[ODE] Re: dSpaceCollide2 - quadspace vs hashspace bug

[Jon Watte]

Does the ODE abstraction really allow you to test for 
specific implementations inside dSpaceCollide2()? At first 
glance, this seems like it would break abstraction. I think 
the "real" solution is making sure that the QuadTree does 
the right thing.

Cheers,

			/ h+


-----Original Message-----
From: ode-bounces at q12.org [mailto:ode-bounces at q12.org]On Behalf Of
Hannes Norda
Sent: Monday, December 13, 2004 1:17 AM
To: ode
Subject: [ODE] Re: dSpaceCollide2 - quadspace vs hashspace bug


Due to the lack of interest in this topic I feel obliged to reply to myself. :)

I just patched in the not-so-beautiful solution here..

In dSpaceCollide2, collision_space.cpp, add these lines:
    if(s1->type == dQuadTreeSpaceClass)
    {
      for (dxGeom *g = s2->first; g; g=g->next) {
        s1->collide2 (data,g,callback);
      }
    }
    else if(s2->type == dQuadTreeSpaceClass)
    {
      for (dxGeom *g = s1->first; g; g=g->next) {
        s2->collide2 (data,g,callback);
      }
    }
    else

... just before these lines:
 // iterate through the space that has the fewest geoms, calling
        // collide2 in the other space for each one.
        if (s1->count < s2->count) {

/Hannes

On Fri, 10 Dec 2004 15:43:01 +0100, Hannes Norda <hannes.norda at gmail.com> wrote:
> Hi!
> 
> I noticed that when using dSpaceCollide2 to collide a hashspace with a
> quadspace, and the hashspace contains more objects than the quadspace,
> no collision occurs.
> 
> The root of the problem seems to be that the quadspace doesn't use the
> variable dxGeom *first; in dxSpace. It is always NULL.
> 
> Look at this piece of code (in dSpaceCollide2, collision_space.cpp):
> // iterate through the space that has the fewest geoms, calling
>        // collide2 in the other space for each one.
>        if (s1->count < s2->count) {
>          for (dxGeom *g = s1->first; g; g=g->next) {
>            s2->collide2 (data,g,callback);
>          }
>        }
>        else {
>          for (dxGeom *g = s2->first; g; g=g->next) {
>            s1->collide2 (data,g,callback);
>          }
>        }
> 
> When (hashspace->count > quadspace->count) there will be no
> collisions, since the for-loop is immediately ended.
> 
> Anyone else stumbled across this and solved it nicely?
> Anyhow, I guess I will do some research and see if I can perhaps
> modify void dxQuadTreeSpace::add(dxGeom* g)  and make it work. Not
> very familiar with quadspace internal structure, so we'll see what
> happens.... :)
> 
> Normally this isn't an issue for me, since I store my static world in
> the quadspace and my dynamic objects in the hashspace, and I've always
> had more static objects up until I made a small test application
> yesterday. Was mildly confused when all of a sudden no collisions
> occured..
> 
> /Hannes
>
_______________________________________________
ODE mailing list
ODE at q12.org
http://q12.org/mailman/listinfo/ode