Inventory Calculation gone wrong

[matony1989]

I have came across this bug. From the replay, there is fight starting at GF 30845. There was one general wandering around from previous fight trying to find way back home. After he went to HQ during assault on that HQ it bugged the privates.HQ showed I have 2 privates available. In my YT video you can see that that number never changes to lower - I had empty military building near borders, but soldiers were not going there. I tried assigning them into HQ to defend, but nothing. So the Privates didn´t exist, but HQ was showing 2 anyway.

https://www.youtube.com/watch?v=VzWjuwHNOoU

2026-03-18_09-07-58.zip

[ottml]

The problem comes from this part of code

In void nobBaseMilitary::CancelJobs() the soldier is removed from the leave_house queue. The entre in the queue becomes an empty entry. When the check in void nobBaseWarehouse::AddActiveSoldier is checking if the soldier is in the leave_house queue it returns false and the inventory visual count for Privates is wrongly increased. This triggers later an assert in void nobBaseWarehouse::HandleLeaveEvent() and leads to the wrong inventory. Showing two Privates which do not exist.

[Flamefire]

Good catch.
That check doesn't actually make sense at all: It can never be true as the leave queue contains unique_ptrs

Working on it. Should be reasonably easy to write a test for that. I'm just surprised our current tests and test replays didn't catch this.