...But INFile(data: urlData, ...) only works if urlData is smaller than 512 MB, so that's not a complete replacement.

Update: As a fix, I have used: if let urlData = try? Data(contentsOf: fileURL) { intentResponse.file = INFile(data: urlData, filename: fileName, typeIdentifier: nil) } but it's still frustrating to encounter such an unexpected behaviour.

Update: the iOS device in question was jailbroken. I'll assume the jailbreaking method had somehow replaced libc++ with another version.

The answer is that I need to use fclose, not close. So: do { fclose(stdout_file) fclose(stdin_file) try stdout_pipe.fileHandleForWriting.close() try stdin_pipe.fileHandleForReading.close() try stdout_pipe.fileHandleForReading.close() try stdin_pipe.fileHandleForWriting.close() }

I can confirm I have the same issue for two of my apps. I also contacted Apple Support.

Update: I was able to make it work by uploading the build using an earlier version of Xcode. You can access earlier versions of Xcode here: https://developer.apple.com/download/all/ It works for me with Xcode 12.1, it doesn't work with Xcode 14.2, I'll let you try with the intermediate versions.

That's interesting, I have the same issue.

It is not a very common issue, so it's hard to infer things from a small sample size. But, from what I understand: if it happens, it happens immediately after installation, for the first command. rebooting does not help. re-installing usually helps, but not always. some users report that re-opening the app after a few days fixed the issue.

It is, yes. It's named a-Shell and it weighs around 1.2 GB once expanded.

One user has reported that the issue solved itself without action (which confirms it's a "weird bug", as you said). Another user has experienced the issue and managed to generate a sysdiagnose log (256 MB). What is the next step for me? Should I file an official bug report? For orders of magnitude, I've had 5 different users reporting this issue over the past month for 10,000 installs. Assuming only one out of 10 users affected takes the time to raise a complaint (it's a free app), it happens roughly once every 200 installs.

I've received feedback from several users that their devices were never jailbroken. So the next question is: how can they produce a useful diagnosis?

Thanks a lot for the feedback. I'll ask the question. In case it's the latter, how do they "get a sysdiagnose log from the user"?

Oh, sorry, it's on iOS. I have reports of the issue occuring on iOS 15.7, 16 and 16.1. Re-installing the app usually fixes the problem.

Here is what it looks like after around 5 hours.

Hi, I have the same symptoms: the app crashes immediately on all devices running iOS 14. I have tried everything recommended in this thread (disabling bitcode, adding libswift_Concurrency, adding "-Wl,-weak-lswift_Concurrency -Wl,-rpath,/usr/lib/swift" to "other linker flags"...) but it still crashes.