DisplayRepresentation.Image(systemName:tintColor:symbolConfiguration:) no longer applies the provided tintColor reliably since iOS 18. Observed behavior by OS version: iOS 17: SF Symbol tint is applied consistently as expected. iOS 18: SF Symbol tint is inconsistent and sometimes appears with incorrect or seemingly random colors instead of the provided tintColor. iOS 26: SF Symbol is rendered without any tint (default monochrome), completely ignoring the provided tintColor. This appears to be a regression in how App Intents renders DisplayRepresentation.Image with tinting across OS versions. iOS17.5: iOS 18.6: iOS26: Code: import AppIntents import UIKit struct CategoryEntity: AppEntity, Hashable { var id: Category.ID var name: String var icon: Int? var color: Int? var parentCategoryName: String? init(from category: Category) { self.id = category.id self.name = category.name self.icon = category.icon self.color = category.parent?.color ?? category.color self.parentCategoryName = category.parent?.name } var displayRepresentation: DisplayRepresentation { DisplayRepresentation( title: "\(name)", subtitle: parentCategoryName.map { "\($0)" }, image: .init( systemName: Icon.sfSymbolName(from: icon), tintColor: ColorTag.from(color) ) ) } static let typeDisplayRepresentation: TypeDisplayRepresentation = "Category" static let defaultQuery = CategoryQuery() } [Documentation API] (https://developer.apple.com/documentation/appintents/displayrepresentation/image-swift.struct/init(systemname:tintcolor:symbolconfiguration:)-3snvy?changes=_5)

Hi everyone, I believe this should be a simple and expected default behavior in a real-world app, but I’m unable to make it work: I have a View (a screen/page in this case) that calls an endpoint using async/await. If the endpoint hasn’t finished, but I navigate forward to a DetailView, I want the endpoint to continue fetching data (i.e., inside the @StateObject ViewModel that the View owns). This way, when I go back, the View will have refreshed with the fetched data once it completes. If the endpoint hasn’t finished and I navigate back to the previous screen, I want it to be canceled, and the @StateObject ViewModel should be deinitialized. I can achieve 1 and 3 using the .task modifier, since it automatically cancels the asynchronous task when the view disappears: view .task { await vm.getData() } I can achieve 1 and 2 using a structured Task in the View (or in the ViewModel, its the same behavior), for example: .onFirstAppearOnly { Task { away vm.getData() } } onFirstAppearOnly is a custom modifier that I have for calling onAppear only once in view lifecycle. Just to clarify, I dont think that part is important for the purpose of the example My question is: How can I achieve all three behaviors? My minimum deployment target is iOS 15, and I’m using NavigationView + NavigationLink. However, I have also tried using NavigationStack + NavigationPath and still couldn’t get it to work. Any help would be much appreciated. Thank you!

Hello, I am experiencing an issue related to SFSafariViewController during App Store review, and I would like to ask whether this could be an OS-level behavior or a known issue. App context Our app supports iPhone only During App Review, the reviewer tested the app on an iPad The app was rejected because the keyboard did not appear Issue description The problematic screen presents a sign-up web page using SFSafariViewController When tapping the nickname input field: The keyboard area appears (the layout shifts as if the keyboard will be shown) However, the actual keyboard does not appear Reproducibility We tested this issue on our own physical devices and on iPad simulators with the same app configuration However, we are unable to reproduce the issue in either environment What we have checked The web page uses a standard HTML There is no custom keyboard handling or keyboard-related logic in the app The app simply presents SFSafariViewController with a URL Locale-related question The URL includes a locale parameter (e.g. LOCALE=US) Could the app’s locale, system language, or region settings affect keyboard behavior in SFSafariViewController? Is it possible that a specific locale or language configuration on iPad could cause the keyboard to fail to appear? Questions Is there any known issue or OS-level behavior where the keyboard may fail to appear in SFSafariViewController on iPad? Could this be related to: Locale / language / region settings iPad multitasking or keyboard modes External keyboard state Specific iPadOS versions Are there any recommended workarounds or defensive implementations for this scenario? Any insights or similar experiences would be greatly appreciated. Thank you.

I’m exploring macOS development, comparing Mac Catalyst apps vs native AppKit/SwiftUI apps. What are the main limitations of Catalyst today? In what scenarios is a native AppKit or SwiftUI app unavoidable? Any insights are much appreciated — I’m trying to understand when Catalyst is sufficient and when going native is worth the extra effort.

I have a program that installs an NSStatusItem button in the Menubar. It registers for both the right and left click events. Before Tahoe, clicking into the statusitem with the mouse at the very top of the menubar would also invoke the Action handler, be it a right or a left click. Now, with Tahoe, where the menubar height was changed from 24 to 30, left clicks at the top edge still trigger the Action event, but right clicks do not - they only work if moving the mouse a few pixels down. This is quite inconvenient - the idea of the menubar being at the edge of the screen was always that one can just push the mouse to the edge and then click to interact. Tahoe has broken this basic GUI concept now partially. And since left clicks still work, it suggests that I don't do anything wrong - the clicks are controlled by the framework, and I'm never asked to provide a frame in which the clicks shall be accepted. It's that Apple's code apparently now uses the smaller statusitem's bounds instead of the full menu bar height for hit-testing. Is that on purpose or a bug? And does someone know a work-around, i.e. can I hook into the click-handler at a deeper level and do the hit testing myself? I've tried adding a subview to the statusitem's window's contentView, but that also only reacts to clicks in the smaller area.

Anytime I launch a view that contains a textfield, as soon as it’s in focus the view dismisses, and I log this warning, I have tried everything I could think of and still no solution, anyone know a workaround to this? note: this happens both on simulator and physical devices -[rtiinputsystemclient remotetextinputsessionwithid:performinputoperation:] perform input operation requires a valid sessionid. input modality = keyboard, input operation = , custom infotype = uiemojisearchoperations

Hello Apple Developer Forum Community, I’ve got a problem with the display of my SwiftUI View, that is tested on my physical iPhone. It’s shown very small (Picture) and on the Xcode Canvas Simulator it get’s shown right. What is the problem with my code?

Is there any reason why the SwiftUI Image hasn’t a direct init(data: Data) like UIImage from UIKit? In my opinion it’s very unintuitive and expensive to create a UIImage in the first step to create a SwiftUI Image with Image(uiImage: UIImage) in the second step. In addition to that, this causes unnecessary UIKit imports. In my opinion this is a very obvious small in the API, so are there any reasons why it is what it is? Best regards

Are we allowed to use AI-generated images for our Swift Student Challenge Project? The instructions say that we can use third party content but must cite/credit it and be transparent about it. Are we allowed to do the same for AI-generated images? Will this dock points on creativity?

I have an iOS application that derives on UITextInput to enter text. I have also overridden pressesBegan() and pressesEnded() in order to have some extra keyboard management (auto-repeat, special actions for arrow keys, function keys...) That works well for single-character languages (most roman languages, such as English, French, etc). For multi-character languages (Chinese, Japanses, Korean, Hindi), I can detect that the keyboard has been set to that language, and switch back to the default version of pressesBegan(): if let keyboardLanguage = self.textInputMode?.primaryLanguage { if (keyboardLanguage.hasPrefix("hi") || keyboardLanguage.hasPrefix("zh") || keyboardLanguage.hasPrefix("ja") || keyboardLanguage.hasPrefix("ko")) { super.pressesBegan(presses, with: event) } } But that strategy fails with the Tiếng Việt Telex keyboard (for Vietnamese language input). The way that keyboard works (as you can see if you open a document in Pages) is that you type as you go: T-i-e-n-g V-i-e-t T-e-l-e-x and the system adds the relevant diacritics once you've finished a word, so typing "Tieng Viet Telex" gives you "Tiếng Việt Telex" on the screen. Is there any documentation on the inner workings of this specific keyboard? What should I do (or not do) in order to make my application compatible with Tiếng Việt Telex?

I’m developing an iOS app that integrates with CallKit. Starting from iOS 26, I’ve noticed that the system automatically presents a top banner / toast-style UI when a CallKit call becomes active (see attached screenshot). This UI appears to be fully managed by the system. On iOS versions prior to iOS 26, this UI did not appear under the same CallKit configuration. What I’ve observed The banner is displayed automatically by the system It appears at the top of the screen, similar to a toast or call status banner It is not a view created by my app I could not find any public API or CallKit configuration related to dismissing or controlling it My questions： Is this top banner an intended system behavior change in newer iOS versions? Is there any public API to dismiss, hide, or customize this UI? If not, is this UI considered non-dismissible by design? Any clarification on the expected behavior or recommended approach would be greatly appreciated.

Has anyone noticed that there are way too many bugs caused by the liquid glass issues? I'm noticing apple adding private layer classes between existing UI views to add glassy effect and no way to disable them. Some of these bugs have no way around without rewriting the code from. For example, search controller doesn't work anymore on iOS 26 but does work on iOS 18. The tab bar controller doesn't show at all on iOS 26 but does on iOS 18. I'm wondering if other people are facing the same issues.

My #1 crash report since iOS 26 is the following: 1 libobjc.A.dylib objc_exception_throw 2 Foundation -[NSAssertionHandler handleFailureInMethod:object:file:lineNumber:description:] 3 UIKitCore -[UINavigationBar layoutSubviews] 4 UIKitCore UIView._layoutSubviewsWithObservationTracking() 5 UIKitCore @objc UIView._layoutSubviewsWithObservationTracking() Thread 0 9 libobjc.A.dylib objc_exception_throw 10 Foundation -[NSAssertionHandler handleFailureInMethod:object:file:lineNumber:description:] 11 UIKitCore -[UINavigationBar layoutSubviews] 12 UIKitCore UIView._layoutSubviewsWithObservationTracking() 13 UIKitCore @objc UIView._layoutSubviewsWithObservationTracking() I tried many things, without success so far. It seemed related to the new view update with observation tracking: link to documentation Any lead on how I should investigate this? Many thanks in advance!!

Hello everybody. I need your help with configuring navigationItem properly. I want to use its appearance properties to configure how the navigation bar looks on per-screen basis without interacting with UINavigationBar directly (the API has been implemented for this specific use case, right?) What I'm confused about is that setting any appearance property resets back button image to the default chevron. What I do now: Inside my app delegate let navigationBarAppearance = UINavigationBar.appearance() navigationBarAppearance.backIndicatorImage = UIImage() navigationBarAppearance.backIndicatorTransitionMaskImage = UIImage() I do this to hide the default chevron to customize back button appearance on per-view-controller basis Then in my presenting view controller: override func viewDidLoad() { super.viewDidLoad() navigationItem.title = "Root" navigationItem.backBarButtonItem = UIBarButtonItem( title: nil, image: nil, primaryAction: UIAction( title: "", image: UIImage(systemName: "square.and.arrow.up"), identifier: nil, discoverabilityTitle: nil, attributes: [], state: .on, handler: { [weak self] _ in self?.navigationController?.popViewController(animated: true) } ), menu: nil ) let appearance = UINavigationBarAppearance() appearance.configureWithTransparentBackground() appearance.backgroundColor = UIColor.red appearance.titleTextAttributes = [ .foregroundColor: UIColor.purple ] appearance.largeTitleTextAttributes = [ .foregroundColor: UIColor.purple ] navigationItem.standardAppearance = appearance navigationItem.compactAppearance = appearance navigationItem.scrollEdgeAppearance = appearance navigationItem.compactScrollEdgeAppearance = appearance } The colors are set up as expected, however the default chevron image appears on the next view controller next to square.and.arrow.up. If I don't touch navigation item appearance properties - only square.and.arrow.up is displayed, as I want, but obviously I loose all the customization. Can anyone please explain how to set things up properly? Thanks!

Hi, I have created an AppIntent in which there is a parameter called price, I have set the default value as 0. @Parameter(title: "Price", default: 0) var price: Int Problem When the shortcut is run this parameter is skipped Aim I still want to price to be asked however it needs to be pre-filled with 0 Question What should I do that the shortcut can still ask the price but be pre-filled with 0?

Overview I have a custom type Statistics that has 3 properties inside it I am trying to return this as part of the AppIntent's perforrm method struct Statistics { var countA: Int var countB: Int var countC: Int } I would like to implement the AppIntent to return Statistics as follows: func perform() async throws -> some IntentResult & ReturnsValue<Statistics> { ... ... } Problem It doesn't make much sense to make Statistics as an AppEntity as this is only computed as a result. Statistics doesn't exist as a persisted entity in the app. Questions How can I implement Statistics? Does it have to be AppEntity (I am trying to avoid this)? (defaultQuery would never be used.) What is the correct way tackle this?

Hello, the Liquid Glass effect of iOS 26 was possible during development, but it disappeared after its release and launch. I haven't found the reason for this, could you please help me clarify

I am building a macOS utility using SwiftUI and Swift that records and displays keyboard shortcuts (like Cmd+C, Cmd+V) in the UI. To achieve this, I am using NSEvent.addGlobalMonitorForEvents(matching: [.keyDown]). I am aware that global monitoring usually requires the app to be non-sandboxed. However, I am seeing some behavior I don't quite understand during development: I started with a fresh SwiftUI project and disabled the App Sandbox. I requested Accessibility permissions using AXIsProcessTrustedWithOptions, manually enabled it in System Settings, and the global monitor worked perfectly. I then re-enabled the App Sandbox in "Signing & Capabilities." To my surprise, the app still records global events from other applications, even though the Sandbox is now active. Is this expected behavior? Does macOS "remember" the trust because the Bundle ID was previously authorized while non-sandboxed, or is there a specific reason a Sandboxed app can still use addGlobalMonitor if the user has manually granted Accessibility access? My app's core feature is displaying these shortcuts for the user's own reference (productivity tracking). If the user is the one explicitly granting permission via the Accessibility privacy pane, will Apple still reject the app for using global event monitors within a Sandboxed environment? Code snippet of my monitor: // This is still firing even after re-enabling Sandbox eventMonitor = NSEvent.addGlobalMonitorForEvents(matching: [.keyDown]) { event in print("Captured: \(event.charactersIgnoringModifiers ?? "")") } I've tried cleaning the build folder and restarting the app, removing the app from accessibility permission, but the events keep coming through. I want to make sure I'm not relying on a "development glitch" before I commit to the App Store path. Here is the full code anyone can use to try this :- import SwiftUI import Cocoa import Combine struct ShortcutEvent: Identifiable { let id = UUID() let displayString: String let timestamp: Date } class KeyboardManager: ObservableObject { @Published var isCapturing = false @Published var capturedShortcuts: [ShortcutEvent] = [] private var eventMonitor: Any? // 1. Check & Request Permissions func checkAccessibilityPermissions() -> Bool { let options: NSDictionary = [kAXTrustedCheckOptionPrompt.takeUnretainedValue() as String: true] let accessEnabled = AXIsProcessTrustedWithOptions(options) return accessEnabled } // 2. Start Capture func startCapture() { guard checkAccessibilityPermissions() else { print("Permission denied") return } isCapturing = true let mask: NSEvent.EventTypeMask = [.keyDown, .keyUp] eventMonitor = NSEvent.addGlobalMonitorForEvents(matching: mask) { [weak self] event in self?.processEvent(event) } } // 3. Stop Capture func stopCapture() { if let monitor = eventMonitor { NSEvent.removeMonitor(monitor) eventMonitor = nil } isCapturing = false } private func processEvent(_ event: NSEvent) { // Only log keyDown to avoid double-counting the UI display guard event.type == .keyDown else { return } var modifiers: [String] = [] var symbols: [String] = [] // Map symbols for the UI if event.modifierFlags.contains(.command) { modifiers.append("command") symbols.append("⌘") } if event.modifierFlags.contains(.shift) { modifiers.append("shift") symbols.append("⇧") } if event.modifierFlags.contains(.option) { modifiers.append("option") symbols.append("⌥") } if event.modifierFlags.contains(.control) { modifiers.append("control") symbols.append("⌃") } let key = event.charactersIgnoringModifiers?.uppercased() ?? "" // Only display if a modifier is active (to capture "shortcuts" vs regular typing) if !symbols.isEmpty && !key.isEmpty { let shortcutString = "\(symbols.joined(separator: " ")) + \(key)" DispatchQueue.main.async { // Insert at the top so the newest shortcut is visible self.capturedShortcuts.insert(ShortcutEvent(displayString: shortcutString, timestamp: Date()), at: 0) } } } } PS :- I just did another test by creating a fresh new project with the default App Sandbox enabled, and tried and there also it worked!! Can I consider this a go to for MacOs app store than?

I am porting my app to SwiftUI and I am hitting a wall when using ScrollView. In my application, I have nested scrollViews to represent a scheduler. outer vertical scroll view inner horizontal scroll view that allows to horizontally scroll multiple columns in the scheduler each column in the inner scroll view is a view that needs to allow for a drag to initiate the creation of a new appointment on macOS, I do a mouse-down drag, so it does not affect the scroll view and works fine on iOS, if I add a drag gesture to the column, it short circuits the scroll view and scrolling becomes disabled. To initiate the drag, there is a long-press, and that gesture is fine, only the subsequent drag gesture is problematic. I have attached URL to a test app. The UI allows you to toggle the drag gesture. Hopefully, someone can help to get it to work since I would eventually like to port the macOS target to Catalyst. Download Test App

I’m trying to understand how SwiftUI List handles row lifecycle and reuse during scrolling. I have a list with around 60 card views; on initial load, only about 7 rows are created, but after scrolling to the bottom all rows appear to be created, and when scrolling back to the top I again observe multiple updates and apparent re-creation of rows. I confirmed this behavior using Instruments by profiling my app. Even though each row has a stable identifier, the row views still seem to be destroyed and recreated, which doesn’t resemble UIKit’s cell reuse model. I’d like clarity on how List uses identifiers internally, what actually gets reused versus recreated, and how developers should reason about performance and view lifetime in this case.