What is the prefered way to use a Newton scene collision?

[godlike]

In some of the Newton demos I see the following scenario:
1. Create a newton scene collision
2. Append a number of sub-collisions to the scene collision
3. Create a body with the scene collision

What I've tried was to add sub-collisions after creating the body (do 2 after 3). It didn't work as I imagined. Adding sub-collisions after the body creation doesn't seem to do anything.

Is this expected?

[Julio Jerez]

yes that should work.

but you should get the collision form the body, CreateBody make a copy of the collision that you pass in, so adding to the shape ate to create the body, add the sub shape to the other scene.

your step should be

1. Create a newton scene collision
2. Destroy collision you passed to create body and Get Collision form the created body
3. Append a number of sub-collisions to the scene collision
4. Create a body with the scene collision

[godlike]

Thanks for the response,

I am not sure I understand the lifetime of the bodies.

1. create initial scene collision "scene0"
2. add 10 sub-collisions to "scene0"
3. create a body "body0" with the "scene0"
4. destroy "scene0"
5. get the collision "scene1" from the "body0"
6. add 5 more sub-collisions to "scene1"
7. ?????

What should I do in step 7? Create a new body with "scene1" and destroy "body0"? Or something else?

[Julio Jerez]

that should work, you can add, remove or move sub shapes in step 7 without destroying the body.
has you implemented debug display

[godlike]

Thanks Julio! That worked fine.

[Julio Jerez]

remember when you add or remove shape to a scene collision or a compound collision you have to do it from a BeginAddRemove / EndAddRemove block

if you do not it, it will added or remove the sub shape, but the broad phase will not be updated.
this is allow because some functionality like vehicle use a compound collision add move the sub shapes, but does no nee to update the broadphase because the body is dynamics.