A crash occurs during the editor shutdown process after running the OodleNetworkTrainerCommandlet.

Development Pipeline & Plugins
, ,
1

Hello, I’m writing to report a crash that occurred while running the OodleNetworkTrainerCommandlet.

After the OodleNetworkTrainerCommandlet finishes and the editor is closing, an error occurs at the following line in

FBlueprintEditor* BlueprintEditor = static_cast<FBlueprintEditor*>(GEditor->GetEditorSubsystem<UAssetEditorSubsystem>()->FindEditorForAsset(BlueprintAsset, /*bFocusIfOpen =*/false));

At this point, UAssetEditorSubsystem is already null.

This crash occurs during the reinstancing process when FlushAsyncLoading() is called during Exit() in LaunchEngineLoop.cpp. Consequently, this is treated as a crash in our Jenkins pipeline, causing the job to be stopped. We are currently solving this problem by adding a validity check for UAssetEditorSubsystem.

Any advice would be greatly appreciated. Thank you.

2

재현 방법

  1. Create an Actor Blueprint.
  2. Attach a component to the Actor Blueprint.
  3. Modify the attached component to make it a target for reinstancing.
  4. Run the OodleNetworkTrainerCommandlet.
3

Hello,

I have put together a bug report for this, which can be tracked here if/when it’s approved for public visibility: Unreal Engine Issues and Bug Tracker (UE\-309482\). There is no ETA as priorities for bugs and features can shift at any time.

We don’t provide updates on UDN, but progress can be followed on that public issue tracker page.

Take care,

John

4

Hello,

We’ve tried to reproduce this issue and need some help. We created an environment that was setup with onlinesubsystem and create session / join session functionality. Adding an actor with a component didn’t crash it when modified and running the OodleNetworkTrainerCommandlet.

If you’re still running into this crash, could you please share logs, dumps and ideally a repro project example, or additional repro steps and project setup details.

Thank you in advance!

Regards

5

I’m sorry. During the server migration process, we deleted the Jenkins logs from before August.