[Pkg-openmpi-maintainers] Bug#405929: mpich2 and openmpi FTBFS: no atomic primitives (but we have some!)

[Thorsten Glaser]

Dixi quod…

>Package: mpich2
>Justification: fails to build from source (but built successfully in the past)
>
>See attached build log. I’ll try building openmpi next.

OK, neither of these two builds (the openmpi problem still
exists and has been reported by cts already).

Atomic primitives are in libatomic-ops-dev and should work.
AFAIK (but please ask the m68k porters) there are no MP (SMP
or otherwise) machines out there, or at least supported by
Linux. You can pretty certainly reuse this information and
make either available on m68k.

Until then, please revert mpi-defaults to lam on m68k, so
that we can have build-dependencies fulfillable (or I just
don’t build a newer mpi-defaults). ISTR lvm2 depending on
this (oh, and boost…).

Thanks,
//mirabilos
-- 
Solange man keine schmutzigen Tricks macht, und ich meine *wirklich*
schmutzige Tricks, wie bei einer doppelt verketteten Liste beide
Pointer XORen und in nur einem Word speichern, funktioniert Boehm ganz
hervorragend.		-- Andreas Bogk über boehm-gc in d.a.s.r