I've been testing my open source libraries with Swift 6.2 and the new Default Actor Isolation concurrency build setting set to MainActor (with Complete strict concurrency turned on). My library Destinations uses protocols extensively, often applying conformance to foundational Swift protocols like Hashable and Identifiable. Many of these basic protocols are not flagged as running on the @MainActor in Beta 1, leading to situations like this: Given this example code: public protocol Contentable: Identifiable { var id: UUID { get } } final class ContentModel: Contentable { let id: UUID = UUID() } I get the warning: Multiline Conformance of 'ContentModel' to protocol 'Contentable' crosses into main actor-isolated code and can cause data races; this is an error in the Swift 6 language mode The fix it suggests is to put a @MainActor before the Contentable protocol declaration in ContentModel, which seems to be a new attribute configuration in Swift 6.2. This solves the warning, but would create a lot of extra noise across the codebase. Was it an oversight or a temporary omission that protocols like Hashable and Identifiable do not run on @MainActor by default, or is there some other reason they are excluded? Considering how often protocols in our code may conform to foundational protocols like this, it seems at odds to the MainActor mode of the Default Actor Isolation setting given that it was created to make concurrency easier and less boilerplate to implement.

Hi, I’m trying to use the new InlineArray type, but noticed that it is unfortunately only available on macOS 26 and not on macOS 15 and others. As this is quite an essential type, I was wondering if this is intended or will this change in later beta’s? Not having it available on older Darwin platforms would severily limit it’s usage in the coming years. Thanks!

After switching our iOS app project from Swift 5 to Swift 6 and publishing an update, we started seeing a large number of crashes in Firebase Crashlytics. The crashes are triggered by NotificationCenter methods (post, addObserver, removeObserver) and show the following error: BUG IN CLIENT OF LIBDISPATCH: Assertion failed: Block was expected to execute on queue [com.apple.main-thread (0x1f9dc1580)] All scopes to related calls are already explicitly marked with @MainActor. This issue never occurred with Swift 5, but appeared immediately after moving to Swift 6. Has anyone else encountered this problem? Is there a known solution or workaround? Thanks in advance!

According to the doc: The value returned is the same as the value returned in the kEventParamKeyCode when using Carbon Events. So where can I find kEventParamKeyCode?

I found a similar problem here https://developer.apple.com/forums/thread/764777 and I could solve my problem by wrapping the call to requestAutomaticPassPresentationSuppression in a call to DispatchQueue.global().async. But my question is if this is really how things should work. Even with strict concurrency warnings in Swift 6 I don't get any warnings. Just a runtime crash. How are we supposed to find these problems? Couldn't the compiler assist with a warning/error. Why does the compiler make the assumptions it does about the method that is declared like this: @available(iOS 9.0, *) open class func requestAutomaticPassPresentationSuppression(responseHandler: @escaping (PKAutomaticPassPresentationSuppressionResult) -> Void) -> PKSuppressionRequestToken Now that we have migrated to Swift 6 our code base contains a bunch of unknown places where it will crash as above.

Hello, I have a json array showing in Xcode debugger (from the line "print(dataString)"): Optional("[{\"id\":\"8e8tfdcssu4u2hn7a71tkveahjhn8xghqcfkwf1bzvtrw5nu0b89w\",\"name\":\"Ameliana\",\"country\":\"France\",\"type\":\"Private\\/Corporate\",\"overallrecsit\":\"Positive\",\"dlastupd\":\"1741351633\",\"doverallrecsit\":\"1546848000\"},{\"id\":\"z69718a1a5z2y5czkwrhr1u37h7h768v05qr3pf1fegh4r4yrt5a68\",\"name\":\"Timberland\",\"country\":\"Switzerland\",\"type\":\"Charter\",\"overallrecsit\":\"Negative\",\"dlastupd\":\"1741351615\",\"doverallrecsit\":\"1740434582\"}, But my JSON decoder is throwing the catch error "Error in JSON parsing" This is the code: super.viewDidLoad() let urlString = "https://www.pilotjobsnetwork.com/service_ios.php" let url = URL(string: urlString) guard url != nil else { return } let session = URLSession.shared let dataTask = session.dataTask(with: url!) { (data, response, error) in var dataString = String(data: data!, encoding: String.Encoding.utf8) print(dataString) if error == nil && data != nil { // Parse JSON let decoder = JSONDecoder() do { let newsFeed = try decoder.decode(NewsFeed.self, from: data!) print(newsFeed) print(error) } catch{ print("Error in JSON parsing") } } } // Make the API Call dataTask.resume() } And this is my Codable file NewsFeed.swift: struct NewsFeed: Codable { var id: String var name: String var country: String var type: String var overallrecsit: String var dlastupd: String var doverallrecsit: String } Please do you know why the parsing may be failing? Is it significant that in the debugging window the JSON is displaying backslashes before the quotation marks? Thank you for any pointers :-)

Using the DebugDescription macro to display an optional value produces a “String interpolation produces a debug description for an optional value” build warning. For example: @DebugDescription struct MyType: CustomDebugStringConvertible { let optionalValue: String? public var debugDescription: String { "Value: \(optionalValue)" } } The DebugDescription macro does not allow (it is an error) "Value: \(String(describing: optionalValue))" or "Value: \(optionalValue ?? "nil")" because “Only references to stored properties are allowed.” Is there a way to reconcile these? I have a build log full of these warnings, obscuring real issues.

I was trying to evaulate let myTuple = ("blue", false) let otherTuple = ("blue", true) if myTuple < otherTuple { print("yes it evaluates") } Ans I got /tmp/S9jAk7P7KW/main.swift:5:12: error: binary operator '<' cannot be applied to two '(String, Bool)' operands if myTuple < otherTuple { My question is why there is no compile time issue in first place where the declaration is let myTuple = ("blue", false) ~~~~~~ something like above

AsyncStream { continuation in Task { let response = await getResponse() continuation.yield(response) continuation.finish() } } In this WWDC video https://developer.apple.com/videos/play/wwdc2025/231/ at 8:20 the presenter mentions that if the "Task gets cancelled, the Task inside the function will automatically get cancelled too". The documentation does not mention anything like this. From my own testing on iOS 18.5, this is not true.

After swapping the -objectAtIndex: method using method_exchangeImplementations, it will cause continuous memory growth. Connect the iPhone and run the provided project. Continuously tap the iPhone screen. Observe Memory; it will keep growing. Sample code

Is there any way to retrieve the memory pressure percentage using native libraries? When I run the memory-pressure command, I can see the percentage of free memory, but I’d like to retrieve the same information using a native library.

NSPredicate(format: "SELF MATCHES %@", "^[0-9A-Z]+$").evaluate(with: "126𝒥ℰℬℬ𝒢𝒦𝒮33") Returns true, and I don't know why. 𝒥ℰℬℬ𝒢𝒦𝒮 is not between 0-9 and A-Z, and why it returns true? How to avoid similar problem like this when using NSPredicate?

I'm having trouble dealing with concurrency with the SFAuthorizationPluginView. Does anybody know how this can be solved? https://developer.apple.com/documentation/securityinterface/sfauthorizationpluginview The crux of it is: If I inherit an object as part of an API, and the API is preconcurrency, and thus is nonisolated (but in reality is @MainActor), how do I return a @MainActor GUI element? https://developer.apple.com/documentation/securityinterface/sfauthorizationpluginview/firstresponder() The longer story: I made my view class inherit SFAuthorizationPluginView. The API is preconcurrency (but not marked as preconcurrency) I started using concurrency in my plugin to retrieve data over XPC. (https://developer.apple.com/documentation/xpc/xpcsession + https://developer.apple.com/documentation/swift/withcheckedthrowingcontinuation(isolation:function:_:)) Once I retrieve the data over XPC, I need to post it on GUI, hence I've set my view class as @MainActor in order to do the thread switch. Swift compiler keeps complaining: override func firstResponder() -> NSResponder? { return usernameField } "Main actor-isolated property 'usernameField' can not be referenced from a nonisolated context; this is an error in the Swift 6 language mode" override func firstResponder() -> NSResponder? { MainActor.assumeIsolated { return usernameField } } "Sending 'self' risks causing data races; this is an error in the Swift 6 language mode" I think fundamentally, the API is forcing me to give away a @MainActor variable through a nonisolated function, and there is no way to shut up the compiler. I've tried @preconcurrency and it has no effect as far as I can tell. I've also tried marking the function explicitly as nonisolated. The rest of the API are less problematic, but returning a GUI variable is exceptionally difficult.

Consider this simple miniature of my iOS Share Extension: import SwiftUI import Photos class ShareViewController: UIViewController { override func viewDidLoad() { super.viewDidLoad() if let itemProviders = (extensionContext?.inputItems.first as? NSExtensionItem)?.attachments { let hostingView = UIHostingController(rootView: ShareView(extensionContext: extensionContext, itemProviders: itemProviders)) hostingView.view.frame = view.frame view.addSubview(hostingView.view) } } } struct ShareView: View { var extensionContext: NSExtensionContext? var itemProviders: [NSItemProvider] var body: some View { VStack{} .task{ await extractItems() } } func extractItems() async { guard let itemProvider = itemProviders.first else { return } guard itemProvider.hasItemConformingToTypeIdentifier(UTType.url.identifier) else { return } do { guard let url = try await itemProvider.loadItem(forTypeIdentifier: UTType.url.identifier) as? URL else { return } try await downloadAndSaveMedia(reelURL: url.absoluteString) extensionContext?.completeRequest(returningItems: []) } catch {} } } On the line 34 guard let url = try await itemProvider.loadItem ... I get these warnings: Passing argument of non-sendable type '[AnyHashable : Any]?' outside of main actor-isolated context may introduce data races; this is an error in the Swift 6 language mode 1.1. Generic enum 'Optional' does not conform to the 'Sendable' protocol (Swift.Optional) Passing argument of non-sendable type 'NSItemProvider' outside of main actor-isolated context may introduce data races; this is an error in the Swift 6 language mode 2.2. Class 'NSItemProvider' does not conform to the 'Sendable' protocol (Foundation.NSItemProvider) How to fix them in Xcode 16? Please provide a solution which works, and not the one which might (meaning you run the same code in Xcode, add your solution and see no warnings). I tried Decorating everything with @MainActors Using @MainActor in the .task @preconcurrency import Decorating everything with @preconcurrency Playing around with nonisolated

Given the below code with Swift 6 language mode, Xcode 16.2 If running with iOS 18+: the app crashes due to _dispatch_assert_queue_fail If running with iOS 17 and below: there is a warning: warning: data race detected: @MainActor function at Swift6Playground/PublishedValuesView.swift:12 was not called on the main thread Could anyone please help explain what's wrong here? import SwiftUI import Combine @MainActor class PublishedValuesViewModel: ObservableObject { @Published var count = 0 @Published var content: String = "NA" private var cancellables: Set<AnyCancellable> = [] func start() async { let publisher = $count .map { String(describing: $0) } .removeDuplicates() for await value in publisher.values { content = value } } } struct PublishedValuesView: View { @ObservedObject var viewModel: PublishedValuesViewModel var body: some View { Text("Published Values: \(viewModel.content)") .task { await viewModel.start() } } }

Title Why doesn’t this async function see external changes to an inout Bool in Release builds (but works in Debug)? Body I have a small helper function that waits for a Bool flag to become true with a timeout: public func test(binding value: inout Bool, timeout maximum: Int) async throws { var count = 0 while value == false { count += 1 try await Task.sleep(nanoseconds: 0_100_000_000) if value == true { return } if count > (maximum * 10) { return } } } I call like this: var isVPNConnected = false adapter.start(tunnelConfiguration: tunnelConfiguration) { [weak self] adapterError in guard let self = self else { return } if let adapterError = adapterError { } else { isVPNConnected = true } completionHandler(adapterError) } try await waitUntilTrue(binding: &isVPNConnected, timeout: 10) What I expect: test should keep looping until flag becomes true (or the timeout is hit). When the second task sets flag = true, the first task should see that change and return. What actually happens: In Debug builds this behaves as expected: when the second task sets flag = true, the loop inside test eventually exits. In Release builds the function often never sees the change and gets stuck until the timeout (or forever, depending on the code). It looks like the while value == false condition is using some cached value and never observes the external write. So my questions are: Is the compiler allowed to assume that value (the inout Bool) does not change inside the loop, even though there are await suspension points and another task is mutating the same variable? Is this behavior officially “undefined” because I’m sharing a plain Bool across tasks without any synchronization (actors / locks / atomics), so the debug build just happens to work? What is the correct / idiomatic way in Swift concurrency to implement this kind of “wait until flag becomes true with timeout” pattern? Should I avoid inout here completely and use some other primitive (e.g. AsyncStream, CheckedContinuation, Actor, ManagedAtomic, etc.)? Is there any way to force the compiler to re-read the Bool from memory each iteration, or is that the wrong way to think about it? Environment (if it matters): Swift: [fill in your Swift version] Xcode: [fill in your Xcode version] Target: iOS / macOS [fill in as needed] Optimization: default Debug vs. Release settings I’d like to understand why Debug vs Release behaves differently here, and what the recommended design is for this kind of async waiting logic in Swift.

Hello, While watching WWDC25: Code-along: Elevate an app with Swift concurrency at timestamp 25:48, I noticed something in the slide/diagram that might be incorrect. The diagram shows ExtractSticker twice, but based on the code context and spoken explanation, I think it was meant to be ExtractSticker and ExtractColor. Reasoning: The surrounding code and narration describe the use of async let and a Sendable Data object. From the flow, one task extracts a sticker while the other extracts a color, so it seems like the diagram is inconsistent. I do understand that with @concurrent, having two ExtractSticker operations on the same Data is technically possible (with two concurrent process executing their respective ExtractSticker) — but that would be a different meaning than what the talk was describing. Since concurrency is already a subtle and error-prone topic, I thought it was worth pointing this out. If I’m mistaken, I’d love clarification. Otherwise, this could be a small correction to keep things aligned and clearer for everyone. Minor point overall, but Swift 6’s concurrency model is doing a fantastic job at helping us write safer code—so thank you to the team for that! (Attaching screenshots for reference)

I'm trying to use FormatStyle from Foundation to format numbers when printing a vector structure. See code below. import Foundation struct Vector<T> { var values: [T] subscript(item: Int) -> T { get { values[item] } set { values[item] = newValue } } } extension Vector: CustomStringConvertible { var description: String { var desc = "( " desc += values.map { "\($0)" }.joined(separator: " ") desc += " )" return desc } } extension Vector { func formatted<F: FormatStyle>(_ style: F) -> String where F.FormatInput == T, F.FormatOutput == String { var desc = "( " desc += values.map { style.format($0) }.joined(separator: " ") desc += " )" return desc } } In the example below, the vector contains a mix of integer and float literals. The result is a vector with a type of Vector<Double>. Since the values of the vector are inferred as Double then I expect the print output to display as decimal numbers. However, the .number formatted output seems to ignore the vector type and print the values as a mix of integers and decimals. This is fixed by explicitly providing a format style with a fraction length. So why is the .formatted(.number) method ignoring the vector type T which is Double in this example? let vec = Vector(values: [-2, 5.5, 100, 19, 4, 8.37]) print(vec) print(vec.formatted(.number)) print(vec.formatted(.number.precision(.fractionLength(1...)))) ( -2.0 5.5 100.0 19.0 4.0 8.37 ) // correct output that uses all Double types ( -2 5.5 100 19 4 8.37 ) // wrong output that uses Int and Double types ( -2.0 5.5 100.0 19.0 4.0 8.37 ) // correct output that uses all Double types

Hello, Please can you tell me how to create an array of dictionaries? This code below should create 4 dictionaries in an array, but I'm getting these errors: For line "var output = [id: "testID", name: "testName"]": cannot find 'name' in scope Type '(any AnyObject).Type' cannot conform to 'Hashable' For line "return output": Type '(any AnyObject).Type' cannot conform to 'Hashable' var quotes: [(id: String, name: String)] { var output = [[(id: String, name: String)]] () for i in 1...4 { var output = [id: "testID", name: "testName"] } return output }

Hello, I was hoping to clarify my understanding of the use of for await with an AsyncStream. My use case is, I'd like to yield async closures to the stream's continuation, with the idea that, when I use for await with the stream to process and execute the closures, it would only continue on to the following closure once the current closure has been run to completion. At a high level, I am trying to implement in-order execution of async closures in the context of re-entrancy. An example of asynchronous work I want to execute is a network call that should write to a database: func syncWithRemote() async -> Void { let data = await fetchDataFromNetwork() await writeToLocalDatabase(data) } For the sake of example, I'll call the intended manager of closure submission SingleOperationRunner. where, at a use site such as this, my desired outcome is that call 1 of syncWithRemote() is always completed before call 2 of it: let singleOperationRunner = SingleOperationRunner(priority: nil) singleOperationRunner.run { syncWithRemote() } singleOperationRunner.run { syncWithRemote() } My sketch implementation looks like this: public final class SingleOperationRunner { private let continuation: AsyncStream<() async -> Void>.Continuation public init(priority: TaskPriority?) { let (stream, continuation) = AsyncStream.makeStream(of: (() async -> Void).self) self.continuation = continuation Task.detached(priority: priority) { // Will this loop only continue when the `await operation()` completes? for await operation in stream { await operation() } } } public func run(operation: @escaping () async -> Void) { continuation.yield(operation) } deinit { continuation.finish() } } The resources I've found are https://developer.apple.com/videos/play/wwdc2022-110351/?time=1445 and https://forums.swift.org/t/swift-async-func-to-run-sequentially/60939/2 but do not think I have fully put the pieces together, so would appreciate any help!