[Gmsh] Is about a bug in Von-Mises vs time representation

Christophe Geuzaine geuzaine at acm.caltech.edu
Tue Feb 24 17:54:35 CET 2004


Jean-Luc FLEJOU wrote:

> 
> Hello,
> 
> I have just met an error (GMSH 1.50) in the representation of the tensors,
> when there is a time list.
> Here a file to highlight the bug
> 
> (See attached file: tenseur_cub8.pos)
> It seems to me that the calculation of Von-Misès do not take into account
> the time list.
> 
> In file "Graphics/PostElement.cpp" procedure "void Draw_TensorElement"
> 
> There is
>   // View->TensorType == DRAW_POST_VONMISES
>   double V_VonMises[8];
>   for(int i = 0; i < nbnod; i++){
>     V_VonMises[i] = ComputeVonMises(V + 9*i);
>   }
> 
> There should be something of that
>   // View->TensorType == DRAW_POST_VONMISES
>   int ts = View->TimeStep;
>   View->TimeStep = 0;
> 
>   double V_VonMises[8];
>   for(int i = 0; i < nbnod; i++){
>      V_VonMises[i] = ComputeVonMises(V + 9*(i + nbnod * ts));
>   }
> 
> and right before leaving this procedure
> 
>   View->TimeStep = ts;
> 

Good catch!

Thanks,

Christophe

-- 
Christophe A. Geuzaine
Applied and Computational Mathematics, Caltech
geuzaine at acm.caltech.edu - http://geuz.org