dolfin team mailing list archive

[Johan Hake <hake@xxxxxxxxx>]

On Wednesday 31 December 2008 09:48:32 Garth N. Wells wrote:
> Anders Logg wrote:
> > There seem to be only a few issues remaining before we can make a new
> > release:
> >
> > 1. Fix problem with shared_ptr in EpetraMatrix.
>
> If Johan can't fix this easily in PyDOLFIN, I can just add an extra
> member function for the time being to EpetraMatrix (as is done in
> EpetraVector).

It is true that I have looked into it, but I havn't got it worked yet. It is 
also notoriously difficult to debug and a small change will trigger a 
regeneration of the swig file... I have a proof of principle code (nothing 
related to DOLFIN) up and running with wrapped shared_ptr, that works in 
python, but as mentioned I have not got it work for PyDOLFIN and I do not 
think I will have anything up and running before the release.

> It would be nice if the Epetra classes could be completed in the near
> future. There is quite some functionality missing.

While at it. The set function in the EpetraVector will segfault from python. 
Uncomment the part that test the Epetra la interface in 

   test/unit/la/python/test.py

for an illustration.

Running this test I also got an error when using the wrapped operator() in the 
PETScMatrix from python. This function just calls the get() function, which 
raises:

   RuntimeError: *** Assertion (A) [at dolfin/la/PETScMatrix.cpp:210 in get()]


Johan

> > 2. Fix problem with using PETSc C++ interface in PETSc classes.
>
> I'll take a look at this.
>
> We could ask the maintainer of the Debian package to add C++ support in
> future packages.
>
> PETSc 3 is out now. Should we move to it after the next release?
>
> > 3. Maybe fix assemble_system. It currently seems to be broken. To
> > test, set the "symmetric" option in the DG demo:
> >
> >   dolfin_set("symmetric", True)
> >
> > Is anyone going to look at assemble_system now?
>
> I looked at it a while ago and the results looked ok for DG demo. I'll
> take a look again. It's an important feature, so it should be fixed
> before a release.
>
> Garth
>
> Otherwise, we can
>
> > leave it as it is. I've added a warning:
> >
> >   warning("assemble_system() may not give correct results for forms
> > involving facet integrals.");
> >
> >
> >
> > ------------------------------------------------------------------------
> >
> > _______________________________________________
> > DOLFIN-dev mailing list
> > DOLFIN-dev@xxxxxxxxxx
> > http://www.fenics.org/mailman/listinfo/dolfin-dev
>
> _______________________________________________
> DOLFIN-dev mailing list
> DOLFIN-dev@xxxxxxxxxx
> http://www.fenics.org/mailman/listinfo/dolfin-dev