantrell> so next meeting would be 10-Aug ?
17:46:21 <decathorpe> +1 I'll keep the lights on
17:46:35 <Eighth_Doctor> yes
17:47:01 <nirik> we would still need someone for then...
17:48:07 <tstellar> I won't be able to make that meeting.
17:48:27 <decathorpe> I should be able to handle August 10
17:48:33 <tstellar> decathorpe: Ok, thanks.
17:48:46 <tstellar> #action decathorpe will chair next meeting
17:48:56 <tstellar> #topic Open Floor
17:49:36 <tstellar> Any other business?
17:49:36 <decathorpe> it was pointed out during the FPC meeting an hour ago that the third-party repo policy could use ... improvements
17:49:39 <decathorpe> c.f. https://docs.fedoraproject.org/en-US/fesco/Third_Party_Repository_Policy/
17:50:26 <mhroncok> wasn't the policy recently updated?
17:50:52 <decathorpe> I can't remember. but we talked about it today so I hope the thing we talked about was the current version
17:51:50 <nirik> just clarifying / wording changes? or something completely missing?
17:52:11 <decathorpe> for example, as I read it, the entire policy document also applies to shipping things that are disabled by default, but that's rather hidden in the small print
17:52:55 <decathorpe> FPC had a request for how to package third-party repo files, and as far as I can tell, all we can say is "policy forbids packaging third-party repos except those approved by FESCo or a WG"
17:53:26 <decathorpe> or am I missing something here?
17:54:09 <decathorpe> (there's also no distinctions between foo-repos packages that are installed by default or only installed on user request)
17:54:52 <nirik> If it's just rewording, I would say file a PR... but if we are changing meanings we should probibly discuss on list first...
17:55:23 <decathorpe> I'm not sure we need to change the meaning, but some things aren't explicit in the policy right now, so who can tell
17:55:30 <mhroncok> agreed with Kevin
17:56:07 <nirik> I've not looked at it in a long while...
17:56:15 <decathorpe> ok ... I can try to come up with a fesco-docs PR with other FPC folks and we can decide whether it needs broader discussion once there's actual changes to talk about?
17:56:51 <Eighth_Doctor> makes sense to me
17:57:13 <tstellar> Ok, any other topics...
17:57:26 <decathorpe> nope
17:57:56 <Eighth_Doctor> nothing from me
17:58:11 <Eighth_Doctor> oh wait, we probably want to cancel next week's meeting
17:58:22 <dcantrell> we did
17:58:39 <mhroncok> :D
17:58:46 <Eighth_Doctor> since Flock is going on at that time
17:59:02 <Eighth_Doctor> oh welp
17:59:03 * Eighth_Doctor is sleep deprived
17:59:09 <decathorpe> hm, salimma on devel says that the calendar still has the fesco meeting in fedora-meeting, can somebody update that?
17:59:30 <nirik> I can.
17:59:37 <salimma> thanks nirik
17:59:39 <decathorpe> thanks! :)
17:59:50 <tstellar> #action nirik will update fesco meeting location in the calendar.
17:59:54 <tstellar> nirik: Thanks.
18:00:12 <tstellar> Alright, if there is nothing else, then we can end the meeting.
18:00:31 <decathorpe> right on schedule 🥳
18:01:04 <tstellar> #endmeeting