#meeting-1:fedoraproject.org log

<@james:fedora.im>

17:02:33

!startmeeting fpc

<@meetbot:fedora.im>

17:02:34

Meeting started at 2025-11-13 17:02:33 UTC

<@meetbot:fedora.im>

17:02:34

The Meeting name is 'fpc'

<@james:fedora.im>

17:02:40

!topic Roll Call

<@tibbs:fedora.im>

17:02:47

Hello.

<@carlwgeorge:fedora.im>

17:02:52

!hi

<@zodbot:fedora.im>

17:02:54

Carl George (carlwgeorge) - he / him / his

<@conan_kudo:matrix.org>

17:03:20

!hi

<@zodbot:fedora.im>

17:03:33

Neal Gompa (ngompa) - he / him / his

<@james:fedora.im>

17:03:38

Hey

<@limb:fedora.im>

17:04:04

!hi

<@zodbot:fedora.im>

17:04:05

Gwyn Ciesla (limb) - she / her / hers

<@james:fedora.im>

17:07:07

Okay, we have 5 and I think this one is easy...

<@james:fedora.im>

17:07:19

!topic FPC PR#1501 https://pagure.io/packaging-committee/pull-request/1501

<@limb:fedora.im>

17:07:42

+1.

<@decathorpe:fedora.im>

17:08:44

!hi

<@zodbot:fedora.im>

17:08:46

Fabio Valentini (decathorpe) - he / him / his

<@decathorpe:fedora.im>

17:08:49

sorry for being late

<@decathorpe:fedora.im>

17:09:26

+1 for the PR

<@tibbs:fedora.im>

17:09:29

+1 (I know zero about ocaml but keping -devel stuff out of regular packages is always good)

<@carlwgeorge:fedora.im>

17:11:46

<@salimma:fedora.im>

17:11:55

<@james:fedora.im>

17:12:27

Seems like a merge :)

<@james:fedora.im>

17:13:09

This one is a little older...

<@james:fedora.im>

17:13:18

!topic FPC#1489 https://pagure.io/packaging-committee/issue/1489

<@tibbs:fedora.im>

17:14:39

Slightly less relevant as F43 is actually released, but my experience is that the page is wrong.

<@salimma:fedora.im>

17:15:03

Zbyszek worked on this right?

<@decathorpe:fedora.im>

17:15:47

yeah it would be good to get his input on this. IIRC this *should* be working on Fedora 42+?

<@tibbs:fedora.im>

17:17:18

Well the %sysusers_*_compat macros definitely do not evaluate to empty.

<@tibbs:fedora.im>

17:17:23

In F42.

<@james:fedora.im>

17:18:30

I feel like the obvious cheap change is to just s/f42/f43/ ... but I'm open to something else.

<@tibbs:fedora.im>

17:18:46

My problem is that I don't know what the expected behavior is.

<@salimma:fedora.im>

17:19:01

it might need to be fixed though even for f42, if the current behavior is a bug

<@tibbs:fedora.im>

17:19:49

I have a package that cannot work with the compat scriptlets in the guidelines, and I ran into that discrepancy while trying to figure out what to do.

<@decathorpe:fedora.im>

17:19:51

should they?

<@decathorpe:fedora.im>

17:20:11

not all packages were / could be converted, so making it a noop would break them, no?

<@tibbs:fedora.im>

17:20:14

The guidelines page says explicitly that they will evaluate as empty in F42.

<@tibbs:fedora.im>

17:20:26

That's the source of my confusion.

<@conan_kudo:matrix.org>

17:21:11

we probably should ask EPSCo about how to reconcile this for EPEL too

<@conan_kudo:matrix.org>

17:21:41

I've been kind of wary about this change because I find it difficult to reason how to handle all the release cases that common packagers have to deal with

<@tibbs:fedora.im>

17:22:24

The change was a mess; those PRs should not have been merged.

<@conan_kudo:matrix.org>

17:22:41

I think I wound up reverting or blocking some that affected KDE

<@conan_kudo:matrix.org>

17:23:02

ah yes, I never merged the one for sddm

<@conan_kudo:matrix.org>

17:24:11

I still don't know what I'm actually supposed to do

<@conan_kudo:matrix.org>

17:24:31

and I have to revisit this as part of packaging plasma-login-manager

<@decathorpe:fedora.im>

17:25:43

I'm all for getting rid of scriptlets, but this one was rushed

<@tibbs:fedora.im>

17:25:52

If you don't use one spec across several releases then it isn't a big issue, as the stuff does appear to work as expected in F43. But it doesn't work in F42 without using the compat stuff, and the guidelines say directly that it will.

<@conan_kudo:matrix.org>

17:26:28

I'd rather just make the scriptlets evaluate to nothing in cases where it isn't needed

<@conan_kudo:matrix.org>

17:26:33

that's what we did for ldconfig scriptlets

<@salimma:fedora.im>

17:27:06

right. but that's what the change proposal said

<@salimma:fedora.im>

17:27:11

the scriptlet *should* evaluate to nothing

<@salimma:fedora.im>

17:27:27

should we raise this to fesco? or ask Zbyszek directly first

<@decathorpe:fedora.im>

17:27:27

s/F42/F43/ in the guidelines should be easy enough, but I'd rather ask zbyszek whether that's actually *supposed* to be this way

<@salimma:fedora.im>

17:27:34

maybe a F42 update was just missed

<@salimma:fedora.im>

17:27:46

if we defer to f43 we should ask what's the best way to fix things on f42

<@tibbs:fedora.im>

17:27:54

It may just be a bug, or maybe I'm just not understanding everything that's supposed to happen.

<@decathorpe:fedora.im>

17:28:03

I don't think it *hurts* that it doesn't evaluate to noop, right? it should be a noop at runtime instead?

<@tibbs:fedora.im>

17:28:34

Maybe?

<@tibbs:fedora.im>

17:29:03

At best it's just a source of confusion, as it was for me.

<@conan_kudo:matrix.org>

17:29:17

same

<@conan_kudo:matrix.org>

17:29:29

at this point I genuinely don't know what I'm supposed to do

<@james:fedora.im>

17:29:42

Anyone else want to comment and ping zbyszek ... if not I can do it.

<@tibbs:fedora.im>

17:30:38

I wasn't sure who to ping when I filed that. I also managed to forget entirely that I filed it.

<@tibbs:fedora.im>

17:30:52

I can investigate and ping.

<@james:fedora.im>

17:31:05

Yeh, n/p ... I had a quick look today and didn't remember seeing it before ;)

<@james:fedora.im>

17:33:51

!topic Open Floor

<@tibbs:fedora.im>

17:34:11

We will need to discuss the R guidelines thing at some point.

<@james:fedora.im>

17:34:12

Those were the two things I'd seen ... does anyone else have anything to bring up or discuss?

<@tibbs:fedora.im>

17:34:24

And we will need to talk about the forge migration.

<@james:fedora.im>

17:34:48

https://pagure.io/packaging-committee/issue/1499

<@tibbs:fedora.im>

17:34:49

I'm happy to do the latter after other teams have migrated.

<@tibbs:fedora.im>

17:35:30

So the R thing seems conceptually reasonable but I haven't dug into the details.

<@tibbs:fedora.im>

17:36:40

It's also odd to see the old wiki workflow back from the dead.

<@tibbs:fedora.im>

17:37:20

There is also the question of whether we need to complete our discussion before the inclusion of R-srpm-macros to the buildroot.

<@tibbs:fedora.im>

17:37:58

Since PRs have already been filed to do that (in rawhide and EPEL9/10 but not any release of Fedora, for whatever reason).

<@tibbs:fedora.im>

17:39:49

One thing I see is ```Version: %R_rpm_version 1.2-3```

<@tibbs:fedora.im>

17:40:04

Doesn't macros in version conflict with some other work being done?

<@james:fedora.im>

17:40:33

autospec?

<@tibbs:fedora.im>

17:40:55

Yes, weren't they switching away from using rpm to parse the spec? This would probably break.

<@james:fedora.im>

17:41:31

Yeh, although I'm not sure how many R packages use autospec

<@carlwgeorge:fedora.im>

17:41:32

yeah that was https://github.com/praiskup/norpm i believe

<@tibbs:fedora.im>

17:43:28

I will ask that question, and obviously we will need a PR instead of a wiki page.

<@tibbs:fedora.im>

17:45:01

But I'm not sure what else to ask at this point. The document looks clean. There may be some slight issues around how license files are done in R.

<@decathorpe:fedora.im>

17:45:48

*ahem* I have a pending PR where I've been waiting for feedback for 2 months :)

<@tibbs:fedora.im>

17:47:32

Feel free to link it....

<@james:fedora.im>

17:48:28

Guessing this one: https://pagure.io/packaging-committee/pull-request/1483

<@decathorpe:fedora.im>

17:48:39

yes, this one

<@james:fedora.im>

17:49:12

Looks fine to me, I commented with the really minor thing and you explained

<@tibbs:fedora.im>

17:50:25

<@carlwgeorge:fedora.im>

17:51:53

<@michel-slm:matrix.org>

17:53:33

<@james:fedora.im>

17:57:26

Okay, I'll merge

<@james:fedora.im>

17:57:29

Anything else?

<@tibbs:fedora.im>

17:58:14

Nothing from me.

<@limb:fedora.im>

17:59:09

Nope.

<@limb:fedora.im>

17:59:23

And +1. Sorry, brief AFK.

<@james:fedora.im>

18:00:42

Okay, normally we'd meet in two weeks ... but that's Thanksgiving day in the US ... so not sure if we want to just skip, or try to bump to a different week?

<@carlwgeorge:fedora.im>

18:02:10

skipping is fine by me

<@limb:fedora.im>

18:02:18

same.

<@james:fedora.im>

18:02:48

Okay, see y'all in four weeks.

<@james:fedora.im>

18:02:50

!endmeeting