dolfin team mailing list archive

[Harish Narayanan <harish.mlists@xxxxxxxxx>]

Garth N. Wells wrote:
> 
> 
> Harish Narayanan wrote:
>> Garth N. Wells wrote:
>>> I've modified petsc.py to work for PETSc 3.0.0 which I guess should
>>> go into simula-scons.
>>
>> I've linked to a patch purports to update DOLFIN to the minor changes in
>> the PETSc interface from 2x -> 3x. It seems to work for me with PETSc
>> 3.0.0, and will break with PETSc 2x. (It doesn't check for the PETSc
>> version and do the appropriate thing.)
>>
>> If anyone is interested in testing it with PETSc 3.0.0, here it is:
>> http://home.simula.no/~harish/files/patches/dolfin-harish-2009-02-07.patch
>>
>>
> 
> Can you add some ifdefs so that it will work with PETSc 2.3.3 and PETSc
> 3.0.0? It doesn't look like there are many interface changes in PETSc.

I will do that. There is one thing I don't know how to handle.
dolfin/la/PETScMatrix.cpp uses the following variables that are no
longer defined in PETSc 3.0.0's petscmat.h.

+//   case spooles:
+//       return MATSEQAIJSPOOLES;
+//   case superlu:
+//       return MATSUPERLU;
+//   case umfpack:
+//       return MATUMFPACK;

I didn't spot them while grepping through the new petsc release either.

Harish