I noticed that beds mess up the pathfinder the most, particularly in the morning and evening. I've noticed colonists stuck on useable tiles like fireplaces or walking into walls (This is why I use doors more than walls). In one other case, I saw a colonist stuck walking into a tree, and another made its way up to a roof area and started pacing like it was in an infinite loop. Sorry if I made it sound like owls were too hard, I actually found them too obnoxious. I recommend looking into Minecraft phantoms, which suffer from the same problems. In particular, most of the owls' kills probably were byproducts of bugs or other things not controllable by the player, if not outright player stupidity. Currently, owls seem defeated by a floating roof. Optional combat or defense orientated nighttime events would probably be a substitute for why I wanted it kept (Padding the resource count/keeping a bed clear). Just a heads up though, blocking resource gathering/building outright at night may frustrate some. A system based on lighting conditions may be better. The simplest way is to expose the priorities to players in some ways. CDDA's base building system, Minecraft's village mechanics, and No Man's Sky's settlements would be good to look into for this as they feature very different approaches. Something I mentioned indirectly (and just now caught it was not directly), a configurable sandbox mode would be immensely useful to test the game properly.
I gave it a shot but the work vs. reward level is not paying off. I spent weeks in my free time overhauling a number of systems including the pathfinding and I added a sandbox mode. The game runs more efficiently and the villagers were able to navigate complex mazes (including multiple levels) confidently but they are still doing weird unpredictable behaviours like walking into trees and I haven't been able to figure out why so the current build would be no fun to try. I am finding it frustrating to work with the old version of Godot and I tried porting to the new version but it is a monster task for uncertain gain. I appreciate your interest but for the sake of my own sanity right now I need to move on. If more people were interested it could be worth the effort but as someone wise once said "I do this for fun, and it's not fun any more". I'm still happy to hear ideas if you have them, especially if you are able to highlight what you like about it that makes the game fun, maybe there is a simpler games I could make instead. It's possible one day I may revisit this (especially if I start from the beginning and design it properly, since this was only intended to be a quick prototype) but for now I wanted to let you know my situation rather than keep you waiting indefinitely. Thanks again!
Tried the sandbox this morning and again hour or so ago. The sandbox and pathfinding works very well, the villagers do reliably get to their destination (and in one case, switch paths to a better one). I did notice three things in my playthroughs though. Firstly, the game crashed during every playthrough, though at different times. The only common factor I know of is that I was building things. More trivially, removing the delay on the delete tool made it function like a raycaster, deleting everything in the players line-of-sight. In addition, the larger villagers now do not fit through doors. A 1×3/2×3 door would be sufficient, or possibly returning the hitbox to the old size.
I fixed a game crash but I'm not sure it was the one affecting you, please test and if it crashes send the log file. I am slowly tackling some of your suggestions whilst also adding other stuff that I feel like. You are welcome to bring an issue or a suggestion to the surface again if you think it is important and I should get to it sooner. Thanks again for your advice and feedback! Also if you take any screenshots or video clips I can use them to freshen up the main page.