Hey I've found a problem which you can see in the tests for joints such as Test6DofJointMotor and TestHingeJointMotor. If you get them to move at the start then everything is fine. If you wait about 10 seconds then the joints seem to lock up and do not respond . It seems to be to do with the joints as the updates are still running and the keys know they are being pressed. Any ideas?
I think the physics start to sleep in oder to conserver cpu,
have you treid setting the nondeactivation flag on the physicnodes? (Sorry not sure how it is named in jmejbullet
There is a method called setSleepingThresholds which if I give 0 or negative numbers I still get the same result. This does seem to be the problem though because if I call the activate method on every physicsNode I don't get the same problem. Am I doing something wrong with the setSleepingThresholds method or is there a way I can set a nondeactivation flag as Empire Phoenix suggested?
Yes, its the deactivation. Just use physicsNode.setActive(true) to re-enable the nodes. I was thinking about putting in an activate() call for both physicsnodes when the force of a joint is changed, dont quite understand why thats not built into bullet…
Cheers,
Normen
to allow sleeping on a hinge for example, cause the force nearly never gets mathematically zero, and always(nearly always) apply a small force, so it would need also a activation Thresholds then
might be an idea to improve your binding, just let us set activation minimum forces.