CGAL users discussion list

[Michael Hemmer <>]

I traced it back, there are several statements that print a normalized plane.
But only in debug mode. That is, planes are never normalized in release mode.
At least not with the code that caused the problem.

Michael

On Mon, Aug 1, 2011 at 10:15 PM, Cody Rose 
<>
 wrote:
> Thank you. Why does this compile in Release mode though?
>
> Cody
>
> On 8/1/2011 6:05 AM, Michael Hemmer wrote:
>>
>> The problem is that the function
>>
>>   template<typename R>  static
>>   CGAL::Plane_3<R>  normalized(const CGAL::Plane_3<R>&  h) {
>>  in file Nef_S2/Normalizing.h
>>
>> assumes that FT is decomposable into a numerator and a denominator.
>> This is not the case if FT is CORE::Expr.
>>
>> The function should first check whether it can decompose FT using the
>> tag that is provided by the Fraction_traits.
>>
>> If FT is not decomposable it is probably best to divide all
>> coefficients by the first coefficient that is not zero.
>> This way one also achieves a normalized, i.e. unique representation,
>> of the plane.
>>
>> Michael Hemmer
>>
>>
>> On Mon, Aug 1, 2011 at 1:20 AM, Cody 
>> Rose<>
>>   wrote:
>>>
>>> In Visual Studio 2010, the code:
>>>
>>> #include "CGAL/CORE_Expr.h"
>>> #include "CGAL/Cartesian.h"
>>> #include "CGAL/Nef_polyhedron_3.h"
>>>
>>> typedef CGAL::Cartesian<CORE::Expr>      K;
>>> typedef CGAL::Nef_polyhedron_3<K>      nef;
>>>
>>> int main() {
>>>    nef foo;
>>> }
>>>
>>> compiles in Release mode but not Debug mode. The error is
>>>
>>> cgal\nef_s2\normalizing.h(307): error C2676: binary '*=' :
>>> 'CGAL::Null_tag'
>>> does not define this operator or a conversion to a type acceptable to the
>>> predefined operator
>>>
>>> many times. Someone else on this mailing list had this problem back in
>>> 2008,
>>> but that was resolved by switching kernels. However, I need square roots,
>>> and therefore (as I understand it) the only number types I can use are
>>> CORE::Expr and leda_real. I'm getting vastly better performance with
>>> CORE::Expr so I would really like to avoid leda_real. The fact that this
>>> works fine in Release mode indicates to me that there must be some way to
>>> get it to work in Debug mode, but the earlier thread didn't discuss why
>>> that
>>> might be.
>>>
>>> Thanks,
>>> Cody
>>>
>>> --
>>> You are currently subscribed to cgal-discuss.
>>> To unsubscribe or access the archives, go to
>>> https://lists-sop.inria.fr/wws/info/cgal-discuss
>>>
>>>
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss
>
>