[ODE] contact joints / attaching them

[Flavien Brebion]

I don't see why. The collision detection functions are completely
independant of the timestep. Notice that the time step parameter
isn't even involved in collisions.

I need this to fix the "tunneling" problem. Say that my "normal"
timestep is 0.1; say i have two objects, X and Y. X doesn't collide
with anything, but i detect a collision at time 0.04 on Y; so i need
to simulate X for 0.1 seconds and Y for 0.04 seconds, then the
remaining 0.06 seconds to synchrozine it with X.

F. Brebion

  -----Original Message-----
  From: Dino Patti [mailto:dino at nordmark.dk]
  Sent: Friday, October 10, 2003 4:24 PM
  To: Flavien Brebion
  Subject: Re: [ODE] contact joints / attaching them


  Well i think its impossible...

  2 objects with diffrent timesteps can't correctly collide...

  But why is it necessary to have one world per object?

  /*Dino*/

    ----- Original Message -----
    From: Flavien Brebion
    To: ode
    Sent: Friday, October 10, 2003 2:42 PM
    Subject: [ODE] contact joints / attaching them


    I'm trying to use one world per dynamic object ( in order to have
different timesteps
    per object ); but ODE doesn't allow collisions between multiple worlds.
Why is it
    so ?

    Specifically, the correct way to use contact joints is, once a contact
has been
    determined, to do something like this:

        dJointID c = dJointCreateContact(m_odeWorld, m_odeContacts, contact
+ i);
        dJointAttach(c, b1, b2);

    Where b1 and b2 are the two bodies.

    If i try to do that, i get an assertion, "b1 and b2 are in different
world".

    So, i wonder if it would be equivalent to do this:

        dJointID c1 = dJointCreateContact(m_odeWorld1, m_odeContacts1,
contact + i);
        dJointAttach(c1, b1, 0);

        dJointID c2 = dJointCreateContact(m_odeWorld2, m_odeContacts2,
contact + i);
        dJointAttach(c2, 0, b2);

    But i do not seem to get an equivalent result. This is driving me mad,
please help :)
    Thanks,

    F. Brebion



----------------------------------------------------------------------------


    _______________________________________________
    ODE mailing list
    ODE at q12.org
    http://q12.org/mailman/listinfo/ode
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://webserver.computershop.calgary.ab.ca/pipermail/ode/attachments/20031010/31939b5a/attachment.htm