CGAL users discussion list

[Peter Hachenberger <>]

> I have reverted back to not using CGAL::SNC_indexed_items, but increased 
> the 
> threshold to 1e-8 instead. This seems to give better results, although it 
> is 
> still possible to get failed assertions. I am catching the exceptions 
> associated with those now, to avoid crashing the program, but then I ran 
> into 
> an even rarer case where a segfault happens. The backtrace is attached, can 
> you see why this would happen?

If this assertion is thrown, the representation of the polyhedron will
not be consistent. So catching the assertion means to proceed with an
incomplete or faulty representation. Why did you revert back from
SNC_indexed_items? They are faster and give you a better chance of
succeeding.

Peter