Hi Daniel,
On Tue, 2004-01-13 at 03:04, Daniel Grest wrote:
> Hello,
>
> i just found a bug in the Volume intersection methods
> (OSGVolumeFunctions.cpp).
>
> For the intersection with the CylinderVolumes/Sphere and Cyl/Box there
> is missing this check:
>
> s1 = (apos - sphere.getCenter()).length();
> s2 = (apos + adir - sphere.getCenter()).length();
>
> // to be added:
> if ((s1<=DBL_EPSILON) || (s2<=DBL_EPSILON)) return true;
>
> otherwise u might divide by zero.
For Sphere/Cyl I can see that, but for the box the only division I see
is by the cylinder axis length, which is not influenced by the above
calculation. Am I missing something. Anyway, thanks for the fix.
> Also i couldn't find the function which is called for
> CylinderVolume.intersect(SphereVolume).
> In the OSGVolumeFunctions.cpp there is a intersect(CylinderVol, Volume),
> but i couldn't find a intersect(CylinderVol, SphereVol) ?
> And i dont see why the intersect(SphereVol, CylinderVol) should be
> called. As far as i saw the intersect(CylinderVol, Volume) should call
> itself for a Sphere, which is not happening... ???
Yes, that's a little tricky. We have some inlined parameter turnaround
functions in OSGVolumeFunctions.inl, like the following
inline
bool intersect(const CylinderVolume &cylinder, const SphereVolume
&sphere)
{
return intersect(sphere, cylinder);
}
These are just to remove the symmetries and reduce the number of actual
intersect functions we need.
Thanks
Dirk
-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
_______________________________________________
Opensg-users mailing list
Opensg-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensg-users
