remove dependency on cgroupfs-mount for Debian trixie

[zen-fu]

Summary

remove dependency on cgroupfs-mount for Debian trixie, because that package isn't available in that release.

Related Issues (if any)

Fixes #1023

Checklist

• 🟢 Spec tests.
• 🟢 Acceptance tests.
• Manually verified. (For example puppet apply)

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[charlyforot]

Suggested change

	'Debian' => $facts['os']['distro']['codename'] ? { 'bookworm' => ['cgroupfs-mount',], 'trixie' => [] },
	'Debian' => $facts['os']['distro']['codename'] ? {
	'trixie' => [],
	default => ['cgroupfs-mount',],
	},

[zen-fu]

Hello, @charlyforot, and thanks for the suggestion!

As i don't see the cgroupfs-mount package available in sid (ref), i suggest we keep a default of no packages, thinking about the future and assuming for now that only bookworm will have that package.

Does that make sense?

If so, then this is what we'd want:

Suggested change

	'Debian' => $facts['os']['distro']['codename'] ? { 'bookworm' => ['cgroupfs-mount',], 'trixie' => [] },
	'Debian' => $facts['os']['distro']['codename'] ? {
	'bookworm' => ['cgroupfs-mount',],
	default => []
	},

How does that sound?

[zen-fu]

As i don't see the cgroupfs-mount package available in sid (ref), i suggest we keep a default of no packages, thinking about the future and assuming for now that only bookworm will have that package.

i just implemented something like that, even though a bit different than in my comment above, to capture the current state of package availability in Debian, so:

• if codename is bullseye (current oldstable) or bookworm (current stable), we require the package
• else, we don't.

Note that in some days bullseye will become oldoldstable, bookworm will become oldstable, and trixie will become stable, so the proposed code should be good at least for the life of the next stable release, that is, a couple of years.

i'm actually unsure about whether we should introduce the requirement for bullseye, as it was not there before and will impact any bullseye installation that uses this module.