Can anyone else confirm (or othewise), my experience that setting a SpriteKit scene as an overlay on a SceneKit scene, via setting overlaySKScene, does not render the SKScene. The same application (same binary), for me, renders the SKScene correctly in 10.11.3, but when run on 10.11.4b4 (and b3) fails to show the SKScene. [Bug Report #24850598]
SceneKit overlay (overlaySKScene) problem - 10.11.4?
I have some problem with overlaySKScene in OS X too, but my problem is Crash 🙂 when trying to set another scene in overlaySKScene. Even simple overlaySKScene = SKScene(size: bounds.size) couses Crash
I have the same issue on the released 10.11.5. Same code renders the overlaySKscene just fine on iOS and tvOS, and used to work fine on 10.11.3. (Not sure I tested it on 10.11.4.)
If you go to your GameView in your storyboard/xib, and look at the View Effects Inspector, and uncheck the Core Animation Layer checkbox, the overlaySKscene becomes visible, but all the textures are wrong --it grabs textures from my scnassets file somehow.
I downloaded the latest Fox sample code and also see the overlaySKscene is grabbing textures from the scnassets (and has the CA Layer checkbox unchecked). Checking the CA Layer checkbox makes the overlaySKscene not render anything at all. The tvOS and iOS builds render the overlaySKscene just fine regardless of this checkbox.
I suspect this is an OS bug, but I also had to move my SSD from a MBP15 (mid-2012) to an older MBP17 (early-2009), so it could be a Metal issue with this older GPU.
Have you found any resolution?
It's been hard to diagnose this. There may be buggy software involved, but my resolution seems to have been hardware related. Two weeks ago my logic board was replaced [same failures after 12 hours of use], with no effect; last week it was replaced again and this time it may be OK [no failures after three days]. I can't be sure that there was a relationship between these hardware failures (kernel reporting unable to restart GPU) and the Sprite/SceneKit errors I was getting, but I suspect there was. My sample program that illustrated the failure now works.
I learned long ago that my programming skills are not prefect, and that blaming the compiler or the hardware wasn't the solution; checking my code more carefully for edge effects, off-by-one errors, etc was always more fruitful. In this case, however, it was a hardware problem.