[R-pkg-team] Bug#983748: dh-r: pkg-r-autopkgtest breaks autopkgtests

[Andreas Tille]

Hi Graham,

On Wed, Mar 03, 2021 at 12:44:34PM +0200, Graham Inggs wrote:
> Hi Andreas
> 
> On Wed, 3 Mar 2021 at 12:12, Andreas Tille <tille at debian.org> wrote:
> > My motivation for the change was that for
> > example in r-bioc-mutationalpatterns (see bug #983027) the
> > autopkgtest-pkg-r script failed to install all needed packages
> > to run its test.  This was not the only package that was affected.
> 
> This sounds like #961138.

Yes.  I admit I failed noticing this totally since I'm relying on
my autopkgtest pbuilder hook ... which is run as root. :-(

> > I'd like to discuss this quickly whether I'm again on the wrong track.
> 
> In my opinion, dh-r 20210301 with the commit [1] reverted should be OK
> for bullseye.  We should be able to see the effect in unstable soon
> after your upload.

Done.
 
> > The effect would be that *a few* tests in r-bioc- packages will become
> > superfluid but that should be bearable for the moment.
> 
> What about skipping pkg-r-autopkgtest on the few affected packages?

That's a maintenance burden since our clever tools keep on re-injecting
it if its missing.

> We would lose a lot of test coverage if you disabled it everywhere.
> Let's hear opinions from the rest of the team.

I think by just droping[1] it should work without my drastical
suggestion.  So I uploaded according to your suggestion and added a hint
to TODO for the remaining packages to install.  I tested
r-bioc-mutationalpatterns which motivated me to the change and
it passes the tests.

I repeat that it was not a clever move to do [1] in that stage of
the release. :-( :-(

Kind regards

     Andreas.

> [1] https://salsa.debian.org/r-pkg-team/dh-r/-/commit/62a4128c07a17de845e9d7260504fae187c3d193
> 
> _______________________________________________
> R-pkg-team mailing list
> R-pkg-team at alioth-lists.debian.net
> https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/r-pkg-team
> 

-- 
http://fam-tille.de