> the rationale for introducing MeshWorker was simplifying integration > loops over cells and faces. Now, currently we do not have faces in 1D, > therefore it did not work. Even if Wolfgang removed part of the code > in 1D, I never bothered about the implications of using MeshWorker in 1D, > since most of the reason for its existence vanishes in 1D. I could > imagine, that some typedefs and class definitions might fail in 1D.
Right. But I think in the long run we'd like to support this sort of thing since it may, for example, support building the system matrices along with the level matrices using only a single assembly function. That would be a win (in 1D as well as in higher dimensions) even if the problem uses continuous elements and consequently no face integration is necessary. The same would be true for example for matrix free methods. Or did I misunderstand what you had in mind? Best W. _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
