Adds checks for container and bootc hosts

[jmarrero]

This changes the is_container() func for is_bootc_host() and updates the logic. This should detect on all ostree and bootc hosts to date that are not using bootc usroverlay or ostree admin unlock for development purposes.

[pep8speaks]

Hello @jmarrero! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2024-07-22 13:47:34 UTC

[jmarrero]

did a quick test on a bootc host and get the message by default, and can use dnf to install software in a transient way when I do bootc usroverlay or run it in a container.

[jmarrero]

This is missing the links to the docs described in: https://issues.redhat.com/browse/SWM-1313

[jlebon]

Nice and simple. Looks sane to me overall.

[travier]

See: #2108

[dcantrell]

@jmarrero can you squash these commits in to a single for the purposes of the PR? In the commit message please explain the fix and reference that it resolves RHEL-49670 and RHEL-49671

[jlebon]

For the docs URL, we could pull it from /etc/os-release. Possibly DOCUMENTATION_URL, which is an existing field, but we could also add a separate field too. This makes it so that different distros can reference different pages. (And also so that we don't have to block on getting docs ready.)

[jmarrero]

@jmarrero can you squash these commits in to a single for the purposes of the PR? In the commit message please explain the fix and reference that it resolves RHEL-49670 and RHEL-49671

done.

[ravanelli]

For the docs URL, we could pull it from /etc/os-release. Possibly DOCUMENTATION_URL, which is an existing field, but we could also add a separate field too. This makes it so that different distros can reference different pages. (And also so that we don't have to block on getting docs ready.)

That’s a thing we would need to align with the RHEL/FEDORA team to have it there? I mean, that’s not something it up to us to decided if we can add it there or not. We may have a no as an answer.
Should we probably try to look at this option asap?

[jmarrero]

Talked to Colin about it and we can add it as part of the build process of our image-mode systems. We should probably split this into it's own issue.
I have a draft here #2112 to rebase once we have merged this PR. We can "bike shed" the message and process to get it there.

[cgwalters]

I don't think this can be the first case where we've wanted to reference docs from software we ship. It is a complex topic. Don't get me wrong, I'm not totally fixated on the idea the docs have to be online - it could also be a man dnf-bootc or something (but, while that makes sense it hits on the problem we don't ship man pages in many targets where want this today).

Many systemd units for example that we ship reference Documentation= pointing to an upstream URL. That upstream URL could be somewhere dnf-upstream owned too.

[ppisar]

Could you also remove "import json" and "import subprocess" lines from dnf/util.py? Your patch removed the only reason for their imports.

[ppisar]

Thanks for renaming the function. Please change the name also in the commit message.

[jmarrero]

Could you also remove "import json" and "import subprocess" lines from dnf/util.py? Your patch removed the only reason for their imports.

done

[jmarrero]

Thanks for renaming the function. Please change the name also in the commit message.

done