Re: Why does excuses list a failed autokpgtest that can never run because of unsatisfied B-D?

Soren Stoutner <[email protected]> Wed, 24 Jun 2026 12:11:17 -0700

Newsgroups	gmane.linux.debian.devel.mentors
Organization	Debian
Message-ID	<12918573.O9o76ZdvQC@soren-desktop>

On Wednesday, June 24, 2026 1:13:06 AM Mountain Standard Time Andrey 
Rakhmatullin wrote:
> On Tue, Jun 23, 2026 at 06:17:33PM -0700, Soren Stoutner wrote:
> >The failure is simply that the Build-Dependencies cannot be satisfied on
> >loong64.  As far as I know, they have never been satisfiable.
> 
> The package is not in testing so "this is not a regression" doesn't apply.
> 
> >None of the other architectures that are B-D unsatisfiable are listed as
> >blockers.
> 
> RT doesn't require passing autopkgtests on all architectures. See
> 
> ∙ ∙ Autopkgtest skipped on ppc64el: not installable (which is allowed)
> ∙ ∙ Autopkgtest skipped on riscv64: not installable (which is allowed)
> ∙ ∙ Autopkgtest skipped on s390x: not installable (which is allowed)
> 
> I assume loong64 *is* required to pass for migration.

But *why* would it be required to pass for migration?  It wasn’t in the past 
(the package was removed from testing due to to a bug in a dependency).  Why 
is loong64 different than all the other architectures where it is B-D 
uninstallable?

-- 
Soren Stoutner
[email protected]

signature.asc (application/pgp-signature, 833 B)