input files for iMesh_unitTest

[Carl Ollivier-Gooch]

Mark Miller wrote:
> Hi Carl,
> 
> Ok, so I am a little puzzled by this.
> 
> Can't we use tags to define some persistent, impl. neutral 'id' for vtx
> entities and then use that 'id' to confirm a specific vtx entity has the
> desired coords or adjacencies? Or, is that what you are basically
> saying?

That works for full implementations.  But I'm thinking about this in the
context of modular testing of partial implementations, which may not
have tags.

So my plan is to match up vertices from an implementation to be tested
(Test) and a reference implementation that is known to be correct (Ref).
  (Yes, this is a floating point comparison, and one has to be careful
with tolerances...)    Mismatches will imply some problem with vertex
coordinate retrieval and/or setup of the data in Test.  If successful,
this comparison process will give me a two-way map between the vertex
handles.

Then it's easy enough to test correctness of downward adjacency to
vertices (remember, Ref is known to be correct, so Test can be compared
to that).  Other adjacency tests can be resolved based on X->vertex in
one way or another.

Carl

> On Tue, 2009-10-27 at 17:43, Carl Ollivier-Gooch wrote:
> 
>> will require that an implementation have some way of 
>> getting its internal state to match some legacy VTK file.  This will 
>> make it possible to test correctness of adjacency calls and vertex 
>> coords, etc.  


-- 
------------------------------------------------------------------------
Dr. Carl Ollivier-Gooch, P.Eng.                   Voice: +1-604-822-1854
Associate Professor                                 Fax: +1-604-822-2403
Department of Mechanical Engineering             email: cfog at mech.ubc.ca
University of British Columbia              http://www.mech.ubc.ca/~cfog
Vancouver, BC  V6T 1Z4                  http://tetra.mech.ubc.ca/ANSLab/
------------------------------------------------------------------------