If I run the following code in didFinishLaunchingWithOptions()
UNUserNotificationCenter.current().requestAuthorization(options: [.alert, .badge, .sound]) { granted, error in
if granted {
DispatchQueue.main.async {
application.registerForRemoteNotifications()
}
}
}
Then the result is that didRegisterForRemoteNotificationsWithDeviceToken() gets called.
However if I change the code to be just:
DispatchQueue.main.async {
application.registerForRemoteNotifications()
}
Or as as its already running on main in this scenario, then just
application.registerForRemoteNotifications()
Then didRegisterForRemoteNotificationsWithDeviceToken() does NOT get called, but also neither does didFailToRegisterForRemoteNotificationsWithError().
Obtaining a push token is supposed to be independent of the user granting notifications permissions, so why am I not observing that behavior?
I only observe this behavior when running on hardware, when running on the simulator both forms of the code work.
Yet its nothing to do with my phone not being able to contact the Apple servers etc. - if I change the code back and forth back and forth between the two then if 100% works when using requestAuthorization() and 100% doesn't when not using it.
There's nothing additional or out of the ordinary with the code, its standard app delete template stuff.
Why isn't it getting a push token when requestAuthorization() isn't used?
(I've tried adding an async delay to calling registerForRemoteNotifications(), but it made no difference).
Selecting any option will automatically load the page
Post
Replies
Boosts
Views
Activity
I've noticed that NSLog() doesn't do anything on iOS 26 when an app is downloaded from TestFlight.
I've got an app uploaded with a liberal sprinkling of NSLog lines in it for debugging purposes - if I download that onto an iOS 18 devices, and use Apple Configurator's console or the Mac's console app, then the logging output is verbose and as expected.
But if I download that same app onto an iOS 26 handset, there's nothing.
Logging is visible in the Xcode console, however sometimes there are situations where you need console logging with a TestFlight build.
What happened? Was this an intentional change? Why. This is a major hindrance in diagnosing issues / bugs etc.
Trying to add some release test notes to a TestFlight build - but after clicking the blue save button nothing happens, nothing is saved.
I tried with two different browsers and two different Apple accounts, same thing with both.
Anybody else experiencing this issue?
The iOS documentation shows notification actions buttons with the text center aligned: https://developer.apple.com/documentation/usernotifications/declaring-your-actionable-notification-types
But there's no apparent way for an app to control this. The buttons are controlled and rendered by the system and the text is always left aligned.
Is there some way to get the text center aligned?
I'm currently finding it impossible to get a text filtering extension to be invoked when there's an incoming text message.
There isn't a problem with the app/extension because this is the same app and code that is already developed, tested, and unchanged since I last observed it working.
I know if there's any history of the incoming number being "known" then the extension won't get invoked, and I used to find this no hindrance to testing previously provided that:
the incoming number isn't in contacts
there's no outgoing messages to that number
there's no outgoing phone calls to the number.
This always used to work in the past, but not anymore.
However, I've ensured the incoming text's number isn't in contacts, in fact I've deleted all the contacts.
I've deleted the entire phone history, incoming and outgoing, and I've also searched in messages and made sure there's no interactions with that number.
There's logging in the extension so I can see its being invoked when turned on from the settings app, but its not getting invoked when there's a message.
The one difference between now and when I used to have no problem with this - the phone now has iOS 18.5 on it.
Its as if in iOS 18.5 there ever was any past association with a text number, its not impossible to remove that association.
Has there been some known change in 18.5 that would affect this call filtering behavior and not being able to rid of the incoming message caller as being "known" to the phone?
Update
I completely reset the phone and then I was able to see the the message filter extension being invoked. That's not an ideal situation though.
What else needs to be done beyond what I mentioned above in order to get a phone to forget about a message's number and thus get an message filtering extension to be invoked when there's a message from that number?
Delete me
Is this setup possible / have a solution:
There is a .xcframework F, which uses a 3rd party library, lets call it L.
There is an app which uses the xcframework
The app also uses L
Both the app and F use SPM to integrate L
F is using L for its own internal purposes. F is providing some functionality to the app. How it implements that ideally should be a black box from the app's perspective.
The app also happens to use L for its own purposes.
I can't get this set up working, always get warnings about duplicate symbols when running the App.
This will presumably be due to the fact there are separate copies of L in both F and A.
So how can that be eliminated?
Can F not statically like to L and use the App's version of L at runtime? If so how can Xcode be configured so that F can actually compile?
Or vice versa - can the App not statically link in its own copy of L and use that in the framework? If so, similar questions, how to configure Xcode to set this up?
I can't believe this is an obscure use case, yet after days of searching and reading documentation I can't find any solution.
Note that I was able to get this going when the app and the framework used Cocoapods to integrate L, but I just can't do similarly if the use of Cocoapods is replaced with SMP.
When using cocoapods, within the frameworks Xcode section, the pods framework is set to Do Not Embed. This is probably the vital difference between the working Cocoapods implementation and the not working SPM solution. However, when using SPM Xcode doesn't present any option to either embed nor not embed the dependency. Why not? Can it somehow be set to not embed?
I just noticed this today while building a React Native app.
Anybody else seen this and know what it means/how to track it down/how to fix it
A message filter extension is only forwarded SMSs by the OS for filtering, iMessages aren't.
But what is the situation with RCS messages? Will they be filterable by a message filtering extension?
If an Xcode project has some compiler flags set in Build Phases / Compile Sources, then is it possible to have those enabled if scheme A is selected and disabled if scheme B is selected.
Same question for things in Build Settings, such as Other Compiler Flags.
I suppose it could be achieved by having two targets, one with things enabled and one without, but for a very large complex project, duplicating targets is not necessarily an easy thing to do.
Topic:
Developer Tools & Services
SubTopic:
Xcode
I've got several Xcode iOS projects and in the Editor menu section there are dozen's of options, as in the diagram.
However if I create a new iOX Project (with Xcode 16.2) look at how sparse the Editor menu is. Why is that, why do they appear for other projects but not for a new one and why are the contents different?
If I create an Xcode (version doesn't matter, 16.N )project of type framework then install some dependencies using SPM.
Then within the Frameworks and Libraries section, the Embed part is empty.
This doesn't happen if the project type is app rather than framework.
If I want to set this to embed or not embed then how can this be done if its not even visible, for that matter how can I tell what it is set to even?
I've got an iOS app with lots of extensions, some of them complex and doing a lot of stuff.
After a bug I'd like to be able to use OSLogStore to get a holistic picture of logging for the app and its extensions and send that to a debugging server to retrospectively view logs for the app and its extensions.
The constructor is OSLogStore.init(scope: OSLogStore.Scope), however scope only has one value .currentProcessIdentifier.
Implying if that is called from within the app it can only get access to logging for its process only. I tried it out to confirm this is the case - if I log something in an extension (using Logger), then run the app with code like this:
let logStore = try! OSLogStore(scope: .currentProcessIdentifier)
let oneHourAgo = logStore.position(date: Date().addingTimeInterval(-3600))
let allEntries = try! logStore.getEntries(at: oneHourAgo)
for entry in allEntries {
look at the content of the entry
Then none of the entries are from the extension.
Is there anyway from within the app I can access logging made within an extension?
Within Xcode's settings location section is a drop down menu to switch between setting the derived data location to be default, relative or custom.
However its a global setting.
I work on more than one project simultaneously, and for one of them I want the location set to relative, but default for all the others.
Is there any way of achieving that?
If an iOS app gets terminated by watchdog due to, for example, hanging the main thread, is that considered to be a crash or something different.
I'm asking because, according to google and AI, Crashlytics can detect and report these,but in my experience it does not.
If I deliberately cause a watchdog termination by for example sleeping on the main thread for a long time then these never appear in Crashlytics.
I know Apple folks here don't comment on non Apple software, so I'm not asking about Crashlytics, just wondering and interested about watchdog timeout terminations and how they differ from a "regular" crash.