CGAL users discussion list

[<>]

I'm experimenting  with using Delaunay_Mesher_2, and I notice that in my 
trials
I get non-deterministic results; running the same input through multiple times
can give drastically different results.

This can be seen with a simple test case using the mesh_class.cpp sample code
provided with CGAL; if I lower the maximum edge length from .5 to .1 in line 
49
and then run the code repeatedly, the number of vertices in the final mesh 
will
vary between 2157 and 2170. 

Is this section of CGAL meant to be deterministic? If not, is there any way I
can try to force it to be? Searching over the mailing list archives I see
people trying to make other sections of CGAL deterministic by removing random
generators from various parts of the code, but I wasn't having any luck
figuring out where to do so to get repeatable results from the 2D mesher.

Thanks,
Micah