BulletX Dynamics Bugs????

Topics: Developer Forum
Jan 17, 2010 at 4:28 PM

If I drop a small box (restitution=1.0) on a large static box (restitution=1.0) (ground) the box sticks to the ground and does not bounce.  (same happens for sphere).  I found where the restitution is forced to zero in SolveGroupCacheFriendly() when penetration velocity is less than the reflection (I assume this is for extra stability).  If I disable this the box bounces but it bounces higher than the original height (should be the same height). 

I can do the same experiment with the c++ version of Bullet and the experiment works perfect (even if I comment out the similar code that forces restitution to zero).

I notice the code for SequentialImpulseConstraintSolver in Bullet is very different than BulletX.  Is this because the way Bullet does it is too expensive in C#.

Do I need to do additional setup in BulletX or is this a bug?

Is Anybody looking into this already?  I will start to dig deeper into the cause for this.

I have also noticed strange behavior relating to friction due to rotation.

 

Jan 17, 2010 at 4:55 PM

I Notice a long pause at initial impact.  This may be related to the problem.