06 may Without that sea monster puzzle weighing on me, I’m happy to start the second-to-last puzzle
Day 24, Part 1
It involves a floor tiled with hexagonal tiles. The tiles have a white side and a black side, and can flip from one to the other. The puzzle involves starting from a center tile, and following directions (east, northeast, west, etc.) to get to another tile, which must be flipped. The answer to the puzzle is how many tiles are flipped after following all the directions.
So! I've never had to work with a hexagonal grid before, but so many games have one, it must be a solved problem. I google “represent hex grid in array” and land on a Stack Overflow question, which leads me to this brilliant page, “Hexagonal Grids” by Amit Patel. This is nothing short of a miracle, telling me everything I need to know in order to do things with hexagonal grids.
After reading https://loansolution.com/title-loans-hi/ that page and thinking about it for a while, I decide that I will use cube coordinates (x, y, z) and that I don't even need to store a grid. I just need to store the destination coordinates that each instruction from the input takes me to. A tile is white at the end, if its coordinates are reached an even number of times (including 0), and a tile is black if its coordinates are reached an odd number of times.
I could store the destination coordinates in a hashmap from coordinate to count, but I wonder if there is a multiset similar to the multimap I used a few days ago. There is. With that, I can write the code for Part 1:
Day 24, Part 2
In a surprise twist, the second part of the puzzle is yet another Conway's Game of Life, this time on the hex grid! So no more storing the coordinates of the flipped tiles in a multiset.