Incorrect layer boundary for 3D Meshes

Heya Dave, thanks for the reply. I did check the Orthogonal Views early on and you’re - ALMOST!  :-D

It’s a case of comparing Apples and Oranges - there is no issue in the orthogonal views - this does not mean all is fine in the perspective views nor should we assume both sets of views perform the same calculations. 

For my case, there is no issue with extruded Texts and Shape Layers but only for 3D Meshes with non-square edges. 

So, with a 2-Up View (Top and Active Cam) - if you start with the 3D Cube Mesh, as is, and then crank up its Bevel  Radius Radius, you will see the edges of the Cube in the Perspective View shrinks inwards regardless of the camera type - and all is well in the Top View.

Now, if you perform a similar routine with extruded Texts and Shape Layers, no such issue. 

My thoughts are this shrinking of the edges is due to tiny primitive shapes used to draw 3D meshes - when rounded or bevelled (which creates roundness, inherently), parts of these primitive shapes disappears but the app does not re-calculate the 3D Rect Boundary and relies on the default, pre-rounded/bevelled shapes. 

On the workaround, I did think about it - theoretically it should work but it’s going to requires tests (for code + test) time and during runtime, it’s going to add compute time. 

So, my thoughts are to double check with engineering - the shrinkage in the visual size of the cube when its bevel size is large (to see the artefact) should mean its 3D Rect Boundary should update. And if this is agreed, then all shapes that are non-rect should also have their default 3D Rect Boundary be re-calculated. 

Of course, I could be imagining all of these and MSU  :-D  

TIA