Re: parallel installability of g77/gfortran lapack/blas/atlas planned?

[Matthias Klose <doko@cs.tu-berlin.de>]

[note that I got rejects from my last emails to Camm, that his mail
server could not be reached]

Kevin B. McCarty writes:
> Hi Camm,
> 
> as I understand it the gfortran-based versions of LAPACK, BLAS and ATLAS
> are still only in experimental, right?
> 
> Do you intend to make it possible (before uploading into unstable) to
> install the runtime library packages for the g77 versions and the
> gfortran versions of these libraries in parallel?  Or are you going to
> use the simpler model originally suggested by Riku, in which the
> gfortran versions have the same soname as the old g77 versions, and
> Conflict against them?

What do you mean by "before uploading into unstable"? If that is not
meant as a final package, then why not install the old and the new
libs in two different chroots?

> I ask because I am wondering whether to make it possible to do such a
> parallel-install of Cernlib libraries, but of course if you are not
> going to permit a parallel install, I don't really have the choice.  (In
> principle I could still do it for the CERN libraries that don't depend
> on LAPACK, but that would not be so useful.)  So I would like to know
> now, before expending much energy on parallel installability.

Non-conflicting packages would mean to upload with libraries with a
changed soname.  I don't think that binaries are that portable, but
does it really make sense to keep both sets of libraries, and for each
package build for g77 *and* for gfortran?

  Matthias

PS: Seems that we are still stalled, so currently I'd prefer anything
that brings us a bit forward...