I'm trying to use SwiftUI views inside an NSToolbarItem, and I feel like I'm fighting the system. The goal is to create custom toolbar controls. I don't want AppKit or SwiftUI to recognize that there's a SwiftUI view in the toolbar and then try to style it automatically, which is what I think is happening. Consider this hierarchy: - NSToolbar - NSToolbarItem - view -> MyCustomNSView - addSubview(NSHostingView<CustomView>) So CustomView is inside an NSHostingView, which is inside an NSView, which is assigned to the .view property of an NSToolbarItem. For a while, I struggled to get this working at all because SwiftUI's layout metrics appeared to be affected by safe area insets. SwiftUI's ignoresSafeArea() only seemed to move the problem around. For now, I've disabled full-size content view, which I don't really need anyway. However, a SwiftUI Button or Menu still doesn't behave or style the way I would expect when placed inside an NSToolbarItem. They don't seem to honor many of the styling attributes or view modifiers that work elsewhere. Is there a way to tell AppKit to "ignore" the children of my NSToolbarItem and not apply toolbar-specific styling to them? Am I even correct in thinking that AppKit is recognizing a Button or Menu in the toolbar and forcibly applying styling to it? Previously, I tried implementing my "custom toolbar" entirely outside of NSToolbar, but then SwiftUI struggled with placing views under the transparent, non-existent toolbar. I'd appreciate any direction or hints here. It would be really nice to use SwiftUI for the controls I have in mind, but at the moment I can't get close to the result I'm aiming for.

Hello, with macOS' and iOS' recent updates (26.4.x), Handoff in my app appears to have become dysfunctional, as I receive the following message in the logs: getContinuationStreamsWithCompletionHandler(), inputStream=(null) outputStream=(null) error=Error Domain=NSPOSIXErrorDomain Code=94 "Bad message" UserInfo={NSLocalizedDescription=Could not set up internal streams} It ONLY works: From a Mac with 15.x to other Macs or iOS devices (on 26.4.x). All other combinations (see below) do NOT work: iOS (26.4.x) > iOS (26.4.x) iOS (26.4.x) > Mac (26.4.x) iOS (26.4.x) > Mac (15.x) Mac (26.4.x) > iOS (26.4.x) The Handoff prompt is shown in the Dock on Mac, in the app switcher on iOS, and in the Dock on iPadOS, but once I click/press it, I get the continuationStreams error. Now, is it my app, or the OS? Thank you kindly, – Matthias

Metal GPU Driver Crash on M5 Pro + macOS 26.5 — kIOGPUCommandBufferCallbackErrorOutOfMemory with <2GB working sets Summary The Metal driver AGXMetalG17X 351.2 on macOS 26.5 (25F71) for the M5 Pro chip crashes with kIOGPUCommandBufferCallbackErrorOutOfMemory (00000008) when running LLM inference workloads with working sets as small as ~1.5GB, despite 24GB of unified memory being available and Apple Diagnostics confirming the hardware is fully functional. This affects multiple tools: MLX, llama.cpp (Metal backend), and native apps using Metal for inference. System Component Value Model MacBook Pro (Mac17,9) Chip Apple M5 Pro (applegpu_g17s) GPU Cores 16 RAM 24 GB LPDDR5 macOS 26.5 (25F71) Metal Metal 4 GPU Driver AGXMetalG17X 351.2 Xcode 26.5 (17F42) Reproduction MLX (Python) pip install mlx mlx-lm python -m mlx_lm.generate \ --model mlx-community/Qwen2.5-3B-Instruct-4bit \ --max-tokens 10 \ --prompt "Hello" Expected: Normal text generation Actual: Crash with: libc++abi: terminating due to uncaught exception of type std::runtime_error: [METAL] Command buffer execution failed: Insufficient Memory (00000008:kIOGPUCommandBufferCallbackErrorOutOfMemory) llama.cpp brew install llama.cpp llama-cli --model model.gguf --prompt "Hello" --n-predict 20 --n-gpu-layers 99 Expected: Fast GPU generation Actual: Process hangs indefinitely Test Results Tool Model Peak Memory Result MLX Qwen2.5-0.5B-4bit 0.36 GB ✅ Works MLX Qwen2.5-1.5B-4bit 0.98 GB ✅ Works MLX Qwen3-1.7B-4bit 1.01 GB ✅ Works MLX Qwen2.5-3B-4bit ~1.5 GB ❌ Metal OOM crash MLX Qwen3-4B-4bit ~2.1 GB ❌ Metal OOM crash MLX Qwen3-8B-4bit ~4.5 GB ❌ Metal OOM crash llama.cpp Qwen2.5-0.5B GGUF ~0.5 GB ❌ Hangs with GPU llama.cpp Qwen2.5-0.5B GGUF ~0.5 GB ✅ Works with CPU only Key Evidence Hardware is healthy — Apple Diagnostics passed all tests Basic Metal works — matmul, array ops work fine CPU inference works — llama.cpp with -ngl 0 runs correctly The error is NOT about actual memory exhaustion — kIOGPUCommandBufferCallbackErrorOutOfMemory means the kernel rejects the Metal memory commit, not that physical memory is full. The system reports 17.76GB available for Metal working set. Crash Log Extract Thread 31 Crashed: 0 libsystem_kernel.dylib __pthread_kill + 8 1 libsystem_pthread.dylib pthread_kill + 296 2 libsystem_c.dylib abort + 148 3 Metal MTLReportFailure.cold.1 + 48 4 Metal MTLReportFailure + 576 5 Metal -[_MTLCommandBuffer addCompletedHandler:] + 104 ... Exception Type: EXC_CRASH (SIGABRT) Termination Reason: Namespace SIGNAL, Code 6, Abort trap: 6 Related Issues ml-explore/mlx#3586 — Metal compiler regression on macOS 26.5 ml-explore/mlx#3534 — M5 float32 precision issue ml-explore/mlx#3568 — M5 random divergence ml-explore/mlx#3539 — Metal residency OOM (M4 Max) Request Please investigate the AGXMetalG17X driver for M5 Pro on macOS 26.5. The driver appears to incorrectly reject Metal memory commits for LLM inference workloads, even when the working set is well within the system's reported limits (1.5GB requested vs 17.76GB available). Happy to provide full crash logs, sysdiagnose archives, or run additional tests.

Hi All, We have released our application in both iOs and Mac Os platform. But Now we have changed our plan to target Mac Os only. So, We need to disable the availability of iOs version of our app. As I have checked in 'Pricing And Availability' Page, There is one option to remove from sale. But What we need to do to disable iOs app availability alone(Not Mac app)? Please give your suggestion to do this. Thanks.

I am seeing a reproducible issue with macOS ControlCenter's per-app menu bar tracking state in a SwiftUI MenuBarExtra app. On launch the app creates its menu bar status item. ControlCenter then reads group.com.apple.controlcenter / trackedApplications, moves the app's status item host to the blocked list, sends NSStatusItemChangeVisibilityAction with visibility=0, and the app terminates because its only status item is removed. Two observations: Changing only the bundle identifier makes the menu bar item appear normally. In trackedApplications, the affected app has its own entry with isAllowed: true; a separate entry for another app has isAllowed: false, and that disallowed foreign entry's menuItemLocations contained an entry referencing the affected app. Removing only the affected app's reference from the other app's disallowed entry fixed the issue: the app launched normally, and ControlCenter no longer sent visibility=0. ControlCenter appears to associate one app's menu item identity with another app's disallowed tracked record, letting the foreign blocked record override the app's own allowed record. Is this expected? And is there a supported way to reset this per-app ControlCenter menu bar state without editing the protected group.com.apple.controlcenter plist directly? Testing Environment: macOS 26.3 (25D125) Xcode 26.3 (17C529) LSUIElement: true Sandboxed app SwiftUI MenuBarExtra Relevant log pattern: Host properties initialized; (bid:[AffectedApp]-Item-0-[pid]) State(applicationItem: true, clientRequestsVisibility: true, neverClip: false) looked up value <private> for key trackedApplications (Domain: group.com.apple.controlcenter) Moving host to blocked list; (bid:[AffectedApp]-Item-0-[pid]) Requesting blocked host to not be visible; (bid:[AffectedApp]-Item-0-[pid]) Sending action(s): <NSStatusItemChangeVisibilityAction: ...> Received action(s): NSStatusItemChangeVisibilityAction 0 agent requesting visibility=0 temporary=0 0 terminating on removal

I have the issue with simply trying to ssh to any other host on the LAN since Tahoe 26.5. last entries from ssh -v answer are debug1: Host 'answer' is known and matches the ED25519 host key. debug1: Found key in /Users/dgerman/.ssh/known_hosts:5 debug1: rekey out after 134217728 blocks debug1: SSH2_MSG_NEWKEYS sent debug1: expecting SSH2_MSG_NEWKEYS debug1: SSH2_MSG_NEWKEYS received debug1: rekey in after 134217728 blocks debug1: SSH2_MSG_EXT_INFO received debug1: kex_ext_info_client_parse: server-sig-algs= debug1: kex_ext_info_check_ver: publickey-hostbound (AT) openssh.com=<0> debug1: SSH2_MSG_SERVICE_ACCEPT received debug1: Authentications that can continue: publickey,password debug1: Next authentication method: publickey No problem from older MacBooks This worked before the update.

We have been having unexplained failures with the codesign tool recently on macosx aarch64 and x64 hosts. Every once in a while when signing an app locally using the following command: /usr/bin/codesign -s - -vvvv --force /home/me/FooBarCalculator.app results in the following error: /home/me/FooBarCalculator.app: timestamps differ by 185 seconds - check your system clock The number of seconds reported in the error message keeps varying (but usually in that range). We have checked the system clock but there isn't anything wrong (from what we can see) with the host. In fact, we have been seeing this error on several hosts now, so it isn't specific to one host. While looking into this issue, we even printed the details of an already signed binary using the following command: codesign -dvvv HelloWorld.app and that prints among other things, similar warning message: ... Timestamp=12 May 2026 at 5:36:0 AM HelloWorld.app: timestamp mismatch: internal time 12 May 2026 at 5:32:59 AM (184 seconds apart) I'm looking for inputs on how we go about debugging this issue and where/how the codesign tool sources these timestamps from (any specific API?) and what value is it comparing against to notice a difference. These affected hosts have different operating system versions some 15.x and some 26.x.

I have a chat app with a floating composer pill (NSGlassEffectView) pinned to the bottom of my content area. Messages scroll behind it. The pill looks fine but content behind it is completely clear. Fully legible text showing through, no frosting at all. Mail, Messages, Notes all have this frosted look where content behind a floating bar is blurred enough that you can't read it. That's what I'm going for. My scroll view has wantsLayer = true. I tried hasVerticalScroller = true + scrollerStyle = .overlay based on a tip I saw in another thread. No change. Is NSGlassEffectView supposed to blur what's behind it, or is the frosting in Apple's apps coming from the scroll edge effect? If it's the scroll edge effect, is there any way to trigger it for a custom floating view at the bottom of a scroll view? I can't find any AppKit API for this. I also spent time trying this in SwiftUI using .glassEffect() on macOS. Same result, the material renders nearly clear with no way to control the frosting level via any public API. If I'm missing something obvious I'd love to know. I've been stuck on this one for a while. Thanks for any advice! I've attached photos of my input bar (clearly transparent) and Apple Messages, which has a style that I see all over iOS and Mac OS 26.

Just posted this feedback regarding macOS 26 "Tahoe" (FB19853155) - please support with additional submissions if you share my view. I will miss the beautiful and individual designed icons of the past! "macOS 26 is enforcing squicles for app icons, falling back to a grey background for 3rd party apps without a compliant AppIcon asset. As a result many original app icons are reduced in size and hard to distinguish because they share the same background color. Although I respect Apple's strive for an iOS-like UI on Macs, a smooth transition path would be more user- and developer-friendly ... e.g. with some info.plist property to opt-out icon migration, potentially ignored by a future macOS version. The current solution causes a bad usability, and makes the system look inconsistent as many - especially free - software will not be updated with new icon designs. Please reconsider this bad design decision!"

Hello! I'm trying to implement an editor app (macOS) that allows the user to write code, which will be compiled and executed, showing the result in the editor window. Imagine it like SwiftUI previews, but the graphic output is created with Metal, not SwiftUI. I found that IOSurface can be used to share that kind of data over XPC, so I would not have to rely on the private NSRemoteView. However, I'm confused if it is, at all, possible for my editor app to connect to an XPC Service, that was NOT bundled with it (but compiled by it at runtime). I succeeded to launch an XPC service defined as: <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd"> <plist version="1.0"> <dict> <key>Label</key> <string>com.myteam.myproject.service</string> <key>MachServices</key> <dict> <key>com.myteam.myproject.service</key> <true/> </dict> <key>Program</key> <string>/Path/to/service/run_my_service.sh</string> </dict> </plist> But the call to let connection = NSXPCConnection(machServiceName: "com.myteam.myproject.service") let proxy = connection.remoteObjectProxyWithErrorHandler { error in continuation.resume(throwing: error) } as? MyServiceProtocol fails with "The connection to service named com.myteam.myproject.service was invalidated: Connection init failed at lookup with error 3 - No such process." I have added <key>com.apple.security.temporary-exception.mach-lookup.global-name</key> <array> <string>com.myteam.myproject.service</string> </array> to my entitlements. Since the tutorials I followed are quite old, I'm wondering if support for something like this was dropped at some point. Thanks for any advice!

In SwiftUI 6 when developing for MacOS in Xcode 26, there does not seem to be a way to provide an automatic increment of the current build number, i.e. "Current Project Version" in settings. Does anyone have a solution that works in the latest version of Xcode. i.e. 26? All the solutions that I have tried are based on older versions that involve running scripts to update the build file, (.xcconfig) or the .info.plist file which simply will not work in the current configuration.

As part of the OpenJDK testing we run several regression tests, including for Java SE networking APIs. These APIs ultimately end up calling BSD socket functions. On macos, starting macos 26, including on recent 26.2 version, we have started seeing some unexplained but consistent exception from one of these BSD socket APIs. We receive a "ENOBUFS" errno (No buffer space available) when trying to construct a socket(). These exact same tests continue to pass on many other older versions of macos (including 15.7.x). After looking into this more, we have been able to narrow this down to a very trivial C code which is as follows (also attached): #include <stdio.h> #include <sys/socket.h> #include <string.h> #include <unistd.h> #include <sys/errno.h> static int create_socket(const int attempt_number) { const int fd = socket(AF_INET6, SOCK_STREAM, 0); if (fd < 0) { fprintf(stderr, "socket creation failed on attempt %d," " due to: %s\n", attempt_number, strerror(errno)); return fd; } return fd; } int main() { const unsigned int num_times = 250000; for (unsigned int i = 1; i <= num_times; i++) { const int fd = create_socket(i); if (fd < 0) { return -1; } close(fd); } fprintf(stderr, "successfully created and closed %d sockets\n", num_times); } The code very trivially creates a socket() and close()s it. It does this repeatedly in a loop for a certain number of iterations. Compiling this as: clang sockbufspaceerr.c -o sockbufspaceerr.o and running it as: ./sockbufspaceerr.o consistently generates an error as follows on macos 26.x: socket creation failed on attempt 160995, due to: No buffer space available The iteration number on which the socket() creation fails varies, but the issue does reproduce. Running the same on older versions of macos doesn't reproduce the issue and the program terminates normally after those many iterations. Looking at the xnu source that is made available for each macos release here https://opensource.apple.com/releases/, I see that for macos 26.x there have been changes in this kernel code and there appears to be some kind of memory accountability code introduced in this code path. However, looking at the reproducer/application code in question, I believe it uses the right set of functions to both create as well as release the resources, so I can't see why this should cause the above error in macos 26.x. Does this look like some issue that needs attention in the macos kernel and should I report it through feedback assitant tool?

I am coding in Perl and I want to create Environment Variables (External to the app) specifically for this app only, however my custom ENV variables are deliberately blocked by the OS for security purposes (as described in documentation).

Hi, I’m pretty new to AppKit and I’m trying to make an NSTextField inside an NSTableView both: Editable Multi-line / wrapping Right now, wrapping works fine until I set: tf.isEditable = true Then the text becomes a single line. How do I make it editable while still wrapping correctly? import AppKit final class ViewController: NSViewController, NSTableViewDataSource, NSTableViewDelegate { let tableView = NSTableView() let text = String(repeating: "A", count: 500) override func viewDidLoad() { super.viewDidLoad() view = tableView tableView.addTableColumn(NSTableColumn()) tableView.usesAutomaticRowHeights = true tableView.dataSource = self tableView.delegate = self } func numberOfRows(in tableView: NSTableView) -> Int { 1 } func tableView(_ tableView: NSTableView, viewFor tableColumn: NSTableColumn?, row: Int) -> NSView? { let cell = NSTableCellView() let tf = NSTextField(wrappingLabelWithString: text) tf.lineBreakMode = .byCharWrapping if let tableColumn { tf.preferredMaxLayoutWidth = tableColumn.width } tf.isEditable = true // comment this out and wrapping works cell.addSubview(tf) tf.translatesAutoresizingMaskIntoConstraints = false NSLayoutConstraint.activate([ tf.leadingAnchor.constraint(equalTo: cell.leadingAnchor), tf.trailingAnchor.constraint(equalTo: cell.trailingAnchor), tf.topAnchor.constraint(equalTo: cell.topAnchor), tf.bottomAnchor.constraint(equalTo: cell.bottomAnchor), ]) return cell } }

Hello, I'm evaluating possibility to use virtualization to setup on-permise parallel testing system for a product I work on. My compatibility range is wide, i.e. it would cover macOS 12 - 26, so any testcase needs to be executed on at least 5 different OS versions (and introducing any parallization to the test execution would mean I deal with higher number of VMs). As far as I understand, there is a constraint in Apple Software License, that limits number of OS VM installations per one physical system to 2 (section 2.B.(iii)) in case of OS downloaded from the Internet or through App Store. Clearly these days a single high-end Mac hardware could sustain more than 2 VMs running in parallel. The license also mentions it is also possible to be in a volume or maintenance license program and then the terms of this program apply instead. So I wonder how do people normally deal with the above limitation? What is the path I need to follow if I want to be able to run more than "2 additional copies or instances of" macOS on my VMs?

In our macOS solution, we have a few processes and a few plugin modules which communicate with each other over XPC. We have recently started enforcing library validation flag along with hardened runtime for all processes and plugins. To enforce that, we are trying to get signing information from the XPC audit token using SecCodeCopySigningInformation with kSecCSDynamicInformation flag. As per documentation, this flag requires a live SecCode not SecStaticCode to be passed to SecCodeCopySigningInformation. However, SecCodeCopySigningInformation explicitly requires SecStaticCode in its parameters. So I am unsure how to pass live SecCode to SecCodeCopySigningInformation without copying SecStaticCode from it using SecCodeCopyStaticCode. Force cast from SecCode to SecStaticCode fails. Is unsafeBitCast a valid option in this case? Note that we support macOS version 12 and later.

In macOS Tahoe 26.4.1 we noticed that when we use our custom authorisation plugin to perform unlock using SF Authorisation Plugin View, the SF window isnt focused. User has to manually click on it to type in the password. We also noticed that this wasnt the case in macOS Tahoe 26.2 . Can you kindly suggest us if any flags have to be enabled for the same? Any help on this issue is highly appreciated

I'm writing a read-only filesystem extension. I see that the documentation for loadResource(resource:options:replyHandler:) claims that the --rdonly option is supported, which suggests that this should be possible. However, I have never seen this option provided to my filesystem extension, even if I return usableButLimited as a probe result (where it doesn't mount at all - FB19241327) or pass the -r or -o rdonly options to the mount(8) command. Instead I see those options on the volume's activate call. But other than saving that "readonly" state (which, in my case, is always the case) and then throwing on all write-related calls I'm not sure how to actually mark the filesystem as "read-only." Without such an indicator, the user is still offered the option to do things like trash items in Finder (although of course those operations do not succeed since I throw an EROFS error in the relevant calls). It also seems like the FSKit extensions that come with the system handle read-only strangely as well. For example, for a FAT32 filesystem, if I mount it like mount -r -F -t msdos /dev/disk15s1 /tmp/mnt Then it acts... weirdly. For example, Finder doesn't know that the volume is read-only, and lets me do some operations like making new folders, although they never actually get written to disk. Writing may or may not lead to errors and/or the change just disappearing immediately (or later), which is pretty much what I'm seeing in my own filesystem extension. If I remove the -F option (thus using the kernel extension version of msdos), this doesn't happen. Are read-only filesystems currently supported by FSKit? The fact that extensions like Apple's own msdos also seem to act weirdly makes me think this is just a current FSKit limitation, although maybe I'm missing something. It's not necessarily a hard blocker given that I can prevent writes from happening in my FSKit module code (or, in my case, just not implement such features at all), but it does make for a strange experience. (I reported this as FB21068845, although I'm mostly asking here because I'm not 100% sure this is not just me missing something.)

Hi everyone, I'm building a macOS SwiftUI app and I'm trying to intercept both: Tab Shift + Tab to perform custom actions (similar to how text editors indent/outdent items). Right now, plain Tab works fine, but Shift + Tab never reaches my .onKeyPress(.tab) handler. Here's what I'm currently trying: import SwiftUI struct ShiftTabNotIntercepted: View { @State private var shiftKeyPressed = false var body: some View { Text("Hello") .focusable() .onKeyPress(.tab) { print("tab pressed with shift: \(shiftKeyPressed)") return .handled } .onModifierKeysChanged(mask: .shift) { old, new in shiftKeyPressed = new.contains(.shift) } } } Behavior: Pressing Tab prints: tab pressed with shift: false Pressing Shift + Tab does nothing — .onKeyPress(.tab) never fires. I also noticed: if a sidebar is visible, Shift + Tab moves focus to the sidebar if no sidebar is visible, it still doesn't trigger the handler So it seems macOS is intercepting Shift + Tab for focus navigation before SwiftUI sees it. My goal is to fully own this keyboard behavior for a custom outline/tree editor UI. Questions: Is there a SwiftUI-native way to intercept Shift + Tab? Is .onKeyPress fundamentally unable to capture this combination? Do I need to drop down to AppKit (NSViewRepresentable, keyDown, etc.) to reliably handle it? Thanks!

I'm trying to setup a macOS 26 build environment in a VM (using UTM and the virtualization framework Apple provides). I have Xcode 26 installed and have logged into my Apple ID and verified that the team and other configuration looks fine in Xcode settings. When trying to build the macOS app, I see errors saying the VM's device ID has not been registered. I have confirmed that the device ID is registered both in the Provisioning portal AND the downloaded .provisionprofiles (in Library > Developer > Xcode > UserData). This problem appears on multiple targets (e.g. the main app and extensions). If I try to manually provision the app, using the Provisioning portal, I can build the product, but it will not launch because of Gatekeeper issues. Finally, signing to run locally doesn't work either. As the app launches, frameworks refuse to load because Team IDs don't match. With ad hoc provisioning, there are no Team IDs. I've come to the conclusion that this just isn't possible. Which is a shame because I need to support products with a build environment on macOS 15 and cannot move over to macOS 26 yet. I suspect many developers outside of Apple are in a similar position.