CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

On 01/29/2013 01:25 PM, Vero wrote:
> Dear Sebastien,
>
> If by making the triangulation invalid you mean that I may get a non
> planar graph with overlapped triangles, my own code is trying to handle
> that, in part with the springs, but to do it properly I have been making
> some assumptions about the way ccw and cw work... concretely I had
> thought that they would return the next vertex as it would have been
> when the triangulation was created.  But, if by "making the
> triangulation invalid" you mean that these kind of operations will not
> behave consistently any more, I suppose I will have to look for another
> solution.
You will find in [1] what is expected for the triangulation data 
structure to be valid.

cw_v and ccw_v are simply combinatorial (there are no predicate 
involved). You may need to use an orientation predicate

> Unfortunately move doesn't help, because, as I understood
> from the documentation, it modifies the triangulation, that would mess
> around the springs and the constants that had been assigned to them may
> no longer work.  Or did I understood wrong what the documentation means?

If by the "it modifies the triangulation" you mean that the Delaunay 
property is preserved while moving vertices then yes. However, the 
information associated to vertices will not change (only that of the 
faces created/destroyed)

Sebastien.

[1] 
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/TDS_2/Chapter_main.html#Subsection_38.1.1

[2] 
http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Kernel_23_ref/Function_orientation.html

> Thanks a lot.
>
>
> On 28 January 2013 12:25, Sebastien Loriot (GeometryFactory)
> <
>  
> <mailto:>>
>  wrote:
>
>     Modifying the triangulation using vh->set_point(newPoint) does not
>     modify the combinatorial structure of the triangulation and can
>     make the triangulation invalid.
>
>     You need to use the move function instead: t.move(vh,newPoint).
>
>     Sebastien.
>
>
>     On 01/28/2013 12:33 PM, Vero wrote:
>
>         Dear members,
>
>         I have been using the Constrained_Delaunay___triangulation_2 for a
>         mass-spring model.  However, I am getting somewhat worried that
>         when the
>         triangles get flipped, the behaviour of ccw and cw may not be
>         returning
>         the vertices according to their original orientation, but
>         according to
>         their current orientation.  I had supposed the opposite.  Could you
>         please give me some help here?  To set the new coordinate of the
>         vertices I have been using vh->set_point(newPoint).
>
>         Thanks,
>
>         --
>
>         Veronica E. Arriola
>
>         Room G5 Chemistry West,
>         School of Computer Science
>         University of Birmingham
>         http://www.cs.bham.ac.uk/~__vxa855/
>         <http://www.cs.bham.ac.uk/~vxa855/>
>
>
>
>     --
>     You are currently subscribed to cgal-discuss.
>     To unsubscribe or access the archives, go to
>     https://sympa.inria.fr/sympa/__info/cgal-discuss
>     <https://sympa.inria.fr/sympa/info/cgal-discuss>
>
>
>
>
>
> --
>
> Veronica E. Arriola
>
> Room G5 Chemistry West,
> School of Computer Science
> University of Birmingham
> http://www.cs.bham.ac.uk/~vxa855/