Any save created on the new game version causes a simulation error with the Deserializer, old saves continue to load fine

[Mikeblatz]

Information​

I have disabled all mods​

Yes

Required​

Summary​

Any save created on the new game version causes a simulation error with the Deserializer, old saves continue to load fine

Expected Result​

The game loading correctly

Actual Result​

The simulation fails to load due to an "Array index-out of range" error on any save created on version 1.18.1-f3

Steps to reproduce​

Loading a save game created on the previous version, and then saving even without changing anything, creates a save that can't be loaded. The same is true when starting a new game on my empty custom map with all mods disabled

Platform​

Steam

Game Version​

1.18.1-f3

Optional​

Description​

As of the update on the 29th of October, I can no longer load newly created saves. Saves created on the previous game version continue to load correctly, and the simulation runs as before once in-game with no new issues.
I have already attempted extensive troubleshooting, such as by comparing the output logs of my most recent save from before the update to a newly saved version of said save, with nothing changed. As the output logs are quite extensive, I have copied the relevant sections here.

Once loading a new save, according to Loading Screen Mod the error occurs when loading TreeManager. I have compared the output logs of the two saves. The error occurs after:
"71,942ms | Saving C:\Users\mikeb\AppData\Local\Colossal Order\Cities_Skylines\userGameState.cgs [Settings - Internal]" (new save)
"53,914ms | Saving C:\Users\mikeb\AppData\Local\Colossal Order\Cities_Skylines\userGameState.cgs [Settings - Internal]" (corresponding section in the old save).
As there is some difference in the way Tree Control loaded, I removed all trees and loaded a save with tree control disabled. Here the identical error occured at:
"64,859ms | Simulation error: Array index is out of range."

The error text is identical between the two failed attempts and reads as follows:
"73,285ms | Simulation error: Array index is out of range.
at (wrapper dynamic-method) TreeManager/Data.TreeManager+Data.Deserialize_Patch0 (TreeManager/Data,ColossalFramework.IO.DataSerializer) <0x005d4>
at ColossalFramework.IO.DataSerializer.Deserialize<SimulationManager/Data> (System.IO.Stream,ColossalFramework.IO.DataSerializer/Mode,ColossalFramework.IO.DataSerializer/LegacyResolverHandler) <0x00134>
at ColossalFramework.IO.DataSerializer.Deserialize<SimulationManager/Data> (System.IO.Stream,ColossalFramework.IO.DataSerializer/Mode) <0x0004b>
at LoadingManager/<LoadSimulationData>c__IteratorB.MoveNext () <0x007d4>
at AsyncTask.Execute () <0x000a8>
[Core]"

Creating a new game on the mostly empty custom map I used for my city creates a similar error at the same time in the loading process, even with all mods exept LSM disabled (also disabling LSM still led to the same result):
"74,777ms | Simulation error: EndDeserialize(VehicleManager) tag mismatch:
at LoadingProfiler.EndDeserialize (ColossalFramework.IO.DataSerializer s, System.String name) [0x00000] in <filename unknown>:0
at VehicleManager+Data.Deserialize (ColossalFramework.IO.DataSerializer s) [0x00000] in <filename unknown>:0
at ColossalFramework.IO.DataSerializer.Deserialize[Data] (System.IO.Stream stream, Mode mode, ColossalFramework.IO.LegacyResolverHandler customResolver) [0x00000] in <filename unknown>:0
at ColossalFramework.IO.DataSerializer.Deserialize[Data] (System.IO.Stream stream, Mode mode) [0x00000] in <filename unknown>:0
at LoadingManager+<LoadSimulationData>c__IteratorB.MoveNext () [0x00000] in <filename unknown>:0
at AsyncTask.Execute () [0x00000] in <filename unknown>:0 [Core]"

Meanwhile starting a new game with my current mod and asset configuration on a vanilla map doesn't throw up this error.
I still had the last version of the game installed on my laptop. Aside from "Assembly-CSharp.dll", I replaced the "Managed", "Mono" and "Plugins" files with those from the last version, and the error still occured. Also replacing "Assembly-CSharp.dll" when attempting to create a new game on my custom map with all mods disabled still caused the same error as above.

Since this deserialization error occurs irrespective of wether mods are enabled, and as previous saves still load and run as before, I suspect that this is a game error occuring whilst creating a save on the new version, and not directly caused by any mod. As I've invested a significant amount of time and effort into my current city and therefore would rather not lose the ability to continue working on it, I'd greatly appreciate any help in finding the cause of this issue. Thanks in advance!

Attached is the save saved on the previous and current version of the game, as well as the Map file and the output logs corresponding to my report above. (Edit: I was unable to upload the saves, presumably due to their file size of over 300 Mb each)

Tags​

• Compatibility
• Map
• Modding
• Save
• Stability
• User Experience

Attachments​

View attachment output_log new map.txtView attachment output_log no tree control.txtView attachment output_log new save.txtView attachment output_log old save.txtView attachment River Delta A.crp

 

[Mikeblatz]

II was able to compress and upload the saves mentioned above

 

[MarkJohnson]

You have over 4,000 subscriptions and over 15,000 workshop items (Each subscription can have several items).

This could take days to resubscribe.

maybe check your memory. You could potentially have up over 64GB of workshop, not counting the possible 32GB the game uses already with all DLC purchased.

Otherwise you may need some help from the Skyve mod.

 

[Mikeblatz]

I've got 96 GB of 5200 MHz RAM installed, in addition to a 128 GB pagefile. According to LSM, while loading the game uses about 26 GB of Pagefile and 70 GB of RAM, which becomes less once the game is running. When creating a new city on the map attached above I had disabled all mods including assets, and also encountered the issue described above. I have additionaly deleted and redownloaded the Workshop folder, which didn't fix the issue (the old save still loaded fine). The output log also doesn't seem to indicate an issue with assets or RAM.
I have also tried loading a newly created save with all mods Skyve inidcates as problematic turned off; that too hasn't worked. Meanwhile, my exact configuration continues to load saves created before the update and new games on vanilla maps just fine.

 

[MarkJohnson]

It sounds like your saved game is corrupt, since all of your other saves load fine.

I have seen this issue happen with mods like, "Exit to Desktop" that kill windows processes. There is a slight chance that if you are saving the game and click exit to desktop that it will immediately stop the saved game, as you told it to kill the save game progress.

Possibly, it could have been a glitch in some other mod(s), that cause it to corrupt the save game process.

But if your other saved games are fine, then you should be fine continuing from one of your other saved games.

FYI, disabling mods does not prevent them from compiling and loaded into memory that can still cause issues. You need to unsubscribe a mod in order to stop it from working completely.

Assets won't function if you disable them and shouldn't harm your game.

 

[Mikeblatz]

The issue with continuing from another save is that any time I save, no matter which old save, the new save is corrupted. I will try it without "Exit to Desktop", but I generally do wait before quitting until the in-game popup about saving has resolved.

 

[MarkJohnson]

Maybe it needs a mod reinstalled? I don't know why it won't save.

Make make a new thread in the "Support & Bug Reports" forums, but first go to this thread and create the documents it asks of you, then attach those files to the new thread you create.

How to best report technical issues and get help!

So you are experiencing some technical issues? Please keep the following in mind before posting in this forum. Reducing non-constructive posts in this forum is for everyone's best as it will help us identify real problems faster and it will be...

forum.paradoxplaza.com