In T2094#98528, @EricBlank wrote:

I ran .529 for about an hour and a half and couldnt trigger the debug script or lose any docked entities. Updating to .531 now
Again in .531 cannot seem to trigger the bug. I got loads of messages about docked entities from traders guild faction fleets losing their docks and having them moved, but none of them triggered the debug script and my fleet had no such problems

-Validated-

In T2118#98526, @33Cav wrote:

2 could be a result of the core counting as 2m cube. In your example did you add 19 or 20 blocks to reach dimension of 21?

I ran .529 for about an hour and a half and couldnt trigger the debug script or lose any docked entities. Updating to .531 now

Just ran .199.531 on a game where I reinstalled starmade from scratch and then added my blueprints and server-database.

2 could be a result of the core counting as 2m cube. In your example did you add 19 or 20 blocks to reach dimension of 21?

misclicked on "closed" instead of "resolved"

fixed Nullpointer. Both fixes 0.199.531 or later

Logs for the nullpointer:

[2017-04-20 01:35:56] [RAIL] Client(1) DISCONNECTED FROM RAIL: Ship[schema_1492644932863](517) LEFT IN RAIL SpaceStation[ENTITY_SPACESTATION_Big shipyard V2_1492644923502(516)]: []
[2017-04-20 01:35:56] java.lang.NullPointerException
[2017-04-20 01:35:56] 	at obfuscated.aaz.a(SourceFile:952)
[2017-04-20 01:35:56] 	at obfuscated.Yc.updateLocal(SourceFile:2731)
[2017-04-20 01:35:56] 	at obfuscated.Yr.updateLocal(SourceFile:403)
[2017-04-20 01:35:56] 	at org.schema.game.common.controller.EditableSendableSegmentController.updateLocal(SourceFile:1168)
[2017-04-20 01:35:56] 	at obfuscated.YA.updateLocal(SourceFile:493)
[2017-04-20 01:35:56] 	at obfuscated.J.update(SourceFile:1333)
[2017-04-20 01:35:56] 	at obfuscated.aFN.a(SourceFile:719)
[2017-04-20 01:35:56] 	at obfuscated.aFT.b(SourceFile:698)
[2017-04-20 01:35:56] 	at obfuscated.V.e(SourceFile:1110)
[2017-04-20 01:35:56] 	at org.schema.game.common.Starter.startMainMenu(SourceFile:1489)
[2017-04-20 01:35:56] 	at org.schema.game.common.Starter.main(SourceFile:1034)

Fixed class cast exception

-QA Testing-

I decided to also test on our server, since I have full access to the logs and could submit them here. I thought it might also be useful to have a snapshot from whatever the issue may have been on the version we are currently running.

In T2094#98467, @EricBlank wrote:

In T2094#98459, @AndyP wrote:

Does replicate mean: you could get the bug to happen again and thus upload the logs again, or does it mean you replicated the situation and the bug did not happen?

Sorry, i mean the server debug script detected something changed and shut itself down. When i went to view the bombers after reloading it all their docked entities are still attached. So i guess the bug didnt happen but the server thought it did?

Thread dump when class cast exception occurs (taken with VisualVM):

And furthermore, when docking a ship to the shipyard and manually undocking, also causes a crash to main menu. When rejoining the server, the ship was also gone.
No trace of it.

Also possibly related. If you enter the design ship (not physical), then bring up the DEL menu and click the "Undock all" button, it will crash you. Upon rejoining, all the docked entities on the design are gone. Unloading the design and reloading it works, but the entities are still undocked.

In T2094#98459, @AndyP wrote:

Does replicate mean: you could get the bug to happen again and thus upload the logs again, or does it mean you replicated the situation and the bug did not happen?

ok thank you. This is likely a false positive. I uploaded another build that will fix that. It will also notify admins, if a ship is waiting for its docks for a bigger amount of time.

Does replicate mean: you could get the bug to happen again and thus upload the logs again, or does it mean you replicated the situation and the bug did not happen?

I replicated the bug again, this time in 199.527. did not get the admin error messages this time. Jumping around on their patrol path when they're about to offload does seem to be fairly reliable.

Optimized Data (disabled for now. need extra checking)

In T2094#98394, @schema wrote:

in 0.199.526 or later

in 0.199.526 or later

Fixed a bug where the fallback would trigger in the wrong circumstances.

thank you. this might help a lot

In T2094#98359, @schema wrote:

Please report if you encounter the admin message:
[ADMIN] (NON FATAL) WARNING: Dock for <ShipName> was not in
the right sector.
It has been moved to its mothership.
Please send in logs since this fallback shouldn't happen."

Recent versions do not seem to be affected anymore. The issue can be close.

Andy you are welcome.

Now the stations are merged.

Thank you for the logs, we will look into them.

The attached file is a zip of the logs from trying the new pre-build.
The first time entering the world I didn't even make it in before error occured.
I tried again. Upon doing so i got many many errors flashing across the screen and red text to small to see all across the screen. So i did some looking to see what I could find before the server shut down.
i took a screen shot of what I seen. I enclosed it in the file as well.
It probably would have been better if I recorded it.
From appearances the errors happened as the ships loaded in. It appears you created a function to test if a child entity is missing from the parent from looking at the error log.
I would bet this is a database race condition as I posted in the thread.

I Know this is not my thread but i want to help.
I have version .524 and get the Admin message.
Here is the Log file.

Please report if you encounter the admin message:

I have a possible fix (0.199.524 or later).

Minion Seven was decomissioned due to losing turrets before these logs. The turrets themselves may have been floating around in the background. The fleet in question had: Minion 01, Minion 02, Underling 01, and Underling 02. (Names partially abbreviated due to my mark/version system).

Oh and another thing. I suppose the turrets aren't part of the fleet directly (as in: you added them to the fleet and then docked them)

Also, did you jump into their path? I mean, did you trigger the loading by jumping in a sector next to them directly, or did you jump into a sector that wasn't close enough for them to load so they went into loading range and loaded?

I saw a false positive in there (debug triggered shutdown because it thought there was a difference but the docking process just didn't finish)

Log zip dragged and dropped per lancake's request. The logs include a bit of unrelated building, the fleet launch, fleet debug maneuvers, a game crash, and fleet movements that include turret loss and turret return.

Could you provide logs of that session where they disappeared? Zip those up so you don't have to upload them individually. (Drag and drop the file into the comment section to upload to phabricator)

In T2094#98298, @Cecil_Effwad wrote:

Could schema please clarify which game build has corrected the problem?
I installed the latest pre-release availabe today and performed the fleet debug mve test described above. At first I thought that the issue was well and truely gone as I watched a fleet of four ships with ten turrets each do a debug mission and keep all their turrets. That worked, so I disabled fleet debug and sent the fleet from one of my bases to another one. When I caught up to them they still had all their turrets. I ordered them home and jumped ahead of them to watch their arrival.
This time they lost all their turrets. Just like before. I turned them back around and followed their flight path until they magically regained their turrets about half-way through the journey.
Was the latest pre-release build the wrong build? Should I install the dev build instead?

Definitely needs resolving. Fed up of finding missing turrets or NPC delivery fleets leaving doors or turrets behind!

-QA Testing-

Could schema please clarify which game build has corrected the problem?

Can go one over allowed dimension. Once that happens only remove will work on that structure.

on mac the caps lock key counts as if the caps lock key was pressed continuously for some reason

Can't reproduce so most likely OSX only.

-QA Testing-

This should be finally resolved.

In T2094#98260, @5ilence wrote:

I think an update to the title of this bug is required - it doesnt only affect NPC fleets, but player fleets aswell. Hopefully this will get some traction, as its currently crippling the use of combat fleets, and its even getting mentioned in YouTube videos (https://www.youtube.com/watch?v=UO74wAJxtPc at 2:40) as 'that fleet turret bug' #FleetTurretBug!

I think an update to the title of this bug is required - it doesnt only affect NPC fleets, but player fleets aswell. Hopefully this will get some traction, as its currently crippling the use of combat fleets, and its even getting mentioned in YouTube videos (https://www.youtube.com/watch?v=UO74wAJxtPc at 2:40) as 'that fleet turret bug' #FleetTurretBug!

That's great news! I'll leave it on resolved for a while in case some special case pops up.

Ok. thank you. I'll probably cause the effect of ffmpeg failing forcefully for my side to try to find all remaining issues.

First: thank you greatly for giving this attention.

Does not crash on ffmpeg now, but does crash when attempting to click the red X and close the empty [Introduction] window; whether launched by vanilla startup config, or by manually clicking the tutorial button. See screenshot. User cannot re-launch game from menu, but must exit to desktop, otherwise user is presented with apparent endless loop of NPEs dialogs.

-QA Testing-

-QA Testing-

It is possibly fixed with this. But I can't be sure.. The reason this seems OS dependent is because this is a running condition, so it really depends on how fast/slow threads are being handled.

Ok ok, perfect then ^_^good work guys and tanks for all !

possible connection to openjdk https://bugs.openjdk.java.net/browse/JDK-8050499

should no longer crash now

Nevermind, it's already been fixed in the pre-builds. Most likely related to T1299
Don't need to set this one on private though, merging tasks.

Can't seem to reproduce, I changed a basic rail into other rail types and orientations, but its connection with the storage chest was still there.
Do you have a blueprint of a ship where this is happening? Upload it here (drag and drop the file into the comment section) and I can take a look at it.