Re: Commit ‘gnu: ghc-8.10: Disable failing test on i686.’ has a cross-compilation bug

[Maxime Devos <maximedevos@telenet.be>]

Lars-Dominik Braun schreef op za 06-11-2021 om 12:07 [+0100]:
> Hi Maxime,
> 
> > https://git.savannah.gnu.org/cgit/guix.git/commit/?id=89de1924cb535fc2c97d3654e21badaebd43518e
> > 
> > > +           ,@(if (string=? "i686-linux" (%current-system))
> > […]
> > 
> > Barring any reports of the contrary, I'd presume the same would
> > apply for the Hurd. Also, %current-target-system seems more
> > appropriate, though here it doesn't matter because it's only
> > for tests.
> the GHC package declares support for x86 and x86-64 on Linux only,
> because we don’t have a bootstrap path for the Hurd (there are
> no prebuilt binaries) and the Hurd is not officially supported by
> upstream. Unless someone puts some effort into that I doubt it’ll
> ever
> run on that platform.

Currently, non-Linux is not supported by the GHC package.  However,
people learn how to package things by example (and by reading
documentation, etc.), so I'd prefer to avoid (accidentally) teaching
people to make their package definitions Linux-specific.

> 
> > I suggest: ,@(if (target-x86-64?) '(...) '())
> I don’t see `target-x86-64?` being defined on master and I assume you
> meant `target-x86-32?`, right?

Yes, I meant target-x86-32?. I don't think it is defined on master,
but it could be ‘back-ported’ from core-updates.

Greetings,
Maxime.
-- 
not hacking on guix for a while, only occassionally looking at IRC logs
and bug reports.  E-mails are unsigned until backup is located.