Matt Keeter

Results 243 comments of Matt Keeter

Yup, I agree that this would be nice! I'm not going to take up the role of flatpak maintainer, but I'd be happy to assist if someone wants to step...

Writing OS-specific metadata seems kinda weird – I think that adding them through the OS makes sense, and your link has a few good directions. If you end up with...

Hi @klpeders ! This isn't working as you'd expect because of how `loft` works under the hood. The `loft` operation does a weighted blend between the two distance fields, with...

There's actually a more fundamental reason against having ternary operators – the algorithms for rendering and meshing are pretty flexible, but require C^0 continuity, i.e. `f(x + ε, y, z)...

@doug-moen That's an interesting function – I have to think about it a little more. > But LibFive works differently: all of the intermediate results must be continuous functions, not...

Wow, that's quite a glitch! I wonder if there are numeric stability issues somewhere in the cone function...

I believe this is because of the branch cut in `atan`, where it breaks continuity and goes from π to -π (with Z negative, X going from positive-to-negative): ![Screen Shot...

@yunfan thanks for the hint – I updated the website to work with the new `sphere` argument order (it used to be `sphere center r`, and is now `sphere r...

This is another function that we should inject into the `Shape` class! I think that the most general-purpose way to handle this is to add `mypy`-style annotations to `gen_py.py` and...

Oh, that's a good point! It would be pretty easy to implement a `Shape.call(fn, *args, **kwargs)`, e.g. `circle(1).call(reflect_xz).call(move, [0, 0, 1])`, something like ``` class Shape: ... def call(self, fn,...