Awesome

Advent of Code 2018

All my Advent of Code repos:

This year, I will do it once again in Nim, and ~~later on maybe in some other language too~~ in October 2021 I practiced Racket by solving some of the puzzles, and some time before that (in 2021) I also solved the first half of the puzzles using Python.

For Nim solutions, my aim was to provide clean, readable, idiomatic code. Python solutions extensively use lambda, and the aim was to provide short (but still readable) solutions. I'm a beginner in Racket, so the solutions here are part of my learning of the language, and consequently might not be idiomatic.

Solutions

Task	Nim Solutions	Python Solutions	Racket Solutions	Comments (on Nim solutions)
Day 1: Chronal Calibration	day01.nim	day01.py	day01.rkt	Dogfooding by using itertools to `cycle` through the input. Using `IntSet` for fast lookups.
Day 2: Inventory Management System	day02.nim	day02.py	day02.rkt	The original solution used `zip`, but that allocates a new sequence.
Day 3: No Matter How You Slice It	day03.nim	day03.py	day03.rkt	No need for regex, `scanf` macro is great for these kinds of inputs. Using smaller integers instead of `int` gives noticeable speed boost. Using templates to keep the main part short and readable, without unnecessary repetitions.
Day 4: Repose Record	day04.nim	day04.py	day04.rkt	[2018-12-04 06:00] Guard narimiran begins shift
Day 5: Alchemical Reduction	day05.nim	day05.py	day05.rkt	4x speed improvement when using the already shortened polymer (first part) for the second part.
Day 6: Chronal Coordinates	day06.nim	day06.py	day06.rkt	The slowest task so far.
Day 7: The Sum of Its Parts	day07.nim	day07.py	day07.rkt	Using `heapqueue` is a no-brainer here.
Day 8: Memory Maneuver	day08.nim	day08.py	day08.rkt	Using recursion is a no-brainer here.
Day 9: Marble Mania	day09.nim	day09.py	day09.rkt	Compile it with `--gc:regions` to get the most performance out of it.
Day 10: The Stars Align	day10.nim	day10.py	day10.rkt	The first usage of Nim templates this year. Using unpack for `<-` sequence unpacking.
Day 11: Chronal Charge	day11.nim	day11.py	day11.rkt	Using summed-area table to have O(n^3) solution (naïve solution is O(n^5)). Using threads gives 2x speed boost.
Day 12: Subterranean Sustainability	day12.nim	day12.py	day12.rkt	Nothing to write home about.
Day 13: Mine Cart Madness	day13.nim			Using complex plane is the obvious choice for the tasks like this, but `complex` in Nim is limited to floats, so I decided to use plain old tuples of integers.
Day 14: Chocolate Charts	day14.nim			Using `int8` to keep the memory usage down.
Day 15: Beverage Bandits				Ain't nobody got time for that.
Day 16: Chronal Classification	day16.nim			Nim bitsets don't have `pop`.
Day 17: Reservoir Research	day17.nim			Recursion keeps things nice and simple. Templates help with the readability.
Day 18: Settlers of The North Pole	day18.nim			Simplified boundary conditions by creating a border around the area.
Day 19: Go With The Flow	day19.nim			Figured out the inner loop, do it "automatically".
Day 20: A Regular Map	day20.nim			The initial solution first created a maze and then DFS-ed through it. Current solution immediately calculates the distances, for 3x performance gain.
Day 21: Chronal Conversion	day21.nim			The most interesting part of the task (figuring out what the instructions really do) was done on paper. Here is a part of it.
Day 22: Chronal Conversion	day22.nim			The first time in four years that I use `A*` algorithm for some AoC task.