Compound to Empty Mesh Demo works differently for Cannon & Omio vs Ammo

Pryme8 · April 4, 2022, 3:37pm

Ammo seems to be the only one that works correctly given this setup the others treat the position of the parent mesh as an offset?

This is the demo scene from the documents “Compound to Empty Mesh”

Deltakosh · April 4, 2022, 3:54pm

cc @Cedric

Cedric · April 5, 2022, 10:27am

PR

github.com/BabylonJS/Babylon.js

fix physics regression

BabylonJS:master ← CedricGuillemet:compoundOimoCannon

opened 10:26AM - 05 Apr 22 UTC

CedricGuillemet

+1 -1

followup https://forum.babylonjs.com/t/compound-to-empty-mesh-demo-works-differe…ntly-for-cannon-omio-vs-ammo/29062 When computing bounding info from scratch, before this change, the min and max position was `mesh.position`. So, in this case: ``` var compoundBody = new BABYLON.Mesh("compoundBody", scene); compoundBody.position.y=3; compoundBody.computeWorldMatrix(true); console.log(compoundBody.getBoundingInfo().boundingSphere.centerWorld.y); console.log(compoundBody.position.y); ``` min and max are (0,3,0), bouding sphere local center is (0,3,0). world bounding sphere center is local center transformed by world matrix and its value is (0,6, 0) which is not correct. It breaks the physics for oimo and cannon. Computation is different with ammo and was not touched by this issue. @RaananW Git blame indicates you changed it last. Do you remember why/what it fixed?

Topic		Replies	Views
Problem with Compound Bodies Questions	1	288	January 12, 2022
Child physics imposter dynamic positioning Questions physics	3	633	February 28, 2022
Physics engine and limitations Questions	2	594	September 7, 2020
Joint object point [edit gravity custom model] Questions	8	1190	March 28, 2019
Ammo joint connected axis change Questions	10	288	March 23, 2022

Compound to Empty Mesh Demo works differently for Cannon & Omio vs Ammo

Related topics