dolfin team mailing list archive

["Johannes Ring" <johannr@xxxxxxxxx>]

On Fri, January 16, 2009 13:38, Johan Hake wrote:
> On Friday 16 January 2009 13:30:42 Johannes Ring wrote:
>> On Fri, January 16, 2009 13:22, Johan Hake wrote:
>> > On Friday 16 January 2009 13:18:23 Johannes Ring wrote:
>> >> On Fri, January 16, 2009 13:06, Johan Hake wrote:
>> >> > On Friday 16 January 2009 13:03:20 Johannes Ring wrote:
>> >> >> On Fri, January 16, 2009 09:08, Kent Andre wrote:
>> >> >> > On to., 2009-01-15 at 19:05 +0100, Martin Sandve Alnæs wrote:
>> >> >> >> I don't know, but I noticed these lines:
>> >> >> >>
>> >> >> >> A_epetra = cpp.down_cast_EpetraMatrix(A).mat()
>> >> >> >> b_epetra = cpp.down_cast_EpetraVector(b).vec()
>> >> >> >> x_epetra = cpp.down_cast_EpetraVector(U.vector()).vec()
>> >> >> >>
>> >> >> >> you should be able to do just
>> >> >> >>
>> >> >> >> A_epetra = down_cast(A).mat()
>> >> >> >> b_epetra = down_cast(b).vec()
>> >> >> >> x_epetra = down_cast(U.vector()).vec()
>> >> >> >>
>> >> >> >> the type is found in a python layer by introspection.
>> >> >> >>
>> >> >> >> Martin
>> >> >> >
>> >> >> > OK.
>> >> >> >
>> >> >> > Trilinos and Dolfin are compiled with different versions of
>> SWIG.
>> >> >>
>> >> >> Anyone
>> >> >>
>> >> >> > have experience with this? Could this be the problem?
>> >> >>
>> >> >> The Trilinos packages at packages.simula.no are now recompiled
>> with
>> >>
>> >> SWIG
>> >>
>> >> >> 1.3.37. They are only available for Hardy at the moment. Let me
>> know
>> >>
>> >> if
>> >>
>> >> >> anyone wants me to build Trilinos packages for Intrepid as well.
>> >> >
>> >> > If you compile PyDOLFIN with swig 1.3.37, (and that could actually
>> be
>> >> > nice),
>> >> > you also have to compiled Trilinos with this version.
>> >>
>> >> I haven't touched the DOLFIN packages, only the Trilinos packages has
>> >> been
>> >> rebuilt.
>> >
>> > Ok, but I suppose you built the hardy DOLFIN with swig 1.3.37?
>>
>> No, the DOLFIN 0.9.0 packages were built with standard Hardy SWIG (i.e.
>> version 1.3.33). This was sometimes last week I think, before you made
>> your changes.
>
> Then will the DOLFIN from this package not be compatible with the trilinos
> package.

Yes, I noticed that now. When running the Trilinos demos I got this:

Checking mesh ordering (finished).
Calling FFC just-in-time (JIT) compiler, this may take some time... done
Creating Python extension (compiling and linking), this may take some
time... done
Traceback (most recent call last):
  File "demo.py", line 30, in <module>
    V = FunctionSpace(mesh, "CG", 1)
  File "/usr/lib/python2.5/site-packages/dolfin/functionspace.py", line
183, in __init__
    FunctionSpaceBase.__init__(self, mesh, element)
  File "/usr/lib/python2.5/site-packages/dolfin/functionspace.py", line
48, in __init__
    self.__dolfin_element = cpp.FiniteElement(self.__ufc_element)
  File "/usr/lib/python2.5/site-packages/dolfin/cpp.py", line 10300, in
__init__
    this = _cpp.new_FiniteElement(*args)
TypeError: in method 'new_FiniteElement', argument 1 of type 'std::string'

I also got similar errors for the poisson Python demo. I guess I'll have
to rebuild DOLFIN with SWIG 1.3.37?

Johannes