Am Mittwoch, 12. November 2008 00:18:54 schrieb Carsten Neumann: > Hello Garry, > > Garry Hoberg wrote: > > I have the same problem as Telmo. The WalkNavigator "walks" me into a > > wall and then gets stuck (it can still turn but not "walk" any more; I > > only noticed actually being inside the wall after setting the near > > clipping plane to 0.0001 and turning 90 degrees). In my case the > > collision model and the drawn one are the same nodes. > > > > Large values didn't change anything. To me it seems like either the > > _wallDistance, _fatness and _width properites are collectively ignored or > > the IntersectAction's _hitT property value is set or returned falsely. > > well, not quite, but the ray shot into the scene was aimed in the wrong > direction and as it turns out from the old position not the one the user > tries to move to (which made it impossible to just go backwards to get > away from a wall) -- why did I never see this when looking at the code > earlier, arghh > I'm confused. mv is still calculated the same, isn't it?. Before, the ray was shot from the old position towards the new one in mv direction. Now the ray is shot from the new position towards the old one in -mv direction. Except for a tiny speed improvement per collision test [(rFrom-_rFrom).length() isn't calculated and added to _fatness+_wallDistance any more] I can't see the difference. Please explain it to me.
> > I managed to work around this but I'd like to be able to fully rely on > > the collision detection. > > hm, the collision detection code is rather simple in what it does, i.e. > it only shoots a few rays into the scene, it does not consider real > volumes, so there are limits to what it can do. For a walkthrough with a > not too demanding model it should do fine though. > > > Since I'm a little behind the schedule for my diploma > > thesis, I won't have the time to rewrite the library to find the bug > > myself. > > Anyways, the changes are in the cvs now (and analogous changes to > OSGWalkEngine.cpp for 2.0 in svn), the patch is attached [1]. Hope it > works better now, good luck with your thesis, > Thank you for both the good wishes and your work on OpenSG. Regards, Garry > regards, > Carsten > > [1] It also cleans the formatting, which makes it hard to see what has > actually changed - sorry about that. ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Opensg-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensg-users
