[vtkusers] 3D cells behave differently

[Andy Miller]

> From: jnorris at uiuc.edu
> Reply-To: jnorris at uiuc.edu
> Subject: [vtkusers] 3D cells behave differently
>
> Hi all,
>
> I've noticed an inconsistancy in behavior of the 3D cells in vtk3.2beta.
> This is probably in older versions as well.
>
> My data is in an unstructured grid.  I need to be able to generate the
> surface (via vtkGeometryFilter) and then add a cutting plane.  I do this
by
> using vtkClipPolyData to clip the surface.  I cap the clipped surface by
> using a vtkCutter on the unstructured grid and appending the outputs of
the
> two filters together via vtkAppendPolyData.  The vtkCutter and
> vtkClipPolyData use the same vtkPlane object.  This works great for
pyramids,
> wedges, voxels, and hexahedrons, but not for tetrahedrons.  With the other

For me, the backside is facing out for both tets and pyramids.  On a related
note,
several memory access violations are generated in GDI32.DLL by the test
program
when I interact with either tets or pyramids (I have Windows 98), but not
the
others.

> 3D cells, the normals of the polygons generated by vtkCutter are opposite
> the normal of the cutting plane, so on my clipped surface they are facing
> out (which is what I want).  But with tets, they apparently are facing in.

Agreed since the cut surface is drawn correctly when BackfaceCulling is off.

> I can't simply pass the output of my vtkCutter through a vtkReverseSense,
> because I have to be able to handle grids with mixed cell types, and only
> the tets are facing in.
>
> Any ideas?
>
> Thanks,
> John
>
> P.S. I had an example that demonstrates this, but it mysteriously quit
> working while I was at lunch.  I'll email/post it once it's working again,
> if anyone is interested.

Works (demonstrates the problem) for me. No ideas yet.

Andy