We would like to minimise the amount of code exposed in the product we have developed and an obvious option is to configure our build setting to use the All Symbols for Strip Style. Some investigation with an AI tool reports there are possible problems doing so. ContentFilter: non-global → all — risky. The NetworkExtension and XPC subsystems load your bundle and look up classes by name through the ObjC runtime, which uses __objc_classlist and friends (not the Mach-O symbol table), so in principle all works. In practice it occasionally breaks bundle loading in edge cases. If you try this, test specifically that systemextensionsctl activates the bundle and the filter actually traps packets. Is it safe to use All Symbols with a content filter or do we need to accept Non-Global symbols? Thanks, Dave

Switched to Xcode 27.2 and it started downloading documentation. After download is finished documentation window is empty, prompting to look up documenation online or download it. Triggering any download does not help – no documentation is available. (Sidenote: Reinstalled Xcode 27.2 and documentation was there. But Xcode started a download automatically and without confirmation and afterwards it is gone)

As the title reads, is a short and simple issue which is simply very frustrating. Once you exist fullscreen it works, Validate can be pressed in fullscreen, but not Distribute on an archived product.

Hello! I'm glad that Xcode 27.2 could fix many StoreKit testing issues (27.2 release notes). While refreshing my unit tests for Xcode 27.2, I noticed a new behavior of "Ask to Buy" transactions. Since Xcode/iOS 27.2 at least, it seems that declining an Ask to Buy request in an SKTestSession makes StoreKit 2 record a purchase. Is this the intended new behavior? To reproduce: let session = try SKTestSession(configurationFileNamed: "ConfigFile") session.askToBuyEnabled = true session.resetToDefaultState() session.disableDialogs = true session.clearTransactions() let result = try await product.purchase() // .pending let pending = session.allTransactions().first! // state .deferred try session.declineAskToBuyTransaction(identifier: pending.identifier) // a few seconds later for await t in Transaction.all { print(t.unsafePayloadValue) }In my case the purchased product is an auto-renewable subscription with a free-trial intro offer, from a local .storekit file, in a unit test run inside a host app. After the decline, I could observe the following iOS 26.2 simulator SKTestTransaction.state = .deferred Transaction.updates : nothing Transaction.all : empty iOS 27.2 simulator SKTestTransaction.state = .failed Transaction.updates : emits a transaction Transaction.all : 1 transaction: ownershipType = .purchased, reason = .purchase, no revocation, expires at the end of the trial. The test session marks the request as failed, yet StoreKit 2 reports an active subscription... My understanding was that a declined "Ask to Buy" request never produces a transaction. With the new behavior, apps relying on Transaction.all or currentEntitlements would grant access that the Apple user refused... Is the new behavior expected? Or is it a regression (and is it limited to SKTestSession) ? Thanks for your help!

I am getting an error whenever I try get a screenshot from the iPhone DUO in the Device Hub. Error is: "Cannot choose between 2 displays."

Hello, I'm having an issue with SwiftUI Preview in Xcode. When I switch between different iPhone models, including iPhone 18 Pro, iPhone 18 Pro Max, iPhone Air, and iPhone Duo, the selected device name changes, but the actual device appearance and preview dimensions remain exactly the same. This happens in all SwiftUI projects, including newly created ones. I've tried reinstalling Xcode, clearing DerivedData and caches, and even performing a clean macOS installation, but the issue still persists. - macOS 27.0.1 - Xcode 27.2 Has anyone experienced this issue or found a solution? Thank you!

在通用架构版本 Xcode 26.3 中，在 Storyboard/Xib、表情键盘和配套的模拟器 iOS 26.2 (23C57) SDK + iOS 26.3.1 (23D8133) Simulator 中存在BUG，无法显示 emoji 表情图案。 虽然现在已经有了更新版本，但是 Xcode 26.3 是 macOS Sequoia 最后支持的版本，也就是有一批设备无法更新 Xcode 软件，只能使用 Xcode 26.3 版本，所以请求官方能修复此问题。 由于更新策略，2019年的苹果电脑最高只能升级到 macOS Sequoia 系统。 从控制台日志来看，不能显示 emoji 表情的 BUG 仅仅是因为缺少 AppleColorEmoji.ttc 和 hapticpatternlibrary.plist 文件。看上去是模拟器的问题，仅更新模拟器镜像文件应该就能修复吧。 FontParser could not open filePath /Library/Developer/CoreSimulator/Volumes/iOS_23D8133/Library/Developer/CoreSimulator/Profiles/Runtimes/iOS 26.3.simruntime/Contents/Resources/RuntimeRoot/System/Library/Fonts/Core/AppleColorEmoji.ttc: [2: No such file or directory] <_UIKBFeedbackGenerator: 0x60000351f700>: Error creating CHHapticPattern: Error Domain=NSCocoaErrorDomain Code=260 "未能打开文件“hapticpatternlibrary.plist”，因为它不存在。" UserInfo={NSFilePath=/Library/Audio/Tunings/Generic/Haptics/Library/hapticpatternlibrary.plist, NSURL=file:///Library/Audio/Tunings/Generic/Haptics/Library/hapticpatternlibrary.plist, NSUnderlyingError=0x600000d29bc0 {Error Domain=NSPOSIXErrorDomain Code=2 "No such file or directory"}}

I have encountered a possible compatibility issue with apps built using Xcode 27.1 and running on iOS 15.8 devices. Development environment Xcode 27.1 macOS 27.2 Build Configuration: Debug iOS Deployment Target: 15.8 Test device iPhone 7 iOS 15.8 Observed behavior When running an app built with Xcode 27.1 on an iPhone 7 running iOS 15.8, button interactions sometimes fail to update the user interface. The button press animation is displayed, indicating that the button is being interacted with. However, the expected UI state changes are sometimes not reflected, even after repeated taps. This behavior affects multiple buttons, rather than a single specific button. Comparison An IPA built with Xcode 26.2 works correctly on the same iPhone 7. An IPA built with Xcode 27.1 exhibits the UI update problem. The app runs correctly in the Xcode 27.1 Simulator. Question Is there a known compatibility issue in Xcode 27.1, its Swift compiler, or the iOS SDK that could cause SwiftUI UI updates to fail on iOS 15 devices? Although iOS 15.8 is available as a deployment target, I would like to know whether apps built with Xcode 27.1 are expected to work correctly on iOS 15 devices. Has anyone encountered similar behavior?

Hello, I'm facing an issue with the Xcode Canvas preview (#Preview). ​No matter which device model I select from the destination list (e.g., switching between iPhone 11 Pro, iPhone 16, or iPhone 18 Pro Max), the preview rendering remains completely identical. It shows the exact same physical frame (with a Dynamic Island) and the layout dimensions don't adjust to reflect the different screen sizes or notch styles of older devices.

i have a macbook pro late 2019 and has stopped take any major update software like mac os 27 i had the version of xcode 26.5 which by the way it was crashing many time before the update that it get today the version 27.2 and it says that it can't be opened because it doesn't support this mac ,my question is what should i do ,i get a new one which i can't financialy or there are any updates that i can update the app and worked just fine and be compatible. For me it's not right to disable apps just because there is a new software version and apple can't supported anymore there rest of us what do we do we become thieves.

It seems it's not possible anymore to view the output of the thinking process of an LLM. In Xcode 27 we were able to enlarge the output of each thinking phase, even though the UX wasn't actually great for doing this. Now, in Xcode 27.2 this is almost completely gone. There's only one line that seems to print the last line of the output. Do I miss something? Observing this content while the LLM is thinking is important, in order to be able to steer the LLM, or help it to find things, stop digressing into irrelevant topics, etc. While this was also not great, since enqueuing a message while the LLM is thinking has been queued – but the thinking phase has never been interrupted. So, enqueuing a message in order to steer the LLM was almost futile anyway. Nonetheless, I would expect newer Xcode versions to improve the user experience. I am unsure why this choice was made and included in a release. In my view, the AI chat and conversation behavior need improvement in several ways, though that is material for a few more posts.

Anyone working with the 27.3 devices beta in Xcode 27.3 on macOS 27.3? Canvas and device hub/simulator are throwing this error while open, several on first load, then periodically (not sure of the periodicity). Any help is appreciated. Not sure if it's impacting at the code or app, but it is interfering with active work (it assumes modal position while working). Can reproduce from Device Hub only (all other apps, including XCode closed), screencast here - demos 27.2 for the time it is started no errors, 27.3 as soon as it is started "identityservicesd quit unexpectedly" errors. Looks like stages in the rendering also - noticed dock not displayed, error, then dock displayed, next - home icons not displayed, error, then home icons displayed, etc. https://app.screencast.com/iK0BaVIVpjt0O

Hi, I’m encountering a consistent boot failure on iOS 27.3 simulator runtime with the new DeviceHub architecture in the latest Xcode. Environment Mac: Apple Silicon macOS version: [macOS Golden Gate 27.2] Xcode version: [27.2] Runtime: iOS 27.3 Simulator Runtime Architecture: DeviceHub (new CoreDevice stack, no old CoreSimulatorService fallback) Error Message

I have a powerful machine (Mac Studio M1 Max, 64 GB). Switching between apps is instantaneous. Except for Xcode. It consistently takes 2+ seconds to switch to Xcode from any app under any memory pressure. (And yes, it's an absolutely enormous amount of time for this task.) The same problem was with Xcode 26. I hoped that Xcode 27 will fix it, but nothing changed after updating. There are days when I need to switch between other apps and Xcode often AND many times (in hundreds). This annoying delay is absolutely unacceptable. I'm using Xcode since version 2 and have never experienced this problem. I wonder if other Xcode users experience the same? WHY there's even a delay? Again, no other app behaves like this - and I have quite a few apps open, including multiple heavy Adobe apps. FB25437949

With Xcode 27.1 (27A9275) and an iOS 27.0.1 (24A446) physical device, the UI-test runner starts, then fails about 60 seconds later with com.apple.dt.XCTest.XCTFuture, code 1000: “Timed out while enabling automation mode.” No business test method starts (0 methods started). What is the officially supported, targeted diagnostic entry point for this XCTest initialization boundary? Specifically: What supported configuration and logging level expose the runner sending an automation-mode RPC, device testmanagerd receiving it, testmanagerd sending a reply, and the runner receiving that reply? Which timestamps and session/request correlation fields associate those events across the runner and testmanagerd? Can collection be limited to one device, one test attempt, these processes, and a short time window? What minimal text fields are necessary? Can unrelated-app data, screenshots, audio/video, runner environment values, credentials, and personal information be excluded? Please specify any required setup and how to disable or remove it afterward. We need a supported entry point and a precise, minimal privacy scope before changing configuration or collecting broader diagnostics. A general CoreDevice or sysdiagnose workflow alone does not establish that these RPC send/receive events will be recorded.

I am looking for current Apple guidance on installing an independently developed, Watch-only watchOS app on an Apple Watch configured with Apple Watch For Your Kids / Family Setup, without changing its existing pairing, Apple Account or Family Setup configuration. Configuration: Apple Watch SE 3 (GPS + Cellular). watchOS 26.6. Managing iPhone running iOS 26.6.2. Xcode 27.1. Family Setup user is a child under 13. The app operates independently of a companion iPhone app. Its conventional-device version has been installed and tested on another, conventionally paired Watch. The isolated variant builds for watchOS with signing disabled, but has not been signed or installed on this Family Setup Watch. A read-only development-device inventory did not identify the Family Setup Watch. We have not changed its settings or attempted installation on it. I found this January 2022 answer from an Apple Frameworks Engineer stating that Family Setup Watches were not supported for Xcode development: https://developer.apple.com/forums/thread/698831 I also reviewed the later discussion about missing Developer Mode, which contains similar reports but no supported installation procedure: https://developer.apple.com/forums/thread/741161 Does the earlier limitation still apply to the configuration above? If direct Xcode development deployment is now supported, what is the exact workflow for discovering the Watch as a development destination, registering it for development, enabling Developer Mode where applicable, and provisioning/installing a Watch-only app while preserving Family Setup? Please identify required trust prompts, restarts or settings changes. If it is supported but the device is undiscoverable, which non-destructive diagnostics should be collected first? If it remains unsupported, is there another Apple-supported registered-device testing route that preserves Family Setup, or is a reviewed standalone Watch App Store release required? TestFlight is not an eligible route for this child's use under Apple's published age restriction. We are not seeking an age/account workaround, private APIs, or instructions to unpair, erase or convert the Watch to conventional pairing. Please link the applicable documentation or explicitly confirm the current platform limitation. This question concerns development installation only.

I have a personal company app for internal use only. Under Xcode26 if would build and I could distrubte it no problem. I updated to Xcode27. Made some minor changes to some of the Text that is displayed in the app and now I can not get the Build function to work. It is grayed out completly.

Versions under discussion: Apple Command Line Tools: 26.5.0.0.1777544298 LLDB: 2100.0.17.108 Components: lldb, LLDB.framework, debugserver I am seeking documented capability boundaries for observing an owned test process on macOS, rather than reporting a reproduced defect. Earliest observation point For a process launched under LLDB/debugserver, what is the earliest guaranteed stopped point? Is there a supported workflow that establishes observation before dyld and image initializers execute, or is observation necessarily established later? Please distinguish an exec stop, a stop at main, and images already present at the first stop. Event completeness and loss detection Are image-load/unload notifications a complete, ordered event stream, including shared-cache/preexisting images? Can library-initializer execution be observed with equivalent guarantees? What documented sequence, overflow, or dropped-event indication lets a client detect an incomplete observation, rather than treating a snapshot or lack of events as proof of completeness? Process ownership and reaping In launch versus attach modes, which process actually creates the target and which process consumes its terminal wait status? What documented behavior applies if the LLDB frontend or debugserver exits unexpectedly? In particular, how can a controller distinguish a termination signal being sent from the target actually being reaped, without competing with another wait-status consumer? Pointers to documentation or source/build references matching these components would be appreciated. If any of these guarantees are unsupported, an explicit limitation would also answer the question.

Environment Mac mini with Apple M4 macOS 26.7.2 (build 25G332) Xcode 26.6 CoreSimulatorService 1051.55 External USB SSD formatted as APFS, mounted read/write, with ownership enabled Source simulator: iPhone 17 Pro running iOS 26.5, shut down Issue I am attempting to preserve a shutdown simulator using the supported simctl clone operation, with the destination CoreSimulator device set located on an external APFS SSD. Command form: xcrun simctl clone <source-UDID> "<preserved-device-name>" "/Volumes/<external-volume>/<device-set-path>" The destination directory is owned by the logged-in user and has ordinary writable POSIX permissions. The volume is mounted read/write with ownership enabled. No restrictive ACL entries or file flags were found on the destination directories. Observed result simctl reports: NSPOSIXErrorDomain Code=22: Device was allocated but was stuck in creation state. CoreSimulator records the preceding data-clone failure as: NSCocoaErrorDomain Code=513 Underlying NSPOSIXErrorDomain Code=1 Operation not permitted The temporary destination simulator entered the Creating state and CoreSimulator automatically rolled it back. At the same timestamp, unified logging attributes the rejected external-volume write to com.apple.CoreSimulator.CoreSimulatorService, rather than to simctl. The TCC and sandbox sequence shows: A check for kTCCServiceSystemPolicyAllFiles. A check for kTCCServiceSystemPolicyRemovableVolumes. Removable-volume access requiring a TCC prompt. Prompting being denied because the requesting CoreSimulatorService is an Apple platform binary. Launch Services being unable to create an application record for the CoreSimulatorService XPC bundle. No consent prompt was displayed. CoreSimulatorService's Info.plist contains an NSRemovableVolumesUsageDescription string specifically describing access when a simulator device set is configured on external storage. This is a development-tool/service issue; no application project or third-party framework is involved. Expected result CoreSimulatorService should either present or arrange the supported removable-volume consent workflow, or use a documented user-controlled Files & Folders permission that allows an external device set. Questions What is the supported procedure for granting SystemPolicyRemovableVolumes access to com.apple.CoreSimulator.CoreSimulatorService when the Apple platform XPC service cannot display the consent prompt? Should CoreSimulatorService appear in System Settings > Privacy & Security > Files & Folders with a Removable Volumes switch? Is cloning into an external APFS CoreSimulator device set supported with Xcode 26.6 on macOS 26.7.2? Is there a supported way to initiate the operation so consent is attributed to Xcode or another user-facing Apple process? Is the Launch Services application-record failure expected, or is this a known CoreSimulator/TCC issue with an available fix? I have not edited the TCC database, reset privacy decisions, granted broad Full Disk Access, disabled security controls, changed permissions, removed extended attributes, manually copied protected simulator records, or retried using an unsupported filesystem workaround. I can provide focused, redacted CoreSimulator, sandboxd, and tccd excerpts privately if an Apple engineer requests them.

I am seeking a supported public diagnostic or prerequisite for a standalone Siri UI test. Feedback: FB25377231. Developer Technical Support directed this question to the forums. Environment: Mac mini (2024), Apple M4, 32 GB, native arm64 macOS 26.7.2 (25G332) Xcode 26.6 (17F113), iOS Simulator SDK 26.5 iPhone 17 Pro Simulator; the original test's xcresult identifies iOS 26.5 (23F73) Classic Siri, en-GB A small test-host project compiled and ran one selected XCTest case. Its core public API call was: XCUIDevice.shared.siriService.activate( voiceRecognitionText: "What is two plus two?" )The exact recognition text reached the testing activation path, but SpringBoard logged: No network. Not starting the request for SASRequestSourceTestingNo arithmetic answer appeared during the configured 30-second wait. This was observed once; a repeatable frequency is not established. This reproducer has no product app, app-specific Siri intent, or CarPlay integration. In the original Simulator test runner, the first unconstrained NWPathMonitor callback was satisfied, with en0/wiredEthernet, but supportsIPv4, supportsIPv6, and supportsDNS were all false. Default-route SCNetworkReachability samples reported reachable. A separate, later diagnostic used the same simulated device with the installed iOS 26.5 (23F77) runtime. Identical public Foundation/Network helpers on the host and Simulator each issued one URLSession HEAD request to: https://www.apple.com/library/test/success.html Both received HTTP 200. The Simulator helper's initial path callback again reported satisfied with IPv4/IPv6/DNS capability flags false. This proves a response only for that helper/API/URL/time. It does not prove that the flags stayed false, that a fresh DNS lookup occurred, that Siri endpoints were reachable, or that the earlier 23F73 failure had the same cause. A setup confounder may matter: additional iOS platform support, including runtime 23F77, was installed while the original 23F73 Simulator remained booted. The arithmetic test ran afterward but still reported 23F73 in xcresult. The later HTTPS control ran separately on 23F77. There has been no further Siri activation. What supported Simulator/XCUISiriService prerequisite or public diagnostic can distinguish the cause of this specific request-start refusal? Is it expected for this configuration, a known issue, or evidence of an unmet prerequisite? The complete original results and a minimal test-host project are retained. Selected evidence and source are already attached to FB25377231; the same material can be supplied privately to DTS. I am seeking a supported approach rather than private preferences, service overrides, or repeated test requests.