Re: Give back petsc [was: Re: Suitesparse transition]
On Wed, 2008-03-19 at 22:52 -0400, Adam C Powell IV wrote:
> On Wed, 2008-03-19 at 20:29 +0100, Rafael Laboissiere wrote:
> > Adam Powell uploaded version 2.3.3-7 of petsc that build-depends
> > alternatively on liopenmpi-dev or lam4-dev. This should fix the Dep-Wait
> > status on some arches, like arm, mips, mipsel, and s390.
> > Will the autobuilds automatically happen on these arches or should the
> > package be given back on them?
> False alarm. The way I tried it didn't work. I'll need to explicitly
> declare arches with openmpi and with lam. I'll upload tomorrow.
> But to answer your question, the autobuilders tried to make it work.
I uploaded new spooles and petsc this morning. Spooles has built
successfully on arm. petsc was tried on hppa, but failed because of
missing lam. lam has not failed on hppa, but 7.1.2-1.3 has not been
attempted; same with armel.
When the hppa and armel autobuilders build lam, then they should be able
to build spooles and petsc as well, unblocking suitesparse.
One other note: libmesh and petsc4py may need to be rebuilt after petsc
goes in, didn't see those on your igloo. Let me know if you want to
bin-NMU them; otherwise I will see about fixing a couple of libmesh bugs
GPG fingerprint: D54D 1AEE B11C CE9B A02B C5DD 526F 01E8 564E E4B6
Engineering consulting with open source tools