CGAL users discussion list

[Sylvain Pion <>]

Daniel Duque wrote:
> I am thinking about efficiently adding info to the edges of a triangulation. 
> E.g., imagine you frequently need their lengths, and do not want to compute 
> them each time you need them, just once when the triangulation is built.
>
> Since edges are not explicitly represented, I had the idea of using 
>
> Cell_circulator  	  tds.incident_cells ( Edge e)  Starts at an arbitrary cell 
> incident to e
>
> to get one cell incident to an edge. The cell could be provided with info 
> relevant to the edge. E.g., a simple way would be to keep a 4x4 matrix, with 
> element i,j giving the info for the vertex with indices i,j.
>
> My question is if this approach is (basically) sensible.

I think it's OK.

If you are happy with a log(n) access time, then of course an external
map<Edge, Info> or map<pair<Vertex_handle, Vertex_handle>, Info>
could work as well.  Same for the unordered_map (hash) variants.

> Also, I am assuming the "arbitrary" cell will always be the same one for a 
> given edge. I would be surprised if this was not the case.
>
> BTW, there was a previous email from "400555 <  >" about this 
> last July, but it did not get any answers.

Yep, it's clear that built-in support for this in CGAL would help a lot.

--
Sylvain Pion
INRIA Sophia-Antipolis
Geometrica Project-Team
CGAL, http://cgal.org/