> But I don't quite understand the patch. It changes a conditional that reads
>   # First up is a twiddle for Darwin/MacOSX
>   ifeq ($(findstring darwin,$(TARGET)),)
>     ...
>   else
> Am I right that the comment doesn't make any sense? We get into the ifeq if
> we are *not* on darwin, right? findstring is defined as

Correct. We have the right logic on darwin but always use petscall.so
otherwise. This was fine until now, but I guess we need to take the
same logic as at the bottom (namely for 3.1+ use libpetsc).

> So if I understand the result of your patch correctly, you bypass
> libpetscall.so in favor of the original libpetsc.so on non-darwin systems,
> right? I think that is the long-term correct fix anyway, now that PETSc
> links itself into a single librarary...

Yes, something I wanted to discuss with y'all in Heidelberg but first
I was trying to collect some evidence. Note that the patch is not my
proposed solution but just a quick hack to get it to work with petsc
3+ on my machine. ;-)

-- 
Timo Heister
http://www.math.tamu.edu/~heister/
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii

Reply via email to