Re: math/suitesparse breakage

[Kevin Bloom <ktnb%netbsd.org@localhost>]

> From thomas.orgis%uni-hamburg.de@localhost  Wed Jun 18 15:24:34 2025
>
> Just here for the typos =E2=80=A6 since those differences could be crucial.
>
> Am Wed, 18 Jun 2025 14:05:41 +0000 (UTC)
> schrieb Kevin Bloom <ktnb%netbsd.org@localhost>:
>
> > If this line is added to suitespare's Makefile, on NetBSD it
> > fails with a linking error to gfortran.so.3
>
> That's so.5, right?

Yes, it is. Sorry!

> >         -lgfortran.5 =3D> /usr/pkg/gcc10/lib/./libgfortran.so.5
> >         -lquadmath.0 =3D> /usr/pkg/gcc10/lib/./libquadmath.so.0
>
> > 2. building math/lapack, math/blas, and susquently math/suitesparse
> >    on linux fails to build.
> > thor@ mentioned that he has this issue when using gcc12 but is able
> > to build it successfully with gcc12.
>
> I am able to build with gcc11 from Ubuntu and with gcc13 from
> self-installed toolchain on Debian 12. The gcc12 on Debian 12 also
> fails.
>
> So it might be not the GCC version itself, but the way it was built,
> and maybe just the binutils associated with it (I build binutils
> in-tree with gcc). This used to work before =E2=80=A6 I'm sure. Do we have =
> an
> overview of past bulk builds to see when math/blas started failing?
>
>

In these two you can see the same error that I was getting:

Debian:
2025-03-18 00:39:00 +0000 UTC
https://releng.netbsd.org/bulktracker/pkg/48765336

Linux 6.1.0
2025-03-20 00:28:00 +0000 UTC
https://releng.netbsd.org/bulktracker/pkg/48884577

This one builds:

Rocky Linux
2025-01-11 10:44:00 +0000 UTC
https://releng.netbsd.org/bulktracker/pkg/46771270

> Alrighty then,
>
> Thomas
>
> --=20
> Dr. Thomas Orgis
> HPC @ Universit=C3=A4t Hamburg
>