CGAL users discussion list

[Sylvain Pion <>]

Panagiotis Foteinos a écrit :
> 1) When a new tetrahedron, say t,  is created: Is a new Cell_handle
> allocated to accommodate t, or Cell_handle might point to a previously
> allocated (but not currently used for any tetrahedron accommodation)
> memory? I am asking because it would be ideal for my purposes, if no new
> tetrahedron takes info() values from an old one...

It is the case in the current implementation, yes.

Some notes, however :
- I think it is not the case during the "bootstrapping phase",
 when going from dimension 1 to 2 and 2 to 3.

- Delaunay_triangulation_2 currently differs here.

- we may change this in the future, for efficiency reasons,
 but it's not clear yet when, nor for what exact behavior.

> 2) When a Vertex_handle, say v, is inserted: Does  a Cell_handle,
> accommodating a tetrahedron outside the cavity/hole of v, point to
> another memory address after v's insertion? The answer should be
> negative, just making sure...

Definitely negative.  Cells outside the conflict hole are only changed
in the neighbor pointers that need to be changed.

> 2) When a Vertex_handle, say v, is deleted: Does  a Cell_handle,
> accommodating a tetrahedron outside the hole to be retriangulated, point
> to another memory address after v's removal? The answer should be
> negative, just making sure...

Negative as well.

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