Aabb support

[JoshuaMKW]

This would be useful to have for fast selection support amongst other potential utilities so I've added a simple calculation for it.

[Sage-of-Mirrors]

Hey, apologies for not getting back to this for so long. Is this still something you would like to contribute?

[JoshuaMKW]

Hey, apologies for not getting back to this for so long. Is this still something you would like to contribute?

I think it is a useful thing to expose through API for various reasons, the most obvious one being a form of determining user selection (I'm of course aware that J3DUltra supports pixel perfect selection nonetheless). Another useful application this could provide is a collision approximator to filter out bogus interactions before doing more exotic collision detection by e.g. a .col model.

So yes, I think it'd be within the interests of J3DUltra to merge this and support the feature.

[LagoLunatic]

Does this implementation currently support adjusting vertex positions for the joints they're rigged to? Or are only unrigged models supported?

I currently do a hack to try to estimate a model's visual bounding box automatically by looking at vertex positions, but vertexes that are rigged to only a single joint being stored as relative to the origin makes this pretty inaccurate a lot of the time. If J3DUltra had a proper implementation that took this into account it would be very useful.