Testing tline3d() + colour tables along with some new userdata operations for picotron 0.3.1

This example uses a batch lerp() to generate edge data that can be fed directly into a batch tline3d call without spending a lot of cpu shunting data around. The example scene has 800 quads and uses 60~80% cpu.

Instead of directly providing a build-in tri() or even draw_scene() function, the idea is to instead work the userdata api into something that can carry implementations in userland wherever possible. I think in the long term this approach can yield a richer ecosystem of middleware and snippets that explore original techniques that extend beyond use cases the api design was tested on.

Brain dump on one possible way to draw quads in 0.3.1:

https://pastebin.com/raw/9MMLsUFb

(it doesn't explain the whole thing, but shows the direction I'm heading)

"the perfect loop doesn't exist-"