So, I don't know if you remember my last post, the one about building the physics for my new non-euclidean game prototype. If you don't I'll refresh you on generally what happened.

Specifically, I said:
[quote]

It works!!
[/quote]

Well, turns out that I was a little wrong.

There are weird issues that come up when right and left don't really mean anything anymore. It turns out that converting a point from view-space to world-space in this game is sort of a lossy conversion. While I can technically convert back and forth points from view-space to world-space, surface normals don't really like to play along.

I *had* been playing a game of "use the world-space normals to infer geometry, and use that to calculate view-space normals," but when left could very well be located downward, it doesn't really work. This is especially apparent on corners, where there is no one distinct normal, but the collision normal is dependent on the location of something else, and that something else changes space rapidly and unexpectedly.

So corner clippings still rampant, I have a short little youtube link showing off about what it looks like at the moment. Excuse the programmer art, the semi-broken physics, the lack of sound...

Actually, on second thought, just excuse everything about it. I'm trying my best to fix it

[media][/media]