[ODE] Collision response for symetrical tri-meshes

[Flavien Brebion]

Jon Watte (ODE) wrote:

>I would assume, at a minimum, that you need to reverse the winding of 
>the triangles in a mirrored trimesh.
>
Makes sense, except.. are you suggesting that i reverse the order of the 
tri-mesh data at creation time ? That wouldn't work, since an instance 
of that tri-mesh can have a "normal" matrix, and another instance a 
symetrical matrix.

The correct way to do it would probably be to add a flag for each 
tri-mesh instance, and to check that flag when the triangle intersection 
function is called. Remains the question of detecting when that should 
be done, simply looking at the rotation matrix, and extracting the scale..

F. Brebion