CGAL users discussion list

[Stefan Salewski <>]

On Thu, 2013-10-10 at 22:54 +0200, Stefan Salewski wrote:
> May the problem be simple that the output loop fails when we have only
> one visible site?

After reading (again) some documentation about CGAL's circulators, I
really guess that there is a bug.

It is clear that I have to test for empty graph, maybe something like
circ == NULL (I read about tests for empty ranges, but I guess that will
not work well for hulls.)

But ++circ should never give invalid values for circulators, from my
understanding of the documentation. But segmentation fault indicates
that there is an undefined value.

So it seems that I really have to handle the case with only one visible
side separately, which makes my code a bit ugly.