#meeting-1:fedoraproject.org log

<@marmijo:fedora.im>

15:31:05

!startmeeting fedora_coreos_meeting

<@meetbot:fedora.im>

15:31:06

Meeting started at 2025-12-10 15:31:05 UTC

<@meetbot:fedora.im>

15:31:06

The Meeting name is 'fedora_coreos_meeting'

<@marmijo:fedora.im>

15:31:19

!topic roll call

<@siosm:matrix.org>

15:31:21

!hi

<@ydesouza:fedora.im>

15:31:24

!hi

<@zodbot:fedora.im>

15:31:25

Yasmin Valim de Souza (ydesouza)

<@zodbot:fedora.im>

15:31:25

Timothée Ravier (siosm) - he / him / his

<@spresti:fedora.im>

15:31:27

!hi

<@zodbot:fedora.im>

15:31:30

Steven Presti (spresti)

<@hricky:fedora.im>

15:31:43

!hi

<@zodbot:fedora.im>

15:31:44

Hristo Marinov (hricky) - he / him / his

<@dustymabe:matrix.org>

15:33:00

!hi

<@zodbot:fedora.im>

15:33:02

Dusty Mabe (dustymabe) - he / him / his

<@ravanelli:matrix.org>

15:33:06

!hi ravanelli

<@zodbot:fedora.im>

15:33:08

Renata Ravanelli (ravanelli)

<@marmijo:fedora.im>

15:34:30

!topic Action items from last meeting

<@marmijo:fedora.im>

15:34:59

!info There are no action items from the last meeting.

<@marmijo:fedora.im>

15:35:11

!topic Review Fedora 44 Release Schedule

<@marmijo:fedora.im>

15:35:20

!link https://fedorapeople.org/groups/schedule/f-44/f-44-key-tasks.html

<@marmijo:fedora.im>

15:36:29

The first Fedora 44 change checkpoint is a week from today. If we have anything related to infrastructure changes, we should work on submitting it soon

<@marmijo:fedora.im>

15:37:24

There's another checkpoint, for system wide changes or those requiring mass rebuilds, in two weeks

<@dustymabe:matrix.org>

15:37:44

PSA: get your change requests submitted for F44

<@marmijo:fedora.im>

15:38:34

!info Fedora 44 change deadlines are rapidly approaching. Please submit your change requests soon!

<@marmijo:fedora.im>

15:39:23

!topic tracker: Fedora 44 changes considerations

<@marmijo:fedora.im>

15:39:35

!link https://github.com/coreos/fedora-coreos-tracker/issues/2063

<@marmijo:fedora.im>

15:40:45

Stand by, i'm reviewing last week's meeting log

<@marmijo:fedora.im>

15:42:14

Okay, we covered only 107 last week. Let's start with 108

<@marmijo:fedora.im>

15:42:18

!link https://fedoraproject.org/wiki/Changes/CMake_ninja_default

<@marmijo:fedora.im>

15:43:24

I don't think we use cmake for our upstream packages.

<@marmijo:fedora.im>

15:44:29

109 is also related to cmake, so I think we can skip it

<@marmijo:fedora.im>

15:44:32

stop me if I'm wrong...

<@marmijo:fedora.im>

15:44:35

let's move on to 110

<@spresti:fedora.im>

15:44:37

well

<@spresti:fedora.im>

15:45:07

We do use the macro "make_install" which that change request says to make sure not to use => example https://src.fedoraproject.org/rpms/ignition/blob/rawhide/f/ignition.spec

<@dustymabe:matrix.org>

15:45:44

good catch

<@siosm:matrix.org>

15:46:03

I don't think that will impact us as this should only matter in the context of cmake builds

<@siosm:matrix.org>

15:46:39

looking at the guidelines: https://docs.fedoraproject.org/en-US/packaging-guidelines/

<@spresti:fedora.im>

15:48:41

Hmm okay, its just a bit concerning that the change request is calling out that macro.

<@siosm:matrix.org>

15:49:08

here it's for our own makefile that we have in https://github.com/coreos/ignition/blob/main/Makefile

<@siosm:matrix.org>

15:49:24

for the cmake case, the Makefile is generated by the CMake tooling itself

<@siosm:matrix.org>

15:49:37

thus the need to use the corresponding macros

<@spresti:fedora.im>

15:50:02

Ahhh I see, thank you for explaining that.

<@siosm:matrix.org>

15:50:36

in this case there won't even be a Makefile as the build will use ninja insteada

<@siosm:matrix.org>

15:50:56

so %make_install will not even work

<@siosm:matrix.org>

15:51:54

I think it's safe to skip for us. I don't think we have CMake based packages

<@siosm:matrix.org>

15:52:17

https://src.fedoraproject.org/rpms/rpm-ostree/blob/rawhide/f/rpm-ostree.spec#_86

<@siosm:matrix.org>

15:52:30

ok we do?

<@siosm:matrix.org>

15:53:19

ok it's for libdnf

<@marmijo:fedora.im>

15:54:47

Thanks for finding that! So we do have some work on these two changes.

<@siosm:matrix.org>

15:55:08

I still don't think so

<@marmijo:fedora.im>

15:56:39

Ok, looks like the change for CMAKE 4.0 only affects packages that specify `cmake_minimum_required`

<@siosm:matrix.org>

15:56:48

let's make an issue to track investigation

<@marmijo:fedora.im>

15:57:00

Will do!

<@siosm:matrix.org>

15:57:03

I'll paste my investigations there

<@jcapitao:matrix.org>

15:57:11

!hi

<@zodbot:fedora.im>

15:57:13

Joel Capitao (jcapitao) - he / him / his

<@marmijo:fedora.im>

15:57:37

110

<@marmijo:fedora.im>

15:57:37

!link https://fedoraproject.org/wiki/Changes/KojiServiceImageBuilderRemoval

<@dustymabe:matrix.org>

16:00:07

We never used that service anyway since we call osbuild directly in COSA

<@ravanelli:matrix.org>

16:00:09

We are using osbuild directly, the only koji packages we use are for Brew

<@ravanelli:matrix.org>

16:00:20

But not these ones

<@marmijo:fedora.im>

16:00:49

Ok, that's what I thought, we use osbuild by other means

<@marmijo:fedora.im>

16:01:00

Ok, that's what I thought, we use osbuild <del>by</del> through other means

<@marmijo:fedora.im>

16:01:23

112:

<@marmijo:fedora.im>

16:01:23

!link https://fedoraproject.org/wiki/Changes/Enforcing_signature_checking_by_default

<@marmijo:fedora.im>

16:03:36

We already rely on signed repos, but will this affect laying abilities?

<@dustymabe:matrix.org>

16:04:59

marmijo: good question

<@dustymabe:matrix.org>

16:05:19

I know `rpm-ostree` uses a forked version of libdnf. I'm not sure if that comes into play here or not

<@siosm:matrix.org>

16:05:43

I think it will likely impact layering of unsigned packages

<@marmijo:fedora.im>

16:06:08

We already rely on signed repos, but will this affect <del>laying</del> layering abilities?

<@siosm:matrix.org>

16:06:23

we'll need a tracker for that upstream in rpm-ostree

<@dustymabe:matrix.org>

16:06:29

yeah, and we probably don't (or won't) have a way to disable it when `rpm-ostree install` is called

<@siosm:matrix.org>

16:07:00

easiest workaround will be to sign the package and import the key locally

<@siosm:matrix.org>

16:07:09

will likely* be

<@dustymabe:matrix.org>

16:07:46

or define your own RPM macro that undoes `%_pkgverify_level` ?

<@marmijo:fedora.im>

16:09:41

I'll open a tracker issue in rpm-ostree and link it to the change considerations tracker so we can investigate the change

<@conan_kudo:matrix.org>

16:09:50

keep in mind we don't offer signed repositories, only repositories with signed packages

<@conan_kudo:matrix.org>

16:10:05

there is a difference, and if you mess the settings up, you could break consuming fedora content

<@conan_kudo:matrix.org>

16:10:59

I _think_ the dnf team adapted dnf4 libdnf for rpm6

<@conan_kudo:matrix.org>

16:11:07

so this should work with the right repo settings

<@dustymabe:matrix.org>

16:11:09

marmijo: would be great to have a tracker issue in the FCOS issue too so we can revisit periodically

<@conan_kudo:matrix.org>

16:12:51

nothing sticks out as adapting for this, so this will need to be raised to the rpm team to prioritize

<@conan_kudo:matrix.org>

16:13:08

this is something y'all will have to do, as I have no power to raise this

<@dustymabe:matrix.org>

16:13:30

Conan Kudo: to prioritize fixing dnf4 libdnf for this?

<@conan_kudo:matrix.org>

16:13:32

once libdnf is adapted, rebasing the submodule will be sufficient

<@conan_kudo:matrix.org>

16:13:34

yes

<@dustymabe:matrix.org>

16:13:48

yeah, fully expect a hard no on that one. and they'd probably be right

<@conan_kudo:matrix.org>

16:14:11

well the other option is telling rpm-ostree to propagate the macro to librpm

<@conan_kudo:matrix.org>

16:14:21

like it does with dbpath

<@conan_kudo:matrix.org>

16:14:32

to set it to the "old" behavior

<@conan_kudo:matrix.org>

16:16:50

also...

<@conan_kudo:matrix.org>

16:16:51

!hi

<@zodbot:fedora.im>

16:16:53

Neal Gompa (ngompa) - he / him / his

<@marmijo:fedora.im>

16:17:04

!link https://fedoraproject.org/wiki/Changes/golang1.26

<@marmijo:fedora.im>

16:17:04

Let's move on to 113

<@marmijo:fedora.im>

16:18:37

This will most definitely affect FCOS. I'm not sure what version(s) of GO we're using in our packages now, but I'll open a tracker for this so we can make sure to get them all upgraded/synced

<@conan_kudo:matrix.org>

16:18:57

hopefully it won't be as painful as 1.24 was

<@marmijo:fedora.im>

16:19:36

I hope not!

<@marmijo:fedora.im>

16:19:47

!link https://fedoraproject.org/wiki/Changes/Hardlink_identical_files_in_packages_by_default

<@marmijo:fedora.im>

16:19:47

change 204

<@conan_kudo:matrix.org>

16:20:22

this shouldn't impact you negatively since rpm-ostree does the same thing implicitly

<@siosm:matrix.org>

16:20:47

yep

<@conan_kudo:matrix.org>

16:20:49

it may result in some packages being broken due to automatic hardlinking, but I doubt any of those are part of the FCOS stack

<@conan_kudo:matrix.org>

16:21:23

heads-up though: you can hit weird sharp edges where koji thinks noarch packages are being build wrong across arches

<@conan_kudo:matrix.org>

16:21:37

because the way hardlinking works might result in subtle differences

<@conan_kudo:matrix.org>

16:21:44

this bit us with some LXQt packages

<@marmijo:fedora.im>

16:23:05

Okay, i'll make a note that we should be stay alert when the change lands, but there should be no impact to FCOS

<@marmijo:fedora.im>

16:23:34

we can probably get to one or two more

<@marmijo:fedora.im>

16:23:41

209

<@marmijo:fedora.im>

16:23:41

!link https://fedoraproject.org/wiki/Changes/cmark-0.31

<@siosm:matrix.org>

16:24:29

I don't think we include that

<@marmijo:fedora.im>

16:25:21

change 211

<@marmijo:fedora.im>

16:25:21

!link https://fedoraproject.org/wiki/Changes/Nix_package_tool

<@siosm:matrix.org>

16:25:27

We don't include Nix :)

<@marmijo:fedora.im>

16:25:33

Looks like a new tool/package being added to Fedora

<@conan_kudo:matrix.org>

16:25:52

yeah, it'll probably be more of a concern for atomic desktops

<@conan_kudo:matrix.org>

16:26:12

you might want to preinstall nix-filesystem though

<@conan_kudo:matrix.org>

16:26:18

so that the directory exists in your images

<@conan_kudo:matrix.org>

16:26:32

some people will do nix on coreos and without that, it'll break

<@siosm:matrix.org>

16:27:07

we've had requests but I'm not a fan

<@siosm:matrix.org>

16:27:15

anyway, for another meeting :)

<@conan_kudo:matrix.org>

16:27:22

I'm not either, but there you go :)

<@hricky:fedora.im>

16:28:12

Neither do I, if that matters at all. ;)

<@marmijo:fedora.im>

16:28:31

116 (not added to the tracker yet)

<@marmijo:fedora.im>

16:28:31

last one...

<@marmijo:fedora.im>

16:28:31

!link https://fedoraproject.org/wiki/Changes/UseKmsconVTConsole

<@conan_kudo:matrix.org>

16:28:51

this does affect you

<@conan_kudo:matrix.org>

16:29:13

and probably deserves a meeting of its own to figure out (or at least a ticket)

<@marmijo:fedora.im>

16:30:21

Yeah, let's cover it in the next meeting. We're at time now

<@marmijo:fedora.im>

16:30:48

!topic Open Floor

<@marmijo:fedora.im>

16:31:26

I'll close the meeting in a few minutes if there's nothing to discuss here :)

<@marmijo:fedora.im>

16:32:36

Thanks all for joining!

<@marmijo:fedora.im>

16:32:46

!endmeeting