View Issue Details
ID | Project | Category | Date Submitted | Last Update | |
---|---|---|---|---|---|
0020417 | AI War 2 | Bug - Gameplay | Oct 1, 2018 6:15 pm | Oct 2, 2018 6:36 pm | |
Reporter | ZeusAlmighty | Assigned To | Chris_McElligottPark | ||
Status | resolved | Resolution | fixed | ||
Product Version | 0.776 A Legible Galaxy Map | ||||
Fixed in Version | 0.778 I'm Talking To HIM, Not You! | ||||
Summary | 0020417: ships automatically going to grav well; not all go through worm hole | ||||
Description | Having issues control my raiders--they act erratically regardless of FDR mode. Gave group command for fleet to go to distant planet and some would stay behind near each worm hole | ||||
Tags | No tags attached. | ||||
|
|
|
Hey Chris, this is a really good reproduction scenario. Load the game, select all the Raid Starships and have them move around that planet grav well (I would send them from one end to another). Every 3 or 4 times I would issue the command they would do something wacky. |
|
I can duplicate it!!! Happy day! ...even after all my fixes so far today on this build, this is still there. But now I can fix it. :) |
|
This save is such a goldmine... I'm so excited! The data in here is so messed up, it's glorious. It's nothing that will make the save unplayable, but it's revealing some cross-threading issues with the pooling. Finally something I can test repeatedly and thus fix! |
|
Is this what a programmers dream christmas is? |
|
Thanks! Finally nailed it. :) * Found and fixed a REALLY nasty bug that was causing entities to get their orders confused and do all sorts of insane things. ** This was a cross-threading issue where two different threads (short term planning and AI order reevaluation) would be both clearing out dud orders and putting the same order back into the pool multiple times. ** Now it is a lot more careful with how things go into and out of the pool, using the appropriate multithreading locks, as well as making it so that in the (very frequent) even that two orders try to go back into the pool at once, it only puts them in there once. ** There is also a lot more error checking and logging in there that throws out clearly-bogus orders and throws visible errors when finding an incorrect case. This is a lot better than it just getting subtly incorrect in a quasi-hidden way! |
|
Finding a juicy bug repro case and being able to fix it? You bet. :D |
Date Modified | Username | Field | Change |
---|---|---|---|
Oct 1, 2018 6:15 pm | ZeusAlmighty | New Issue | |
Oct 1, 2018 6:15 pm | ZeusAlmighty | File Added: ships moving bug.savemet | |
Oct 1, 2018 6:15 pm | ZeusAlmighty | File Added: ships moving bug.save | |
Oct 1, 2018 6:27 pm | BadgerBadger | Note Added: 0049667 | |
Oct 2, 2018 5:25 pm | Chris_McElligottPark | Note Added: 0049684 | |
Oct 2, 2018 5:25 pm | Chris_McElligottPark | Assigned To | => Chris_McElligottPark |
Oct 2, 2018 5:25 pm | Chris_McElligottPark | Status | new => assigned |
Oct 2, 2018 5:59 pm | Chris_McElligottPark | Note Added: 0049685 | |
Oct 2, 2018 6:14 pm | RocketAssistedPuffin | Note Added: 0049686 | |
Oct 2, 2018 6:29 pm | Chris_McElligottPark | Status | assigned => resolved |
Oct 2, 2018 6:29 pm | Chris_McElligottPark | Resolution | open => fixed |
Oct 2, 2018 6:29 pm | Chris_McElligottPark | Fixed in Version | => 0.778 I'm Talking To HIM, Not You! |
Oct 2, 2018 6:29 pm | Chris_McElligottPark | Note Added: 0049688 | |
Oct 2, 2018 6:36 pm | Chris_McElligottPark | Note Added: 0049689 |