dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Apr 20, 2009 at 03:02:57PM +0200, Ola Skavhaug wrote:
> On Mon, Apr 20, 2009 at 2:21 PM, Anders Logg <logg@xxxxxxxxx> wrote:
> > On Mon, Apr 20, 2009 at 02:10:09PM +0200, DOLFIN wrote:
> >> One or more new changesets pushed to the primary dolfin repository.
> >> A short summary of the last three changesets is included below.
> >>
> >> changeset:   6005:68b2dc927f857e686f4990c7125c212eef24c012
> >> tag:         tip
> >> user:        "Ola Skavhaug <skavhaug@xxxxxxxxx>"
> >> date:        Mon Apr 20 14:11:49 2009 +0200
> >> files:       dolfin/mesh/MeshPartitioning.cpp sandbox/passembly/Poisson.h
> >> description:
> >> More work on parallell numbering of mesh entities.
> >
> > Could we do something like this to avoid using entities[0].size():
> >
> > uint i = 0;
> > while (i < values.size())
> > {
> >  const uint entity_index = values[i++];
> >  const uint num_vertices = values[i++];
> >  std::vector<uint> entity;
> >  for (uint j = 0; j < num_vertices; ++j)
> >    entity.push_back(values[i++])
> >  ...
> > }
> >
> > ?
> 
> Yes, this is possible. Why do you want that? Any plans for supporting
> more advanced heterogenous
> meshes?

Yes, at some point. But I just thought entities[0] looked ugly. :-)

If we look at entities[0] we don't need to communicate the number of
vertices in each entity (second value in data) but we might as well do
it anyway.

-- 
Anders

Attachment: signature.asc
Description: Digital signature