<@mhayden:fedora.im>
19:30:00
!startmeeting FESCO (2024-03-04)
<@meetbot:fedora.im>
19:30:01
Meeting started at 2024-03-04 19:30:00 UTC
<@meetbot:fedora.im>
19:30:02
The Meeting name is 'FESCO (2024-03-04)'
<@mhayden:fedora.im>
19:30:04
!meetingname fesco
<@salimma:fedora.im>
19:30:04
!hi
<@zodbot:fedora.im>
19:30:06
Michel Lind (salimma) - he / him / his
<@mhayden:fedora.im>
19:30:11
Chairs: @conan_kudo:matrix.org, @ngompa:fedora.im, @nirik:matrix.scrye.com, @humaton:fedora.im, @zbyszek:fedora.im, @sgallagh:fedora.im, @jistone:fedora.im, @dcantrell:fedora.im, @mhayden:fedora.im, @tstellar:fedora.im
<@mhayden:fedora.im>
19:30:16
!topic Init Process
<@conan_kudo:matrix.org>
19:30:17
!hi
<@mhayden:fedora.im>
19:30:18
!hi
<@zodbot:fedora.im>
19:30:19
Neal Gompa (ngompa) - he / him / his
<@zodbot:fedora.im>
19:30:19
Major Hayden (mhayden) - he / him / his
<@jvanek:fedora.im>
19:30:25
hello!
<@smilner:fedora.im>
19:30:28
!hi
<@zodbot:fedora.im>
19:30:29
None (smilner)
<@humaton:fedora.im>
19:30:31
!hi
<@zodbot:fedora.im>
19:30:32
Tomáš Hrčka (humaton) - he / him / his
<@nirik:matrix.scrye.com>
19:30:48
morning
<@jistone:fedora.im>
19:30:58
!hi
<@zodbot:fedora.im>
19:30:59
Josh Stone (jistone) - he / him / his
<@mhayden:fedora.im>
19:31:01
man, i hit the enter button right at :30 past. small wins...
<@sgallagh:fedora.im>
19:31:20
!hi
<@zodbot:fedora.im>
19:31:21
Stephen Gallagher (sgallagh) - he / him / his
<@pmikova:fedora.im>
19:31:29
!hi
<@zodbot:fedora.im>
19:31:30
Petra Alice Mikova (pmikova)
<@tstellar:fedora.im>
19:31:31
!hi
<@zodbot:fedora.im>
19:31:33
Tom Stellard (tstellar)
<@zbyszek:fedora.im>
19:31:38
!hi
<@zodbot:fedora.im>
19:31:39
Zbigniew Jędrzejewski-Szmek (zbyszek)
<@mhayden:fedora.im>
19:31:43
welcome back nirik
<@mhayden:fedora.im>
19:32:45
okay, we have quorum as i counted to five on one hand and then ran out of fingers 🤣
<@mhayden:fedora.im>
19:32:59
i think i counted 7
<@mhayden:fedora.im>
19:33:37
!topic #3166 Clarify voting process for non-responsive maintainers
<@mhayden:fedora.im>
19:33:43
!fesco 3166
<@zodbot:fedora.im>
19:33:43
**fesco #3166** (https://pagure.io/fesco/issue/3166):**Clarify voting process for non-responsive maintainers**
● **Opened:** a month ago by jistone
● **Last Updated:** 5 hours ago
● **Assignee:** Not Assigned
<@zbyszek:fedora.im>
19:33:53
This was actually merged already…
<@sgallagh:fedora.im>
19:33:55
This should be done. I merged the MR today
<@mhayden:fedora.im>
19:33:56
i think this one just needs more attention on the PR
<@conan_kudo:matrix.org>
19:34:03
It's all done.
<@mhayden:fedora.im>
19:34:11
oh nice, you beat me to it, Stephen Gallagher
<@zbyszek:fedora.im>
19:34:17
So I guess we're in the mode "holler if you see issues".
<@zodbot:fedora.im>
19:34:25
mhayden has already given cookies to sgallagh during the F39 timeframe
<@zbyszek:fedora.im>
19:34:32
Stephen Gallagher++, thanks.
<@zodbot:fedora.im>
19:34:34
zbyszek has already given cookies to sgallagh during the F39 timeframe
<@mhayden:fedora.im>
19:34:35
okay, next one
<@sgallagh:fedora.im>
19:34:38
zbyszek: Or possibly "Feel free to scream into the void"
<@mhayden:fedora.im>
19:34:45
!topic #3157 Request to have FESCo be the group that sets the priority for CPE Infrastructure Projects
<@mhayden:fedora.im>
19:34:50
!fesco 3157
<@zodbot:fedora.im>
19:34:51
**fesco #3157** (https://pagure.io/fesco/issue/3157):**Request to have FESCo be the group that sets the priority for CPE Infrastructure Projects**
● **Opened:** a month ago by amoloney
● **Last Updated:** 5 hours ago
● **Assignee:** amoloney
<@mhayden:fedora.im>
19:35:03
I think Aoife Moloney was interested in talking about 3157
<@smilner:fedora.im>
19:35:13
I as well :-)
<@zodbot:fedora.im>
19:35:29
mhayden gave a cookie to smilner. They now have 7 cookies, 2 of which were obtained in the Fedora 39 release cycle
<@mhayden:fedora.im>
19:35:31
go for it!
<@zodbot:fedora.im>
19:35:36
zbyszek gave a cookie to smilner. They now have 8 cookies, 3 of which were obtained in the Fedora 39 release cycle
<@smilner:fedora.im>
19:35:45
Hello everyone! I’m Steve Milner, a Fedora community member and one of the managers for the CPE team. Today I am presenting the Infrastructure Projects Proposal in short form. In https://pagure.io/fesco/issue/3157 Aoife asked for some time today to give a tl;dr as well as answer questions. Reminder: this is a _proposal_ which we can work together to refine and/or accept/reject at a later date :-)
<@smilner:fedora.im>
19:36:03
The high level tl;dr of the process (https://fedoraproject.org/wiki/CPE_Infrastructure_Projects: Note: some updates may not be reflected there yet such as switching from Pagure issues to discourse discussions) is as follows...
<@jvanek:fedora.im>
19:36:09
!hi
<@zodbot:fedora.im>
19:36:10
jiri vanek (jvanek)
<@amoloney:fedora.im>
19:36:18
!hi
<@zodbot:fedora.im>
19:36:19
Aoife Moloney (amoloney)
<@smilner:fedora.im>
19:36:30
1. Someone proposes a project for prioritization
2. One or more community members help refine the proposal in discourse (done to help proposals be in a state ready for FESCo to review)
3. CPE Product Owner (PO) takes it to FESCo
4. FESCo decides if it goes on the list or not, as well as what it’s priority on the list should be
5. CPE PO updates and publishes the list
<@smilner:fedora.im>
19:36:47
The guiding goals we came up with for this...
<@smilner:fedora.im>
19:36:58
Pull together two sets of infrastructure/service project lists into one centralized, living, and prioritized list for all interested people
<@smilner:fedora.im>
19:37:02
Pull together two sets of infrastructure/service project lists into one centralized, living, and prioritized list for all interested people
<@smilner:fedora.im>
19:37:12
Priority to come from a community entity (Example: FESCo)
<@zodbot:fedora.im>
19:37:15
mhayden has already given cookies to amoloney during the F39 timeframe
<@smilner:fedora.im>
19:37:20
Encourage an increase of people working together towards common goals that benefit the Fedora Project
<@smilner:fedora.im>
19:37:24
Increase visibility of high impact contribution opportunities and outcomes
<@smilner:fedora.im>
19:37:37
We also came up with some anti-goals to be aware of ...
<@smilner:fedora.im>
19:37:47
A central list of work the community _must_ do
<@smilner:fedora.im>
19:37:59
Disincentivize anyone from working on things not on the list
<@smilner:fedora.im>
19:38:00
Add undue process hoops for community members
<@smilner:fedora.im>
19:38:16
We also have a short odp presentation (which are bullet points for items brought up here) as well as a summary document. I’ll be working with Aoife Moloney to make sure they are shared in the right place :-).
<@smilner:fedora.im>
19:38:41
I know that's a lot of info in a short amount of time. Please know discussion can continue and this is just a proposal!
<@smilner:fedora.im>
19:38:52
Are there any questions, comments, concerns, ideas, at this time?
<@zbyszek:fedora.im>
19:39:15
This sounds reasonable. I'm happy to try the process. If it turns out to work, great. If not, we can iterate.
<@zbyszek:fedora.im>
19:39:32
This sounds reasonable. I'd be happy to try the process. If it turns out to work, great. If not, we can iterate.
<@sgallagh:fedora.im>
19:39:50
So to summarize (maybe?): someone from CPE will come to us and ask us two questions: 1) Is this important enough to spend limited paid resources on and 2) how important relative to the things we already decided to have them work on?
<@smilner:fedora.im>
19:40:36
Stephen Gallagher: Close! But it's more along the lines of a list of work anyone (including paid resources) can work on. Hopefully more collaboration between all people :-)
<@mhayden:fedora.im>
19:40:41
i'm with zbyszek here -- it's worth trying a process and making some adjustments if needed. i'd hate for us to become a bottleneck to improvements
<@nirik:matrix.scrye.com>
19:41:16
do note that the deliverable here is a curated list. CPE isn't promising to work only only things in this priority or whatever. It's for everyone to decide what to work on based on all the things they would need to decide that
<@sgallagh:fedora.im>
19:41:41
Steve: I think we need to make the decisions based on whether it's worth expending our paid resources and be grateful if other resources turn up.
<@sgallagh:fedora.im>
19:42:38
I'm worried that otherwise, the list will end up just a FESCo rubber-stamp.
<@nirik:matrix.scrye.com>
19:42:48
This list is not a promise to expend paid resources. ;) It's just a list of 'what is important to the community/fesco'
<@jistone:fedora.im>
19:43:29
not just a stamp, but with ranking too
<@zbyszek:fedora.im>
19:43:35
I think that the fact that there's a priority order is important here. I expect that there'll always be a tail of things which would be nice to have, but unlikely to happen. What matters the most is what at the top of the list.
<@nirik:matrix.scrye.com>
19:43:37
(oh and a place where CPE and anyone else can coordinate/work on those things)
<@amoloney:fedora.im>
19:43:47
the ask for FESCo is to act as an advisory group and assist in reviewing and prioritising this work list based on the benefit and impact to the Fedora project
<@smilner:fedora.im>
19:43:51
Correct. It's less about "who is going to do what" and more so having people aware of, and feel safe to join important sets of work as defined by Fedora.
<@sgallagh:fedora.im>
19:44:01
zbyszek: Sure, but I'm worried about the list growing unmanageable.
<@sgallagh:fedora.im>
19:44:24
Realistically, we'll be unlikely to look at more than maybe 10 items. If the list grows longer than that, chances are good that items 11+ won't ever make it to the top
<@smilner:fedora.im>
19:44:32
And combining two different streams of work. Fedora Infra Projects and CPE Initiatives
<@amoloney:fedora.im>
19:44:32
CPE, and others, can pull from this list and work on those items that are known
<@humaton:fedora.im>
19:44:57
I think that there the CPE PO role comes in when the person is responsible for the list to move
<@jistone:fedora.im>
19:45:21
I think we may need rough buckets rather than absolute priorities. We're not likely to have a firm idea about one "tail" item vs another.
<@zbyszek:fedora.im>
19:45:56
Yeah, and also obviously people work asynchronously and in parallel on things…
<@nirik:matrix.scrye.com>
19:46:00
The list definitely needs to be a living thing... not just a fire and forget
<@smilner:fedora.im>
19:46:23
Stephen Gallagher: I _think_ the priority framework laid out should help avoid a lot of time from FESCo. Items may fall off as it's a living list and that's OK! Priorities change with need.
<@amoloney:fedora.im>
19:46:25
Good thing you are in both worlds jednorozec :)
<@smilner:fedora.im>
19:46:53
But it's still good feedback! Thank you Stephen Gallagher 🙏
<@amoloney:fedora.im>
19:47:12
What does 'tail' items mean? I think I understand the sentiment here, but want to make sure
<@sgallagh:fedora.im>
19:47:36
I'm not trying to say "no". I'm just concerned about how list items will be managed as time goes on.
<@jistone:fedora.im>
19:47:39
Aoife Moloney: I mean the long tail of lower priority items
<@amoloney:fedora.im>
19:48:11
Ah, thats actually *not* what I thought that meant 🤣
<@sgallagh:fedora.im>
19:48:23
If Steve and Aoife are planning to curate it and drop (or ask to drop) items that have fallen out of favor or reasonable expectation of completion, then I'm happy
<@amoloney:fedora.im>
19:48:27
but I think the buckets approach could be interesting!
<@mhayden:fedora.im>
19:48:41
so do we need to take a vote on this one?
<@amoloney:fedora.im>
19:48:50
Im not actually involved in the list per se
<@humaton:fedora.im>
19:48:59
That will be my job
<@zbyszek:fedora.im>
19:49:08
I think we should vote, just to have a formal resolution.
<@humaton:fedora.im>
19:49:11
outside of FESCO responsibilities...
<@smilner:fedora.im>
19:49:21
mhayden: I don't believe so. This was meant to add clarification to the proposal we made earlier this year and allow further discussion by FESCo moving forward.
<@amoloney:fedora.im>
19:49:26
Im here to facilitate the information passing from one group to another and help with process is all
<@sgallagh:fedora.im>
19:49:28
mhayden: I'm not sure that's warranted. This definitely feels like something that qualifies as part of FESCo's charter to guide the technical direction of the Fedora Project.
<@sgallagh:fedora.im>
19:49:34
But I'm fine if people want to reaffirm that
<@zodbot:fedora.im>
19:49:58
sgallagh gave a cookie to humaton. They now have 53 cookies, 4 of which were obtained in the Fedora 39 release cycle
<@amoloney:fedora.im>
19:50:43
Well a vote that you will accept or reject this additional task would probably be good to have for propriety
<@mhayden:fedora.im>
19:50:48
thanks for the detailed proposal here 👏
<@nirik:matrix.scrye.com>
19:50:55
(and we should update fesco docs, etc)
<@amoloney:fedora.im>
19:51:01
indeed
<@smilner:fedora.im>
19:52:16
As a follow up I'll work with Aoife Moloney to ensure we have the process updated on the wiki for review as well as the tldr document/presentation for it. Thank you all for providing the time, feedback, and discussion ❤️
<@amoloney:fedora.im>
19:52:20
may I suggest, if you are willing to accept this as something FESCo will 'do', maybe accept it conditionally - with a trial period and feedback points
<@zbyszek:fedora.im>
19:52:40
mhayden: I'm fine with thatever procedure. You're the boss today.
<@sgallagh:fedora.im>
19:52:42
Proposal: FESCo agrees to assume responsibility for reviewing and aiding in the prioritization of Fedora Infrastructure Initiatives. The list will be maintained and presented to FESCo by the CPE team as needed.
<@tstellar:fedora.im>
19:52:57
+1
<@zbyszek:fedora.im>
19:53:02
+1
<@humaton:fedora.im>
19:53:04
+1
<@mhayden:fedora.im>
19:53:04
Proposal for vote is to accept the additional task of reviewing priorities for the CPE Infrastructure Projects as defined at: https://fedoraproject.org/wiki/CPE_Infrastructure_Projects ?
<@jistone:fedora.im>
19:53:43
+1
<@mhayden:fedora.im>
19:53:43
I've just been notified that I must depart at the top of the hour to do an unexpected school pickup 😱
<@nirik:matrix.scrye.com>
19:53:49
+1
<@conan_kudo:matrix.org>
19:54:07
my only real quibble is using discourse for discussion, but that can be shopped later
<@sgallagh:fedora.im>
19:54:09
mhayden: I'll take over chair if you like
<@conan_kudo:matrix.org>
19:54:13
let's give it a shot +1
<@mhayden:fedora.im>
19:54:36
i'm +1 on this (and thank you for offering Stephen Gallagher , i will take you up on that)
<@mhayden:fedora.im>
19:54:53
i see (+4, 0, 0) so far
<@sgallagh:fedora.im>
19:54:55
I count +8
<@mhayden:fedora.im>
19:55:11
oof, i missed the highest votes up there
<@sgallagh:fedora.im>
19:55:44
The only one I don't see is from dcantrell, but he's not here.
<@mhayden:fedora.im>
19:55:50
!agreed FESCo accepts the task of reviewing priorities for CPE Infrastructure Projects (+8, 0, -0)
<@sgallagh:fedora.im>
19:56:27
OK, I'll take over so mhayden can go rescue his kid(s)
<@mhayden:fedora.im>
19:56:33
Stephen Gallagher: can i pass you the torch? the only thing i had left was the F40 stuff
<@mhayden:fedora.im>
19:56:45
!topic #3173 F40 Change Proposal Status: Incomplete Changes
<@jvanek:fedora.im>
19:56:51
...only...
<@mhayden:fedora.im>
19:56:52
.fesco https://pagure.io/fesco/issue/3173
<@adamwill:fedora.im>
19:56:58
do we still have quorum? we need some decisions here
<@humaton:fedora.im>
19:57:06
we do
<@sgallagh:fedora.im>
19:57:10
adamw: We have 7 remaining
<@mhayden:fedora.im>
19:57:30
passing the chair to Stephen Gallagher -- thanks all! 👋 dad taxi underway
<@conan_kudo:matrix.org>
19:57:36
I've got some slightly unhappy things about this too
<@sgallagh:fedora.im>
19:58:03
OK, adamw do you want to go through them in any particular order?
<@sgallagh:fedora.im>
19:58:13
Or Conan Kudo, do you have a specific one you want to address first?
<@conan_kudo:matrix.org>
19:58:28
yeah mine should be quick ish
<@adamwill:fedora.im>
19:58:29
so, we had blocker review a few hours ago, and two of the outstanding unfinished Changes requested FEs to land late (very late)
<@adamwill:fedora.im>
19:58:50
we approved both conditionally upon FESCo's decision, since it seems like Changes landing after the freeze should require both FESCo approval and an FE, to me
<@adamwill:fedora.im>
19:59:04
they are https://fedoraproject.org/wiki/Changes/LLVM-18 and https://fedoraproject.org/wiki/Changes/Java21
<@conan_kudo:matrix.org>
19:59:14
let's address that first and then come to my thing
<@sgallagh:fedora.im>
19:59:24
oof, those are two not-small Changes.
<@humaton:fedora.im>
19:59:33
:/
<@jvanek:fedora.im>
19:59:45
java21 people are here .. jsut in case you need some answers
<@nirik:matrix.scrye.com>
20:00:01
llvm had the same thing last time iirc... due to the release cycle upstream...
<@adamwill:fedora.im>
20:00:01
the Java update has been filed, but I see it failed tests. I will have to see why. https://bodhi.fedoraproject.org/updates/FEDORA-2024-129d8ca6fc
<@adamwill:fedora.im>
20:00:06
the LLVM update I think is not filed yet.
<@tstellar:fedora.im>
20:00:07
The LLVM-18 proposal says the update will be done as a beta freeze exception, so this is not unexepected.
<@zbyszek:fedora.im>
20:00:09
jiri vanek++, thank you for coming here and participating in the discussion.
<@zodbot:fedora.im>
20:00:11
zbyszek gave a cookie to jvanek. They now have 5 cookies, 1 of which were obtained in the Fedora 39 release cycle
<@adamwill:fedora.im>
20:00:20
Tom Stellard: ah, i had missed that bit.
<@adamwill:fedora.im>
20:00:25
František Zatloukal: is around for the LLVM change
<@sgallagh:fedora.im>
20:00:26
Yeah, the LLVM one we knew about well in advance.
<@zbyszek:fedora.im>
20:00:51
František Zatloukal++, thank you for coming here and participating in the discussion too.
<@zodbot:fedora.im>
20:00:54
zbyszek gave a cookie to frantisekz. They now have 45 cookies, 1 of which were obtained in the Fedora 39 release cycle
<@sgallagh:fedora.im>
20:01:12
So that one, I'm inclined to just wave on through unless a FESCo member motions to reconsider.
<@adamwill:fedora.im>
20:01:12
i'll note on the Java one that fesco didn't really provide any strong guidance on it. last week you just info'ed it as "INFO: Changes/Java21 is in progress, we'll revisit after Beta Freeze starts", which feels a bit weak, honestly
<@sgallagh:fedora.im>
20:01:38
adamw: I think we were hoping it would land ahead of the Freeze :)
<@conan_kudo:matrix.org>
20:01:56
everything got done at this point in a megaupdate, right?
<@jvanek:fedora.im>
20:02:03
we were hoping too:(
<@adamwill:fedora.im>
20:02:05
okay, the f40 update is failing for the exact same reason the rawhide update failed. i guess they did not rebuild dogtag-pki against java-21 in the update properly.
<@jvanek:fedora.im>
20:02:27
adamw: I had filled PRs agaisnt dogtag
<@jvanek:fedora.im>
20:02:31
to fxit
<@pmikova:fedora.im>
20:02:40
adamw: we wanted to wait for the results of the discussion before merging
<@jvanek:fedora.im>
20:02:47
I can merge it and add immediately
<@pmikova:fedora.im>
20:02:51
but there are PRs for that in the repo
<@sgallagh:fedora.im>
20:02:58
jiri vanek: Just to level-set: what is the practical effect if we were to say "sorry, defer this to F41"?
<@conan_kudo:matrix.org>
20:02:58
!link https://src.fedoraproject.org/rpms/dogtag-pki/pull-request/9
<@zbyszek:fedora.im>
20:03:04
Please somebody correct me if I'm wrong, but my understanding is that both of those changes are unlikely to cause problems for the live cds and other deliverables. The only problem would be if they break the compose, but that should be easy to fix.
<@conan_kudo:matrix.org>
20:03:04
!link https://src.fedoraproject.org/rpms/dogtag-pki/pull-request/10
<@conan_kudo:matrix.org>
20:03:25
It affects Fedora Server because FreeIPA is partially Java.
<@adamwill:fedora.im>
20:03:26
zbyszek: the primary impact of Java is on FreeIPA, which is a release-blocking part of Fedora with a lot of stuff written in Java
<@sgallagh:fedora.im>
20:03:32
zbyszek: Dogtag is part of FreeIPA which is a blocking feature of Server Edition
<@adamwill:fedora.im>
20:03:47
from Rawhide testing we know at least the bits of that path that openQA tests are working now, but that testing certainly doesn't cover anything close to all of FreeIPA's functionality
<@jvanek:fedora.im>
20:03:50
Stephen Gallagher: We will need to undo few commits (not all) and double check maven
<@zbyszek:fedora.im>
20:04:01
I think that deferring at this point, where the work is essentially done, would be extremely punitive.
<@jvanek:fedora.im>
20:04:08
...rebuild maven+jpakagetools and that stuff
<@adamwill:fedora.im>
20:04:42
and just in case anyone needs a refresher on the schedule: the early release target is next tuesday, which means we need to build images tomorrow or latest wednesday to hit that date. there *are* other risk factors there, though (we still don't have backgrounds, and we're waiting on one arm blocker, and we still don't have new signed shim)
<@adamwill:fedora.im>
20:04:55
zbyszek: i would like to argue against that somewhat
<@sgallagh:fedora.im>
20:05:00
jiri vanek: I don't necessarily mean the actual work required to defer. I mean what would the other effects be: would this block other work or force us to carry an EOL version of Java, etc.
<@adamwill:fedora.im>
20:05:10
we cannot normalize the situation where you get a pass to land your change after freeze. that is not how this is supposed to work
<@jvanek:fedora.im>
20:05:35
No EOLed versions of java should beenforced, jdk17 will be maintained for a year at least
<@adamwill:fedora.im>
20:05:36
i would like us to set and communicate stronger expectations for the f41 cycle and later where you should expect that if your change is not substantially done by beta freeze *it is not going into the release*
<@nirik:matrix.scrye.com>
20:06:02
yeah.
<@jvanek:fedora.im>
20:06:02
It will block some progress we wanted to do in java
<@zbyszek:fedora.im>
20:06:07
adamw: I can support such a stance going forward.
<@conan_kudo:matrix.org>
20:06:12
I'm also starting to wonder if we need to adjust the whole schedule for spring releases
<@adamwill:fedora.im>
20:06:14
this is already really how it's supposed to be, but we don't seem to be holding that line strongly enough
<@conan_kudo:matrix.org>
20:06:27
we keep having problems with spring releases
<@nirik:matrix.scrye.com>
20:06:45
also, I think if its possible to miss, bringing that up sooner is much better thank asking forgiveness after...
<@jvanek:fedora.im>
20:06:53
Conan Kudo: I guess thats because most of the work should be done during christmas :)
<@adamwill:fedora.im>
20:06:53
Conan Kudo: there is no magic release cycle where everything lands perfectly the week before freeze. there are just going to be things that don't get in
<@conan_kudo:matrix.org>
20:07:18
adamw: no, but having three weeks dead in the water does massively impact things
<@nirik:matrix.scrye.com>
20:07:29
also, I think if its possible to miss, bringing that up sooner is much better than asking forgiveness after...
<@sgallagh:fedora.im>
20:07:38
adamw: It's very much a balancing act, because there's also negative press associated with shipping with older tech. But I agree, the line as it stands today is too fuzzy.
<@humaton:fedora.im>
20:08:04
having more time is always better. but we also need to enforce the rules about submitting changes
<@conan_kudo:matrix.org>
20:08:06
and this release was particularly bad because the mass build problems we had
<@conan_kudo:matrix.org>
20:08:22
so it might also be that we would not have as much slippage _everywhere_ if it weren't for that
<@sgallagh:fedora.im>
20:08:40
Added to that was the sheer amount of stuff being shoveled in to make the cut before CentOS Stream 10 forked.
<@sgallagh:fedora.im>
20:08:57
So this is definitely a Perfect Storm of problems this cycle
<@conan_kudo:matrix.org>
20:09:02
yes, the tight coupling of RHEL and Fedora schedules does not help
<@conan_kudo:matrix.org>
20:09:38
we should try to see if we can do something about it before Fedora 46...
<@frantisekz:fedora.im>
20:09:42
apart from these potential clarifications, should we say beta freeze is the hard deadline, it would make sense not somehow monitor the period between the beta release (unfreeze) and final freeze, because, in my opinion, prohibiting some late major change from landing during the ebta freeze and allowing it before the final freeze starts brings its own problems...
<@sgallagh:fedora.im>
20:09:52
But let's try not to get too caught up in how we got here and let's focus on how to proceed on these specific cases.
<@adamwill:fedora.im>
20:10:17
yeah, as part of this i want it to be clear that "land it after beta" is not an allowable alternative
<@jvanek:fedora.im>
20:10:29
Stephen Gallagher: one more point - the java reelase cycle is now every two years. One of the main reasons to pushit now, is to have at least three fedoras to ready to next one. Two maybe to few
<@sgallagh:fedora.im>
20:10:36
František Zatloukal: That would not happen (or it would be forcibly reverted)
<@nirik:matrix.scrye.com>
20:10:45
I'm +1 to letting both thru this time, but I'm also in favor of being more strict moving forawrd and communicating that out
<@zbyszek:fedora.im>
20:11:08
nirik: same here
<@conan_kudo:matrix.org>
20:11:14
I think we also need to be firmer with both the toolchain team and the Java folks that stuff needs to come in _earlier_
<@conan_kudo:matrix.org>
20:11:36
at least particularly with the toolchain folks, they've been reluctant to ship stuff earlier, and I think we need to be firmer on them doing it anyway
<@jvanek:fedora.im>
20:11:58
If you check my history of system jdk bumps, all were done in time
<@conan_kudo:matrix.org>
20:12:00
because it's chaotic when it arrives so late
<@sgallagh:fedora.im>
20:12:04
Conan Kudo: I've had conversations with the gcc and glibc folks about getting their Changes and plans filed earlier. I fully expect that to go more smoothly next time
<@jvanek:fedora.im>
20:12:04
this time sicknesses stepped in
<@jistone:fedora.im>
20:12:13
In the spring, it doesn't help that upstream *also* can't move schedules up much due to winter breaks, even when we have that influence.
<@conan_kudo:matrix.org>
20:12:17
jiri vanek: that's totally okay
<@jvanek:fedora.im>
20:12:25
As small lift off, this one was the smoothest one
<@conan_kudo:matrix.org>
20:12:40
don't get me wrong, jiri vanek you did pretty well here
<@jvanek:fedora.im>
20:12:46
as jdk21 as pretty compatible with 17 (unlike 17 to 11 and disasterous 11 to8)
<@sgallagh:fedora.im>
20:12:58
jiri vanek: I'm sorry people got sick, but raising that as a concern about delivery would have been helpful.
<@jistone:fedora.im>
20:12:59
and LLVM also has library ABI to worry about, if we push too early a release candidate
<@jvanek:fedora.im>
20:13:20
Conan Kudo: no wrong taken, thank you for kudos
<@conan_kudo:matrix.org>
20:13:38
Josh Stone: a concern to raise upstream is that maybe RCs shouldn't have unstable ABI
<@jvanek:fedora.im>
20:13:43
Stephen Gallagher: right you are. point noted
<@conan_kudo:matrix.org>
20:13:53
I don't know if LLVM does betas, but it should consider doing so instead
<@frantisekz:fedora.im>
20:13:58
I guess the process would then to be to do the rebuilds when bumping to a later rcs for llvm, but not sure if the llvm has got all the manpower they'd need to do so
<@conan_kudo:matrix.org>
20:14:02
for the ABI unstable part
<@frantisekz:fedora.im>
20:14:23
I guess the process would then to be to do the rebuilds when bumping to a later rcs for llvm, but not sure if the llvm team has got all the manpower they'd need to do so
<@conan_kudo:matrix.org>
20:14:37
but regardless, rebuilds to accept RCs are less painful than everything breaking because massive churn
<@conan_kudo:matrix.org>
20:15:07
they suck, but it's less bad because stakeholders like Mesa and clazy and others get more notice everything is breaking
<@jistone:fedora.im>
20:15:19
Rust makes LLVM ABI rebuilds even harder, since it bootstraps with whatever it used before
<@sgallagh:fedora.im>
20:15:19
Josh Stone: Correct me if I'm wrong, but RCs are *intended* to be ABI-stable, but history shows that there's usually one or two unexpected changes?
<@zbyszek:fedora.im>
20:15:28
A rebuild, if the changes are small, and no new FTBFS are expected, is actually not much of a problem.
<@tstellar:fedora.im>
20:15:41
Stephen Gallagher: They aren't intended to be ABI stable.
<@zbyszek:fedora.im>
20:15:45
Esp. if the rebuild doesn't cause a cascade of futher required rebuilds.
<@conan_kudo:matrix.org>
20:15:57
RCs that aren't :D
<@zbyszek:fedora.im>
20:16:22
I.e. an ABI break is not much of a problem. An API break could be.
<@sgallagh:fedora.im>
20:16:30
Tom Stellard: Honestly, I'm inclined to suggest that we don't rebase to newer LLVM's until the autumn Fedora release, then.
<@amoloney:fedora.im>
20:16:36
So, communication went out ahead of both Testable and Complete deadlines, to both RHEL and Fedora lists. However, human nature and all, I am happy to take an action to be more direct in reaching out to change owners in the bz trackers. This might not solve unresponsiveness, but I am willing to acknowledge I can do more and try to improve the situation for F41 onwards
<@amoloney:fedora.im>
20:16:50
So, communication went out ahead of both Testable and Complete deadlines a week in each case, to both RHEL and Fedora lists. However, human nature and all, I am happy to take an action to be more direct in reaching out to change owners in the bz trackers. This might not solve unresponsiveness, but I am willing to acknowledge I can do more and try to improve the situation for F41 onwards
<@conan_kudo:matrix.org>
20:16:56
right, I'm actually okay with ABI breaks
<@tstellar:fedora.im>
20:17:02
Stephen Gallagher: We can discuss this when I submit the LLVM-19 change proposal.
<@sgallagh:fedora.im>
20:17:24
Right, we DID already pre-approve that one landing late.
<@conan_kudo:matrix.org>
20:17:25
my concern is things don't build and nobody has time to fix them
<@conan_kudo:matrix.org>
20:17:29
because the API changed
<@sgallagh:fedora.im>
20:17:39
So let's focus on Java 21
<@tstellar:fedora.im>
20:18:10
Also, the LLVM updates provides a compat version so things can continue to work with the old version.
<@adamwill:fedora.im>
20:18:20
Aoife Moloney: what i'm hoping for in future is a bit stronger feedback from fesco and the change check-in messages - not just "we'll check in again in a week!" but "if you don't get this landed by freeze, assume it is going in the next release"
<@tstellar:fedora.im>
20:18:38
Also, the LLVM updates provide a compat version so things can continue to work with the old version.
<@frantisekz:fedora.im>
20:18:59
afaik the llvm compat packages are also full-blown from now on (incl. clang compat binaries, correct me if I am wrong here)
<@jvanek:fedora.im>
20:19:02
adamw: Does it imply no possible beta delay?
<@tstellar:fedora.im>
20:19:16
František Zatloukal: They are.
<@amoloney:fedora.im>
20:19:18
Understood. I am just willing to help in other areas too where I can
<@sgallagh:fedora.im>
20:19:27
jiri vanek: I'm not sure I understand your question.
<@sgallagh:fedora.im>
20:19:36
Beta should only be delayed for unresolved blocker bugs.
<@sgallagh:fedora.im>
20:19:46
Not getting a new feature in is NOT a blocking event
<@jvanek:fedora.im>
20:20:03
I see
<@nirik:matrix.scrye.com>
20:20:19
I think its swung back and forth over time... but I agree we have swung back to the too easygoing side of late.
<@adamwill:fedora.im>
20:20:36
jiri vanek: the messaging "you should assume it's going in next release" doesn't entirely lock off the possibility of us granting an exception to land a change late. but it clearly sets the expectation that this is not the default
<@conan_kudo:matrix.org>
20:20:38
we got used to things working out
<@conan_kudo:matrix.org>
20:20:47
and now that they aren't, we need to recalibrate
<@adamwill:fedora.im>
20:20:56
i would like us to still have the option to, as a special case, land things late, where it really makes sense. but it can't be *expected*.
<@jvanek:fedora.im>
20:21:00
Stephen Gallagher: adamw ok, understood
<@sgallagh:fedora.im>
20:21:32
jiri vanek: Is the Java 21 stuff /finished/ and just waiting for FE approval, or does work remain?
<@sgallagh:fedora.im>
20:21:40
It sounded above like dogtag was still not ready
<@jvanek:fedora.im>
20:21:42
it is finished
<@jvanek:fedora.im>
20:21:57
If you will ack it, I will merger PR for dogtail, build it and add it to the update
<@jvanek:fedora.im>
20:22:00
that shoudl be all
<@adamwill:fedora.im>
20:22:07
that means it's not finished
<@jvanek:fedora.im>
20:22:22
I can merge it right away, before the ack. It jsut seemd weird
<@adamwill:fedora.im>
20:22:37
i mean, it's no weirder than you already having done the builds and the update
<@sgallagh:fedora.im>
20:22:48
Can I see the changes to dogtag? How invasive are they?
<@adamwill:fedora.im>
20:22:50
since you went ahead and did those anyway, you may as well make them *right*
<@conan_kudo:matrix.org>
20:23:09
Stephen Gallagher: I linked them earlier
<@pmikova:fedora.im>
20:23:10
https://src.fedoraproject.org/rpms/dogtag-pki/pull-requests
<@jvanek:fedora.im>
20:23:12
Stephen Gallagher: not at all
<@jvanek:fedora.im>
20:23:17
they had hardcoded jdk17
<@jvanek:fedora.im>
20:23:26
Adam fixed it to hardcode 21
<@sgallagh:fedora.im>
20:23:28
adamw: I understand your frustration, but your tone is becoming a little confrontational.
<@jvanek:fedora.im>
20:23:43
I had improved it today by making it properly version-less jdk
<@adamwill:fedora.im>
20:23:57
sorry, i didn't mean it that way, but i agree it sounded that way :) i just meant that fixing the problem in the thing that is already there doesn't make anything worse
<@adamwill:fedora.im>
20:24:20
jiri vanek: did you check with the dogtag team that is actually what the ywant?
<@conan_kudo:matrix.org>
20:24:21
the fix jiri vanek made is the correct and proper fix
<@conan_kudo:matrix.org>
20:24:31
you shouldn't be locking on to specific java versions in Fedora anyway
<@adamwill:fedora.im>
20:24:32
my understanding is that they do not *want* to bump to new java releases until they've tested them upstream
<@nirik:matrix.scrye.com>
20:24:43
perhaps also messaging should include "going to the next release. if you would like an exception, ask, but DO NOT DO ANY work on it yet"?
<@adamwill:fedora.im>
20:24:45
i don't know whether that's sensible
<@conan_kudo:matrix.org>
20:25:05
I don't think that's a sensible stance to have
<@conan_kudo:matrix.org>
20:25:18
we don't generally allow that for any other stack
<@adamwill:fedora.im>
20:25:21
Stephen Gallagher: anyway, the change is small in one sense - it's just changing the package's build dependencies - but large in another sense - the dogtag team has not yet tested to their satisfaction that dogtag actually works okay with java 21 (AIUI).
<@nirik:matrix.scrye.com>
20:25:29
well, it depends on how hard it is to back out I guess.
<@adamwill:fedora.im>
20:25:48
we know from rawhide testing that it works well enough that the openQA smoke tests pass.
<@sgallagh:fedora.im>
20:25:57
adamw: Looking at the diff, it's changing from explicitly depending on 21 to depending on whatever's available.
<@sgallagh:fedora.im>
20:26:06
So I assume they already made the transition to 21
<@sgallagh:fedora.im>
20:26:14
https://src.fedoraproject.org/rpms/dogtag-pki/pull-request/10#request_diff for reference
<@adamwill:fedora.im>
20:26:30
Stephen Gallagher: no, the change to 21 was one i did just recently
<@adamwill:fedora.im>
20:26:33
to make rawhide stop failing
<@sgallagh:fedora.im>
20:26:38
... oh
<@jvanek:fedora.im>
20:26:44
..right
<@sgallagh:fedora.im>
20:27:36
Conan Kudo: That approach only works if the interpreter lands much MUCH earlier in the cycle
<@nirik:matrix.scrye.com>
20:28:11
so where are we here? still gathering info?
<@conan_kudo:matrix.org>
20:28:14
Stephen Gallagher: outside of this incident, it generally _has_
<@adamwill:fedora.im>
20:28:20
the PR is actually a bit weird, since if you look at what's actually *in* the f40 branch right now, it says 17, not 21. https://src.fedoraproject.org/rpms/dogtag-pki/blob/f40/f/dogtag-pki.spec#_66
<@sgallagh:fedora.im>
20:28:25
Fair enough
<@adamwill:fedora.im>
20:28:33
so i dunno how that PR thinks the base is -21, but anyhow.
<@zbyszek:fedora.im>
20:28:37
proposal: Late merging of Changes/Java21 is accepted, provided that it happens before Friday March 8th. Otherwise the change will be postponed to F41.
<@sgallagh:fedora.im>
20:28:55
Sadly, based on this discussion I'm inclined to vote -1
<@conan_kudo:matrix.org>
20:29:00
zbyszek: +1
<@jvanek:fedora.im>
20:29:07
adamw: of coourse it says 17, as system jdk in f40 is 17
<@sgallagh:fedora.im>
20:29:09
I think it's too big, too late. I wish it were otherwise.
<@nirik:matrix.scrye.com>
20:29:13
why friday?
<@jvanek:fedora.im>
20:29:18
with that PR, it will work with both
<@adamwill:fedora.im>
20:29:39
jiri vanek: the weird thing is the diff in https://src.fedoraproject.org/rpms/dogtag-pki/pull-request/10#request_diff showing the original lines as java-21, not java-17. no idea how that's happening.
<@nirik:matrix.scrye.com>
20:29:56
is it on top of the other pr?
<@jvanek:fedora.im>
20:30:05
adamw: the merge commit
<@adamwill:fedora.im>
20:30:15
ugh, I hate merges.
<@jvanek:fedora.im>
20:30:22
So do I .
<@jvanek:fedora.im>
20:30:31
I can take cherrypick
<@zbyszek:fedora.im>
20:30:37
nirik: just a semi-random date that is this week, so it's "short", but actually there's some time to merge the PRs calmly and rebuild.
<@adamwill:fedora.im>
20:30:49
zbyszek: friday does not work for the release process
<@nirik:matrix.scrye.com>
20:30:57
so that means we miss the eary date for sure
<@jvanek:fedora.im>
20:30:57
but dogtag-maintainers were living in pretty linera history, I would like to come to agreement with them
<@zbyszek:fedora.im>
20:31:01
adamw: when do we need it then?
<@adamwill:fedora.im>
20:31:02
to actually get this in, if we don't slip for anything else, it needs to be pushed stable by tomorrow or wednesday at the latest
<@conan_kudo:matrix.org>
20:31:10
it would have to be tomorrow
<@adamwill:fedora.im>
20:31:13
the chances that we slip for something else are not zero, but we have to *assume* we won't
<@sgallagh:fedora.im>
20:31:21
nirik: Tom Stellard Josh Stone : Votes?
<@zbyszek:fedora.im>
20:31:21
OK, let me redo the proposal.
<@adamwill:fedora.im>
20:31:33
that should be possible, rebuilding dogtag takes a half hour or so
<@jistone:fedora.im>
20:31:40
Stephen Gallagher: was holding for the clarifications
<@zbyszek:fedora.im>
20:31:47
proposal: Late merging of Changes/Java21 is accepted, provided that it happens no later than Wednesday March 6th. Otherwise the change will be postponed to F41.
<@adamwill:fedora.im>
20:31:51
then openqa tests take another hour or so. so we can have this landed in two hours. just whether we want to accept the risk.
<@sgallagh:fedora.im>
20:32:07
I'm still -1
<@conan_kudo:matrix.org>
20:32:22
zbyszek: +1
<@conan_kudo:matrix.org>
20:32:46
with firm finger wagging for everyone we will NOT do this again
<@tstellar:fedora.im>
20:32:47
0
<@nirik:matrix.scrye.com>
20:32:55
+1 i guess... perhaps I am falling prey to the sunk cost fallicy. ;)
<@jistone:fedora.im>
20:33:04
+1
<@zbyszek:fedora.im>
20:33:23
(Sorry for "Friday". I know it was mentioned earlier in the chat, but I'm always a bit fuzzy on both the days of the week and details of the compose process.)
<@humaton:fedora.im>
20:33:52
my keyboard does not want to write it but +1
<@sgallagh:fedora.im>
20:34:50
I count (+4, 1, -0) so far. We need one more +1 for it to pass.
<@zbyszek:fedora.im>
20:35:05
I count 5.
<@zbyszek:fedora.im>
20:35:08
I count +5.
<@sgallagh:fedora.im>
20:35:17
Sorry I count (+4, 1, -1)
<@jistone:fedora.im>
20:35:19
you didn't count your own -1?
<@jistone:fedora.im>
20:35:22
ah
<@sgallagh:fedora.im>
20:35:27
Mistyped
<@conan_kudo:matrix.org>
20:35:28
if you assume zbyszek votes for his own proposal
<@conan_kudo:matrix.org>
20:35:35
then it's +5
<@zbyszek:fedora.im>
20:35:42
That's the rule.
<@sgallagh:fedora.im>
20:35:47
Ah, I missed jednorozec
<@sgallagh:fedora.im>
20:36:14
#agreed Late merging of Changes/Java21 is accepted, provided that it happens no later than Wednesday March 6th. Otherwise the change will be postponed to F41. (+5, 1, -1)
<@jvanek:fedora.im>
20:36:16
I was never merging a udpate in betafreez with fesco exception. What to do with that?
<@adamwill:fedora.im>
20:36:34
jiri vanek: you don't need to do anything different
<@jvanek:fedora.im>
20:36:45
fair enough...
<@adamwill:fedora.im>
20:36:47
you just need to make sure the update is complete and passes tests
<@adamwill:fedora.im>
20:37:05
i (with my hat as the quality team person who sends the releng team push requests) deal with most of the rest
<@adamwill:fedora.im>
20:37:15
oh, and make sure the update is marked as fixing the bug, too
<@jvanek:fedora.im>
20:37:17
fuuu
<@jvanek:fedora.im>
20:37:18
ok
<@adamwill:fedora.im>
20:37:33
that is, https://bugzilla.redhat.com/show_bug.cgi?id=2267486
<@adamwill:fedora.im>
20:38:07
was there an agreement on llvm already?
<@sgallagh:fedora.im>
20:38:08
#topic Next Week's Chair
<@adamwill:fedora.im>
20:38:18
Stephen Gallagher: you want !, not # :D
<@humaton:fedora.im>
20:38:23
i can chair
<@sgallagh:fedora.im>
20:38:25
Dammit
<@sgallagh:fedora.im>
20:38:36
!agreed Late merging of Changes/Java21 is accepted, provided that it happens no later than Wednesday March 6th. Otherwise the change will be postponed to F41. (+5, 1, -1)
<@adamwill:fedora.im>
20:39:05
did I miss an agreement on llvm?
<@sgallagh:fedora.im>
20:39:07
Regarding LLVM, we (FESCo) previously approved the late landing of LLVM. I asked for anyone to speak against it and no one did
<@adamwill:fedora.im>
20:39:16
ah, okay, so that is considered approved. cool
<@conan_kudo:matrix.org>
20:39:39
yeah, we did so because they're always late
<@sgallagh:fedora.im>
20:39:43
I didn't see a reason to re-vote since no one expressed a concern
<@sgallagh:fedora.im>
20:40:16
!topic Next Week's Chair
<@sgallagh:fedora.im>
20:40:30
!action jednorozec will chair the 2024-03-11 meeting
<@sgallagh:fedora.im>
20:40:54
!topic Open Floor
<@zbyszek:fedora.im>
20:41:17
Conan Kudo: had something?
<@sgallagh:fedora.im>
20:41:23
I have a hard stop at the top of the hour (actually a few minutes prior).
<@conan_kudo:matrix.org>
20:41:28
yes, I did
<@sgallagh:fedora.im>
20:41:35
Oh, darn. Right, I forgot there were other late Changes.
<@adamwill:fedora.im>
20:41:52
sorry i hijacked that a bit
<@conan_kudo:matrix.org>
20:42:08
no worries that's why I let you go first :)
<@sgallagh:fedora.im>
20:42:11
!topic Late F40 Changes (part deux)
<@conan_kudo:matrix.org>
20:43:04
So this is about the MySQL change this cycle...
<@conan_kudo:matrix.org>
20:43:08
!fesco 3092
<@zodbot:fedora.im>
20:43:10
**fesco #3092** (https://pagure.io/fesco/issue/3092):**Change: MariaDB & MySQL repackaging**
● **Closed: Accepted** 4 months ago by zbyszek
● **Opened:** 4 months ago by amoloney
● **Last Updated:** 4 months ago
● **Assignee:** mschorm
<@conan_kudo:matrix.org>
20:43:37
So as we discussed in our previous meeting, I was doing the review request for the rename
<@conan_kudo:matrix.org>
20:43:44
in rhbz#2252763
<@conan_kudo:matrix.org>
20:43:45
!bug 2252763
<@zodbot:fedora.im>
20:43:47
RHBZ#2252763 (https://bugzilla.redhat.com/2252763): [Package Review]: Rename Request: mysql - MySQL client programs and shared libraries
<@conan_kudo:matrix.org>
20:44:16
The problem is that they decided to circumvent the rename process by using the carve out we have multiversioned packages
<@conan_kudo:matrix.org>
20:44:42
even though no base mysql package exists in Fedora right now, a mysql8.0 multiversion package was requested and granted
<@humaton:fedora.im>
20:44:48
huh
<@jvanek:fedora.im>
20:45:00
thanx all for participating, I will now discconect to my normal timezone
<@jvanek:fedora.im>
20:45:09
Thank you for continuing doing fedora what it is
<@conan_kudo:matrix.org>
20:45:11
by jiri vanek
<@conan_kudo:matrix.org>
20:45:14
bye jiri vanek
<@conan_kudo:matrix.org>
20:45:17
*bye
<@jvanek:fedora.im>
20:45:18
and hoping to finish the jdk21 wihtout issues
<@jvanek:fedora.im>
20:45:21
bye!
<@conan_kudo:matrix.org>
20:45:26
good luck!
<@conan_kudo:matrix.org>
20:45:39
anyway... I don't know how to respond to https://bugzilla.redhat.com/show_bug.cgi?id=2252763#c18
<@conan_kudo:matrix.org>
20:46:02
and I figured I'd bring it up here because I'm really not sure how to handle this scenario
<@sgallagh:fedora.im>
20:46:06
Conan Kudo: Meaning they didn't perform a re-review?
<@conan_kudo:matrix.org>
20:46:13
yes
<@conan_kudo:matrix.org>
20:46:17
they skipped it
<@sgallagh:fedora.im>
20:46:21
Or that they did, didn't like the feedback, and bypassed?
<@nirik:matrix.scrye.com>
20:46:21
Ugh.
<@conan_kudo:matrix.org>
20:46:37
it's been bypassed, yes
<@zbyszek:fedora.im>
20:46:47
Ugh.
<@conan_kudo:matrix.org>
20:47:08
I decided not to respond in the BZ because... well, my emotions were getting too hot.
<@nirik:matrix.scrye.com>
20:47:16
So, is the mysql8.0 package ok? or we dont know yet? and they no longer want a unversiond 'mysql' package?
<@conan_kudo:matrix.org>
20:47:21
And I wanted to ask the group here what we or I should do.
<@conan_kudo:matrix.org>
20:47:34
nirik: no, mysql package still is supposed to exist, but it doesn't.
<@salimma:fedora.im>
20:47:42
for completeness - is there a rel-eng ticket for creating the repo? or was it created by hand
<@conan_kudo:matrix.org>
20:48:37
they instead used mysql8.0 to Provide mysql
<@conan_kudo:matrix.org>
20:48:46
which effectively bypassed what was supposed to happen
<@sgallagh:fedora.im>
20:48:49
nirik: They've only got `community-mysql` right now.
<@sgallagh:fedora.im>
20:49:05
Conan Kudo: It's sketchy, but I'm not sure it's clearly a violation.
<@humaton:fedora.im>
20:49:10
so they requested it and got exception https://pagure.io/releng/fedora-scm-requests/issue/60131
<@humaton:fedora.im>
20:49:19
and the new spec is bundling boost :/
<@humaton:fedora.im>
20:49:31
missing some hints from the review
<@nirik:matrix.scrye.com>
20:49:36
oh man. yeah, they have a mysql8.0 package that makes mysql binary packages.
<@conan_kudo:matrix.org>
20:50:19
yeah, you can guess that was not happy when I realized the extent of this
<@sgallagh:fedora.im>
20:50:22
The bypass was approved by jnsamyak it seems
<@tstellar:fedora.im>
20:50:22
nirik: DIdn't we recommend they do this?
<@conan_kudo:matrix.org>
20:50:24
yeah, you can guess that I was not happy when I realized the extent of this
<@salimma:fedora.im>
20:50:29
to be fair, making mysql from mysqlX.Y itself is fine - Python does this too. but doing this without review is... hmm
<@nirik:matrix.scrye.com>
20:50:37
I don't recall? perhaps?
<@sgallagh:fedora.im>
20:51:00
They definitely violated the spirit of the policy, if not the letter.
<@jistone:fedora.im>
20:51:07
the comments on 3092 suggested multi-versioning
<@nirik:matrix.scrye.com>
20:51:07
FWIW, I really think the exceptions tickets should have some info from the requestor... it's really hard for releng to dig in to each one
<@zbyszek:fedora.im>
20:51:18
I think it's a violation. This is a rename, not a compat package.
<@tstellar:fedora.im>
20:51:19
nirik: Some of us did at least: https://pagure.io/fesco/issue/3092#comment-881964
<@sgallagh:fedora.im>
20:51:19
And bundling boost is... uncomfortable
<@conan_kudo:matrix.org>
20:51:57
I feel that's going to bite us later
<@nirik:matrix.scrye.com>
20:51:58
alright.
<@tstellar:fedora.im>
20:52:01
But anyway there shouldn't be massive changes to the package if it's using the compat exception.
<@sgallagh:fedora.im>
20:52:21
OK, so how do we want to proceed?
<@nirik:matrix.scrye.com>
20:52:35
proposal: tell them in the review that... it still needs reviewed.
<@conan_kudo:matrix.org>
20:52:36
Tom Stellard: the problem is that there's nothing to be compat _of_...
<@sgallagh:fedora.im>
20:52:42
Block the package from composes until a review is completed? (Effectively deferring to F41)?
<@nirik:matrix.scrye.com>
20:53:15
I don't think we need to block it, but it should get reviewed and make corrections/fixes based on that review...
<@humaton:fedora.im>
20:53:21
I would say block the package until the rename process is complete
<@nirik:matrix.scrye.com>
20:53:28
I realize there's a lot less incentive
<@sgallagh:fedora.im>
20:53:29
Can I hand over the chair to someone else? I have a parent-teacher meeting in five minutes. I can rejoin after that if we're still going.
<@humaton:fedora.im>
20:53:31
there is rename process there in packaging guidelines isnt it?
<@conan_kudo:matrix.org>
20:53:35
yes
<@tstellar:fedora.im>
20:53:55
Conan Kudo: I think if they just took the identical spec file and put it in a compat package that would be OK, but that doesn't sound like what happens.
<@conan_kudo:matrix.org>
20:53:57
!link https://docs.fedoraproject.org/en-US/package-maintainers/Package_Renaming_Process/
<@zbyszek:fedora.im>
20:53:58
Stephen Gallagher: yes, go ahead. I or somebody else will take over.
<@tstellar:fedora.im>
20:54:11
I also have to go in ~7 minutes.
<@zbyszek:fedora.im>
20:54:20
https://docs.fedoraproject.org/en-US/packaging-guidelines/ReviewGuidelines/#_package_review_process is also relevant.
<@adamwill:fedora.im>
20:54:23
there are a couple of other late changes that probably need looking at
<@zbyszek:fedora.im>
20:54:44
> exceptions: The package is being created so that multiple versions of the same package can coexist in the distribution (or coexist between EPEL and RHEL). The package MUST be properly named according to the naming guidelines and MUST NOT conflict with all other versions of the same package.
This is not this case.
<@nirik:matrix.scrye.com>
20:54:44
blocking this could/will break things.
<@conan_kudo:matrix.org>
20:54:49
I have to go soon too, but I just wanted to solicit some help here, because I'm kind of broken.
<@conan_kudo:matrix.org>
20:55:41
thanks
<@tstellar:fedora.im>
20:55:47
I think at minimum the package should be reviewed and they should address the feedback.
<@zbyszek:fedora.im>
20:56:03
+1 to Tom Stellard
<@nirik:matrix.scrye.com>
20:56:10
How about I ask them to do that and we revist next week in case they don't / we need to do something else?
<@tstellar:fedora.im>
20:56:15
If we can't block packages for technical reasons, we should at least have them follow the process as much as possible.
<@conan_kudo:matrix.org>
20:56:21
nirik: +1
<@sgallagh:fedora.im>
20:56:27
Proposal: if they do not complete a review, the package will be blocked from release in F40. Decision to be finalized a Final Freeze?
<@tstellar:fedora.im>
20:56:27
nirik: +1
<@zbyszek:fedora.im>
20:56:36
niri
<@zbyszek:fedora.im>
20:56:40
nirik: +1
<@sgallagh:fedora.im>
20:56:43
(and now I'm gone, sorry)
<@nirik:matrix.scrye.com>
20:56:45
if we block this we probibly need to unblock community-mysql and... possibly make a bunch of changes in other packages?
<@conan_kudo:matrix.org>
20:56:50
Stephen Gallagher: +1
<@nirik:matrix.scrye.com>
20:56:54
-1
<@conan_kudo:matrix.org>
20:57:16
either nirik or Stephen's proposals work for me
<@conan_kudo:matrix.org>
20:57:29
I just... I don't have the willpower to force anything right now
<@zbyszek:fedora.im>
20:57:34
Are we allowed to vote +2? Or +11?
<@tstellar:fedora.im>
20:57:38
We don't need FESCOs approval to as them to get a package review.
<@nirik:matrix.scrye.com>
20:57:47
We are in freeze. Right before trying to make rcs after landing a bunch of stuff late. I am -1000 to blocking or unblocking things in f40 right now. ;)
<@tstellar:fedora.im>
20:57:54
We don't need FESCOs approval to ask them to get a package review.
<@conan_kudo:matrix.org>
20:57:55
🤣
<@jistone:fedora.im>
20:58:04
we need it to be clear that they can't just skip process for a deadline
<@humaton:fedora.im>
20:58:16
they just skipped the process
<@tstellar:fedora.im>
20:58:34
It doesn't seem like we are going to be able to agree on a proposal right now since people have to leave and we'll lose quorum.
<@humaton:fedora.im>
20:58:48
that is not nice :/ but nirik is right blocking stuff now will make things harder for QA and releng
<@nirik:matrix.scrye.com>
20:58:51
and there's other changes too?
<@adamwill:fedora.im>
20:58:57
the other changes are https://fedoraproject.org/wiki/Changes/KiwiBuiltCloudImages and https://fedoraproject.org/wiki/Changes/ArmMinimalImageOSBuild
<@conan_kudo:matrix.org>
20:59:10
ugh
<@adamwill:fedora.im>
20:59:11
in a similar vein. affecting the compose, which is already affected by lots of other things
<@conan_kudo:matrix.org>
20:59:24
everything I can do about kiwi built cloud images is done
<@conan_kudo:matrix.org>
20:59:31
I'm just waiting for compose tooling deployment
<@tstellar:fedora.im>
20:59:33
Do we need to have another FESCO meeting later in the week maybe?
<@adamwill:fedora.im>
20:59:36
for the former, we're at the stage of reviewing the pungi PR. for the latter, we've been waiting on a pungi-fedora PR since fesco approved the change late.
<@nirik:matrix.scrye.com>
20:59:57
we also need a new pungi version on compose hosts right?
<@conan_kudo:matrix.org>
21:00:07
yes
<@nirik:matrix.scrye.com>
21:00:09
and just hope it works. what could possibly go wrong?
<@adamwill:fedora.im>
21:00:12
nirik: for which? for kiwi, yeah, once the pungi PR is approved and merged
<@tstellar:fedora.im>
21:00:26
I have to go now, sorry.
<@adamwill:fedora.im>
21:00:37
for arm, i *think* all the pungi/osbuild bits are in place and we only need an image build definition. assuming the hard disk size thing i saw isn't a problem, somehow
<@humaton:fedora.im>
21:00:40
the kiwi change is scary at this time of release cycle
<@humaton:fedora.im>
21:00:52
redeployng new composers in freeze :/
<@nirik:matrix.scrye.com>
21:01:06
yes
<@adamwill:fedora.im>
21:01:14
i ran initial openqa tests on the test compose for the kiwi change and they passed except that chrony is missing (which causes two tests to fail)
<@adamwill:fedora.im>
21:01:27
but, of course, openqa is only a smoke check
<@conan_kudo:matrix.org>
21:01:27
I don't know why chrony is missing
<@conan_kudo:matrix.org>
21:01:30
it should be in the comps
<@adamwill:fedora.im>
21:02:02
anyhow...we need to decide if we try to hero these in late along with everything else or say nope, f41
<@conan_kudo:matrix.org>
21:02:40
I would like to try to get the kiwi stuff in this cycle with a finger wag to myself (even though it's not my fault)
<@nirik:matrix.scrye.com>
21:02:43
oh and we need pdc changes too I guess.
<@humaton:fedora.im>
21:03:05
haha
<@humaton:fedora.im>
21:03:14
PDC is not getting new changes
<@conan_kudo:matrix.org>
21:03:17
because otherwise we'll be stuck with imgfac for two more cycles because we have to wait out the EOLs for cloud biweekly builds
<@adamwill:fedora.im>
21:03:42
nirik: what needs changing in pdc?
<@nirik:matrix.scrye.com>
21:03:42
well, whatever it is that has the image types. Because if we don't, toddlers loops on those compose messagesa forever
<@adamwill:fedora.im>
21:04:01
nirik: i think for both these changes they're intended to 'take over' existing images
<@humaton:fedora.im>
21:04:05
righ this is doable in the toddler
<@adamwill:fedora.im>
21:04:06
so they *should* have the same metadata
<@conan_kudo:matrix.org>
21:04:15
that's the idea
<@humaton:fedora.im>
21:04:23
and if not we can modify the data on the fly in toddler
<@adamwill:fedora.im>
21:04:25
(that's part of what i was trying to get straightened out for kiwi)
<@nirik:matrix.scrye.com>
21:04:45
Conan Kudo: you mean the daily ones? but we could switch those whenever I would think....
<@conan_kudo:matrix.org>
21:05:15
there are people that depend on those, like the copr folks
<@conan_kudo:matrix.org>
21:05:21
and the kubevirt people
<@conan_kudo:matrix.org>
21:05:48
I'm sorry, this is all suck. :(
<@nirik:matrix.scrye.com>
21:06:05
interesting... we don't actually advertise them anywhere do we?
<@conan_kudo:matrix.org>
21:06:26
I'm pretty sure we do
<@conan_kudo:matrix.org>
21:06:33
enough people come by and ask about them when they fail
<@conan_kudo:matrix.org>
21:06:48
the regular failures and missing updated images drove the need to replace it
<@nirik:matrix.scrye.com>
21:07:03
huh because I can't recall anytime anyone has cme by releng about them failing...
<@conan_kudo:matrix.org>
21:07:35
I think that's mostly our fault for not wanting to bother you when we know imgfac is a pile of broken with duct tape
<@zbyszek:fedora.im>
21:08:22
OK, so… We still have 5 voting members, so we can make some decisions.
<@zbyszek:fedora.im>
21:08:36
Somebody who groks the situation, please make a proposal.
<@nirik:matrix.scrye.com>
21:08:43
ok... I'm a bit puzzled by it... would be nice to know anyone cares about those. :)
<@jistone:fedora.im>
21:09:19
I'm here, but not well informed on this infrastructure...
<@conan_kudo:matrix.org>
21:10:12
at this point, the PR for pungi looks pretty done, just needs merging and landing
<@conan_kudo:matrix.org>
21:10:22
!link https://pagure.io/pungi/pull-request/1732
<@nirik:matrix.scrye.com>
21:10:31
I'm willing to try and bodge it in... but if this is the last thing we need before a beta, I'd prefer to drop it than slip...
<@conan_kudo:matrix.org>
21:10:58
the only reason it wasn't merged before is that the CI on pungi is broken on some random py2.7 thing
<@conan_kudo:matrix.org>
21:11:09
and lsedler said he'd do some pinning to fix that
<@conan_kudo:matrix.org>
21:11:31
but it has been tested on staging koji and adamw has okay'd the output
<@conan_kudo:matrix.org>
21:11:46
https://pagure.io/pungi/pull-request/1720#comment-199412
<@conan_kudo:matrix.org>
21:11:55
!link https://pagure.io/pungi/pull-request/1720#comment-199412
<@nirik:matrix.scrye.com>
21:12:29
yes, but then we need them to build packages, then deploy them also and then update the config... so it's not like just merging the pr is it. ;)
<@nirik:matrix.scrye.com>
21:13:48
anyhow, I'm willing to try... adamw? or should we just vote or ?
<@humaton:fedora.im>
21:14:01
but at this time I am not sure about it
<@adamwill:fedora.im>
21:14:11
nirik: ugh. if you're willing to try i guess we can too.
<@nirik:matrix.scrye.com>
21:15:24
and I thought this might be a relaxing week for once... my mistake
<@humaton:fedora.im>
21:15:54
well you can always say nope
<@conan_kudo:matrix.org>
21:16:02
I'm sorry :(
<@conan_kudo:matrix.org>
21:16:30
if it makes you both feel better, I'm definitely going to try harder next time to avoid this situation
<@conan_kudo:matrix.org>
21:16:39
it's not great for me either
<@humaton:fedora.im>
21:16:51
I am here to help out anyway I can to package or test stuff
<@humaton:fedora.im>
21:16:56
or fix composes :D
<@adamwill:fedora.im>
21:17:35
as for the arm one...we kinda need peter or someone to come up with a recipe, i think. or we can try mine from the issue on stg and just see what happens?
<@nirik:matrix.scrye.com>
21:17:37
so, shall we just leave it at that then and close out? or any objections?
<@conan_kudo:matrix.org>
21:17:38
and I'm here to help wherever I'm empowered and capable to do so
<@zbyszek:fedora.im>
21:18:43
WFM.
<@conan_kudo:matrix.org>
21:18:47
nirik: sure. I think the only immediate action to do is an agreed statement on the package review thing
<@conan_kudo:matrix.org>
21:19:06
then we can close out the meeting
<@adamwill:fedora.im>
21:19:56
i think we agreed to try for kiwi?
<@zbyszek:fedora.im>
21:20:00
I think we had +5 votes for nirik's "proposal".
<@conan_kudo:matrix.org>
21:20:32
which one, he made two
<@conan_kudo:matrix.org>
21:20:44
are we talking about poking the mysql bz?
<@nirik:matrix.scrye.com>
21:20:55
yeah, I am not sure either. ;)
<@zbyszek:fedora.im>
21:21:08
Hmm, I think we only had +4 for the bugzilla poking thing.
<@nirik:matrix.scrye.com>
21:21:09
I'm happy to try and get some traction on the bug tho
<@jistone:fedora.im>
21:21:26
I don't think I voted, but +1 on poking review
<@nirik:matrix.scrye.com>
21:21:32
and I don't guess that really needs fesco approval, although it would perhaps carry more weight?
<@conan_kudo:matrix.org>
21:22:09
!agreed The mysql rename review needs to be properly completed (+5, 0, -0)
<@nirik:matrix.scrye.com>
21:22:26
alright, I can update the bug...
<@conan_kudo:matrix.org>
21:22:56
!action nirik will poke bug 2252763 to get the mysql package rename review to finish
<@zbyszek:fedora.im>
21:23:45
And what about kiwi? There was no formal proposal.
<@nirik:matrix.scrye.com>
21:24:09
did we agree to let it land late last time?
<@conan_kudo:matrix.org>
21:24:56
I think we just did a status update where we said everything's ready and waiting for pungi
<@conan_kudo:matrix.org>
21:25:09
and moved on
<@nirik:matrix.scrye.com>
21:25:39
ok, how about:
<@nirik:matrix.scrye.com>
21:26:36
trying to word it so they can land, but if they are the last thing we have the option to punt... not sure how to word that.
<@nirik:matrix.scrye.com>
21:26:50
I guess just Kiwi and Arm minimal changes may still land
<@nirik:matrix.scrye.com>
21:26:52
?
<@conan_kudo:matrix.org>
21:26:56
works for me
<@conan_kudo:matrix.org>
21:27:26
!proposal kiwi change and minimal ARM image change may still land
<@conan_kudo:matrix.org>
21:27:35
proposal: kiwi change and minimal ARM image change may still land
<@nirik:matrix.scrye.com>
21:27:41
votes? counter proposals?
<@zbyszek:fedora.im>
21:27:42
+1
<@nirik:matrix.scrye.com>
21:27:47
+1 from me.
<@conan_kudo:matrix.org>
21:27:56
+1
<@jistone:fedora.im>
21:27:59
+1
<@humaton:fedora.im>
21:28:21
+1
<@conan_kudo:matrix.org>
21:28:46
!agreed kiwi change and minimal ARM image change may still land (+5, 0, 0)
<@conan_kudo:matrix.org>
21:28:59
!topic Open Floor
<@conan_kudo:matrix.org>
21:29:19
I think we're finally done here, so I'm going to end for us all now!
<@conan_kudo:matrix.org>
21:29:22
😩
<@conan_kudo:matrix.org>
21:29:30
!endmeeting