On Nov 16, 2007, at 3:52 PM, asreeve wrote:
Python script is now attached.
On Fri, 16 Nov 2007, asreeve wrote:
I ended up modifying the GmshImporter2D to allow it to import
quads. I've attached a script containing the changes along with a
gmsh mesh file to test it. There are still some changes I'd like
to make to this to allow this to handle a grid with different
types of elements (mix of triangles, quads).
Thnaks for the patch, Andy. It looks reasonable, but I'll make Daniel
take a closer look at it when he comes in tomorrow. You're right,
gmshImporter.py was written specifically to handle triangles and
tets. Our assumption at the time was that square meshes were easier
to generate by hand. If you had anything complicated, then we assumed
that you'd let gmsh do everything, using triangles or tets because
they are relatively isotropic and are difficult to generate by hand.
Unlike you, we're not often starting with strongly gridded initial
data, so there's no intrinsic benefit for us to have square cells
within an irregular boundary (we're not usually interested in
irregular boundaries to begin with).
Still, there's no reason *not* to support the other meshes that gmsh
can export; it's just a question of somebody who actually needs it
finding the time to write the code. Thanks again for the contribution.
