Matt Keeter

Good catch, this should be fixed in 8af59dcabbed2f9cee9e39d53cc9e565ddd7d55c!

Yup, that's a little too much "fun". Meshing assumes that shapes are bounded, and 2D shapes are effectively infinite on the Z axis, so you're seeing meshes that are strange...

Not sure, probably just a historical artifact. The C API is basically an accumulation of "whatever I want to call from a higher-level language", so it's not particularly orthogonal. My...

Yup, looks very similar to what I did. Here's the patch from when I was recording the videos (several years out of date at this point): ``` commit 54cbac1baf6e8375602dd9dbdc96021836143526 (render-video)...

That's an interesting question, and a cool model – are you building a mechanical keyboard? My first instinct is to use the `blend` function, which does blending between distance fields....

This is going a little off the rails, so I'm going to hide the unrelated comments in this thread and #209. @sene-je, your writing style seems familiar, so I think...

Nope! What are you trying to do? Have you succeeded in opening Studio and seeing the example file in there?

Yup, that syntax creates a `vec2` or `vec3`. You can get components with the `.x/y/z` functions, e.g. ```scheme (define v [1 2 3]) (.x v) ; returns 1 ```

Cool! You're correct – this is what the oracle interface is for. The oracle interface is C++-only, as it's very low-level; there aren't any public oracles besides basic building blocks....

Very weird! One easy thing to check is whether the shape is producing the same optimized f-rep, e.g. replacing the last clause with ``` str(cylinder_round_y(cylR,cylL).\ call(blend, sphere(cylR).\ call(blend, sphere(cylR), 0.12),...