Undesired movement when adding a joint

[klem]

Hi,

I'm very new to BabylonJS, but I need to admit it's a great framework, very polished. And the playground is a delicious toy

I'm currently playing with the physics plugin (cannon) and I'm trying to understand joints. In fact I've got an issue when I add a joint between 2 meshes : there is a force applied directly on those meshes. The result is an undesired pseudo-random movement at startup. I tried things with setLimit or setMotor on joint, but there is always a kind of big impulse at launch.

It looks like a bug, but I'm not really sure Here's the playground : http://www.babylonjs-playground.com/#24JRLD#1

In this playground, 2 cubes are "hinge-joined" but the initial movement is undesired. If you comment the addJoint call (line 112), cubes are placed correctly.

Is there a way to avoid this undesired movement? Other point, when you remove the joint, the cube (top) never stop moving ! is it a special property to avoid that ? I missed something?

Also I think I found a bug (another?) with the setMass function : joints are removed after setting a new mass on mesh. I prepared a demo with the bug here : http://www.babylonjs-playground.com/#24JRLD (same as before, but when user clicks, a setMass function call breaks the joint).

Any help would be greatly appreciated!

klem

[adam]

http://www.babylonjs-playground.com/#24JRLD#2

http://www.babylonjs-playground.com/#24JRLD#6

1 hour ago, klem said:

Other point, when you remove the joint, the cube (top) never stop moving ! is it a special property to avoid that ? I missed something?

I'm not sure what's going on there. I tried changing the friction and the sleep properties without success.

1 hour ago, klem said:

Also I think I found a bug (another?) with the setMass function : joints are removed after setting a new mass on mesh.

If you set the mass directly to the body, it doesn't break:

http://www.babylonjs-playground.com/#24JRLD#7

[klem]

Hey Adam,

nice shot (https://www.diffchecker.com/rtqR93yv), thanks!

And the workaround for the setMass is good to know