Testing CI RPM build will be here: https://src.fedoraproject.org/rpms/rpkg/pull-request/38

I agree this is an improvement for fedpkg. The automatic koji config might be required to use rhpkg and rhpkg users might be surprised by this change. centpkg also appears to use the koji config (but I'd prefer if it used the centos-stream-9 config on disk).

Ideas?

Maybe the default could be determined on the frontend level?

So keep the koji_config=None, keep the old local config detection mechanism by default ON, and disable it in fedpkg? Sounds OK.

I can take a look at this, but this will require another "protected" method defined here
that will be designed to be overridden in Fedpkg. Before I start hacking on this, and
on fedpkg, I'd like to have ACK from rpkg maintainers that this actually makes sense.

Or anyone else wants to take a look at the patch?

I'm not an rpkg maintainer, but this makes sense to me. Moreover, fedpkg should grow some capability to let the user set the base distro for EPEL and EPEL Next targets by configuring the prefix to look up.

A simple example:

[fedpkg.epel-distros]
base = "alma"
next = "centos-stream"

That will allow for fedpkg to automatically look up the correct configs based on our {distro}+epel(-next)-<releasever>.cfg naming convention.

OK, I updated this anyway ... that was not so huge pile of work, though I bet cleaning that up will take more work...

Corresponding patch for fedpkg: https://src.fedoraproject.org/rpms/fedpkg/pull-request/35

Testing packages are in dnf copr enable praiskup/fedpkg-and-no-epel-8-config

Can I do something about the code? Fedora community thinks that it's better to block Mock in updates-testing, rather than have broken fedpkg mockbuild in epel8 branch. So it would be nice to find a way out from this quickly (CentOS 8 mirrors will be off soon).

The code looks well (except for little flake8 complaints that I will fix). I can merge it.

Commit 83232c55 fixes this pull-request