Sandy Maguire

Add option for terse errors

4

Today I tried `core-warn` on a project that produced thousands of warnings. It'd be nice to have a terse error option, since the majority of the error message is a...

https://github.com/JonathanLorimer/core-warn/blob/cddf3bcfa51f4730134df4e89ccec36786f13c05/src/Warn/Dictionary.hs#L73

Add support for nested parens, e.g. "Definition (O(n) runtime)."

1

Took this todo out of the source. `pandoc-theorem` looks like an excellent project, and this is a reminder to myself for something to work on during #hacktoberfest

rotateExtrude behaves oddly when given noncontinuous functions

1

```haskell test :: SymbolicObj3 test = rotateExtrude 4.0 (Right $ \case 3.1847910293530797->V2 3.0 (-6.681629751553052) _->V2 0.0 0.0 ) (Right $ const 0) (polygon [V2 9.0 (-3.3500040166499234),V2 0.0 (-16.711593776250552),V2 0.0 0.0])...

No other combinators do. Instead, we should provide `center :: Object obj vec => obj -> obj` which moves the center of the bounding box to the origin.

Slow performance

4

Rendering nontrivial meshes takes forever at small resolutions.

eg: https://github.com/colah/ImplicitCAD/blob/742638a33e92d814267481b0fd7418b9a7afc050/Graphics/Implicit/ObjectUtil/GetImplicitShared.hs#L56-L57 Because the let binding is inside the lambda here, the runtime is going to call `getImplicit` on every child for every point it samples. If we instead pull...

@cartazio pointed me at https://www.iquilezles.org/www/articles/distfunctions/distfunctions.htm, which has lots of great primitives. We should steal them --- in particular `twist`, which would allow us to get rid of the ad-hoc twist...

Because of this, `rminimum` and thus `getImplicit . UnionR` can never possibly be associative either.

I have a sneaking suspicion that `-XStrictData` and `-funbox-strict-fields` enabled globally would aggressively improve the performance of `implicit` by eliminating all the pointers inside of `SymbolicObj`. #286 will further improve...