Re: Build failures blocking testing

[Colin Watson <cjwatson@debian.org>]

On Mon, Jun 17, 2013 at 04:21:34PM +0200, Joachim Breitner wrote:
> I think this should be escalated to GHC, they know their code better and
> maybe your analysis is enough to make them say „ah, right, that part of
> the code is fishy, try this patch“.
> 
> (I guess I am overly optimistic.)

Well, they might at least be able to suggest ways to narrow it down.

  http://hackage.haskell.org/trac/ghc/ticket/7993

> >   * hxt-relaxng fails to build on mipsel [3] and s390 [4], with what
> >     looks like an out-of-memory failure.  There are no
> >     reverse-dependencies; remove?
> 
> I gave them back, just to confirm that this is reproducible (and hoping
> that another buildd can cope with it). If it does not work: Removing is
> the only option.

Still fails, so I've filed for removal.

  http://bugs.debian.org/712611

> >   * lens fails to build on kfreebsd-i386 [5]; I filed an upstream bug
> >     [6].  Should I just temporarily disable the tests on that
> >     architecture along with hurd-i386 where it also fails, or perhaps
> >     just disable that one test on all architectures since it's clearly
> >     unreliable?  
> 
> Yes, removing the test is the easiest here.

As luck would have it, upstream fixed this just as I was working on it,
so I backported their patch.

> BTW, it fails on sparc due to -N missing:
> https://buildd.debian.org/status/fetch.php?pkg=haskell-lens&arch=sparc&ver=3.9.0.2-2&stamp=1371444377
> I guess we can simply disable the tests here as well.

lens.cabal had flags to control the test suites in question, so I turned
just those ones off on sparc.

Getting there!

Thanks,

-- 
Colin Watson                                       [cjwatson@debian.org]