Single Player bugs and crashes v43 plus (SVN) - After January 2023

Check the amount of enemy workers. I remember a similar stupid lag being caused by having too many automated workers, EACH ONE thinking "too hard" what to do per turn.
That was about MY workers, but the same problem obviously applies to enemy workers as well - I mean, they ARE "automated".
Stupid PATHING, lol.

I agree that the worker AI is currently quite time-consuming.
I’ll look into this slowdown issue once I’ve finished my update on City production.

I’ve recently added a number of checks and improvements to AI behavior to make merging and attacking more effective, but I knew this would come with some processing cost. Optimization will be the next step—so I ask for a bit of patience in the meantime.


@ingvarr777, if you could also share a save file, that would help me compare and confirm whether we’re facing the same issue or different ones.
 
So I nuked a one tile island with a BBM, causing it and the surrounding tiles to be infected with Infectious Fog. Many turns later I conquered the city, and I was able to remove the fog in the city tile with a regular worker. But my constructor ships don't have the option to remove the smog from the water tiles. Am I missing something?
Thanks for the report.
I'm working on other themes, but It's added in the to-do list
 
I mean, I remember seeing it with my own eyes, lol. Between literally one turn and the other, automating a drove of workers increased MY end-turn time like 5 times. And de-automating them stopped it just as obviously as it first activated it. Was... disappointing, honestly.
 
Check the amount of enemy workers. I remember a similar stupid lag being caused by having too many automated workers, EACH ONE thinking "too hard" what to do per turn.
That was about MY workers, but the same problem obviously applies to enemy workers as well - I mean, they ARE "automated".
Stupid PATHING, lol.
This shouldn't really be a factor in my game.
The total number of AI cities is around twenty.
They didn't produce hundreds of workers.
 
This shouldn't really be a factor in my game.
The total number of AI cities is around twenty.
They didn't produce hundreds of workers.
I had this issue with as few as 15-20 workers already, if automated and "being confused about pathing". Easily achievable by AI. Not sure what causes it to escalate, though.
 
I had this issue with as few as 15-20 workers already, if automated and "being confused about pathing". Easily achievable by AI. Not sure what causes it to escalate, though.
I don't know.
I didn't have that.
When my automated workers got "confused," they just stood around stupidly somewhere on the periphery, "waiting for an escort," whatever that means.
 
I don't know.
I didn't have that.
When my automated workers got "confused," they just stood around stupidly somewhere on the periphery, "waiting for an escort," whatever that means.
Nah, I mean when they (the screen) FREEZE while CALCULATING their pathing. I sometimes literally SAW it during my end turns.
Oh, I obviously mean those with "many move points", obviously. Short-ranged ones calculate it much faster, duh.
 
Nah, I mean when they (the screen) FREEZE while CALCULATING their pathing. I sometimes literally SAW it during my end turns.
Oh, I obviously mean those with "many move points", obviously. Short-ranged ones calculate it much faster, duh.

Yeah, and with Space maps, it's possible that the Path algorithm get stuck or lost. I can confirm the very long time consuming turn, but What I see in profiling log is only pathing. I'll have to dig deeper to find the problem
 
The Quantum Lab and the Asteroid mine on Earth have a limited number of buildings, and this is not stated in their properties.
Is this a bug or did someone just forget to enter it?
 
In the Transhuman era, the game started up like a nail again.
At first I thought it froze, but after 10-15 minutes it continued to run.
There was always a delay at the beginning of the rounds when switching between units, which gradually increased as the game progressed, but this time it happened from one round to the next.
For now, I'm testing whether this current state is permanent or a temporary glitch.

Hi, I think I’ve found one of the causes of the new slowdown. I added the ability for a worker to search for another city to improve. It was checking every city on the entire map to see if it could reach them. On space maps, since there are no connections between Earth and the stars, the pathfinding algorithm often failed and consumed a lot of time. I’ve now limited that ability to the worker’s current area. This results in far fewer generatePath.Failure calls and therefore less processing time.

1758891191239.png


Here a DLL with the modifications, so you can try.
 

Attachments

Hi, I think I’ve found one of the causes of the new slowdown. I added the ability for a worker to search for another city to improve. It was checking every city on the entire map to see if it could reach them. On space maps, since there are no connections between Earth and the stars, the pathfinding algorithm often failed and consumed a lot of time. I’ve now limited that ability to the worker’s current area. This results in far fewer generatePath.Failure calls and therefore less processing time.

View attachment 743489


Here a DLL with the modifications, so you can try.
We need to somehow create a faster way to check cities outside of the area to see if they can be reached via a bridging factor to that other area. This actually would be something we need to eventually apply to a lot of things to fix numerous area limited searches that are otherwise too slow. It was a 'to-do' project but getting into the later game becomes rather important to check flight capabilities, tunnels, and so forth in a faster way than just pathing, right? Nothing to be tooo worried about yet but at this rate we are approaching this need fast.
 
The new Coredll has improved the slowdown a lot, but by the middle of Galactic Era, a round still requires several minutes of waiting due to the breaks in between/at the beginning of the round.
So it's playable, with compromises.
 
Getting a Python error:
 

Attachments

  • IMG_20250927_201858.jpg
    IMG_20250927_201858.jpg
    2.1 MB · Views: 11
  • IMG_20250927_201932.jpg
    IMG_20250927_201932.jpg
    1.1 MB · Views: 12
When I try to enter the BUG menu, the game freezes.
The music continues, but the screen freezes.
After half an hour, I gave up and quit.
 
When I try to enter the BUG menu, the game freezes.
The music continues, but the screen freezes.
After half an hour, I gave up and quit.
Unfortunately, it’s a side effect of the DLL I provide, which includes some changes to that menu (changes that are in other files too). You’ll either have to use the GitHub version or wait for the SVN update.
 
Sajnos ez a mellékhatása az általam biztosított DLL-nek, ami magában foglal néhány változtatást az adott menüben (olyan változtatásokat, amelyek más fájlokban is megtalálhatók). Vagy a GitHub-os verziót kell használnod, vagy meg kell várnod az SVN frissítését.
OK, thanks.
 
When starting with the new dll, there are no animals.
Not at all.
The old one is copied back.
Then, overwriting it with the new one, they are back in the loaded position.:blush:
 
I'm playing one of the recent v45 beta SVN builds, and I got a bug that always happens when using the A-Bomb unit (which I think comes from the Advanced Nukes game option) where using the unit plays the animation but then results in not exiting from being locked in the animation and basically softlocking me into not being able to move any other units or end turn until I press Esc and load a previous save.
 
I'm playing one of the recent v45 beta SVN builds, and I got a bug that always happens when using the A-Bomb unit (which I think comes from the Advanced Nukes game option) where using the unit plays the animation but then results in not exiting from being locked in the animation and basically softlocking me into not being able to move any other units or end turn until I press Esc and load a previous save.
Hi, it was a problem in the unit definition in xml, it wasn't set as suicidal. I think I've fix it a month ago, but I didn't test it with the last SVN
 
Back
Top Bottom