[ODE] Re: Box v. Trimesh, Fishy or Insane

[Panayoti]

In regards to the Box v. Trimesh questions, as I stated in my earlier
post, I've noticed the problem with the ODE tests, specifically
moving_trimesh (although the standard test doesn't use StepFast, but it
strikes me as a collision problem, not a collision resolution problem).

Bust out the moving trimesh problem and throw a couple spheres in there,
not one of them will misbehave. Do the same with boxes and you'll notice
they jut out unrealistically (usually because a corner has penetrated
the trimesh and doesn't know how to fix itself, I imagine).

This behavior is specifically what I am talking about, and what I was
wondering about (whether the behavior was considered "normal" or if
there is a bug in the box v. trimesh collision code that someone might
be working on).

Thanks!

Panayoti