Hi Markus, > I will have a look at it, too, but I think there the problem should not > occur, since, as far as I have understood, the issue only occurs with > nonhomogeneous boundary conditions.
What I meant is that in the case of nonhomogeneous boundary conditions the code was clearly wrong. But in the case of homogeneous ones we should follow Martin's implementation and only add a new constraint if the DoF isn't already constrained. I've changed this in compute_no_normal_flux_constraints now. You may want to do the same in the curl conforming function. > I have seen the DoFTools::make_hanging_node_constraints function in > connection with VectorTools::interpolate_boundary_values in steps 22 and > 27. I think Martin already changed step-22. I'll take a look at step-27. > As there is already a section on Dirichlet boundary conditions right > under the section on hanging node constraints, I think it would be a > good idea to just extend the section on Dirichlet boundary conditions. I've decided to create a doxygen module on constraints into which I just stuck most of the information that was previously in the ConstraintMatrix documentation. I now think this is the proper place for this sort of information. Best W. ------------------------------------------------------------------------- Wolfgang Bangerth email: [email protected] www: http://www.math.tamu.edu/~bangerth/ _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
