Re: Bug#1116026: transition: octave-10

[Adrian Bunk <bunk@debian.org>]

On Fri, Oct 03, 2025 at 11:24:28AM +0200, Emilio Pozuelo Monfort wrote:
> On 03/10/2025 11:15, Rafael Laboissière wrote:
> > * Adrian Bunk <bunk@debian.org> [2025-10-03 12:00]:
> > 
> > > On Fri, Oct 03, 2025 at 09:36:32AM +0200, Emilio Pozuelo Monfort wrote:
> > > > 
> > > > I think octave needs to add breaks on octave-statistics (<< 1.7.5-4). That
> > > > looks to be the last blocker.
> > > 
> > > The octave-abi-* dependency already has the same effect
> > > (that's a benefit of a non-smooth transition).
> > 
> > Indeed, octave-statistics 1.7.5-3 depends on octave-abi-59 (provided by
> > octave 9.4.0-1, currently on testing), while octave-statistics 1.7.5-4
> > depends on octave-abi-60 (provided by octave 10.2.0-3, currently on
> > sid).
> 
> Ah, I didn't notice the abi dependency. Then the test is fine [1], it used
> octave/sid and octave-statistics/testing, but since
> octave-statistics/testing was uninstallable, it installed octave-statistics
> 1.7.5-4:
> 
>  30s The following packages have unmet dependencies:
>  30s  octave-io : Depends: octave-abi-59
>  30s  octave-statistics : Depends: octave-abi-59
>  30s E: Unable to correct problems, you have held broken packages.
>  30s autopkgtest: WARNING: Test dependencies are unsatisfiable with using
> apt pinning. Retrying with using all packages from unstable
> [...]
>  75s Preparing to unpack .../559-octave-statistics_1.7.5-4_amd64.deb ...
>  75s Unpacking octave-statistics (1.7.5-4) ...
> 
> But somehow the test failed, although I can't see the reason:
>...

It installed octave-statistics/sid binaries, but was still using 
octave-statistics/testing sources:

 29s Get:1 http://deb.debian.org/debian testing/main octave-statistics 1.7.5-3 (dsc) [2,434 B]
 29s Get:2 http://deb.debian.org/debian testing/main octave-statistics 1.7.5-3 (tar) [1,350 kB]
 29s Get:3 http://deb.debian.org/debian testing/main octave-statistics 1.7.5-3 (diff) [9,868 B]

> Cheers,
> Emilio
>...

cu
Adrian