Day 19: Beacon Scanner
Haskell: Part 1 (10:32:56, rank 4356), Part 2 (10:51:54, rank 4218)
Ah, the true descent into madness begins.
By far the hardest part 1 so far for me. Really, I actually did spend 10 and a half hours on part 1. What I was doing seemed like it should work, but I messed up my repeatF function for repeatedly applying cube rotations to the scanner.
Even after finding that, it still wouldn't work even for the sample input. I was still not sure that the rotations were correct, right up until I made a seemingly small change that solved the problem. At this point I'm not even sure what fixed it, but it was something to do with how I returned candidate lists of a scanner's points transformed to be from scanner 0's frame of reference.
This was a great puzzle but quite an unpleasant experience which had me questioning my basic intelligence and ability to solve problems like this.
Part 2 was relatively trivial after part 1 was working -- I was already passing back the offset of each scanner after it had been rotated to its correct orientation, but was discarding them somewhere. So part 2 just involved keeping those offsets (which I called loc) in the state returned back up to the match' function and pushed onto the accumulator. The Manhattan distance comparison itself is easy, just the maximum distance calculator between all pairs of scanner locations.
I don't even know. The program grew and ended up with debugging expressions everywhere which didn't help much. Guess I can't draw too much from this, but I'm pretty disappointed in how it went. It was hard but I don't think it's really "get stuck for 11 hours" hard. Maybe I needed to take more breaks and look at it all again piece by piece.