Pathing shenanigans with raiders and mechanoids

[ghost]

Been trying to figure out how to reproduce raiders/mechanoids going through doors because I've seen people comment it on the workshops comments. For now I've only found this error from a raider deciding to pick up and steal a blast door. https://git.io/fhOnB

Raiders will extremely often just take down the blast door to steal it and run away with it, that seems silly.

As far as I can tell raiders will gladly go through Doors Expanded doors that are held open. But raiders that are fleeing will not, and sometimes mechanoids will not go through them either during their attack and instead spam the logs. Here's a log and picture: https://git.io/fhOC0

Seems like if they are trapped and their only path is through a held open Doors Expanded door, they stand still and start spamming the log. Going to try to mess around more with this.

EDIT 5: Nope not just trapped. If mechanoids are raiding you and hitting your walls and I force my blast door to be Held Open, they all stop and start lagging my game with their log spamming.

He's standing in front of a blast door that's held open and is instead spamming my logs because he can't go through. It's not just blast doors either, I've tested and can confirm this happening with every single Doors Expanded door, while they happily waltz through vanilla doors with no errors.

Built this on a brand new save, made the blast door Held Open and started a mechanoid raid and they immediately start spamming logs. https://git.io/fhO8t

[SineSwiper]

Commented on an issue in Steam about the inverse of this problem. Raiders seem to love to try to break through blast doors no matter what, because they calculate that it takes no effort at all to do so:

• First frame is a bunch of Door Expanded doors
• Second frame is adding a vanilla door
• Third frame is what it actually looks like, after turning off the Draw Paths debug feature.

I don't remember having this much trouble with blast doors prior to the April updates.

[lbmaian]

Should be addressed in #23 - can you give that a try?