Up until some point relatively recently, I have been able to use Safari's web inspector to connect to the iOS simulator in order to debug our web application in development at http://localhost:8088. Now, the web inspector still OPENS, but it opens in a broken state. The context is available to select from Safari's "Develop" menu: Develop > "iPhone 16 Pro (Simulator)" > "localhost - login". It appears under the Safari heading if I have navigated to the web app in the browser, or under the Expo heading if I am accessing it through the webview in our React Native wrapper app. When I select it, the web inspector window does appear. However, once it opens, the Elements pane is empty, the Console pane is empty, expressions entered into the console are not evaluated, there's no content in Sources, Network, Storage, etc. Important notes: This broken state happens at http://localhost:8088 as well as http://127.0.0.1:8088, and it seems that the insecure context is the issue. The web inspector DOES work for HTTPS sites. If I navigate to, e.g., https://example.com in the simulator and connect the web inspector, everything works fine. The web inspector also works fine in Safari on macOS (OUTSIDE the simulator) when accessing non-HTTPS sites. It's only a problem for non-HTTPS sites when connecting to the simulator. A coworker has the same problem, so it is not isolated to my machine. I would enable TLS locally as a workaround, but this web app is very complex, and I know from experience that it is very difficult for various reasons to set it up properly for our project in development, and it will take significant non-trivial work to do so. So... Why is this happening? Is this expected behavior? Is there a way that I can debug my site on localhost without HTTPS?

We're observing new crashes specifically on iOS 18.4 devices with this pattern: Exception Type: SIGTRAP Exception Codes: fault addr: 0x000000019bc0f088 Crashed Thread: 0 Thread 0 0 libsystem_malloc.dylib _xzm_xzone_malloc_from_tiny_chunk.cold.1 + 36 1 libsystem_malloc.dylib __xzm_xzone_malloc_from_tiny_chunk + 612 2 libsystem_malloc.dylib __xzm_xzone_find_and_malloc_from_tiny_chunk + 112 3 libsystem_malloc.dylib __xzm_xzone_malloc_tiny_outlined + 312 4 CoreGraphics CG::Path::Path(CG::Path const&) + 132 5 CoreGraphics _CGPathCreateMutableCopyByTransformingPath + 112 6 CoreGraphics _CGFontCreateGlyphPath + 144 7 CoreGraphics _CGGlyphBuilderLockBitmaps + 1112 8 CoreGraphics _render_glyphs + 292 9 CoreGraphics _draw_glyph_bitmaps + 1116 10 CoreGraphics _ripc_DrawGlyphs + 1464 11 CoreGraphics CG::DisplayList::executeEntries(std::__1::__wrap_iter<std::__1::shared_ptr<CG::DisplayListEntry const>*>, std::__1::__wrap_iter<std::__1::shared_ptr<CG::DisplayListEntry const>*>, CGContextDelegate*, CGRenderingState*, CGGStack*, CGRect const*, __CFDictionary const*, bool) + 1328 12 CoreGraphics _CGDisplayListDrawInContextDelegate + 340 13 QuartzCore _CABackingStoreUpdate_ + 612 14 QuartzCore ____ZN2CA5Layer8display_Ev_block_invoke + 120 15 QuartzCore -[CALayer _display] + 1512 16 QuartzCore CA::Layer::layout_and_display_if_needed(CA::Transaction*) + 420 17 QuartzCore CA::Context::commit_transaction(CA::Transaction*, double, double*) + 476 18 QuartzCore CA::Transaction::commit() + 644 19 UIKitCore ___34-[UIApplication _firstCommitBlock]_block_invoke_2 + 36 20 CoreFoundation ___CFRUNLOOP_IS_CALLING_OUT_TO_A_BLOCK__ + 28 21 CoreFoundation ___CFRunLoopDoBlocks + 352 22 CoreFoundation ___CFRunLoopRun + 868 23 CoreFoundation _CFRunLoopRunSpecific + 572 24 GraphicsServices _GSEventRunModal + 168 25 UIKitCore -[UIApplication _run] + 816 26 UIKitCore _UIApplicationMain + 336 27 app _main + 132 28 dyld __dyld_process_info_create + 33284 Key Observations: Crash occurs during font glyph path creation (CGFontCreateGlyphPath) Involves memory allocation in malloc's xzone implementation 100% reproducible on iOS 18.4, not seen in prior OS versions Occurs during standard CALayer rendering operations Not tied to any specific font family or glyph content Questions for Apple: Is this crash signature recognized as a known issue in iOS 18.4's CoreGraphics? Could changes to xzone memory management in iOS 18.4 interact poorly with font rendering? Are there specific conditions that might trigger SIGTRAP in CGPathCreateMutableCopyByTransformingPath? Any recommended mitigations for text rendering while awaiting system updates?

PDFKit PDFPage.characterBounds(at: Int) is returning incorrect coordinates with iOS 18 beta 4 and later / Xcode 16 beta 4. It worked fine in iOS 17 and earlier (after showing the same issue during the early iOS 17 beta cycle) It breaks critical functionality that my app relies on. I have filed feedback (FB14843671). So far no changes in the latest betas. iOS release date is approaching fast! Anybody having the same issue? Any workaround available?

I have an iPhone 14 Pro. I downloaded the iOS 26 beta and had a SERIOUS error, rendering the phone unusable. I charged it to 60% and kept it plugged in while updating. While updating, I restarted several times at the Apple logo, then at the Welcome screen, and it had quite a few bugs with low battery warnings. When I turned it on, I noticed I had 1% (I thought it was strange). When it was plugged in, it wouldn't charge; it only had 1% left, and it also restarted every 2 minutes. Off-plugged, it did exactly the same thing. In the end, I had to go back to iOS 18.5; I had no problems with this version.

We operate a native iOS app that authenticates users via the system browser using OIDC against a 3rd party SaaS authentication broker, which then performs authentication against the individual customer IdPs using SAML within the browser session, and then completes the OIDC login between the broker and our application. Our application initiates the OIDC login using ASWebAuthenticationSession, using the broker’s library, against the broker, and at that point the authentication workflow is handled by the broker. At the beginning of the login session, the broker sets a session-identifying cookie for their domain, before redirecting the user to their company’s identity provider to authenticate, which then redirects the user back to the broker. Intermittently, Mobile Safari does not include that previously set cookie on the final communication with the broker, when being redirected from the IdP as the final step of the SAML portion of the login workflow. When the cookie is missing, the broker cannot correlate the authentication response to the original request and the login fails, even though the user successfully authenticated at their identity provider. The same user can sometimes find success after retrying on the same device minutes later, without any changes. When we first started diagnosing this issue, we were seeing about a 20% rate of these errors across all Mobile Safari logins, which we can identify fairly conclusively (from provider logs, based on their guidance) as being caused by the session cookie set in one request not being provided on the subsequent request to the same domain. Our authentication broker provider has indicated, based on their server logs, and logs from an affected device, that this behavior is caused by Mobile Safari’s Intelligent Tracking Protection (ITP) causing Mobile Safari to not send the login session cookie to the broker when handling the SAML redirect from the IdP back to them. Our authentication broker provider recommended that we switch the SAML Request binding setting from POST to Redirect in the SAML configurations against our customer IdPs, which reduced the rate of these errors to about 5% for most of our customers. However, we have at least one customer which is still seeing about a 20% rate in these errors for Mobile Safari logins after this change, and even a 5% error rate seems high. Our authentication broker has not been able to suggest any further remediation options, and has suggested we contact Apple for assistance. Our questions are: Is it reasonable to assume that ITP is causing this issue? Is there any way to confirm, conclusively, that ITP has caused a previously-set cookie to not be provided on a subsequent request to the same domain, i.e. via device logs? If so, are there any steps which can be taken to reduce or eliminate this error? Changes to how ASWebAuthenticationSession is invoked? Changes to the Authentication Broker which would reduce the chance of ITP being triggered? Changes to the Customer IdP configuration which would reduce the chance of ITP being triggered?

Hi everyone, I’m encountering a critical reliability issue with message passing in my Safari Web Extension on iOS 18.4.1 and iOS 18.5. In my extension, I’m using the standard messaging API. The background script sends a message to the content script using browser.tabs.sendMessage(...), and the content script registers a listener via: browser.runtime.onMessage.addListener(handler); This setup has been working reliably in all prior versions of iOS. However, after updating to iOS 18.4.1 and 18.5, I’ve noticed the following behavior: ✅ The content script is successfully injected, and onMessage.addListener is registered (I see logging confirming this). ✅ The background script sends the message using the correct tabId (also confirmed via logs). ❌ The content script’s onMessage listener is not consistently triggered. ⚠️ This issue is intermittent, sometimes the message is received, sometimes it is silently dropped. ❌ No exceptions or errors are thrown in either script, the message appears to be sent, but not picked up from the content script message listener.

Area: Software Update Type of Feedback: Application Bug Description Device: iPhone 13 Pro running iOS 26 Build environment: Xcode 16.4 Problem description: When a text field has secureTextEntry = YES and Password Autofill / Passkeys is active, the autofill panel is not included in the rect reported from the keyboard notifications (UIKeyboardFrameEndUserInfoKey or others). As a result, when calculating the offset to move the screen up and reveal the hidden input field, the field is not displayed correctly because the reported keyboard height is smaller than the actual visible height. Observed behavior: This only occurs on devices running iOS 26 built with Xcode 16.4. On previous versions of iOS, with the same settings (secureTextEntry and Autofill active), the rect correctly includes the autofill panel height, and the UI works as expected. I tested with both UIKeyboardDidShowNotification and UIKeyboardWillChangeFrameNotification, and in both cases the behavior is the same: the height is incorrect (smaller than expected with the autofill panel). What I expect / questions: That UIKeyboardFrameEndUserInfoKey (or the related notification) correctly reports the total area covered by the keyboard, including any password autofill panel, when secureTextEntry is active. That the new behavior in iOS 26 be documented if this omission is intentional, or otherwise considered a bug if it is not. If there is any official workaround suggested by Apple for developers affected by this issue while a fix is provided. Thank you for your support.

Hi all, I’ve got a usability question about accessibility navigation. My app has a lot of carousels (horizontally scrolling lists of content with far more elements than can fit on the screen). Often, these are just images, but sometimes, they’re cards with multiple subelements. In our previous implementation, each card was a single accessibility element, and we exposed the subelements as accessibility custom actions. Despite this, users frequently mentioned navigating with VoiceOver as a pain point. It takes a long time to navigate through and navigate past these carousels. To solve this, I converted my carousels into a single adjustable element, so users can navigate through it with one swipe, and they can still access the elements by adjusting the values up and down. I got this advice from this 2018 WWDC talk. Is this still the recommended advice? Or is there a new, preferred way to do this? Additionally, I had to get a little creative with the second carousel, the one with multiple subelements. Some of these were interactive (imagine a card with a description, an upvote button, and a downvote button). Adjustable elements override the accessibility custom actions VoiceOver gesture, so I can’t expose the individual buttons as actions. Instead, I made each subelement in each card in the carousel one of the adjustable values. Swiping up would go from description 1 to upvote button 1 to downvote button 1 to description 2, etc. Double tapping with VoiceOver would perform whatever action the carousel is currently on. So if I adjust the value to the element at index 2 (say, downvote 1), double tapping would trigger the downvote button’s action. Does this make sense? Is there a better way to do this? This seemed to be the best compromise between screenreader navigation speed, exposing all actions, and the existing UI.

I’m integrating the Declared Age Range feature to tailor our app’s experience based on a user’s age range. I’m currently in the testing phase and would like to repeatedly test the consent flow and different outcomes from AgeRangeService.shared.requestAgeRange(...). However, once I go through the consent flow and choose to share, the age-range sharing sheet no longer appears on subsequent attempts—so it’s hard to validate edge cases (e.g., changed gates, declined flow, re-prompt behavior). Could you advise on the recommended way to reset or re-prompt during development? In particular: Is there a supported way to clear per-app consent so the system prompts again? Under what conditions should the “Share Age Range Again” control appear in Settings, and is there an equivalent way to trigger it for testing? Are there best practices for QA (e.g., using Ask First at the system level, testing on real devices vs. Simulator, using a separate bundle ID for dev builds, or other steps)? Any other guidance for validating different requestAgeRange results (e.g., declined/not available) would be appreciated.

We are using a TabView as the TabBarController in our app for main navigation. On one of the tabs we have a view that consists of a TabView with .tabViewStyle(.page) in order to scroll horizontally between pages inside of that specific tab. The .tabBarMinimizeBehavior(.onScrollDown) works on all the other TabItem views, but for this one it does not recognise any vertical scrolling in any of the pages, in order to minimize the TabBar. I believe this is a bug? If we don't wrap the views inside the TabView with .page style, we are able to get the expected behaviour using the tabBarMinimizeBehavior. Please let us know if this is going to be fixed in a future iOS 26 beta release.

I have an odd issue I'm trying to troubleshoot. I have an app that is deployed to our enterprise and works on almost all of our devices. Lately, I have devices (very small number) where the app installs from our MDM and it crashes upon launch, it does not get past the launch screen. If I remove and reinstall the app from the MDM it still crashes. I decided to put the devices in developer mode and if I run the App from Xcode on the device it works fine. If I stop it and run it right from the device it also works. If I do a final build of the app and install it using "Devices and Simulators", the app crashes upon launch. Using "Devices and Simulators" I check for crash log for the app and no file. I am stumped as to whats going on.

After updating to Xcode 26 my XCUITests are now failing as during execution exceptions are being raised and caught by my catch all breakpoint These exceptions are only raised during testing, and seem to be referencing some private internal property. It happens when trying to tap a button based off an accessibilityIdentifier e.g. accessibilityIdentifier = "tertiary-button" ... ... app.buttons["tertiary-button"].tap() The full error is: Thread 1: "[<UIKit.ButtonBarButtonVisualProvider 0x600003b4aa00> valueForUndefinedKey:]: this class is not key value coding-compliant for the key _titleButton." Anyone found any workarounds or solutions? I need to get my tests running on the liquid glass UI

I set iOS 26 to install overnight, put my iPhone 16 Pro on the MagSafe charger, watched it charge just fine, and went to sleep. When I woke up the iPhone showed the “plug into power” dead battery screen. I took it off MagSafe and put it back on. A half hour later the phone was warm but still wouldn’t power on, just showed the battery screen with a little red in it. I took it off MagSafe and plugged it into my iPad charging brick with USB cable to give it more power, still it did not turn on. I tried holding all the buttons to try to force a restart but didn’t work. For anyone else encountering this, do this to enter DFU mode and restore it. I had to do it a few times before I got the timing right. Plug into your Mac and open Finder (or apparently a PC with Apple Devices or iTunes) Press and quickly release volume up Press and quickly release volume down Press and hold right side button When the battery disappears and screen goes black, hold volume down and continue holding side button After a couple seconds release the side button and continue holding volume down A prompt to allow connecting to the iPhone should appear after a couple seconds, click Allow, and it’ll say the iPhone entered DFU mode - proceed to restore the firmware

I'm developing a web extension for Safari on iOS using MV3. The extension is working fine in Chrome, but in Safari I experience some seemingly random issues. I would like to debug it, but here is my problem. I have my iPhone connected via cable to Mac, and it works fine with XCode, so I assume this part is OK. I open Safari or Safari Tech Preview (doesn't matter) on my Mac, developers options are enabled, and in the Develop menu, under my iPhone section, there are things I can debug. There is an entry "[Ext name] - Extension Service Worker" but when I click it, it's empty. Web inspector pops up, but there are no network requests, no logs, nothing. I know the extension is working, because I can stream log to my HTTP server, but I don't see them here at all. I can use console to trigger commands like chrome.storage.local.get(null, console.log) and it shows my local store, so why I don't see any logs? Also, the background script is not visible in the Sources tab, just one weird request: navigator.serviceWorker.register('safari-web-extension://E3449EA7-EC25-4696-8E6C-[ID HERE]/background.js'); </script> Any ideas what went wrong? The entire team of 4 people has the same issue and we can't move forward because of that. Also, the Develop => Service workers or any other menu section doesn't show my service worker. Logs for websites running on my phone are visible and in general web inspector for them works fine.

I am using below code to change navigationBar bg colour, but the text is hidden in large title. It works fine in previous versions. Kindly refer below code and attached images. Code: override func viewWillAppear(_ animated: Bool) { super.viewWillAppear(animated) navigationController?.navigationBar.prefersLargeTitles = true navigationItem.largeTitleDisplayMode = .always let appearance = UINavigationBarAppearance() appearance.backgroundColor = UIColor( red: 0.101961, green: 0.439216, blue: 0.388235, alpha: 1.0 ) navigationController?.navigationBar.standardAppearance = appearance navigationController?.navigationBar.scrollEdgeAppearance = appearance navigationController?.navigationBar.compactAppearance = appearance } Referenced images:

Hello. Is there a good SwiftUI approach on getting the TextEditor cursor position? I have a TextEditor and sometimes when we have a longer text inside it, the cursor is not seen because the keyboard is above covering the bottom of the TextEditor. I would like to somehow detect the position of the cursor, and if it's on the last line of the TextEditor, scroll to the bottom. I've already checked a bit and didn't find any good method of doing this in SwiftUI. If you have any ideas on how to do this, or even a different method any help would be highly appreciated. Thank you!

I know there are several existing threads on this topic but things keep changing. The release notes for Xcode 26 beta 3 have the following statements for a couple of resolved issues: Asset Catalog Fixed: Unable to set Icon Composer icon as alternate iOS icon (153305178) (FB18025356) Icon Composer Fixed: Icon Composer icons back deploy to older versions of iOS, macOS, and watchOS with inconsistent rendering. (152258860) I had a working solution under beta 1 and beta 2 for both of these. But under beta 3, I am now seeing the new glass icons for my app when running on a simulated iOS 18 device. This is happening for both the main app icon and any alternates. This contradicts the statement that beta 3 fixes this issue. There is no documentation (that I can find) describing how you are supposed to support old icons for iOS 18 and new glass icons for iOS 26. There is no documentation for how to support alternate glass icons for iOS 26. What I'm doing at the moment (that worked before beta 3) was to have the normal iOS 18 app icons in the Asset catalog and to have the new glass icons added to the project. The filenames for the glass .icon files have the same name as the app icons in the Assets catalog. This worked under beta 1 and beta 2. And despite the Xcode 26 beta 3 release notes stating that Icon Composer icons no longer back deploy to iOS 18, I'm seeing the opposite. Beta 3 now does the opposite of that statement. Does anyone have a working solution that supports old iOS 18 app icons and new iOS 26 glass icons using Xcode 26 beta 3? Note, all of my testing is with simulated iOS devices and I'm running Xcode 26 beta 3 under macOS 15.5. Maybe that's an issue?

When using the .glassEffect modifier on a button in swiftui combined with the .interactive() modifier, the button continues to show the interactive animation even when it’s covered by another element. Example: ZStack { Button { print("Button overlayed by ZStack") // Does not trigger, but interactive animation still plays } label: { image } .glassEffect(.regular.interactive()) Rectangle().fill(.black.opacity(0.7)) } This occurs with overlays, ZStacks, and even if the overlay is a button. Example below: EDIT: It seems like rocketsim's gif recording doesnt show the bug for some reason... really strange... Edit 2: reuploaded gif, recorded as mp4 and converted to gif seems to have worked... Feedback ID: FB22054300 I've attached this sample app to my feedback ticket to help with debugging the issue. It doesn't look like I can share it in this post though.

Hi everyone! I've considered this — what if Apple added a native system-wide feature in all of iOS, iPadOS, and macOS called “CrossRun” where you can natively execute non-App Store software like Windows or Linux apps natively on your device? But not in a sluggish emulator—this would use intelligent Apple-signed Just-In-Time (JIT) compilation inside the virtual containers, and the experience would actually perform fast and feel natural. This is my vision for CrossRun: Every developer, student, creative professional, and enterprise user who relies on specialized software—whether it’s legacy Windows tools, Linux-only applications, or vintage DOS and Classic Mac utilities—feels the pain of platform lock‑in. Artists can’t run niche Linux‑based graphics programs on their iPads. Engineers can’t test x64‑only binaries on Apple Silicon without juggling emulators. Retro‑gaming fans miss their favorite DOS titles. Even enterprises struggle to standardize on Apple hardware because critical Windows‑only applications won’t run seamlessly. If we don’t push for CrossRun now, the Apple ecosystem remains siloed: iPads and iPhones will continue limited to App Store apps, Macs will still need multiple third‑party VM tools, and countless workflows stay fragmented across devices. That means slower development cycles, extra licensing costs for virtualization software, and lost opportunities for education, creativity, and business efficiency. Without CrossRun’s universal runtime, we’ll still be rebooting into different environments or paying for separate virtualization apps—year after year. Apple already provides the building blocks: Rosetta 2, Virtualization.framework, Apple Silicon—and QEMU thrives as open‑source, battle‑tested code. With the next wave of Apple Silicon devices on the horizon, demand for cross‑architecture support, legacy‑app compatibility, and enterprise containerization is only growing. Delaying another year will cost developers, businesses, and users real time and money. Let’s show Apple that the community is ready for a truly universal, system‑integrated solution—right now. Key features we should demand in CrossRun: Built‑in Apple‑signed QEMU for all ISAs (x86, ARM, RISC‑V, PowerPC, 68k, MIPS, etc.) Rosetta 2 JIT for seamless macOS and Windows x64 support Metal‑backed 3D GPU passthrough and Vulkan→Metal / Direct3D→Metal translation Downloadable OS and app containers via the App Store or verified repositories (Ubuntu, Windows ARM/x64, Android, Haiku, ReactOS, FreeBSD, retro OSes) Predictive ML pre‑warm cache to speed cold starts Dynamic resource scaling (CPU, GPU, RAM) per container iCloud‑synced snapshots and shareable VM links for cross‑device continuity Customizable on‑screen controls (D‑pad, virtual buttons, trackpad, keyboard) on iPhone, iPad, and macOS Secure sandboxing via Virtualization.framework with VM disk encryption and MDM policy enforcement Virtual LAN and VPN passthrough for container networking Developer tooling (crossrunctl CLI, Xcode debugger integration, CI/CD support) Plugin ecosystem and container SDK for community‑published templates and translation layers Let Apple know it’s time to bake CrossRun into the system and unlock a universal runtime for every app, past and future, across iOS, iPadOS, and macOS.

The documentation for isEligibleForAgeFeatures states: Use this property to determine whether a person using your app is in an applicable region that requires additional age-related obligations for when you distribute apps on the App Store. But what does "region" mean? Is this going to return true if the user has downloaded the app from the US App Store? Or will it go further and geolocate the user and identify them as being within a particular relevant state within the US?