Thanks, Quinn - that's good advice. I don't have the capability to test all of the things that you've listed, although I did start in on observing and checking as much as I could via packet captures. In the tech note, it says "In some circumstances NSURLSession may retry such requests automatically" — is there any information on the behaviour when this happens? I assume it's a limited set of retries with an time interval based backoff?

Thank you! This workaround is fine for now, and it's great to hear that SwiftData's support for building with Swift 6 is still in progress.

Sorry that it's taken so long - I've just filed FB9750037. It's pretty light on details, because I can't share the framework itself to demonstrate the issue.

Yes, that is correct - the XCFramework that I'm building is for use in a macOS app, built using AppKit and SwiftUI. It won't be used on iOS.

It’s great to hear that it sounds like the issue you’re investigating, Quinn! The dynamic libraries are built by conan (https://conan.io/), and then copied into an Xcode framework that bundles them into the Frameworks directory of the framework. The framework is then archived via Xcode, and I grab the export from that archive for inclusion in my application’s project.

The final product is a standalone macOS app - this framework is the only framework dependency, but I do have 3-4 other Swift Packages integrated. You're correct that exporting the app from the Organiser did the right thing. The binary and dylibs are all signed as expected.