Hi,

Quoting Antoine Beaupré (2022-05-26 16:19:53)
> On 2022-05-26 16:03:04, Christian Kastner wrote:
> > On 2022-05-25 17:40, Antoine Beaupre wrote:
> >> Setting up dose-distcheck (7.0.0-1+b1) ...
> >> Setting up sbuild-build-depends-dose3-dummy (0.invalid.0) ...
> >> (I)Doseparse: Parsing and normalizing...
> >> (I)Dose_deb: Parsing Packages file -...
> >> 
> >> ... and then it's just *stuck* there forever. In fact, even getting
> >> rid of that thing is hard: just one good old control-c doesn't cut
> >> it. You need to ^C the *heck* out of it and it eventually succeeds.
> >> 
> >> I would expect this to fail early. In fact, there *is* an error in apt
> >> earlier, but it seems the error is ignored as it just carries on from
> >> there:
> >> 
> >> The following packages have unmet dependencies:
> >>  sbuild-build-depends-main-dummy : Depends: python (>= 2.7) but it is not 
> >> installable
> >> E: Unable to correct problems, you have held broken packages.
> >> apt-get failed.
> >> E: Package installation failed
> >
> > I've seen this super-hard-to-kill hang with dose encountering
> > uninstallable packages before (although it seems that I didn't file a
> > bug about this).
> >
> > The simple solution that I found was to use apt as the build dependency
> > resolver. To this end, sbuild-qemu passes
> >
> >     --bd-uninstallable-explainer apt
> >
> > to sbuild. And on my host, gbp cloning userdir-ldap and running
> >
> >   $ sbuild-qemu *.dsc
> >
> > terminates just fine with an error message.
> >
> > What's odd is that this sbuild option seems to get lost on your end. Are
> > you perhaps overriding this option somehow?
> >
> > Otherwise, I would suspect an option parsing issue similar to the
> > --overlay-dir issue that we have been discussing privately.
> >
> > Will look into this, but if you have any info on a possible override of
> > --bd-uninstallable-explainer to share, please do.
> 
> I don't believe I do have an override like this. Attached is my (messy)
> .sbuildrc.

your sbuildrc contains some stuff that I find weird but nothing that should
have the effect you are seeing.

Since I'm also maintaining dose3, I'm very interested in reproducing the
problem. Unfortunately, I cannot build the package you are building as it uses
debhelper compat 5 which is not only deprecated since 2016 but also completely
removed from current debhelper. I fixed this locally and am unable to reproduce
your problem:

    Setting up sbuild-build-depends-dose3-dummy (0.invalid.0) ...
    (I)Doseparse: Parsing and normalizing...
    (I)Dose_deb: Parsing Packages file -...
    (I)Dose_common: total packages 66626
    (I)Dose_applications: Cudf Universe: 66626 packages
    (I)Dose_applications: --checkonly specified, consider all packages as 
background packages
    (I)Dose_applications: Solving...
    output-version: 1.2
    native-architecture: amd64
    report:
     -
      package: sbuild-build-depends-main-dummy
      version: 0.invalid.0
      architecture: amd64
      status: broken
      reasons:
       -
        missing:
         pkg:
          package: sbuild-build-depends-main-dummy
          version: 0.invalid.0
          architecture: amd64
          unsat-dependency: python:amd64 (>= 2.7)
    
    background-packages: 66625
    foreground-packages: 1
    total-packages: 66626
    broken-packages: 1

Given that you seem to work with compatibility level 5, are you sure that you
are working on an up-to-date system? Your bug report says you are on stable but
you seem to be mixing in testing and unstable....

Christian already suggested to use --bd-uninstallable-explainer=apt. If you
don't care about the explainer you can also completely disable it with
--bd-uninstallable-explainer=none.

Thanks!

cheers, josch

Attachment: signature.asc
Description: signature

Reply via email to