Robert, are you doing to do any more with this? I don't mind pushing it forward (we need it, because the problem is breaking some tests for us), but I don't think I've grasped all your ideas.
thanks andy -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Andy Skinner Sent: Tuesday, February 19, 2008 1:17 PM To: OpenSceneGraph Submissions Subject: Re: [osg-submissions] try doubles formathinLineSegment::intersect() methods OK, I've looked at this some more. Yeah, we definitely need to switch the types of the local variables to use the correct value type. I think that, plus what you've already done, would answer my changes. I switched a bit around in the last stuff to "optimize" a couple of things that were done twice (a-b and b-a), but I don't know that it was worth doing, compared to the readability. I still have to look to see what I think about having different methods for different types. Specifically, would we get the new one, or would we have to write a new visitor that calls the double version? I want to make sure we benefit from the double precision. andy -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Robert Osfield Sent: Monday, February 18, 2008 9:07 AM To: OpenSceneGraph Submissions Subject: Re: [osg-submissions] try doubles for mathinLineSegment::intersect() methods The first step towards Vec3d support in LineSegment I have introduced a typedef into LineSegment: typedef Vec3d vec_type; typedef vec_type::value_type value_type; I have also changed various methods and intenals of LineSegment to use these typedefs rather than just Vec3. Its a first step as I haven't changed all the internals to use floats - they should now use the value_type method. We should also have several intersect functions, ones with float/Vec3f and ones with double/Vec3d to ensure that calling code can remain using the types that suit the calling code. I have checked these changes as they stand to get the ball rolling. Andy's changes need to be reviewed against mine, and a new set of changes made to complete the work above. Robert. _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegr aph.org _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegr aph.org _______________________________________________ osg-submissions mailing list [email protected] http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org
