I have the same error, and I could not fix it doing the same thing as you, i don’t know why it is happening but it was not the case using 4.15.2 …
Fatal error: [File:F:\dev\UE4_16\Engine\Source\Runtime\CoreUObject\Private\Serialization\AsyncLoading.cpp] [Line: 1335]
No outstanding IO, no nodes in the queue, yet we still have 515 'AddedNodes' in the graph (with 5 boot nodes).
what other things should i be checking?
Eventually found only by trial and error that the problem was triggered by a specific BTService that the Behaviour Tree referenced. Deleting the reference to it didn’t resolve the problem but recreating the Behaviour Tree without it stopped it from happening. Then I recreated the BTService in question from scratch and renamed it at the problem started happening with that. But if I deleted all the graph contents from the BTService the problem still occurs, seems highly temperamental. Maybe this should be marked as not resolved and reported as a bug
Hi, we’re having a crash on launching a packaged game on 4.16.
Fatal error: [File:D:\Build\++UE4+Release-4.16+Compile\Sync\Engine\Source\Runtime\CoreUObject\Private\Serialization\AsyncLoading.cpp] [Line: 1334]
No outstanding IO, no nodes in the queue, yet we still have 342 'AddedNodes' in the graph (with 5 boot nodes).
HyperJam_Win64_DebugGame!FDebug::AssertFailed()
HyperJam_Win64_DebugGame!FEventLoadGraph::CheckForCycles()
HyperJam_Win64_DebugGame!FAsyncLoadingThread::CheckForCycles()
HyperJam_Win64_DebugGame!FAsyncLoadingThread::TickAsyncLoading()
HyperJam_Win64_DebugGame!FlushAsyncLoading()
HyperJam_Win64_DebugGame!LoadPackageInternal()
HyperJam_Win64_DebugGame!ResolveName()
HyperJam_Win64_DebugGame!StaticLoadObjectInternal()
HyperJam_Win64_DebugGame!StaticLoadObject()
HyperJam_Win64_DebugGame!UObjectPropertyBase::FindImportedObject()
HyperJam_Win64_DebugGame!UObjectPropertyBase::ParseObjectPropertyValue()
HyperJam_Win64_DebugGame!UObjectPropertyBase::ImportText_Internal()
HyperJam_Win64_DebugGame!UObjectProperty::ImportText_Internal()
HyperJam_Win64_DebugGame!UClassProperty::ImportText_Internal()
HyperJam_Win64_DebugGame!UProperty::ImportText()
HyperJam_Win64_DebugGame!UArrayProperty::ImportText_Internal()
HyperJam_Win64_DebugGame!UProperty::ImportText()
HyperJam_Win64_DebugGame!UProperty::ImportSingleProperty()
HyperJam_Win64_DebugGame!UScriptStruct::ImportText()
HyperJam_Win64_DebugGame!UStructProperty::ImportText_Internal()
HyperJam_Win64_DebugGame!UProperty::ImportText()
HyperJam_Win64_DebugGame!UObject::LoadConfig()
HyperJam_Win64_DebugGame!FObjectInitializer::PostConstructInit()
HyperJam_Win64_DebugGame!FObjectInitializer::~FObjectInitializer()
HyperJam_Win64_DebugGame!UClass::CreateDefaultObject()
HyperJam_Win64_DebugGame!UObjectForceRegistration()
HyperJam_Win64_DebugGame!ProcessNewlyLoadedUObjects()
HyperJam_Win64_DebugGame!FEngineLoop::PreInit()
HyperJam_Win64_DebugGame!GuardedMain()
HyperJam_Win64_DebugGame!GuardedMainWrapper()
HyperJam_Win64_DebugGame!WinMain()
HyperJam_Win64_DebugGame!__scrt_common_main_seh() [f:\dd\vctools\crt\vcstartup\src\startup\exe_common.inl:253]
kernel32
ntdll
The issue seems similar to this one:
Though that was apparently fixed in 4.16, whereas this issue only appears in 4.16 and not in 4.15 and my call stack doesn’t contain anything about Gameplay Tags.
It might still be related to Gameplay Tags however because the end of the log before the crash contains:
[2017.06.09-06.32.46:781][ 0]LogStats: UGameplayTagsManager::ConstructGameplayTagTree: Construct from data asset - 0.000 s
[2017.06.09-06.32.46:782][ 0]LogStats: UGameplayTagsManager::ConstructGameplayTagTree: ImportINI - 0.000 s
[2017.06.09-06.32.46:782][ 0]LogStats: UGameplayTagsManager::ConstructGameplayTagTree: GameplayTagTreeChangedEvent.Broadcast - 0.000 s
[2017.06.09-06.32.46:996][ 0]LogWindows: Windows GetLastError: The operation completed successfully. (0)
Unfortunately I have no idea how to investigate this or reproduce this in a blank project, and also am not able to access engine debug symbols when debugging the packaged game.
Any advice greatly appreciated!
After a weekend debugging this with compiled engine source I found that it was being caused by ini config containing references to classes that no longer exist.
Still not sure why it was fine in 4.15, but this was the issue on my end at least.
Can you post your whole stack trace?
I’m also getting the same error after upgrading to 4.16, although in the shipping build strangely it was the one a few lines down (“No outstanding IO, no nodes in the queue, yet we still have ‘PackagesWithNodes’ in the graph.”) rather than the AddedNodes one, but in a DebugGame package it is the same as you. Attaching my stack trace for debug game package. I’ve just got setup so I can recreate the problem while debugging in VS, any pointers of how to find the problem? Thanks!
Did you have any luck resolving your problem?
Hmm, all I can suggest is looking into how your AIKartComponent interacts with BehaviorTree since it looks like it’s related to those.
Coincidentally I’ve just started getting this issue again today for another class, but this time it does actually exist so I have no idea what the problem is…
Fun times
Ah bad luck, thanks for the response, will keep looking into it and let you know if I have any luck
Nice, glad you solved it. The second instance of the issue for me was again referencing a blueprint class in config that I thought existed, but was not being cooked. Ended up switching to use the asset manager instead.
Will reopen this issue because I agree it should be reported.
i got this issue with a circular dependency on blueprints on ue5. and solved by referencing the base class of one of them (it was a simple cast check).
this is still an issue and i was able to avoid it, but seems like it can become a big issue in the future.