Bad calculation of barycenter in CompoundCollisionShape

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Add a node from a box shape
2. Add another node from a compoundcollisionshape where is only the same box.
3. Compare results

A sample is provided as attachment.

What is the expected output? What do you see instead?
I suppose that both simple box and box in CompoundCollisionShape have the same 
behavior.
but obviously, they have not.

What version of the product are you using? On what operating system?
I'm using jme3 in trunk (r7250) on Windows 7

Please provide any additional information below.
I also tried the CompoundCollisionShape with the box without moving it when 
adding.
It seems that calculateLocalInertia provides the same point with or without 
moving.

Why :
I want to provide robot part where origin collides with construction map.

Original issue reported on code.google.com by Isam...@gmail.com on 17 Apr 2011 at 8:53

Attachments:

• TestMassEffect.java

[GoogleCodeExporter]

You assign an offset location to the box when you add it to the compound 
collision shape in your code.. Please use the forum if you have no bugfix to 
commit.

Original comment by normen667 on 7 Jun 2011 at 11:39

• Added labels: ****
• Removed labels: ****

[GoogleCodeExporter]

Original comment by normen667 on 7 Jun 2011 at 11:39

• Changed state: Invalid
• Added labels: ****
• Removed labels: ****