I’m running into an issue with TextField focus behavior in SwiftUI. By default, when I set focus to a TextField programmatically (using @FocusState), SwiftUI behaves like AppKit — the entire contents of the text field are selected. This is causing problems for my use case, because I want the caret placed at the end of the text without selecting everything. How I solved this in AppKit In AppKit, I worked around this by subclassing NSTextField and overriding becomeFirstResponder to adjust the editor’s selection: override func becomeFirstResponder() -> Bool { let responderStatus = super.becomeFirstResponder() // Ensure caret is placed at the end, no text selected if let editor = self.currentEditor() { let selectedRange = editor.selectedRange editor.selectedRange = NSRange(location: selectedRange.length, length: 0) } return responderStatus } This successfully prevented AppKit from auto-selecting the entire string when focus changed. The problem in SwiftUI Now I see the same auto-select behavior in SwiftUI when I toggle focus with @FocusState. But unlike AppKit, SwiftUI doesn’t expose the underlying NSTextView or UITextField APIs, so I can’t directly adjust the selection or caret position. Questions: Is there a way in SwiftUI to control the caret/selection behavior when a TextField becomes focused? Is there a built-in modifier or @FocusState trick I’m missing? Has anyone found a reliable SwiftUI-idiomatic approach to ensure the caret is placed at the end of the text instead of selecting all text? update: adding my swiftUI code below: struct TextFieldUI: View { @ObservedObject var pModel:TextFieldModel @FocusState private var pIsFocusedState: Bool var body: some View { VStack(spacing: 20) { TextField(pModel.placeholder, text: $pModel.text) .textFieldStyle(RoundedBorderTextFieldStyle()) .padding() .focused($pIsFocusedState) .onChange(of: pModel.isFocused) { old, newValue in pIsFocusedState = newValue } .onChange(of: pIsFocusedState) { old, newValue in pModel.isFocused = newValue } .onAppear { pIsFocusedState = pModel.isFocused } Toggle("Secure Mode", isOn: $pModel.isSecure) .padding() } .padding() } }

In iOS 18 the following code works to set a state variable when you hold your finger on the Text() field (well, the ScrollView()), but it doesn't work in iOS 26: @State private var pressed: Bool = false ... ScrollView { VStack { Text("Some text goes here") }.frame(maxWidth: .infinity) } .onTapGesture {} // This is required to allow the long press gesture to be recognised .gesture( DragGesture(minimumDistance: 0) .onChanged({ _ in pressed = true }) .onEnded({ _ in pressed = false }) ) .background(pressed ? .black.opacity(0.4) : .clear) I've tried changing this to: var dragGesture: some Gesture { DragGesture(minimumDistance: 0) .onChanged({ _ in self.pressed = true }) .onEnded({ _ in self.pressed = false }) } ... ScrollView { VStack { Text("Some text goes here") }.frame(maxWidth: .infinity) } .gesture(dragGesture) .background(pressed ? .black.opacity(0.4) : .clear) And this: var longPress: some Gesture { LongPressGesture(minimumDuration: 0.25) .onChanged({ _ in self.pressed = true }) .onEnded({ _ in self.pressed = false }) } ... ScrollView { VStack { Text("Some text goes here") }.frame(maxWidth: .infinity) } .gesture(longPress) .background(pressed ? .black.opacity(0.4) : .clear) Neither works. Any ideas? Thanks.

Hi everyone, I have the following issue that I have tried to tweak every possible modifier of ScrollView and still got the same result in iOS 26. Description: Create a SwiftUI ScrollView with scrollTargetBehavior of paging, also create a bottom UI view below the ScrollView. If the starting index is not 0, the position of current page will be off with part of previous page shown above it. It only happens on iOS 26, not on iOS 18. Also if bottom UI view (text view in this case) is removed, it also works fine. I want to see if there is a solution for it or it's an iOS 26 bug. Thanks! import SwiftUI struct ContentView: View { @State private var currentPageIndex: Int? = 3 var body: some View { VStack { scrollView Text("Bottom Bar") .frame(maxWidth: .infinity) .frame(height: 80) .background(.red) } .background(.black) } @ViewBuilder var scrollView: some View { VerticalPagerView( currentPageIndex: $currentPageIndex, itemCount: 10, content: Array(0...9).map { index in content(for: index) } ) } @ViewBuilder private func content(for index: Int) -> some View { // Empty view with random background color Color( red: Double((index * 25 + 0) % 255) / 255.0, green: Double((index * 25 + 80) % 255) / 255.0, blue: Double((index * 25 + 160) % 255) / 255.0 ) } } struct VerticalPagerView<Content: View>: View { @Binding private var currentPageIndex: Int? private let itemCount: Int private let content: [Content] init( currentPageIndex: Binding<Int?>, itemCount: Int, content: [Content] ) { self._currentPageIndex = currentPageIndex self.itemCount = itemCount self.content = content } var body: some View { GeometryReader { geometryReader in ScrollViewReader { reader in ScrollView(.vertical) { LazyVStack(spacing: 0) { ForEach(0 ..< itemCount, id: \.self) { index in content[index] .id(index) .containerRelativeFrame(.vertical, alignment: .center) .clipped() } } .frame(minHeight: geometryReader.size.height) .scrollTargetLayout() } .scrollIndicators(.hidden) .onAppear { guard let currentPageIndex = currentPageIndex else { return } reader.scrollTo(currentPageIndex, anchor: .center) } } .scrollPosition(id: $currentPageIndex, anchor: .center) .ignoresSafeArea() .scrollTargetBehavior(.paging) .onChange(of: currentPageIndex) { oldIndex, newIndex in } } } }

Hi community: I noticed that each closure is counted as lines in code coverage (unit tests) (Xcode 14.1.0) in a swiftUI File. I mean, If you coded and VStack that involves another HStack, and HStack contains 4 lines, and the VStack contains 6 lines counting the HStack. The total executable lines should be 6 (6 lines in the file). But Xcode count 10, counting twice the HStack lines. Is it a bug, or is it correct? You know, I don't know if Apple has another concept about executable lines. Also, Is it possible to remove previews with any configuration from code coverage or constant files? Thanks for all.

I am working on a SwiftUI project using Core Data. I have an entity called AppleUser in my data model, with the following attributes: id (UUID), name (String), email (String), password (String), and createdAt (Date). All attributes are non-optional. I created the corresponding Core Data class files (AppleUser+CoreDataClass.swift and AppleUser+CoreDataProperties.swift) using Xcode’s automatic generation. I also have a PersistenceController that initializes the NSPersistentContainer with the model name JobLinkModel. When I try to save a new AppleUser object using: let user = AppleUser(context: viewContext) user.id = UUID() user.name = "User1" user.email = "..." user.password = "password1" user.createdAt = Date()【The email is correctly formatted, but it has been replaced with “…” for privacy reasons】 try? viewContext.save() I get the following error in the console:Core Data save failed: Foundation._GenericObjCError.nilError, [:] User snapshot: ["id": ..., "name": "User1", "email": "...", "password": "...", "createdAt": ...] All fields have valid values, and the Core Data model seems correct. I have also tried: • Checking that the model name in NSPersistentContainer(name:) matches the .xcdatamodeld file (JobLinkModel) • Ensuring the AppleUser entity Class, Module, and Codegen are correctly set (Class Definition, Current Product Module) • Deleting duplicate or old AppleUser class files • Cleaning Xcode build folder and deleting the app from the simulator • Using @Environment(.managedObjectContext) for the context Despite all this, I still get _GenericObjCError.nilError when saving a new AppleUser object. I want to understand: 1. Why is Core Data failing to save even though all fields are non-nil and correctly assigned? 2. Could this be caused by some residual old class files, or is there something else in the setup that I am missing? 3. What steps should I take to ensure that Core Data properly recognizes the AppleUser entity and allows saving? Any help or guidance would be greatly appreciated.

My app hangs and I get this error when using an inspector view on macos tahoe. Changing to a different type of view resolves the issue. The window has been marked as needing another Update Constraints in Window pass, but it has already had more Update Constraints in Window passes than there are views in the window. <SwiftUI.AppKitWindow: 0x8897e0000> 0xb4d8 (46296) {{1004, 182}, {1254, 948}} en Future marking as needing Update Constraints in Window might be ignored. FAULT: NSGenericException: The window has been marked as needing another Update Constraints in Window pass, but it has already had more Update Constraints in Window passes than there are views in the window. <SwiftUI.AppKitWindow: 0x8897e0000> 0xb4d8 (46296) {{1004, 182}, {1254, 948}} en; (user info absent)

Reported as FB20357097 In SwiftUI, an empty .safeAreaInset modifier attached to a Map causes the map to become zoomed out to planet level. Minimal reproduction: import SwiftUI import MapKit @main struct map_region_safe_area_inset_bugApp: App { var body: some Scene { WindowGroup { Map { // Any Map content MapCircle(center: .init(latitude: 35.6895, longitude: 139.6917), radius: 1000) } .safeAreaInset(edge: .top) { // No content, `EmptyView()`, `Color.clear` } } } } Note: ZStack { } inside the safeAreaInset prevents the bug. Empty safeAreaInset (bug) Non-empty

So I have a perplexing situation. I'm loading multiple SwiftUI AsyncImages according to spec (see code below). For some reason, my 1MB app has over 400+ MBs of documents & data. When I view the app's container, I can see that it is caused by a massive number of the images as .tmp files in the "tmp" folder all starting with the name "CFNetworkDownload". It seems that every time an image is loaded, it's stored in here, but does not delete. This makes the app get bigger every time it's opened. What can I do about this issue? Thanks so much! (P.S. I've tried to condense my code down as much as possible for readability, but there's a few files included because I'm not sure where the problem lies.) Main app Swift file: @main struct MyApp: App { let monitor = NWPathMonitor() @State private var isConnected = true var body: some Scene { monitor.pathUpdateHandler = { path in if path.status == .satisfied { if !isConnected { isConnected.toggle() } } else { if isConnected { isConnected.toggle() } } } let queue = DispatchQueue(label: "Monitor") monitor.start(queue: queue) return WindowGroup { isConnected ? AnyView(ContentView()) : AnyView(ContentViewFailed()) } } } The ContentView that's loaded inside the above WindowGroup: struct ContentView: View { var body: some View { TabView { HomeView() .tabItem { Image(systemName: "house.fill") Text("Home") } . . . } } } And finally, the HomeView where the images are being loaded: struct HomeView: View { var body: some View { let urlString = "https://www.example.com/Home.json" if let url = URL(string: urlString) { if let data = try? Data(contentsOf: url) { do { items = try JSONDecoder().decode([Item].self, from: data) } catch { print(error) } } } return NavigationView { List { ScrollView { VStack(alignment: .leading) { ZStack { VStack(alignment: .leading) { Spacer() HStack { AsyncImage(url: URL(string: "https://www.example.com/images/example.png")) { image in image .resizable() .aspectRatio(contentMode: .fill) .shadow(color: Color(red: 0, green: 0, blue: 0, opacity: 0.25), radius: 1) } placeholder: { ProgressView() .progressViewStyle(.circular) } .frame(width: 202, height: 100) } . . . } } } } } } } } I really appreciate your time. Not sure if this could just be a bug with AsyncImage itself.

Hello! I have experienced a weird bug in iOS 26 Beta (8) and previous beta versions. The safe area inset is not correctly aligned with the keyboard toolbar on real devices and simulators. When you focus a new textfield the bottom safe area is correctly placed aligned the keyboard toolbar. On real devices the safe area inset view is covered slightly by the keyboard toolbar, which is even worse than on the simulator. Here's a clip from a simulator: Here's the code that reproduced the bug I experienced in our app. #Preview { NavigationStack { ScrollView { TextField("", text: .constant("")) .padding() .background(Color.secondary) TextField("", text: .constant("")) .padding() .background(Color.green) } .padding() .safeAreaInset(edge: .bottom, content: { Color.red .frame(maxWidth: .infinity) .frame(height: 40) }) .toolbar { ToolbarItem(placement: .keyboard) { Button {} label: { Text("test") } } } } }

I’m trying to transform user keyboard input in a TextField so that, for example, whenever the user types the letter "a" it is stored and shown as the Greek letter "α". I created a custom Binding to intercept and modify the typed text before saving it to my observable model. Here’s a simplified version of my code: import SwiftUI class User: ObservableObject { @Published var username: String = "" } struct ContentView: View { @ObservedObject var user = User() var usernameBinding: Binding<String> { Binding( get: { user.username }, set: { newValue in // Replace all "a" with "α" user.username = newValue.replacingOccurrences(of: "a", with: "α") } ) } var body: some View { TextField("Username", text: usernameBinding) .padding() .onChange(of: user.username) { newValue in print("username changed to:", newValue) } } } When I type "a", I can see in the console that the onChange handler prints the transformed string ("α"), and the model (user.username) is updated. However, the TextField on screen still shows the original "a" instead of updating to "α" immediately. I expected the text field to update its displayed value whenever the bound property changes (since username is @Published on an ObservableObject), but that doesn’t seem to happen when I modify the text in the binding’s set closure. Is this a known limitation of SwiftUI TextField? Is there a better way to transform user input so the field shows the transformed text based on some processing? Any advice or explanation would be appreciated.

Basically, take just the Xcode 26 AR App template, where we put the ContentView as the detail end of a NavigationStack. Opening app, the app uses < 20MB of memory. Tapping on Open AR the memory usage goes up to ~700MB for the AR Scene. Tapping back, the memory stays up at ~700MB. Checking with Debug memory graph I can still see all the RealityKit classes in the memory, like ARView, ARRenderView, ARSessionManager. Here's the sample app to illustrate the issue. PS: To keep memory pressure on the system low, there should be a way of freeing all the memory the AR uses for apps that only occasionally show AR scenes.

Hi, I've just moved my SpriteKit-based game from UIView to SwiftUI + SpriteView and I'm getting this mesage Adding 'GCControllerView' as a subview of UIHostingController.view is not supported and may result in a broken view hierarchy. Add your view above UIHostingController.view in a common superview or insert it into your SwiftUI content in a UIViewRepresentable instead. Here's how I'm doing this struct ContentView: View { @State var alreadyStarted = false let initialScene = GKScene(fileNamed: "StartScene")!.rootNode as! SKScene var body: some View { ZStack { SpriteView(scene: initialScene, transition: .crossFade(withDuration: 1), isPaused: false , preferredFramesPerSecond: 60) .edgesIgnoringSafeArea(.all) .onAppear { if !self.alreadyStarted { self.alreadyStarted.toggle() initialScene.scaleMode = .aspectFit } } VirtualControllerView() .onAppear { let virtualController = BTTSUtilities.shared.makeVirtualController() BTTSSharedData.shared.virtualGameController = virtualController BTTSSharedData.shared.virtualGameController?.connect() } .onDisappear { BTTSSharedData.shared.virtualGameController?.disconnect() } } } } struct VirtualControllerView: UIViewRepresentable { func makeUIView(context: Context) -> UIView { let result = PassthroughView() return result } func updateUIView(_ uiView: UIView, context: Context) { } } class PassthroughView: UIView { override func hitTest(_ point: CGPoint, with event: UIEvent?) -> UIView? { for subview in subviews.reversed() { let convertedPoint = convert(point, to: subview) if let hitView = subview.hitTest(convertedPoint, with: event) { return hitView } } return nil } }

Prior to MacOS 26, Multiple Pickers could be laid out with a uniform width. For example: struct LayoutExample: View { let fruits = ["apple", "banana", "orange", "kiwi"] let veg = ["carrot", "cauliflower", "peas", "Floccinaucinihilipilification Cucurbitaceae"] @State private var selectedFruit: String = "kiwi" @State private var selectedVeg: String = "carrot" var body: some View { VStack(alignment: .leading) { Picker(selection: $selectedFruit) { ForEach(fruits, id: \.self, content: Text.init) } label: { Text("Fruity choice") .frame(width: 150, alignment: .trailing) } .frame(width: 300) Picker(selection: $selectedVeg) { ForEach(veg, id: \.self, content: Text.init) } label: { Text("Veg") .frame(width: 150, alignment: .trailing) } .frame(width: 300) } } } Renders like this, prior to MacOS26: But now looks like this under MacOS 26: Is there anyway to control the size of the picker selection in MacOS 26?

I don't think the .searchable in the bottom bar / glass look works when you have TabView that has NavStack.

When I dismiss a view presented with .navigationTransition(.zoom), the source view gets a weird background (black or white depending on the appearance) for a couple of seconds, and then it disappears. Here’s a simple code example. import SwiftUI struct NavigationTransition: View { @Namespace private var namespace @State private var isSecondViewPresented = false var body: some View { NavigationStack { ZStack { DetailView(namespace: namespace) .onTapGesture { isSecondViewPresented = true } } .fullScreenCover(isPresented: $isSecondViewPresented) { SecondView() .navigationTransition(.zoom(sourceID: "world", in: namespace)) } } } } struct DetailView: View { var namespace: Namespace.ID var body: some View { ZStack { Color.blue Text("Hello World!") .foregroundStyle(.white) .matchedTransitionSource(id: "world", in: namespace) } .ignoresSafeArea() } } struct SecondView: View { var body: some View { ZStack { Color.green Image(systemName: "globe") .foregroundStyle(Color.red) } .ignoresSafeArea() } } #Preview { NavigationTransition() }

Environment: — Xcode Version: [Current version] — iOS Version: [Target iOS version] — Swift Version: 6.0.3 — Platform: iOS Summary: Direct assignment of new arrays to @State properties sometimes fails to trigger UI updates, even when the arrays contain different data. The assignment appears to succeed in code but the UI continues to display stale data. Presumably when lengths of both arrays are the same. Assigning first empty array and then the new array fixed the issue. Expected Behavior: When assigning a new array to a @State property (self.stateArray = newArray), SwiftUI should detect the change and update the UI to reflect the new data. Actual Behavior: The assignment self.stateArray = newArray executes without error, but the UI continues to display data from the previous array. Debugging shows that self._stateArray (the underlying property wrapper) retains the old data despite the assignment. Minimal Reproduction Case: struct ContentView: View { @State private var items: [String] = ["Old Item"] var body: some View { VStack { ForEach(items, id: \.self) { item in Text(item) } Button("Update Items") { // This assignment may not trigger UI update self.items = ["New Item", "Another Item"] // Workaround: Clear first, then assign // self.items = [] // self.items = ["New Item", "Another Item"] } } } } Workaround: Force the state update by clearing the array before assignment: self.items = [] // Clear first self.items = newArray // Then assign new data Additional Context: — This issue was discovered in a production app where item data loaded from cache wasn't updating the UI. — The same data loading pattern worked in one view (which is modal and doesn't reload data) but failed in another (which needs to be refreshed). — Console logs showed fresh data was loaded but UI displayed stale data. — Debugger showed self._items instead of self.items, which might suggest property wrapper issues. Impact: This causes significant user experience issues where the UI doesn't reflect the actual application state, leading to confusion and apparent data inconsistency. Request: Please investigate why direct @State array assignment sometimes fails and consider fixing the underlying cause, or at minimum document this behavior and recommended workarounds.

On iOS 26 I’m seeing a small stutter when dismissing a SwiftUI .sheet with the swipe-down gesture. The same code was smooth on iOS 18. Has anyone else experienced this issue?

When using a contextMenu in SwiftUI to show a preview of a PHAsset’s full-size image via PHCachingImageManager.requestImage(), memory usage increases with each image preview interaction. The memory is not released, leading to eventual app crash due to memory exhaustion. The thumbnail loads and behaves as expected, but each call to fetch the full-size image (1000x1000) for the contextMenu preview does not release memory, even after cancelImageRequest() is called and fullSizePreviewImage is set to nil. The issue seems to stem from the contextMenu lifecycle behavior, it triggers .onAppear unexpectedly, and the full-size image is repeatedly fetched without releasing the previously loaded images. The question is, where do I request to the get the full-size image to show it in the context menu preview? STEPS TO REPRODUCE 1/ Create a SwiftUI LazyVGrid displaying many PHAsset thumbnails using PHCachingImageManager. 2/ Add a .contextMenu to each thumbnail button with: .onAppear that triggers requestImage() for a high-resolution preview. .onDisappear that calls cancelImageRequest() and sets the image @State to nil. 3/ Tap on several image previews 4/ Monitor memory usage as it increases and eventually crashes

I'm having a heck of a time getting this to work. I'm trying to add an event notification at the end of a timeline animation to trigger something in code but I'm not receiving the notification from RC Pro. I've watched that Compose Interactive 3D Content video quite a few times now and have tried many different ways. RC Pro has the correct ID names on the notifications. I'm not a programmer at all. Just a lowly 3D artist. Here is my code... import SwiftUI import RealityKit import RealityKitContent extension Notification.Name { static let button1Pressed = Notification.Name("button1pressed") static let button2Pressed = Notification.Name("button2pressed") static let button3Pressed = Notification.Name("button3pressed") } struct MainButtons: View { @State private var transitionToNextSceneForButton1 = false @State private var transitionToNextSceneForButton2 = false @State private var transitionToNextSceneForButton3 = false @Environment(AppModel.self) var appModel @Environment(\.dismissWindow) var dismissWindow // Notification publishers for each button private let button1PressedReceived = NotificationCenter.default.publisher(for: .button1Pressed) private let button2PressedReceived = NotificationCenter.default.publisher(for: .button2Pressed) private let button3PressedReceived = NotificationCenter.default.publisher(for: .button3Pressed) var body: some View { ZStack { RealityView { content in // Load your RC Pro scene that contains the 3D buttons. if let immersiveContentEntity = try? await Entity(named: "MainButtons", in: realityKitContentBundle) { content.add(immersiveContentEntity) } } // Optionally attach a gesture if you want to debug a generic tap: .gesture( TapGesture().targetedToAnyEntity().onEnded { value in print("3D Object tapped") _ = value.entity.applyTapForBehaviors() // Do not post a test notification here—rely on RC Pro timeline events. } ) } .onAppear { dismissWindow(id: "main") // Remove any test notification posting code. } // Listen for distinct button notifications. .onReceive(button1PressedReceived) { (output) in print("Button 1 pressed notification received") transitionToNextSceneForButton1 = true } .onReceive(button2PressedReceived.receive(on: DispatchQueue.main)) { _ in print("Button 2 pressed notification received") transitionToNextSceneForButton2 = true } .onReceive(button3PressedReceived.receive(on: DispatchQueue.main)) { _ in print("Button 3 pressed notification received") transitionToNextSceneForButton3 = true } // Present next scenes for each button as needed. For example, for button 1: .fullScreenCover(isPresented: $transitionToNextSceneForButton1) { FacilityTour() .environment(appModel) } // You can add additional fullScreenCover modifiers for button 2 and 3 transitions. } }

Summary As presented in the SwiftUI WWDC video, the new tabViewBottomAccessory should allow for unique contents for .inline. This is what was presented as being used for the Apple Music miniplayer. However, the functionality seems to be either missing or unintuitive. As seen in the photos attached, not only does .inline functionality not seem to do anything, but the inline accessory also has misaligned elements that cannot be fixed by conditionally modifying the contents. Build Target iOS 26.0 Details This problem recurs on physical devices, simulators, and Xcode previews. Here is a view I've constructed for use as a tabViewBottomAccessory: struct FitnessToolbarAccessory: View { @Environment(\.tabViewBottomAccessoryPlacement) var placement var body: some View { if (placement == .inline) { Text("hello") } else { HStack { HStack { Image(systemName: "dumbbell.fill") VStack(alignment: .leading) { Text("Active Workout") Text("Push Day - Chest") .font(.system(size: 13)) } Spacer() Image(systemName: "pause.fill") } .padding() } } } } Here is the result, working as expected in expanded mode: And here is the result in inline mode after minimizing the tabViewBottomAccessory: The content of this inline accessory is clearly incorrect, as it was specified to contain a Text view containing "hello". Additionally, the contents seem to have some incorrect alignment. This occurs regardless of the contents of the accessory, even plain text.