Re: [Libmesh-devel] Destruction!

[Roy S. <roy...@ic...>]

On Thu, 29 Jun 2006, Derek Gaston wrote:

> Hmmm... it would appear that calling
> MeshTools::Modification::all_tri() _after_ doing a uniform refinement
> is a bad idea....

Yes.  Yes it is.  I don't think all_tri() knows anything about
refinement hierarchies.  Would you throw the line:

         assert ((*el)->parent() == NULL);

in there, right after etype is declared, and make sure that assert
catches all_tri() on refined meshes but doesn't break all_tri() on
coarse meshes?
---
Roy