> 
> Unfortunately, this new cull behavior would brake my existing application 
> (heavily relaying on original behavior), too. Anyway, I don't see any harm 
> doing this kind of optimization on the existing intersectors, though. 
> Luckily, the osg can be extended in many ways, and with that being said, one 
> might be able to simply subclass original CullVisitor and other classes of 
> interest and do the special handling of the bounding volumes there. The 
> custom CullVisitor can be installed on init with:
> 
> osgUtil::CullVisitor::prototype() = new CullVisitorPrivate;
> 


Hi Robert! That's a nice idea, unfortunately the isCulled(const 
BoundingSphere&) is inlined (understandably) so I can't override it. I could 
override each accept(Node&) type method and do the check there but that results 
in a bit of code duplication. It's easier for me to patch osg locally with the 
two line change since I already do my own builds anyway to switch off double 
precision.

@The other Robert: Thanks for your time! I understand why the change is 
problematic. I will look into an independent fix for the NaN issue in the 
intersection visitor.

Cheers,
Jannik

------------------
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=65719#65719





_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to