FinalBlocker +1
font should be protected from removal, otherwise DE breaks, that is too important to go through.

This will be tricky. Our criteria are quite extensive in this case:
https://fedoraproject.org/wiki/Fedora_43_Final_Release_Criteria#Installing,_removing_and_updating_software

The only potentially matching sentence is this one:

The package manager must never make the system enter an inconsistent or unbootable state. (E.g. damage the local software database, remove wrong system files, break the bootloader, etc).

However, this really seems to be intended to prevent a fully unbootable state, not "unbootable to desktop". But it's the only place where an argument can be made, I don't think we have anything better.

At this moment, it seems that we should give this FinalBlocker -1. And if we want to block on this, we have to come up with an additional requirement for the criterion above.

Fortunately, this doesn't seem to be easy to trigger. In the Installed tab in gnome-software, I haven't found an item to uninstall the desktop. It's possible through fonts, but they are not listed in Installed, you have to find them manually. I would say this is quite difficult to trigger. It should definitely be documented, if not fixed.

FinalFE +1
Let's see more discussion before casting a blocker vote.

However, this really seems to be intended to prevent a fully unbootable state, not "unbootable to desktop". But it's the only place where an argument can be made, I don't think we have anything better.

I think "unbootable to desktop" is a lucky case here. If the protected packages functionality is not working, it's possible that some other thing could accidentally cause a bigger chunk of dependencies to be removed, and "oops no systemd".

At this moment, it seems that we should give this FinalBlocker -1. And if we want to block on this, we have to come up with an additional requirement for the criterion above.

How about: "If the package manager has a feature intended to protect the system from accidental removal of packages marked as critical, that feature must work"?

I think we should block on this and update the criteria.

FinalBlocker +1

I think "unbootable to desktop" is a lucky case here. If the protected packages functionality is not working, it's possible that some other thing could accidentally cause a bigger chunk of dependencies to be removed, and "oops no systemd".

That might be true, but usually we ask people to demonstrate it before we make it a blocker.

How about: "If the package manager has a feature intended to protect the system from accidental removal of packages marked as critical, that feature must work"?

Sounds reasonable. We already have this sentence in the criterion:

The default graphical package manager for a given software type must appropriately:

so I think we just need to add another bullet point:

** Honor definitions of system software protected from removal, and act in those cases accordingly*

A footnote already contains:
All requirements apply only if such a functionality is supported by the package manager; it does not imply that the package manager must support all this functionality.

I feel like this makes it too easy for a user to accidentally break their system.

FinalBlocker +1
FinalFE +1

Kamil says it is "quite difficult to trigger" above and he outlines how he tried, while @nielsenb claims "too easy to accidentally break the system", but w/o noting how.

If, as Matthew says above, it is somehow possible to accidentally nuke systemd (!) off the system, this deserves block the final. But I don't have a test to prove that at hand, so I'll refrain from blocking final, however:

FinalFE+1

FinalFE +1 for now.

I mean, technically, it's very easy to trigger. Just try and remove systemd directly, or anything systemd depends on.

I think the question is more "is there a realistic scenario in which someone would do that?"

For gnome-shell we had the realistic scenario in the bug report - someone wanted to remove some fonts. It's reasonable to think that removing fonts should be a safe operation. You could argue that we'd need to show a dependency of systemd that looks "sufficiently unimportant" that somebody might reasonably try to remove it and not look too closely at the list of deps that would be removed (or use -y).

Or, alternatively, you could say 'people do funny stuff' and consider it a violation just on the basis that somebody might try.

Never mind, I was missing two things: the criterion doesn't technically apply to dnf, and more importantly, this bug only affects PackageKit, so only really GNOME Software and Discover. So we need to demonstrate that removing something that one or both of those things lets you remove can break the system.

AGREED AcceptedFinalFreezeException
AGREED AcceptedFinalBlocker

Discussed at the 2025-10-06 (blocker / freeze exception) review meeting:

This is accepted on the proviso that we also amend the FInal criterion as proposed by kparal in https://pagure.io/fedora-qa/blocker-review/issue/1952#comment-988437 . this will be accepted as a violation of that amended criterion, but it is also accepted as a Freeze Exception to ensure it gets fixed for release even if it's not a blocker.

https://meetbot-raw.fedoraproject.org//blocker-review_matrix_fedoraproject-org/2025-10-06/f43-blocker-review.2025-10-06-16.00.txt

The following votes have been closed:

• Accepted FinalBlocker (+3, 0, -0)
  • +1 by augenauf, lruzicka, nielsenb
• Accepted FinalFreezeException (+2, 0, -0)
  • +1 by kparal, nielsenb

The criterion change is now proposed here:
https://discussion.fedoraproject.org/t/release-criteria-change-proposal-honor-protected-packages-in-gui-package-managers/167616

Please post your feedback.

Release F43 is no longer tracked by BlockerBugs, closing this ticket.