Is anyone else having issues with watchOS 3 SpriteKit in beta 3? I have a simple SpriteKit animation for a fitness app similar to the progress ring in the workout app. It worked fine in beta 2, but in beta 3 the app crashes in a few seconds with "gpus_ReturnNotPermittedKillClient" error. The app run fine without the animation.
I've submitted a bug report with example project for this #27432871
EDIT: my bug report was marked as a duplicate of #27333513
Seeing this as well, using scenekit + spritekit overlay. I guess I'll try turing off the overlay and see if that helps.
disabled spritekit overlay and still getting gpus_ReturnNotPermittedKillClient.
It mostly happens for me when the watch sleeps
more details.
I've been able to reproduce this fairly consistantly in another experimental app I created using only spritekit and animations.
start an animation, then raise/lower wrist a bunch of times to sleep and wake the watch.
The crash happens usually on sleep and before any of the lifecycle callbacks, so am unable to attempt to prevent by stopping the animation.
I've submitted a bug report with example project for this #27432871
EDIT: my bug report was marked as a duplicate of #27333513
Just an update. This SpriteKit crash still happens in Beta 4 unfortunately. Here's the stack trace:
libGPUSupportMercury.dylib`gpus_ReturnNotPermittedKillClient:
0x349f7342 <+0>: movw r1, #0xbeef
0x349f7346 <+4>: movs r0, #0x1
0x349f7348 <+6>: movt r1, #0xdead
-> 0x349f734c <+10>: str r1, [r0]
0x349f734e <+12>: bx lr
Anyone have an update?
The same crash is still happening to me in Beta 5, but it seems to be improved. It now happens after willAppear is called, but not every time.
Does SpriteKit still crash for you? Any update on the status of 27333513?
Thanks,
Jeff
Still getting crashes, but it seems to go in cycles. I can go quite a while without any issue, then it gets into another state and starts crashing again.
I see the same crashes even with no animations, just updating nodes will sometimes trigger it.
#27333513 was marked as closed about a week ago.
I just spent several hours using my app with no crashes, so I feel pretty confident that it is a transient issue. Hopeful to see it go away soon.
This problem is pretty bad in Beta 6. I didn't actually ever get it in Beta 5, but it happens a lot for me now. Trigged by the raise and lower wrist as you were saying. What happens for me is it will work nicely for 10 minutes or so, then I'll raise my wrist and the Sprite Scene is blank. Then poof. App gets killed.
There is note in the beta 6 release notes though:
"Attempting to reload an InterfaceController object that contains a SpriteKit or SceneKit control while in the background causes a crash.
Workaround: Mark SpriteKit and SceneKit controls as paused in Interface Builder, and resume them only when the extension is marked as being in the foreground."
I've done this and ensured they are paused until when in background. Doesn't work though.
I also attempted to pause SpriteKit unless the extension is in the foreground and still had intermittent issues that were triggered by a raise of the wrist. My use of SpriteKit was pretty simple, an animated progress ring, so I decided to pull the feature and use animated images instead.
I've tried to pause the SpriteKitScene on the ExtensionDelegate, but is still randomly crashing.
After extensive testing I think this is a system bug, I've opened a new bug report with a simple test app project 28043815
I will keep you posted once the bug report state changes
On the GM the scene interface is usually paused in willActivate. I have to unpause it or the screen will not update.
I think this might be how there are solving this issue.
I have not seen any gpus_ReturnNotPermittedKillClient crashes yet, but only used it a short time.
Everything feels better. Animations are smoother than beta versions.
Drove me crazy until I figured out to set non pause state within the WKInterfaceSKScene. Finnaly I solved it!
I'm still seeing this crash with SpriteKit on watchOS 3 GM. I tried doing the trick of pausing and unpausing on willActivate and didDeactivate correspondingly but did not fixed the issue. It may have help though... I’m not sure since the crash is very random and happens only after a while.
Is anyone else still having problems?
