Physics behaviour - inworld prims vis-a-vis uploaded mesh

[anselm Hexicola]

(Very possibly this has been covered in previous threads, but I couldn't find it ...)

I uploaded what I thought was a perfect reproduction of the LL prim cube, ie. 0.5x0.5x0.5. I maxed out the upload in the Uploader window, giving all LOD's and the Physics shape the same dae. file.

My screenshot shows 2 sets of physics-set cubes - the plywood ones are the ones available from the inworld Build menu.

The upper one sits nicely flush on top of the lower one, even when nudged slightly sideways.

 

The white ones are the mesh uploads; the upper one "floats" up; that is to say, it will not rest flush on the lower one.

 

( I only did this exercise because I was I was being thwarted by similar behaviour in a set of objects I had uploaded.)

Any advice gratefully received ..

[arton Rotaru]

That's the legendary "Gap" which all prims used to have. It has been removed for a couple of prim types with the introduction of Mesh in 2011.

Unfortunately meshes will have the gap still.

[ChinRey]

I'm afraid if you used hull physics (that is if you clicked on the "Analyze" button), this is what you should expect. In this particular case using triangle (that is unanalyzed) physics should fix it.

[anselm Hexicola]

Well, as it happens I hadn't pressed the "Analyse" button; anyhow I re-uploaded paying attention to whether the Uploader gave info about triangles or hulls ( definitely the former).

Also, I played with the settings in the inworld Build menu, trying both Convex Hull and Prim in the Features tab. I still had

no luck :-

(The blue ones were my latest attempt.)

Could I be so bold as to ask you to confirm that you can accomplish the desired physics behaviour with a mesh upload?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[arton Rotaru]

It doesn't matter which buttons you click or not in the uploader. You won't get rid of the gap this way.

[ChinRey]

I've only noticed this with hull physics but have to admit I've nver really looked into the matter. I just assumed it was the same issue that causes the common problem with elevation on walkable meshes with hull physics. If it also happens with triangle physics, it must be some other cause.

Is there any point creating a JIRA for this?I can't imagine LL is willing to spend a lot of time and effort correting this bug but they may agree to do something if it turns out to be an easy fix.

[arton Rotaru]

The gap is by design. It's the collision radius of the Havok physics engine. Indeed that radius is adjustable, and it has been tweaked for certain prim types in 2011. That's why physics-set prim cubes sit flush on top of each other, but mesh cubes doesn't. lt's there for performance reasons, which is why the chance that it will be tweaked for meshes as well is rather small.