[ODE] Trimesh2

[LR]

Jon:
Thank you for the response,
I  changed the "maximum correcting velocity" and it's better but not reflect
reality as well as box/trimesh for exemple ...
I will try the convext colision system.


Jason:
I'm using the trunk branch. I can't send you an application because so eavy
...
But, i saw something interesting (perharps it's normal ?) 
   
  /\
 /  \
/    \
------

______
|    |
------

Test1: 
A trimesh cone is falling on a trimesh box: (box with a short rotation to
avoid planar coliding)
Note: All cone triangles have 2 base points and the same top point.

The tutching cone triangle generate 3 contacts points !! (even on the cone
top point ??) I don't understand why ... 

Test 2:
I replace the trimesh box for a dGeomBox (with the same short rotation).
The tutching cone triangle generate the 2 needed contact points  (no
contacts at the cone top)


It's not easy to explain, i hope i give you efficient information:) if you
want i can sent you the JPG picture ...

Thank you 

Laurent.







<What version are you using? Trimesh collisions improved a great deal since
0.5. That said, they are still not perfect. If you can come up with a small
test application that shows an issue we <might be able to figure out what
the problem is.

Jason


-----Message d'origine-----
De : Jon Watte (ODE) [mailto:hplus-ode at mindcontrol.org] 
Envoyé : vendredi 8 septembre 2006 17:15
À : LR
Cc : ode at q12.org
Objet : Re: [ODE] Trimesh2



LR wrote:
> Does i miss something to have a good trimesh/trimesh coliding ?
>

Trimesh/trimehsh has always been quite sensitive.

Make sure that you update the "last position" of the trimeshes each frame,
to give the collider a chance of figuring out what the appropriate contacts
are.

Also, try setting the "maximum correcting velocity" parameter to avoid
really large forces.

Cheers,

          / h+