Simon Lindholm

Suurballe's algorithm is just mincost-maxflow with capacity = 2, we shouldn't include it (or maybe we should just make a note of that fact). Note that the real Suurballe's algorithm...

I think tinyKACTL included something to do with this.

Yeah, somehow that part of the pdf is prone to screwing up when things change size e.g. because of different LaTeX packages. It's some bug in multicol that comes up...

Thanks for the note, @Tyilo. I committed that change at 390bc1626e2ea6853a30e56289296fc2e12181bc. It seems to improve the display marginally on my system as well.

> I personally find 1d fenwick trees to be nearly completely useless - especially since we have the bottom-up segtree that runs nearly as fast as the fenwick tree. How...

> On the topic of shorting we could also decrease the font size for surrounding .tex code -- the algorithms are currently of rather smaller size than the math, while...

Added FastInput and GomoryHu, will look at the rest tomorrow.

Added CirclePolygonIntersection. LineProjectionReflection feels unnecessary since projection is pretty easy if you know your dot products: `a + (p - a).dot(b - a)*(b - a)/(b - a).dist2()` Maybe we shouldn't...

Is the only fix to the precision issue the new lineIntersection? Naively, that would seem to fix only the case with integer coordinates

Apparently the implementation is a bit slow; it got within 66% of the TL on https://ncpc20.kattis.com/problems/bigbrother. Probably due to calling atan2 in the comparator.