CGAL users discussion list

[Marius Kintel <>]

On May 30, 2012, at 07:59 AM, Marc Glisse wrote:
> 
> Anything based on llvm is broken, but we have code in CGAL to work around 
> that (at least in recent versions).
> 
>> to test the basic rounding mode stuff you could try the attached test case.
> 
> That doesn't account for our workarounds, you really want to test 
> Interval_nt. And there is already a run-time test checking that rounding 
> works, which gets executed by every CGAL program.
> 

OK, makes sense. That run-time check appears to pass on all platforms, as 
most CGAL programs I've run actually runs without problem. My issues are (for 
me) limited to Nef_polyhedron_3.

Do you think this is just a bug in CGAL which happen to trigger on some llvm 
compilers?
Is there any way I can report this properly, or provide enough data so that 
the right people can pinpoint and take a look at this?

Cheers,

 -Marius