dolfin team mailing list archive

[Andre Massing <massing@xxxxxxxxx>]

Den 22.11.10 17.40, skrev Johan Hake:
> On Monday November 22 2010 07:36:25 Andre Massing wrote:
> > Hi!
> >
> > I am not entirely sure, whether there was a similar discussion about that,
> >    but I just recognized that using a ufl file like
> >
> > element = FiniteElement("Lagrange","tetrahedron", 1)
> > w = Coefficient(element)
> > M = w*w*dx(0)
> >
> > and a meshfunction with 0, 1, 2 values
> > results into the assembly of a function on *entire* domain and not only
> > cells that a marked with 0.
> > Can anybody confirm this behaviour?
> Nope...
>
> Just confirmed that I got the expected behaviour. Are you sure you passed the
> MeshFunction through the cell_domains argument?
Yes (C++ interface), but encountered something strange behaviour here. 
Have to dig a bit more into it ...
Thanks for confirming..
Andre

> We discussed various ways of implementing the behaviour you report but we
> ended up not doing anything ;)

> Johan