I might give up on day 15 unless we get more test ...
# advent-of-code
k
I might give up on day 15 unless we get more test cases that are blessed as correct by Saint Nicholas himself. Right now I'm trying to debug based on comments on reddit, but some of them directly contradict each other ­čśľ
g
I just debugged mine thanks to @karelpeeters when there are multiple targets at the same distance you should pick the path to the target closest in reading order, not the path that starts with a move closest in reading order. Thats just a bit different, and for none of the tests this difference is relevant.
j
It's actually both - first you need to select the the target closest in reading order, and then (if there are multiple paths of equal length to that point), select the first step according to the read order of all possible first steps