Strange Run Time Crash

[Wavesonics]

I am applying a force to a rigid body and am getting the following run time crash some time later:

Assertion (m_Adiag > btScalar(0.0)) failed in "btJacobianEntry.h", function "",
line 53

Any idea why this would be happening?

[Dirk Gregorius]

Usually this happens when two linked bodies (either contact or joint) have infinite mass.

[Wavesonics]

Oh strange. Well I do have 1 body which has a mass of 0 and it is connected to a dynamic body with a mass of 10, is that ok?

[John McCutchan]

Bodies should either have an infinite mass or a mass > 0.0

Thanks,
John

(edit)

NOTE: In Bullet to specify that a body should have an infinite mass you set it's mass to 0.0

[sparkprime]

I thought static objects should have a mass of 0?

Even internally, an "inverse mass" of 0 means a static object.

I've not seen infinite mass used / mentioned anywhere.

[Wavesonics]

Ya I thought that same thing, I'm using a mass of 0 for all my non-dynamic objects.

How do I set it to be infinite?

[John McCutchan]

Sorry for the confusion. I wasn't specifically referring to the Bullet API. It is valid to set the mass of a rigid body in Bullet to 0.0. This will cause Bullet to treat the rigid body as static.

Thanks,
John

[Wavesonics]

Oh ok, got ya.

But do you have any idea what could be causing this problem?

I do have a dynamic object attached to a static (mass = 0) object, but that should be ok shouldn't it?