Set up a mine on the east side of a cove that is about 10 chunks wide.
Blocked a biter base from spawning enemies, and then let it absorb some pollution from a nearby iron mine sometime in the past (313 pollution absorbed). Located at {-520, 2705}
Pollution cloud from the miners currently does not cover any bases. It has polluted the blocked base sometime in the past.
The pollution cloud from the miners reaches about 4 chunks to the west from the mine, out over open water.
The one nest already created in the new expansion on the west side of the cove has not absorbed any pollution, and should not do so in the time that is relevant.
The Player is currently (relatively) close to 0,0.
The save provided is a stripped down save, in hopes of reducing the amount of variables and save size.
What happened?
An ongoing expansion on the other side of a approximatly 7-8 chunk wide cove will do one of 3 things:
(Located at -938, 2659, far south of the players position, and currently not within radar coverage, but the chunk north of the gathered biters was in radar coverage when the expansion started (daylight coverage only).)
The expansion groups "Stop Gathering" timer has underflowed. (In an earlier save it was positive, but by the time the expansion (sometimes) fails it's a very large number.
IMPORTANT: To make the attack happen, look only with map mode, if the player (in editor mode or otherwise) is too close, the behavior changes. This is a bit annoying, as I have not been able to find a way to step a single tick from within map mode (+ editor mode).
Once the Remaining Building time hits 0 a worm spawns, deleting the single medium spitter in the group.
The attached save is 9 ticks before this happens to make it easier to debug, I can provide a save about 5 minutes earlier if this would help, however all that happens during that time that I can see is the Remaining Building time ticking down.
1. Biters fail to complete the expansion.
This is shown for a single tick, when "Remaining Building Time" reaches 0.
Code: Select all
Group Behaviour:
BaseBuildingBehavior:
State: Failed
Current building position: {empty}
Remaining Building time: 0
Unit Group: 1217639
Number: 121...
State: Moving
Type: CommandDriven
Stop Gathering: 40830647 (*a VERY large number (21 digits maybe)* from now)
Number of Members: 7 (All visible)
Attached To Spawner: FalseThen the expansion group becomes a standalone group, containing 1 Behemoth biter and a mix of spitters.
Code: Select all
Group Behaviour:
State: Waiting
Unit Group:1217639 (same number as before)
Number: 121...
State: Moving
Type: CommandDriven
Stop Gathering: 40830647 (*a VERY large number (21 digits maybe)* from now)
Number of Members: 7 (All visible)
Attached To Spawner: FalseAfter that 2-4 minutes later they (usually) attempt to attack.
In the cases where the attack succeeds, it's targeting the miners about 10 chunks to the east.
and one of the two following things happen, depending on what they player has done.
(It should be safe to skip forward 120 ticks, but I have not tested to get the exact amount of ticks to skip)
1a. Attack the miners. This makes little to no sense to me.
This will be quite visible if the pathfinding debug settings are turned on, at least if stepping tick by tick as it's visible for a few ticks. If time is left running, the pathfinding "path" can be used as a indicator.
1b. Biters don't attack, but create a area of "Bad Attack Chunks" after having completed more of the expansion before they fail. This also doesn't make sense, as why would they even attempt to attack simply because an expansion fails. Sadly I was unable to see anything about what they tried to attack, as it doesn't show the pathfinding for any tick that I could see. Is there some other debug info that would give this information?
This can happen if all miners are deleted.
Or if the blocked base is deleted.
And in some other cases as well. (Player position might matter.)
A short time after that they attach to the nearby nest.
OR
2. Biters fully complete the expansion. This would be a "normal" case.
To make this happen, move down in editor mode to the mine and then let the time run, this can make the expansion succeed.
What did you expect to happen instead? It might be obvious to you, but do it anyway!
Biters not (attempting to) attacking a "distant" target when a expansion (partially) fails. And instead attaching to the closest nest(s).
Does it happen always, once, or sometimes?
Always*
It depends on user action. But it can be reliably reproduced by doing nothing.
The result can be changed by user actions as described above.
The save file is 50Mb, so it's >30Mb limit, so here is a link: https://www.dropbox.com/scl/fi/8956isfk ... if7tp&dl=0
Images are attached showing a bunch of debug info. Most of the things I believe are relevant are already typed out above.
