[ODE] Convex/TriMesh collisions?

[Rodrigo Hernandez]

Native Convex/Convex is not functional, I started working on it, but 
couldn't finish it, the code is there so if someone else wants to 
continue the work, he/she doesn't have to start from
scratch, or for me to continue working on it some time in the future, I 
doubt it will work anytime soon, sorry.

The idea behind Convex/Convex is to implement the separating axis 
theorem for convex objects descrived on Dave Eberly's Game Physics book, 
I believe I got the face-vertex and face-face
in there, but edge-edge is a special case that requires a special bsp 
structure.

Anyway, I believe I commented the file convex.cpp enought to give you an 
idea of whats going on in there, take a look at it.

Mark Williams wrote:
>   
>> I don't think this is well supported at the moment,
>> but one of the reasons of releasing ODE 0.8 is to prepare for addition of 
>> Bullet collision detection,
>> which supports convex versus trimesh.
>>
>> Erwin
>> http://bulletphysics.com
>>
>>     
>
> OK, thanks for the info!
>
> I've also noticed some problems with Convex/Convex collisions, but it's 
> trickly to tell whether or not I'm passing the correct data to 
> dCreateConvex as I can't find any documentation for it.
>
> I'm using qhull to generate my hulls, which produces plane 
> normals/offsets, points, and polygon lists, but I've no idea if the 
> vertex IDs for the polys need to be specified in any particular order. 
> I've tried the order that comes directly from qhull's FOREACHvertex_, 
> and the reverse order, but neither seem to give results which are 
> anything close to being stable.
>
> Should I expect convex/convex collisions to work in 0.8?
>
> Thanks,
>
> Mark
>
> _______________________________________________
> ODE mailing list
> ODE at ode.org
> http://mooshika.org/mailman/listinfo/ode
>
>