Any chance the issue correlates with enabling iCloud Private Relay? Take a look at a problem I have (link) which resembles yours in the broad strokes: old BSD socket code working perfectly for years before suddenly failing for a few “random” users on iOS 15. In my case the problem is the same on both Wi-Fi and cellular (as I had enabled Private Relay at the global level) but it’s worth a look.

Are you in the paid developer program? I don’t know if the free level has fewer offerings, but right now the full list in the Applications tab includes: Xcode 13.2 beta Apple Remote Desktop 3.9.5 beta Classroom for Mac 2.5 beta 2 macOS Server 5.12 beta 3 Apple Configurator 2.15 beta 2 TestFlight 3.2 beta Schoolwork 2.2 beta 2 Reality Converter beta 3

And the smoking gun is... iCloud Private Relay. If I disable iCloud Private Relay (and reboot if the network stack was already busted) then everything is rock solid. If I enable iCloud Private Relay, then the severe misbehavior described above will eventually occur. It does seem less bad on iOS 15.1 (the issue takes longer to show up) but it still does occur. I haven’t tested on 15.2 beta yet.

I’m curious why you would expect to “get the accuracy up by a notch” above what Core Location considers the “best” it can do. Those kids at Apple are pretty smart. I bet they read the GPS module data sheet and everything. ;-) As for why the API design philosophy is different from Android, I imagine an interaction something like this: Question: “What is my location?” Android: “Here's a mishmash of low level APIs you can try. They may not work consistently across our frustratingly fragmented hardware base and it may take longer to implement a robust solution to your problem but at least you get lots of options to try out.” Apple: “Here’s your location.”

That file is part of the macOS SDK, but it’s not part of the iOS SDK. The conversation now turns to: what exactly are you trying to accomplish using that file on iOS?

It’s the new 3rd checkbox in the options page when uploading or validating an archive. See this thread for earlier discussion of the issue and a link to a screenshot. Also the WWDC video linked from that thread is worth watching.

After seeing various reports of this behavior, it appears the new upload-time validation logic considers only the first number in a dotted version string (truncating the rest), so in this case it treats both your existing 2.1.194 and new 2.1.195 as equivalent to simply 2, which it increments to produce 3. You can prevent this by un-checking the automatic build number option when uploading. But going forward it’s probably easiest to just switch to an integer build number rather than using a dotted format that has a redundant app version on the front. (Especially since you are now stuck incrementing from 3 for your current version.) You can reset the build number to 1 if you want for each new app version.

Thanks! I've entered FB9712270 for the wonky location authorization behavior.

Can you post a small (cropped) screenshot showing this "debug banner"? You're saying this isn't part of your app but was added by the system? Or if it is part of your app (such as to designate an internal debug build) then just remove it. As for the other issues in the review, they seem pretty self-explanatory.

Just use this forum's search feature for the word "tobacco" and the fate of your proposed app will become clear.

I know, I know. Of course right now I'm scrambling to rewrite this whole thing using the modern Network framework. Just never got around to it before because the old socket-based code has been working great for over 10 years. Until this week, that is.

The message pretty much says it: you need to use a full developer program account, not a free "personal team" account.

According to the specs on Apple's web site, the iPhone 12 added "GLONASS, Galileo, QZSS, and BeiDou" while earlier models don't list those extra ones. Some news sites say earlier models sold in Japan also had QZSS. Either way, there's no API to get at the low level data. Core Location is a much higher level abstraction over the hardware. Bonus fact: If you are plugged into CarPlay, then iOS can use your vehicle's own navigation system data too.

This would be a question for the manufacturer of the printer. But I happen to have one too (somewhere) and the current app for it appears to be called NiiMbot in the US App Store at least. I also have an app called LabelCloud from the same people. One of those may be the right app for your printer.

The build string is available via sysctlbyname(3) (man page) for the name "kern.osversion".