And we do expect people to use package layering and not feel bad about it!

Every time I layer a package, a small piece of me dies. :laughing:

LGTM!

And we do expect people to use package layering and not feel bad about it!

If that's the case then I really don't see why we are including origin-clients.

If that's the case then I really don't see why we are including origin-clients.

Yep, we could take that all the way to "why include any container tech at all?" "why include Firefox"?

I am quite conflicted on this myself, but the rationale I have here is that we want something where you install the ISO and have some commands you can use "out of the box" in order to bootstrap. Further, things that are installed by default are versioned and tested iwth the OS.

Obviously if we had livefs by default it'd change the equation some...but even then, the problem is things that aren't installed by default "fade" into the "sea of packages".

I guess things will get more interesting down the line if we try to split up the rpm-md repos some...basically "highlight" some packages that even if they aren't installed by default are still a focus for testing?

I think I'm slightly in favour of not baking in origin-clients because in the end most developers who use Workstation are not writing OpenShift apps, right? OTOH, we definitely need some container runtime technologies because that's how we envision people bootstrap their dev environments, as you said. (Actually, I'd be surprised if one couldn't get oc cluster up to work in a container, so no reboot/livefs required).

Or is this about catering to those users that do want to write OpenShift apps and marketing on a good OOTB experience?

because in the end most developers who use Workstation are not writing OpenShift apps, right?

On this topic... https://fedoraproject.org/wiki/Workstation/Workstation_PRD is worth reading.

I am inclined to say we should at least propose installing origin-clients in traditional Workstation by default too.

I think "why include firefox" should be revisited indeed, but maybe it is best to wait until we have flatpaks building in fedora and can pre-install flatpaks.

Closing this as obsolete for now.