Help after upgrading to Newton 2

[Skarik]

Just upgraded project from Newton 1.53 to Newton 2.10.

It's, um, working awesome except for one thing.

When stuff hits the ground, the * hits the fan. Litterally.

I got a box, and it's bouncing on Steroids. What's up, what's possible causes?

[Julio Jerez]

that kind of vage it need more detail.
what are you doing?

[Skarik]

I have no idea.

Code: Select all

// This is the function you change to change gravity
// If you want zero gravity volumes, you'll have to
// ask me to do it for ya. It takes a bit of list shifting
// to get to the volume.
// BTW, this function was copied straight from Irrlicht.
// All I did was make it use U3D's coordinate system.
void PhysUtilApplyGravity( const NewtonBody * pNewtonBody, dFloat timestep, int threadIndex )
{
   dFloat Ixx;
   dFloat Iyy;
   dFloat Izz;
   dFloat mass;

   // Set gravity force
   NewtonBodyGetMassMatrix( pNewtonBody, &mass, &Ixx, &Iyy, &Izz );
   CVect vectForce (0.0f, 0.0f, mass * -98.1f );

   // Get user defined forces ( Not from irrlicht )
   if ( NewtonBodyGetUserData( pNewtonBody ) != NULL )
   {
      CBaseObject * pTargetObject;
      pTargetObject = (CBaseObject * ) NewtonBodyGetUserData( pNewtonBody );
      vectForce += *(pTargetObject->GetSuperControlEverythingForce());
   }

   // Add the force
   NewtonBodyAddForce ( pNewtonBody, vectForce.GetVectorPtr() );
}

That's what I look at.

Now, I'm thinking it's the solver in 2.x, since this is the same * I used to get with Newton. The weird part is, though, that the player doesn't do that, possible as a result of the pin vector constraint, which rocks by the way.

Hey, I'll do some more hacking around with the solver types and see if anything works.

I'm also wondering about thread index. What is that? Also, there's a new argument with shapes, what is that?

[Skarik]

I was right.

I changed the solver model, and it works perfectly now. I'd like to know, though, what has changed in that function since I can't seem to find a good help file for Newton 2.x. Also, my questions from above still persist.

[Skarik]

I take it back, the object still goes crazy when another object - the player - crashes into it.

Edit:

I don't think it's possible for a box to jump by itself that many times of its own accord.

[JernejL]

I don't think it's possible for a box to jump by itself that many times of its own accord.

It isn't, you definetly apply some force to cause that, it may be some broken upgraded forcetorque or joint code.

[Skarik]

Right, I'll tell the guys working on the physics to check the force applying functions.

[Skarik]

Right, it's still doing the same. Gravity works fine, it's only when colliding with a static object, it goes effing nuts.

[JernejL]

Provide a run-able demo or something, we can't solve a problem from just words "it doesn't work", especially when it's very likely that this is triggered by a bug in your app code.

[Skarik]

Ya I know, but I gotta put it together, and I'm just oh so lazy.

Blast, I'll put a demo or something at least before the week is over.