Hi!
We’ve encountered an issue related to texture streaming that results in a crash on both Android and iOS platforms. We have a cutscene with a close-up camera fly-by, as the camera passes a specific object, its texture is streamed in, which causes a noticeable visual pop. To address this, we tried setting the NeverStream flag on the texture. However, after doing so, the application crashes on both platforms just after opening the level that contains the modified texture and the cutscene:
libUnreal +0x13d20f54 UTexture::GetResourcePostInitState (Texture.cpp:1686) libUnreal +0x13d238f4 UTexture2D::CreateResource (Texture2D.cpp:1044) libUnreal +0x13d1e578 UTexture::UpdateResource (Texture.cpp:318) libUnreal +0x13d22ca4 UTexture2D::UpdateResource (Texture2D.cpp:746) libUnreal +0x0df15078 UObject::ConditionalPostLoad (Obj.cpp:1269) libUnreal +0x0dd2bd0c FAsyncPackage2::Event_DeferredPostLoadExportBundle (AsyncLoading2.cpp:6912) libUnreal +0x0dd20618 FEventLoadNode2::Execute (AsyncLoading2.cpp:4319) libUnreal +0x0dd20618 FAsyncLoadEventQueue2::PopAndExecute (AsyncLoading2.cpp:4460) libUnreal +0x0dd2d91c FAsyncLoadingThread2::ProcessLoadedPackagesFromGameThread (AsyncLoading2.cpp:7158) libUnreal +0x0dd324c0 FAsyncLoadingThread2::TickAsyncLoadingFromGameThread (AsyncLoading2.cpp:7420) libUnreal +0x0dd386b4 FAsyncLoadingThread2::ProcessLoadingFromGameThread (AsyncLoading2.cpp:8567) libUnreal +0x0dd3c148 ProcessAsyncLoading (AsyncPackageLoader.cpp:314) libUnreal +0x0e0af090 StaticTick (UObjectGlobals.cpp:816) libUnreal +0x12f7ca54 UGameEngine::Tick (GameEngine.cpp:1719) libUnreal +0x148bd9c0 FEngineLoop::Tick (LaunchEngineLoop.cpp:5819) libUnreal +0x148b67f0 AndroidMain (LaunchAndroid.cpp:613) libUnreal +0x148c5d00 android_main (LaunchAndroid.cpp:859) libUnreal +0x148ee6d4 android_app_entry (android_native_app_glue.c:226) libc +0x000e6a00 <unknown> libc +0x00084c6c <unknown>
We’d like to know if using the NeverStream flag is the correct approach for preventing streaming during such cutscenes, or if there’s a recommended alternative. Also, could this crash be caused by an incorrect setup on our end? Any guidance would be appreciated.
Best regards,
Daniil