Hi Robert Many thanks - your fix to ArrayDispatcher works for me - both with my test data and the original more complex scenegraph that the edge3.osg test case was extracted from.
best Mike On 21 October 2011 17:42, Robert Osfield <robert.osfi...@gmail.com> wrote: > Hi Mike, > > Your comments about ArrayDispatcher and slow paths was most > illuminating. I had a look at the ArrayDispatchers.cpp and how it > integrates with the GLBeginEndAdapter that is used as a fallback when > slow paths have to be used and found that it didn't contain the > handling of Vec3d's so I've added this and checked it into svn/trunk. > With this change your edges3.osg model works with the vertex data > passed in as a Vec3dArray. > > Could you try out svn/trunk and let me know how you get on. > > Cheers, > Robert. > > On Fri, Oct 21, 2011 at 1:26 PM, Mike Connell <michael.conn...@gmail.com> > wrote: > > Hi Filip, > > > > Thanks for your comments. It appears the warnings are indeed from the > > TriangleFunctor which is used by the DriveManipulator & > TerrainManipulator. > > As we don't use those pieces of code in this application I don't think > they > > are the problem here. > > > > I've dug a little deeper in the rendering code and found a solution. The > > geometry object which isn't working has PER_PRIMITIVE and OVERALL > bindings > > and so was shunted off the fast path. I now suspect that there is a bug > > within the ArrayDispatcher code where the Vec3d vertex data is discarded. > > However, if I update the geometry so that fast paths are used, everything > > works as expected. > > > > I am going to take the easy road and update our code to stay on the fast > > path route, but in case anyone else has problems with double precision > > geometry, the non fast path rendering code may be the culprit. > > > > best wishes > > > > Mike > > > > beleive that the problem lies within the ArrayDispatcher code, but > > > > On 18 October 2011 17:11, Filip Arlet <fili...@seznam.cz> wrote: > >> > >> Hi, > >> > >> PolytopeIntersector uses osg::TriangleFunctor< T > and as you can see in > >> code (TriangleFunctor header). > >> Constructor called with Vec*d does nothing (prints warning). > >> In member function TriangleFunctor::drawArrays OSG is accesing Array, > that > >> doesnt exists - and that's where the problem is I think. It looks like > it's > >> unfinished. For my app using double I created TriangleFunctorDouble < T > > by > >> replacing every float type variable to double type, but it's more like a > >> workaround than a sollution. > >> > >> Cheers, > >> Filip > >> > >> ------------------ > >> Read this topic online here: > >> http://forum.openscenegraph.org/viewtopic.php?p=43449#43449 > >> > >> > >> > >> > >> > >> _______________________________________________ > >> osg-users mailing list > >> osg-users@lists.openscenegraph.org > >> > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > > > > > _______________________________________________ > > osg-users mailing list > > osg-users@lists.openscenegraph.org > > > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org > > > > > _______________________________________________ > osg-users mailing list > osg-users@lists.openscenegraph.org > http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org >
_______________________________________________ osg-users mailing list osg-users@lists.openscenegraph.org http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
