On Tue, Jun 17, 2008 at 08:36:19PM +0200, Anders Logg wrote: > On Tue, Jun 17, 2008 at 05:27:36PM +0100, Garth N. Wells wrote: > > > > > > Anders Logg wrote: > > > On Tue, Jun 17, 2008 at 12:27:23PM +0100, Garth N. Wells wrote: > > >> Something went wrong at changeset 4224, > > >> > > >> http://www.fenics.org/hg/dolfin?cs=c5a7da38ff2c > > >> > > >> Garth > > > > > > ok, good. I will take a look. > > > > > > > Problem appears to be related to the local facet number. I'm taking a look. > > > > Garth > > Yes! Thanks for the tip. > > The problem is that the local facet numbers are now created during the > call to the constructor of DirichletBC. Between then and the > assembly, the mesh gets reorerdered (by a call to order()) so the > facet numbers are wrong when we apply the boundary conditions. > > We need two things: > > 1. Make sure that the mesh is ordered when computing BCs. > > 2. Add a flag that allows checking if the mesh has been ordered > already (to prevent extra work).
ok. Both these are fixed now. -- Anders
signature.asc
Description: Digital signature
_______________________________________________ DOLFIN-dev mailing list [email protected] http://www.fenics.org/mailman/listinfo/dolfin-dev
