CGAL users discussion list

[yanyajie <>]

Dear S.

 

I fixed the problem by sustituting int with double in the algebraic_kernel, but intersection routine just doesn’t return objects whose type is what I expect. In the following code, the intersection definitely happens, and two objects are inserted in vector isecs. If nothing goes wrong, they should be two Point_3_SK instances, right? However, object_cast<Point_3_SK> only yields null pointer.

-------------------------------------------------------------------------------------

    typedef SphericalK::Line_3 Line_3_SK;

    typedef SphericalK::Sphere_3 Sphere_3_SK;

    Line_3_SK line (Point_3_SK(0, 0, 0), Point_3_SK(0, 0, 1));

    Sphere_3_SK sphere(Point_3_SK(0, 0, 0), 5);

    vector<CGAL::Object> isecs;

    CGAL::intersection(line, sphere, std::back_inserter(isecs));

    const Point_3_SK *isec_p = CGAL::object_cast<Point_3_SK>(&isecs[0]);

-------------------------------------------------------------------------------------

 

Where does the problem lie?

 

Thank you for your time!

 

 

Yan.

 

From:

Sent: Friday, May 13, 2011 10:32 AM

To:

Subject: Re: [cgal-discuss] precondition violation when intersecting a line with a sphere

 

Dear S.

 

O! I was so naïve, as I thought “int” should be the type used to instantiate Algebraic_kernel when I saw a RT in

CGAL::Algebraic_kernel_for_spheres_2_3<RT>

OK, thank you! I’ll see if this correction could solve the problem.

 

Yan.

 

-----原始邮件-----

From: Sebastien Loriot (GeometryFactory)

Sent: Friday, May 13, 2011 1:45 AM

To:

Subject: Re: [cgal-discuss] precondition violation when intersecting a line with a sphere

 

Algebraic_kernel_for_spheres_2_3<int>

should be Algebraic_kernel_for_spheres_2_3<double>

 

You also should consider using an exact type.

Please read:

http://www.cgal.org/FAQ.html#inexact_NT

 

 

yanyajie wrote:

> Hi,

> 

> I encounterd a precondition violation error when I was trying to find

> the intersection between a line and a sphere.

> -----------------------------------------------------------------------------------

> CGAL error: precondition violation!

> _expression_ : !p.degenerated()

> File       :

> d:\codes\lib\cgal-3.7\include\cgal\algebraic_kernel_for_spheres\int

> ernal_functions_on_roots_and_polynomial_1_3_and_2_3.h

> Line       : 212

> Explanation:

> Refer to the bug-reporting instructions at

> http://www.cgal.org/bug_report.html

> -------------------------------------------------------------------------------------

> 

> I provide necessary info in the following to help reproduce the error:

> 

> ------------------------------------------------------------------------------------------------------------------------------------------------------

> typedef CGAL::Cartesian<double> K;

> typedef CGAL::Spherical_kernel_3<K,

> CGAL::Algebraic_kernel_for_spheres_2_3<int> > SphericalK;

> typedef SphericalK::Point_3 Point_3_SK;

> 

> // to compute the intersection, I construct a line and a sphere using

> the spherical kernel as follows:

> CGAL::Line_3<SphericalK> line(Point_3_SK(-0.1,-0.1,-0.1),

> Point_3_SK(-0.1,-0.1,0.0));

> CGAL::Sphere_3<SphericalK> s(Point_3_SK(0, 0, 0), 0.1); // c is a Point_3

> vector<CGAL::Object> isecs;

> CGAL::intersection(line, s, std::back_inserter(isecs));

> ------------------------------------------------------------------------------------------------------------------------------------------------------

> 

> The assertion error is issued from within CGAL::intersection, when it

> tries to do the confirmation,

> actually, from this function:

> 

> template < class AK, class OutputIterator >

> inline

> OutputIterator

>     solve(const typename AK::Polynomials_for_line_3 &p,

>           const typename AK::Polynomial_for_spheres_2_3& s,

>      OutputIterator res )

> {

> ...

> // we must have a line

> CGAL_kernel_precondition(!p.degenerated());

> ...

> }

> 

> 

> Does this failure of assertion indicates I constructed a degenerate

> line? But from the code you can see the line should be non-degenerate.

> 

> Thanks if anyone can point out the problem!

> 

> 

> Yan.

 

 

--

You are currently subscribed to cgal-discuss.

To unsubscribe or access the archives, go to

https://lists-sop.inria.fr/wws/info/cgal-discuss