 |
There are two parts to the suspicious door code: 1 - the open door causes the guard to search 2 - the door-handling code lets the guard close the door afterward 1 doesn't involve handling the door. 2 does. But since the guard can't use the display case as a regular door, the code probably aborts near the beginning of door handling. This would work if a switch controlled the display case, but I don't think that's the case here. I'll look at this in a week or two, to make sure there's nothing about vertically sliding doors that causes a glitch. I'll also look into adding a special case for closing doors (normal or sliding) the AI can't walk through if I can't find anything wrong. |
|
|
Rule: AI will not use a door if they can't fit through it. This rule is in place to prevent AI from opening cabinet doors. Unfortunately, in order for an AI to close a suspicious door after searching around it, he needs to be able to handle it. All the door-handling code is written on the premise that the AI can walk through the door. If I had noticed this during an alpha or beta test, I would have changed the mechanic to the following: 1 - Write a script that keeps an eye on the AI's alert level, and the moment it drops from Searching to Suspicious, do the following: a - check for distance from the sliding door. if not close by, quit. if close by, continue with the next step 'b'. b - if the door is open, walk up to it, interact with it (raise hand toward it), and close it. |
|
|
I'd like to close this as 'not a problem'. |
|
|
Can I close this? It can't be done with the current door-handling code, it's difficult to add to the current code, and the best solution would be my suggestion in my Dec 5 note above. |
- Anonymous
