As pointed by Adam with this link.

BetaBlocker +1

BetaBlocker +1

BetaBlocker +1

Definitely feels like a violation of the cited criteria.

BetaBlocker +1

So, I'm not sure about this one. On the whole I'd say the most important bug here is https://bugzilla.redhat.com/show_bug.cgi?id=2234516 , the second most important is https://bugzilla.redhat.com/show_bug.cgi?id=2234928 , and this is the third most important. oldUI did also let you do this, it only showed a fairly light warning; we were mostly relying on US being the default keyboard layout to prevent people from doing this, I think.

I think the resolution for this one specifically will be just to add the warning back, and I'm not sure that really makes it a blocker. Definitely an FE, though.

BetaFE +1

Let's say this is at least accepted as a Beta freeze exception issue, so anaconda can fix it on Monday before the meeting if they want to. We can discuss at the meeting whether it needs to be a blocker.

AGREED AcceptedBetaFE

The following votes have been closed:

• Accepted BetaFreezeException (+1, 0, -0)
  • +1 by adamwill

BetaBlocker +1

The warning will not be enough. For non-latin languages, there is no way how one could use ascii only characters in the new setup. And if the new setup should replace the old one, we should make sure it works better than the old setup. So, if it also was possible in the old way, seems kinda irrelevant to me.

If I understand this ticket correctly, it is just about showing a warning. Nothing else. The dual keyboards are voted on elsewhere, the blivet crash is voted on elsewhere (see Adam's comment).

So yeah, the warning is definitely a good idea, because the non-ascii unlock might or might not work, depending on keyboard configuration. But I don't think we can block the release just because "a warning is missing". We have to find a case where the installer allows you to input a passphrase which then can't be used for disk unlocking. Only after that, we can say that the installer allows people to use a non-functional configuration, and block on that (of course, the likely solution for this is to add the warning).

At this moment, I think this is
BetaBlocker -1

Let's say this issue is about "anaconda's behaviour in this area, assuming the crash in blivet is fixed and g-i-s is fixed to add the US English layout any time a non-ASCII-capable layout is selected".

Discussed during the 2023-08-28 blocker review meeting: [0]

The decision to delay the classification of this as a blocker bug was made as currently, votes are +4 / -4, with the remaining +1s from earlier ticket votes. We will describe the situation more clearly in the ticket and ask if anyone who voted +1 wishes to change their vote

[0] https://meetbot.fedoraproject.org/fedora-blocker-review/2023-08-28/f39-blocker-review.2023-08-28-16.07.txt

To describe the situation more clearly:

There are other bugs covering the problems "gnome-initial-setup does not configure a US layout alongside the native one when the native one cannot input ASCII characters" (that's https://bugzilla.redhat.com/show_bug.cgi?id=2234516 ) and "libblockdev crashes on encryption passphrases containing non-ASCII characters" (https://bugzilla.redhat.com/show_bug.cgi?id=2234928 ).

This bug is only for the anaconda behaviour here, assuming the g-i-s and libblockdev bugs are fixed. The difference between oldUI and newUI here is only that oldUI showed some warning text when you entered a passphrase containing non-ASCII characters - see screenshot - while newUI does not. Folks during the meeting felt that wasn't serious enough to constitute a release blocker.

@imsedgar @tflink @nielsenb @lruzicka do any of you want to reconsider your votes?

There are other bugs covering the problems "gnome-initial-setup does not configure a US layout alongside the native one when the native one cannot input ASCII characters" (that's https://bugzilla.redhat.com/show_bug.cgi?id=2234518 )

I suppose you mean https://bugzilla.redhat.com/show_bug.cgi?id=2234516 ?

It seems to me that this bug depends on the two other bugs.
If these are solved, then the problem with the non-ASCII characters doesn't appear,
and then no warning is necessary.

I think the installation system (anaconda) should prevent users and administrators accidentally configure the new installed system in an unusable state. So, if the problem of the other two bugs still exist, there should be at least a warning (or better, deny such passphrases).

Since beta releases are not for production purposes, I think it would be enough to declare it a FinalBlocker instead of a BetaBlocker, and depending it on the other bugs (If they are solved, then no warning is required).
What do the others think about this?

You can assume the other two bugs will be fixed.

However, a warning is still desirable even if those bugs are fixed, because we cannot be sure that any non-ASCII character included in a passphrase will actually be typable when decrypting partitions on boot. (I mean, technically, we can't be 100% sure any character in a passphrase will be typable when decrypting, but it's far more likely to be an issue with non-ASCII characters). That's why old-UI shows this warning.

Thanks, Adam.
Assuming the other two bugs are fixed, a freeze exception is enough.

BetaBlocker -1

BetaFE +1

BetaBlocker -1

Ok, reading all the last arguments I'll change my vote here.

BetaBlocker -1

BetaBlocker -1
(I voted at the meeting last week, but just to get my vote in the ticket count)

AGREED RejectedBetaBlocker

Discussed during the 2023-09-04 blocker review meeting [1]:

we agreed that anaconda not displaying a warning here is not in itself serious enough to be a blocker bug, but fixing it would be useful. The libblockdev crash and gnome-initial-setup keyboard layout selection issues are being addressed separately

[1] https://meetbot-raw.fedoraproject.org/fedora-blocker-review/2023-09-04/f39-blocker-review.2023-09-04-16.00.log.txt

The following votes have been closed:

• Rejected BetaBlocker (+3, 0, -5)
  • +1 by tflink, nielsenb, lruzicka
  • -1 by geraldosimiao, imsedgar, kparal, bcotton, adamwill

Release F39 is no longer tracked by BlockerBugs, closing this ticket.