Hmmm, yes, good thinking about start or end of ray being inside the object.
I wonder if the LineSegmentIntersector::clone() function needs some
adjustment because it is responsible for calculating new start and end
points for the line segment from transforms? Stack trace shows that function
being called inside IntersectionVisitor::apply(Camera&) as a result of
IntersectionVisitor::push_clone()
Chris.
----- Original Message -----
From: "Robert Osfield" <[EMAIL PROTECTED]>
To: "Chris Denham" <[EMAIL PROTECTED]>
Sent: Thursday, November 06, 2008 3:09 PM
Subject: Re: [osg-users] Problem with computeIntersections for nodes below
post render camera
On Thu, Nov 6, 2008 at 2:51 PM, Chris Denham <[EMAIL PROTECTED]> wrote:
Hi Robert,
Just tested at this end and can confirm the problem resulting from
setting
the viewMatrix seems to be fixed. Thankyou!
But I am still worried that something is still not quite right though,
because the object below the nested camera seems to quite often return a
single intersection. By my reckoning, a ray should always pass though
zero
or two polys of an enclosed convex shape. Indeed, it always seems to
return
two intersections per object when there is no nested camera is involved,
so
what's going on?
One intersection would suggest either the start or end point is within
the object. Playing around with osgpick --relative-camera-scene
illustrates the one pick problem, but some positions the cube does
pick twice.
I can't think of any explanation on this from the top of my head.
Robert.
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
