first try at actions of suiteparse/openmpi transition (was: Re: SuiteSparse 3.2.0->3.4.0 transition)

[Rene Engelhard <rene@debian.org>]

Kurt Roeckx wrote:
> > petsc itself is a problem, though, segfaults on hppa :(.
> > (Can this be ignored please given the general hppa problems and this
> > stuff hinted in?)
> 
> Given back too, this ussually helps.

It failed to some new error (#535276). NMUed it.
Unfortunately it failed on hppa again[1].. (#529485, this
according to the maintainer seems spuious, though)

Can the RM team then please:
 a) add a hint

hint freemat/3.6+dfsg-8+b1 illuminator/0.11.0-3 hypre/2.4.0b-2 openmpi/1.3.2-4 lp-solve/5.5.0.13-6 octave3.0/1:3.0.5-6+b1 octave3.0/1:3.0.5-6+b2 openoffice.org/1:3.1.0-5 petsc/3.0.0.dfsg-5 trilinos/9.0.3.dfsg-1 python-scipy/0.7.0-1+b1 pysparse/1.0.1-5.1 pysparse/1.0.1-5.1+b1 [2]

(no idea about the correct version specifiers for bin-NMUs and what do to
in the ocave3.0 case where different archs have different +bX)

together with the (already existing)

remove libmesh/0.6.2.dfsg-1

Maybe that even should be removed from sid or orphaned?

$ grep-excuses libmesh
libmesh (0.6.2.dfsg-1 to 0.6.3.dfsg~rc1-1)
    Maintainer: Debian Scientific Computing Team 
    306 days old (needed 10 days)
    libmesh0.6.3/i386 unsatisfiable Depends: libscotch-5.0.6
    libmesh0.6.3/alpha unsatisfiable Depends: libscotch-5.0.6
    libmesh0.6.3/amd64 unsatisfiable Depends: libscotch-5.0.6
    libmesh0.6.3/ia64 unsatisfiable Depends: libscotch-5.0.6
    libmesh0.6.3/powerpc unsatisfiable Depends: libscotch-5.0.6
    libmesh0.6.3/sparc unsatisfiable Depends: libscotch-5.0.6
    libmesh (source) has new bugs!
    Updating libmesh introduces new bugs: #513814, #534057, #529477
    Not considered

I tried to NMU it but without knowleeg of the build system....

And maybe even age-days mono/petsc/trillinos....

 b) block packages involved in the transition from being uploaded to *sid*?

freemat
illuminator
hypre
openmpi
lp-solve
octave3.0
openoffice.org [2]
petsc
trilinos
python-scipy
pysparse
mono
libmesh

and
 c) try to get petsc into a working/built state or just force in without...

Grüße/Regards,

René

[1] can we please ignore hppa failures for at least failures which are not
    the packages' fault?
[2] Thankfully, due to consequent use of -Wl,--as-needed the 3.1.1 püackages
    won't have direct deps on libcolamd2.7.1 anymore, see
    http://packages.debian.org/experimental/openoffice.org-calc
-- 
 .''`.  René Engelhard -- Debian GNU/Linux Developer
 : :' : http://www.debian.org | http://people.debian.org/~rene/
 `. `'  rene@debian.org | GnuPG-Key ID: 248AEB73
   `-   Fingerprint: 41FA F208 28D4 7CA5 19BB  7AD9 F859 90B0 248A EB73