The key feature for part 2 is changing up what the Dijkstra solver views as an "adjacent" node, injecting the appropriate depth as we traverse inner/outer portals. The main thing that tripped me up here from part 1 was that I needed to be carrying more data along with portals than just their locations, and I could no longer get away with storing the portals as pairs of their inner/outer locations, so a small refactor was needed. Once I made those corrections, it was mostly a matter of ironing out the "get neighbors" function to adhere to part 2's rules, which took me a lot of debugging to get just right.
Parnic
dbc4b5cf6b
