Day 4

Invested another hour or so in collision resolution. I implemented a quick AABB algorithm I've become fond of, where I scan the mob's path's bounding box for potential edges, and compute the collision for the one it would hit first. Unfortunately, it still can't resolve certain corner cases, a problem made more difficult due to the integral coordinates I'm using. (See picture.)

The hackish solution is to discount edges that are between two wall tiles, but I think I will save that for when getting stuck on a wall actually starts to bother me. It's rare enough for now.