Still getting this Error message... As I'm looking for a strange behavior, this was one of my suspects and I lost some time...

About 1 week ago, I wanted to submit my App for review and I found this thread while trying to solve the impossible (5.5" screenshots of my app, which needs iOS 17+). I decided to not fake these screenshots somehow, but created an image with Photoshop, consisting of a white background and the text "Devices with this screen resolution are not supported. Minimum requirement: iOS 17." My App was rejected because of some other rules. Yesterday I requested the App review again and guess what: Screenshots of a 5.5" iPhone aren't required any more! I don't know, if Apple did a regular cleanup because of the iOS18 launch, or if my "screenshot" caused the removal, but finally they recognized, that it doesn't make sense to request screenshots of an antique device...

I have the same issue. My new App supports iOS 17 (and later). How shall I create a 5.5" screenshot. There are no old devices or simulators with 5.5" display supporting iOS 17...

Same problem here. Simple SwiftUI App on M1 under Rosetta. Now graphics code at all. Very annoying.

It seems like now I have found the real reason for this strange behaviour... My App was a AppKit application before the migration to SwiftUI. After some searching in the project settings I found under the macOS build target in "Info / Custom macOS Application Target Properties" the key "Main nib file base name" which contained "MyApp.nib". It looks like sometimes during the start of the App this file was searched and not found and sometimes maybe the start of the App was faster so it was not searched for... However: After deleting the value, the error didn't reappear. Let's hope this is the final solution...

Unfortunately the problem was NOT solved... I've got the problem again and it happens most of the time on my 2021 MacBook Pro (M1 Max) and not so often on my 2020 MacBook Pro (M1). So it looks like a timing problem...

I found the problem :-) During the initialization of the model of the @main view I did some longer running tasks to initialize some background processes. Now I run them on a background task using perform{} and the bug is gone. The reason why the bug appeared on my new, faster hardware obviously was the addition of more data structures to be initialized, so it took longer.