On 2/14/26 2:19 PM, Patrick O'Callaghan wrote:
On Fri, 2026-02-13 at 15:38 -0800, Samuel Sieb wrote:
On 2/13/26 3:43 AM, Patrick O'Callaghan wrote:
I ran 'dnf upgrade' this morning and got a bunch of proposed
upgrades,
one of which failed due to a missing public key (it was Chrome).
See
separate thread.
The odd thing is that *none* of the upgrades applied, even things
completely unrelated to Chrome. When I re-ran the upgrade and added
'--
exclude google-chrome-stable', then it all worked.
This looks like a regression. IIRC in the past dnf would have
installed
everything that didn't depend on the failing item. Am I mistaken?
Any error in the checking will abort the entire transaction.
Otherwise,
it would have to go back and redo all the dependency checking again.
I don't recall it doing that before but I may be wrong. AFAIK it
doesn't happen if some repo is inaccessible, which at least
superficially would seem to be a comparable problem.
It's not. The repo data isn't available so it won't be used for
calculating the update transaction. The resulting transaction is still
consistent. There is a repo option to make it required, in which case
dnf will stop if it can't access the repo, but again that's still before
calculating the transaction.
--
_______________________________________________
users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it:
https://forge.fedoraproject.org/infra/tickets/issues/new
