[vtkusers] Error in vtkQuad associated with problem with vtkCellDerivatives

[Bill Lorensen]

Richard,

Looking at the test coverage for vtkQuad.cxx
http://public.kitware.com/vtk/quality/Coverage/common/vtkQuad.txt
I see that the Derivatives method is not tested by our regression suite.
I'm guessing that it is incorrect. If you have an example in tcl or C++ that
can validate your suggested change, we'll add the fix to the code and your test
to the regression suite.

Bill

At 08:34 AM 9/7/00 -0600, Richard Strelitz wrote:
>I have been having problems with vtkCellDerivatives in that it 
>unfailingly returns values of "0" regardless of the actual values
>of the data.  I was using it with a unstructured grid from which
>I had extracted a layer.  Each cell was thus a vtkQuad.  
>
>  The computation of the derivatives depends on the method
>     cell->Derivatives(subid, pcoords, scalars, dim, derivs)
>  In the process of executing, on line 678  in vtkQuad, the value
>  of the sum[1] is computed as += funcDerivs(3+i)*values[dim*i+j]
>
>**** I am pretty sure that it should be funcDerivs(4+i)  *** 
>but, I am also pretty sure that that is not the sole source of my trouble.
>Has anyone else had dealings with this??
>
>(Just for the record, I am using the latest donwload, but the vtkQuad.cxx
>   module is Version:  $Revision 1.6.3  $  of 2000/02/04 17:03:44 $
>
>Thanks,
>   Richard Strelitz,
>   SciVis Staff , LANL CCS (nee ACL)