[fluxus] ground-plane / passive-box collision confusion

[dave]

On Fri, 2006-07-28 at 13:05 +0100, Claude Heiland-Allen wrote:
> dave wrote:
> > On Thu, 2006-07-20 at 02:35 +0100, Claude Heiland-Allen wrote:
> >> Hi all,
> >>
> >> I think I found a physics bug - if an active-box collides with a 
> >> ground-plane, then (has-collided mypassive-box) returns true, even if 
> >> the passive-box is nowhere near the ground-plane or active-box.  See 
> >> attached script, the cube on the right shouldn't change colour, but it does.
> > 
> > the problem here is that has-collided shouldn't work on passive objects
> > - they are just definitions of bounding areas in ode, and don't know
> > when they are colliding (afaict).
> 
> Aarrrgh!  That's annoying, will have to rethink how to implement this
> game thingy that I'm making then.
> 
> Is there a way to tell which objects have collided?  Would be cool to be
> able to get a collision list: a list of pairs of object ids, each pair
> being a collision between two objects.  This would enable e.g.
> bullet/asteroid collisions to do something but asteroid/asteroid or
> bullet/bullet collisions to do nothing.

you can still detect collisions for active objects - and if you need
active objects to behave like static ones, you can pin them with a fixed
joint.

but yep, something would have to be implemented to allow you to pair up
collisions.... 

fwiw - I did the asteroids.scm example without any physics at all - it
makes some things a lot easier that way.

you might be able to write a collision system on top of ode - ie if an
object has collided, get the nearest object to it and assume it's that
one that's generated the collision...

cheers,

dave

> 
> 
> > there is a bug in the reporting code
> > in fluxus which that script shows up though.
> > 
> > cheers,
> > 
> > dave
> > 
> 
> Thanks,
> 
> 
> Claude