Audun Skaugen

I'd advice using the layout modifiers in XMonad.Layout.Fullscreen instead if you want this behavior.

This would be great for us at cognite. We have an intersection operation that amounts to taking a big `Geometry` object, and intersecting it with a bunch of different `Line`s....

> Yeah, I think we can just do this https://docs.rs/prost-build/latest/prost_build/struct.Config.html#method.disable_comments but for debug impl? Should be pretty easy to add and isn't a breaking change. One would think that, but...

> Ah you're right, maybe we can have a attribute option that sets no debug and then that can be enabled via prost-build. Yes, I am investigating this option.

If I cyclically permute the coordinates, there is no self-intersection. So this is probably related to handling of the exterior being a linear ring. Edit: Nope, one more cyclic permutation...

I guess the `cartesian_intersect` function does not detect intersections with endpoints of lines?

You can use the points from the simplified geometry above: ```rust l1 = Line::new((1., 1.), (7., 0.)); l2 = Line::new((1., 0.), (1., 4.)); assert!(cartesian_intersects(l1.start_point(), l1.end_point(), l2.start_point(), l2.end_point())); ``` The point...

On a side note, I ended up implementing the algorithm described in this paper instead, but optimized using an r* tree in the same way as SimplifyVwPreserve: https://doi.org/10.1016/j.cageo.2013.08.011 Would there...

Got a failure with this linestring on your branch: ```rust let ls = line_string![ (x: -4., y: 3.), (x: 4., y: 3.), (x: 1., y: 2.), (x: 20., y:-3.), (x:...

Sure! I guess it just means that this algorithm tries hard to preserve topology, but can't quite guarantee it. Should probably document that fact :)