CGAL users discussion list

[Stephane Tayeb <>]

Ianic wrote:
> To make things clearer, I am talking here about the "File_medit.h"
> functionality. Basically, when I output the Vertices section of a .mesh
> file, I want to exclude the nodes for which ( get(vertex_pmap, vit) == -1 ).
> I wanted to do this because I thought these vertices would be unreferenced
> in the sense that they would not belong to any cell of c3t3 complex. But
> this is not the case and I wonder why. 
>
> Unfortunately, the File_medit.h file is overall too cryptic for my level of
> C++ proficiency, but I believe that get(vertex_pmap, vit) function should
> give a subdomain index of any cell incident to the vertex. In my case it
> sometimes gives "-1" but I have checked that some of those "-1" vertices
> belong to cells of c3t3 for which get(cell_pmap, cit) gives "2".

Hi,

I did some tests, and you I found the same behavior as you: some 
vertices of the c3t3 of dimension 2 are not incident to any facet. This 
seems to be a bug, I will look at it.

As a quick fix, I could say that you could just ignore the vertices tags 
of the .mesh output, as they are not very accurate:
 + for dim=3 vertices, the tag is the tag of any of its adjacent cell
 + for dim=2 vertices, the tag is the tag of *one* of its adjacent 
surface facet (one 'surface' vertex could be incident to surface facets 
with different indices)

However, this behaviour raised up in my case when I used 
CGAL::perturb_mesh_3. Could you please provide me more details about how 
you generate your mesh ?

Thank you,
Best,
Stéphane.

--
Stephane Tayeb
Software engineer - INRIA Sophia Antipolis
Geometrica Project-Team