Bug#520141: mpi-defaults: wrong arch constraints in Build-Depends
dato at net.com.org.es
Wed Mar 18 20:12:23 UTC 2009
* Manuel Prinz [Wed, 18 Mar 2009 00:08:16 +0100]:
> Am Dienstag, den 17.03.2009, 23:21 +0100 schrieb Adeodato Simó:
> > If, as you say, the sparc failure is expected to be fixed soon, I don’t
> > understand if there would be any problem with uploading with the
> > alpha/LAM change now, keeping sparc with openmpi. If you are concerned
> > that mpi-defaults will fail to build on sparc, that should not be a
> > concern: can get built when openmpi is built on sparc, but in the
> > meantime it will give us mpi-default-dev on alpha. Or is something else
> > to take into account that I overlooked?
> No, I just read your email as the upload was kind of urgent. If it is,
> I'd propose we go for alpha and sparc using LAM. If it's not, I can fix
> alpha and sparc in openmpi and upload mpi-defaults with alpha and sparc
> using Open MPI. I'd prefer this, though I do not know if it's reasonable
> since we could also go back to Open MPI with a later upload.
> What is your preferred option?
Well, my preference would be to get a working mpi-defaults-dev in all
architectures as soon as possible, but if you have grounds to believe
openmi will be available on alpha and again on sparc soon, I don’t have
any inconvenient to wait a few days. Let’s do that.
Thanks, and sorry for all the hassle.
- Are you sure we're good?
-- Rory and Lorelai
More information about the debian-science-maintainers