AttachColliderMesh lost in the space !

[danledan]

Hello,

I try some test on the mechtest app., and I have trouble with the AttachColliderMesh function.

For exemple: I just want to change the actor sphere with a custom object ( like a flattened sphere) and I add a MeshCollider insteat of a standard sphere collider..

in mechtest.cpp I changed the object

Code:

// pcmesh->SetMesh ("Sphere", "sphere");
 pcmesh->SetMesh ("platsphere", "platsphere");

with this, I can see the new mesh object but I want to have the collision based on the mesh

I change in CreateDynActor()

Code:

//  pcmechobject->AttachColliderSphere (.42f, csVector3 (0, 0, 0));
  pcmechobject->AttachColliderMesh() ;

 
but my "platsphere" fall down.. (like if they don't have collision with the ground)
After that, I try without comment pcmechobject->AttachColliderSphere..  and now the ground interact with the sphere collider..  but other objects (like walls and balls) interact with the "platsphere" Mesh..

Any idea ?!
Thanks

[jorrit]

Hello,

I try some test on the mechtest app., and I have trouble with the AttachColliderMesh function.

For exemple: I just want to change the actor sphere with a custom object ( like a flattened sphere) and I add a MeshCollider insteat of a standard sphere collider..

in mechtest.cpp I changed the object

Code:

// pcmesh->SetMesh ("Sphere", "sphere");
 pcmesh->SetMesh ("platsphere", "platsphere");

with this, I can see the new mesh object but I want to have the collision based on the mesh

I change in CreateDynActor()

Code:

//  pcmechobject->AttachColliderSphere (.42f, csVector3 (0, 0, 0));
  pcmechobject->AttachColliderMesh() ;

 
but my "platsphere" fall down.. (like if they don't have collision with the ground)
After that, I try without comment pcmechobject->AttachColliderSphere..  and now the ground interact with the sphere collider..  but other objects (like walls and balls) interact with the "platsphere" Mesh..

Any idea ?!
Thanks

There is currently a bug with plane/mesh colliding. I assume your ground is a plane collider? In that case that is the reason. You might have better luck with a box collider on the ground but the best thing to do is to avoid mesh colliders at all and try to use primitives as much as possible.

Greetings,

[danledan]

Thanks for you short and quick reply..

but, I have an other "stange physical" problem, I change the code to have my flattedsphere with the meshcollider on a box build with

Code:

  csOrthoTransform t;
  t.SetOrigin(csVector3(0.0f,-1.5f,0.0f));
  pcmechobj->MakeStatic (true);
  pcmechobj->AttachColliderBox (csVector3 (200.0f, 1.0f, 200.0f), t);

 

At the begining all look perfect, collision look perfect, but after few test, the physical ground (the box) seem to go down. . if I set  the Mass of all the wallbox to 1000 for example ..   the ground fall down in few second .. . with 10 for Mass the ground fall down slowly (cool effect.. but I don't want this) . . I look into the code of the function MakeStatic and the doc of ODE but all seem to be OK. .. .  if I try it with the standard PlaneCollider for ground, this problem don't exist.

Any other idea ?
Thanks

[danledan]

Finaly I use this function to create a ground who interact with mesh, and this ground is completely locked on the world system, not like the last try.

Code:

void MechTest::CreateGround (csVector3 size, csOrthoTransform t)
{
  iCelEntity* entity_room (pl->FindEntity("room"));

  csRef<iPcMechanicsSystem> pcmechsys = CEL_QUERY_PROPCLASS_ENT (entity_room,
        iPcMechanicsSystem);

  iDynamicSystem* dynsys;
  dynsys=pcmechsys->GetDynamicSystem();
  dynsys->AttachColliderBox (size, t, 1, 1,1);
}

Maybe that will help someone !!