spoilers for AoC 2023!
Nothing fancy this year in form of util and structure.
Part 2 was harder than expected. Tried using regex to solve it, but the problem was overlapping words such as: xtwoney which could be two and also one. The problem was when it was the last number so it would match two and consume the characters.
Solved it first with a super ugly "reverse" regex and only getting the first and the reversed (last) first number.
Feels like I should have parsed the data into real structs, but it was a simple enough problem.
I started making a plan to do a lookup in a hashmap for the coordinates, but then I forgot about that and just iterated over everything. Would be interesting to see the performance differences...
Was a pretty nice and standard task. This time building structs to handle the input.
Started building HashMaps, but that quickly became to hard so with my own structs it was more manageble. Still it took to long for part 2 so I added Rayon to use all cores. It still took almost 2 min on my MacBook M1 and 40s on my UM790 Pro.
Tried to save all processed ranges and ignore seeds that was in that range. It was a bit faster, but not significant...
This feels like it was the simplest task this AoC. I was worried that it would be to big to brute-force, but it took only 27ms on my UM790 Pro. The only thing I needed to change was to use u64.
For me this was the hardest task because I first implemented real poker-rules for determining the value. Everything became much simpler when I read the instructions correctly, but it meant many things were designed for a more advanced case which took some time to undo. It was nice to be able to lean on Rust's type system and deriving Ord to get the value of a hand.
Part 1 was simple, but for part 2 I struggled since I was unsure if it really was a stable cyclic graph. I went for it in the end anyway, but it was not a straight line...