ufl team mailing list archive

[Kristian Oelgaard <k.b.oelgaard@xxxxxxxxxx>]

Quoting Martin Sandve Alnæs <martinal@xxxxxxxxx>:

> I have zero knowledge of discontinuous Galerkin methods,
> so there are no checks at all. What should we check for?
>
> Do you mean (implementation must be different):
>
> if integral is on interior facets:
>     for f in basis function and functions in integrand:
>         assert that f is restricted
> ?

Yes, that's what I would expect. Constants must also be restricted but they're
defined on DG elements right? So this is not a problem.

> We should probably propagate ('+') and ('-') up
> the tree to reach the leaves (terminals), right?

Yes, such that grad(u)('+') == grad(u('+'))

Kristian

> Martin
>
>
>
> On Thu, May 14, 2009 at 7:52 AM, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
> > FFC used to throw an error if functions were integrated on interior
> > facets but not restricted to the '+' or '-' side, but UFL doesn't. If I do
> >
> >     element = FiniteElement("Discontinuous Lagrange", triangle, 1)
> >     v = TestFunction(element)
> >     u = TrialFunction(element)
> >
> >     # Integrate on interior facet
> >     a = v*u*dS
> >
> > what result should I expect? I had anticipated an error.
> >
> > Garth
> > _______________________________________________
> > UFL-dev mailing list
> > UFL-dev@xxxxxxxxxx
> > http://fenics.org/mailman/listinfo/ufl-dev
> >
> _______________________________________________
> UFL-dev mailing list
> UFL-dev@xxxxxxxxxx
> http://fenics.org/mailman/listinfo/ufl-dev
>