Hi everyone, I’ve run into a consistent issue on multiple Apple Vision Pro devices where downloads using URLSessionConfiguration.background are between 4× and 10x slower than when using URLSessionConfiguration.default. This issue is systematic and can easily be reproduced. This only happens on device, in the simulator, both configurations download files at the expected speed with respect to the network speed. Details: Tested on visionOS 26.0.1 and 26.1 (public releases) Reproduced across 2 Vision Pro (currently testing on a third one) Reproduced on 2 different Wi-fi networks (50mb/s and 880mb/s) From my tests this speed issue seems to affects multiple apps on my device: Stobo Vision (our app), Immersive India, Amplium Not server-related (reproduces with Apple CDN, S3, and DigitalOcean) I’ve built a small sample project that makes this easy to reproduce, it downloads a large file (1.1 GB video) using two managers: One with URLSessionConfiguration.default One with URLSessionConfiguration.background You can also try it with your own file url (from an s3 for example) Expected behavior: Background sessions should behave similarly to default sessions in terms of throughput, just as they do in the simulator. To be clear I am comparing both config when running in the foreground, not in the background. Actual behavior: Background sessions on Vision Pro are significantly slower, making them less usable for large file downloads. On this screenshot it's even reaching 27x slower than the expected speed. Default config takes ~97s to download and Background config takes ~2640s. I do now have the fastest internet connection but 44min to download 90.5MB is extremely slow. Has anyone else seen this behavior or found a workaround? Or is this an expected behavior from URLSessionConfiguration.background? If I'm doing something wrong please let me know Repo link: https://github.com/stobo-app/DownloadConfigTesting

For bookmark creation and resolving, there's a NSURLBookmark{Creation,Resolution}MinimalBookmark enum value. What does this value imply on macOS and iOS? Specifically, to create security scoped bookmarks on macOS, we use NSURLBookmarkResolutionWithSecurityScope, but that is not available on iOS. Are iOS bookmarks always security scoped? Does the minimal enum imply security scoped bookmarks? Is 0 a valid value to bookmarkDataWithOptions, and does that give an even less scoped bookmark than NSURLBookmarkCreationMinimalBookmark`? We are also using NSURLBookmarkCreationWithoutImplicitSecurityScope on both iOS and macOS, to avoid any implicit resolution of bookmarks we resolve, so that we can explicitly control access by explicit calls to start/stopAccessing. How does NSURLBookmarkCreationWithoutImplicitSecurityScope relate to the enum values discussed above? Thanks! (https://mothersruin.com/software/Archaeology/reverse/bookmarks.html provides some really interesting insights, but doesn't discuss the minimal bookmarks.)

If I'm reading this guide correctly, macOS will automatically/implicitly startAccessingSecurityScopedResource for files opened via drag to Dock icon or NSOpenPanel. And I'm expected to call stopAccessingSecurityScopedResource to balance it once I'm done. Does that mean that if I don't, I'm leaking kernel resources as per the docs? If you fail to relinquish your access to file-system resources when you no longer need them, your app leaks kernel resources. If sufficient kernel resources leak, your app loses its ability to add file-system locations to its sandbox, such as with Powerbox or security-scoped bookmarks, until relaunched. What is this limit in practice for macOS and iOS? I've seen number ranging from 1000-2500. From my testing, iOS does not provide the same implicit startAccessingSecurityScopedResource when using UIDocumentPickerViewController. Is this a correct observation/per design? Now, in the cases where I'm creating an NSURL by resolving a saved bookmark, I'm expected to explicitly startAccessingSecurityScopedResource. Based on this, from what I can tell, this means that I can't universally call startAccessingSecurityScopedResource whenever I access a resource by URL, balanced with a stopAccessingSecurityScopedResource when done, as depending on how I got the URL it might already be implicitly started. Is this a correct observation? Do I need to explicitly check whether I'm on iOS (never implicit?), or macOS (sometimes implicit?), and selectively startAccessingSecurityScopedResource based on every call site that may give me a security-scoped file? If so, is there a complete list of the entrypoints that may give me such files (file dialog, drag and drop, etc)? Thanks!

I would like to change the name of the app displayed in the pop-up box during the download process.

var testTwo: Double = 0 testDouble = 80 testTwo = 200 var testThree: Int = 0 testThree = Int(testTwo/testDouble) var testDate: Date = .now var dateComponent = DateComponents() dateComponent.day = testThree var newDate: Date = Calendar.current.date(byAdding: dateComponentwith a thread error , to: testDate)! This code works in a playground. However, when I try to use it in Xcode for my app it fails with the following error: Thread 1: Fatal error: Double value cannot be converted to Int because it is either infinite or NaN I printed the value being converted to Int and it was not NAN or infinite.

When my Share Extension receives an image from the macOS Photos app on Tahoe, the NSItemProvider passes a URL to an image file in a temporary location. All attempts to read that file fail silently, such as with NSImage(contentsOfFile) I can see that the file does exist in Finder. This code did work in previous macOS versions. It seems like a permissions issue, but startAccessingSecurityScopedResource had no effect. Other platforms work, other apps work, such as shares from Finder, which shares via data instead of a url. I'm really stuck. Has anyone else run into this? // make sure provider has a conforming item if itemProvider.hasItemConformingToTypeIdentifier(imageType) { do { let data = try await itemProvider.loadItem(forTypeIdentifier: imageType, options: nil) // data may be image NativeImage, Data, or a URL to an image file. // figure out which by casting, then attempt to load uiImage if let image = data as? NativeImage { print("found NativeImage") self.images.append(image) } else if let data = data as? Data { print("found Data") if let image = NativeImage(data: data) { self.images.append(image) } } else if let url = data as? URL{ print("found URL") if let image = NativeImage(contentsOfFile: url.path) { print("loaded from URL") self.images.append(image) } } }catch{ print("⛔️ Share Extension Error: \(error.localizedDescription)") } }

I'm working on an API client for a REST service that uses a custom token-based authentiation scheme. The app hits a specificed authentication endpoint with a username and password, said endpoint returns a token that's good for X amount of time, and the app passes that token along with every subsequent request. When that token expires, we start over.Most literature out there tells me to manually set the Authorization header on my request, but official Apple documentation discourages this, as that header is meant to be 'owned' by the built-in HTTP loading system. That said, official documentation on the 'correct' way to do this is shockingly lacking, and the standard didReceiveChallenge callbacks seem better suited for non-custom Basic/Digest/etc authentication schemes. One thought I had was registering my own URLProtocol subclass to handle our custom flow. However, while I haven't had a chance to sit down and take a crack at that yet, my understanding from skimming these forums is that it's suffering from some bit-rot right now, so it 'might' (?) not be the best choice. That, and it's also not clear to me whether the rules around the Authorization header change when a custom URLProtocol is in play.So, community (paging eskimo in particular!), what's the correct way for me to go about this?

I regularly receive emails from my customers asking why my app, which copies files from a source folder to a destination folder, gives an error during the copy operation, usually when some kind of NAS is involved. I then ask if copying the same file works in the Finder and they usually say yes. Then it’s up to me to contact the NAS vendor and ask why their NAS doesn’t work correctly with the copyfile function which my app uses to copy files, and provide them with a sample Xcode project that reproduces the issue. It would be so much easier for me, and probably other developers as well, if my customers could test their NAS with the Finder, then if it doesn’t work, contact their NAS vendor and solve the issue, or if it works, rest assured that file copies will work for any other app that uses the same API as the Finder. I could spend my time doing much more productive and interesting work rather than checking why every other NAS doesn’t work with my app which, after all, uses the most basic file copy mechanism available on macOS. Is there any chance that in the future (hopefully sooner than later) there will be an API to copy files like the Finder with progress information, or is there a reason why this won't ever be possible? I created FB20929181.

Hi, We can easily get drive throughput using the iostat command, but it only outputs plain text that needs to be parsed, and I’m not sure if the format or column order is consistent across macOS versions. Is there any API that provides the same disk I/O metrics that iostat reports, but in a way that can be safely called from a notarized app?

Hello, I have a question about data deserialization using NSKeyedUnarchiver in iOS SDK development. Current Situation: Previously, we were using the NSKeyedUnarchiver.unarchiveObject(with: Data) function We have changed to using the NSKeyedUnarchiver.unarchivedObject(ofClasses:from:) method to deserialize complex objects stored in UserDefaults We need to include all types in the ofClasses parameter, including Swift primitive types as well as various custom classes and structs within the project Questions: Implementation Approach: Is it correct pattern to include all classes defined in the project in the ofClasses array? Is this approach recommended? Runtime Stability: When using this approach, is there a possibility of runtime crashes? Are there any performance issues? Alternative Methods: If the current approach is not the correct pattern, what alternatives should we consider? Current Code Structure: All model classes conform to the NSSecureCoding protocol We use the requiringSecureCoding: true parameter We use a whitelist approach, explicitly listing only allowed classes I would like to know if this structure is appropriate, or if we should consider a different approach. Thank you.

Hi, I have a macOS app distributed through the App Store that uses an app group to share data with app extensions. The group identifier has the form: .group. In Xcode 26 I am now asked to convert the project to profile based app groups (like on iOS). My question is: Can I convert the project to profile based app groups and will the existing app group (which is prefixed with the Team ID) continue to work (and will exiting users still be able to access their data). If yes, should I add the app group with or without the Team ID prefix to the profile.

Hi, I've tried adding Extensions (using ExtensionFoundation) to one of my iOS apps. They work using Xcode on my devices, but when I upload them to the AppStore, I get the following error: ITMS-90349: Invalid Info.plist value - The value of the EXExtensionPointIdentifier key, AsheKube.app.a-Shell.localWebServer, in the Info.plist of “a-Shell.app/Extensions/localWebServer.appex” is invalid. Please refer to the App Extension Programming Guide at https://developer.apple.com/library/content/documentation/General/Conceptual/ExtensibilityPG/Action.html#/apple_ref/doc/uid/TP40014214-CH13-SW1. I could not find relevant information on the web page pointed. I've tried different values for the "EXExtensionPointIdentifier", but none of them work. This value was generated by Xcode, since I did set the user-defined build setting EX_ENABLE_EXTENSION_POINT_GENERATION = YES, as explained in the instructions. At this point, I have no idea how to fix this issue.

Hi, I'm trying to add an extension to my app on iOS 26. I've followed the instructions on https://developer.apple.com/documentation/extensionfoundation/adding-support-for-app-extensions-to-your-app and made it as far as being able to launch the extension: let monitor = try await AppExtensionPoint.Monitor(appExtensionPoint: .localWebServerExtension) currentIdentity = monitor.identities.first if let currentIdentity = currentIdentity { let myConfig = AppExtensionProcess.Configuration(appExtensionIdentity: currentIdentity, onInterruption: { NSLog("extension was terminated") }) myProcess = try await AppExtensionProcess(configuration: myConfig) myConnection = try myProcess?.makeXPCConnection() } None of these calls throw, and when I examine myProcess from inside that code, it seems to be normal (there's a pid, for example). Yet the code inside my extension seems to not be executed: breakpoints are not triggered, NSLog() calls do not appear on the console. The onInterruption() callback is also not triggered, or at least it does not appear on the console either. I've probably missed something obvious, but what could it be?

I am working with the rich TextEditor introduced in iOS 26, but I am having trouble preserving AttributedString formatting when converting to/from RTF. Here is my exporting logic in my view model (AttributedString to RTF) let nsAttrStr = NSAttributedString(self.text) // text is an AttributedString (bound to the TextEditor input) let range = NSRange(location: 0, length: nsAttrStr.length) let options: [NSAttributedString.DocumentAttributeKey: Any] = [ .documentType: NSAttributedString.DocumentType.rtf ] guard let data = try? nsAttrStr.data(from: range, documentAttributes: options) else { return nil } let rtfBase64 = data.base64EncodedString() When I inspect the result, it seems to lose the font, size, boldness, etc which is being correctly rendered in the TextEditor. When I convert back from RTF to an AttributedString, it reverts to the default text formatting applied in the TextEditor. Any ideas what could be going wrong?

I'm working in Xcode for the first time in over a decade (coming from VS Code and Zed), so bear with me on this one. I've just started up a new project and got things running, and I'm adding the following struct to my project: struct HttpRequest: Codable { let body: String? let headers: [String: String] let id: UUID let method: HttpMethod let url: String } The UUID type is from Foundation, of course. However, when I hit Ctrl+Space to auto-complete it, it says "Foundation is not imported." If I hit enter to select that entry, it does not auto-import Foundation, and gives me an error that it "Cannot find type 'UUID' in scope". I can click that error and it then click "import Foundation" to add the import, but this seems like a lot of effort just to "automatically" add an import that Xcode is already aware I need to add. On the other hand, I found a number of threads online during my search for answers that suggested I shouldn't even need this import, despite Xcode's insistence. TL;DR, my questions are: 1) Do I need the Foundation import to use UUID in Swift? and 2) If I do need it, how can I auto-import the Foundation module when I auto-complete the UUID type?

crash.txt

What’s the recommended way to recursively walk through a directory tree using File Coordination? From what I understand, coordinating a read of a directory only performs a “shallow” lock; this would mean that I’d need to implement the recursive walk myself rather than use FileManager.enumerator(at:includingPropertiesForKeys:options:errorHandler:) plus a single NSFileCoordinator.coordinate(with:queue:byAccessor:) call. I’m trying to extract information from all files of a particular type, so I think using NSFileCoordinator.ReadingOptions.immediatelyAvailableMetadataOnly on each file before acquiring a full read lock on it (if it’s the right file type) would make sense. Am I on the right track?

Hi, We are running into issues with iOS app prewarming, where the system launches our app before the user has entered their passcode. In our case, the app stores flags, counters, and session data in UserDefaults and the Keychain. During prewarm launches: UserDefaults only returns default values (nil, 0, false). We have no way of knowing whether this information is valid or just a placeholder caused by prewarming. Keychain items with kSecAttrAccessibleAfterFirstUnlockThisDeviceOnly are inaccessible, which can lead to broken business logic (the app can assume no session exists). No special launch options or environment variables appear to be set. We can reproduce this 100% of the time by starting a Live Activity in the app before reboot. Here’s an example of the workaround we tried, following older recommendations: __attribute__((constructor)) static void ModuleInitializer(void) { char* isPrewarm = getenv("ActivePrewarm"); if (isPrewarm != NULL && isPrewarm[0] == '1') { exit(0); // prevent prewarm launch from proceeding } } On iOS 16+, the ActivePrewarm environment variable doesn’t seem to exist anymore (though older docs and SDKs such as Sentry reference it). We also tried listening for UIApplication.protectedDataDidBecomeAvailableNotification, but this is not specific to prewarming (it also fires when the device gets unlocked) and can cause watchdog termination if we delay work too long. Questions: Is there a supported way to opt out of app prewarming? What is the correct way to detect when an app is being prewarmed? Is the ActivePrewarm environment variable still supported in iOS 16+? Ideally, the UserDefaults API itself should indicate whether it is returning valid stored values or defaults due to the app being launched in a prewarm session. We understand opting out may impact performance, but data security and integrity are our priority. Any guidance would be greatly appreciated.

Hi there, We are facing some issues regarding TLS connectivity: Starting with iOS 26, the operating system refuses to open TLS sockets to local devices with self-signed certificates over Wi-Fi. In this situation, connection is no longer possible, even if the device is detected on the network with Bonjour. We have not found a workaround for this problem. We've tryied those solutions without success: Added the 'NSAppTransportSecurity' key to the info.plist file, testing all its items, such as "NSAllowsLocalNetworking", "NSExceptionDomains", etc. Various code changes to use properties such as "sec_protocol_options_set_local_identity" and "sec_protocol_options_set_tls_server_name" to no avail. Brutally import the certificate files into the project and load them via, for example, "Bundle.main.url(forResource: "nice_INTERFACE_server_cert", withExtension: "crt")", using methods such as sec_trust_copy_ref and SecCertificateCopyData. Download the .pem or .crt files to the iPhone, install them (now visible under "VPN & Device Management"), and then flag them as trusted by going to "Settings -> General -> Info -> Trust". certificates" The most critical part seems to be the line sec_protocol_options_set_verify_block(tlsOptions.securityProtocolOptions, { $2(true) }, queue) whose purpose is to bypass certificate checks and validate all of them (as apps already do). However, on iOS26, if I set a breakpoint on leg$2(true),` it never gets there, while on iOS 18, it does. I'll leave as example the part of the code that was tested the most below. Currently, on iOS26, the handler systematically falls back to .cancelled: func startConnection(host: String, port: UInt16) { self.queue = DispatchQueue(label: "socketQueue") let tlsOptions = NWProtocolTLS.Options() sec_protocol_options_set_verify_block(tlsOptions.securityProtocolOptions, { $2(true) }, queue) let parameters = NWParameters(tls: tlsOptions) self.nwConnection = NWConnection(host: .init(host), port: .init(rawValue: port)!, using: parameters) self.nwConnection.stateUpdateHandler = { [weak self] state in switch state { case .setup: break case .waiting(let error): self?.connectionDidFail(error: error) case .preparing: break case .ready: self?.didConnectSubject.onNext(Void()) case .failed(let error): self?.connectionDidFail(error: error) case .cancelled: self?.didDisconnectSubject.onNext(nil) @unknown default: break } } self.setupReceive() self.nwConnection.start(queue: queue) } These are the prints made during the procedure. The ones with the dot are from the app, while the ones without are warnings/info from Xcode: 🔵 INFO WifiNetworkManager.connect():52 - Try to connect onto the interface access point with ssid NiceProView4A9151_AP 🔵 INFO WifiNetworkManager.connect():68 - Connected to NiceProView4A9151_AP tcp_output [C13:2] flags=[R.] seq=215593821, ack=430284980, win=4096 state=CLOSED rcv_nxt=430284980, snd_una=215593821 nw_endpoint_flow_failed_with_error [C13 192.168.0.1:443 in_progress channel-flow (satisfied (Path is satisfied), viable, interface: en0[802.11], dns, uses wifi, LQM: unknown)] already failing, returning nw_connection_copy_protocol_metadata_internal_block_invoke [C13] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C13] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_connected_local_endpoint_block_invoke [C13] Client called nw_connection_copy_connected_local_endpoint on unconnected nw_connection nw_connection_copy_connected_remote_endpoint_block_invoke [C13] Client called nw_connection_copy_connected_remote_endpoint on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C14] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C14] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_connected_local_endpoint_block_invoke [C14] Client called nw_connection_copy_connected_local_endpoint on unconnected nw_connection nw_connection_copy_connected_remote_endpoint_block_invoke [C14] Client called nw_connection_copy_connected_remote_endpoint on unconnected nw_connection [C14 192.168.0.1:443 tcp, tls, attribution: developer] is already cancelled, ignoring cancel [C14 192.168.0.1:443 tcp, tls, attribution: developer] is already cancelled, ignoring cancel nw_connection_copy_protocol_metadata_internal_block_invoke [C15] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C15] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_connected_local_endpoint_block_invoke [C15] Client called nw_connection_copy_connected_local_endpoint on unconnected nw_connection nw_connection_copy_connected_remote_endpoint_block_invoke [C15] Client called nw_connection_copy_connected_remote_endpoint on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C16] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_protocol_metadata_internal_block_invoke [C16] Client called nw_connection_copy_protocol_metadata_internal on unconnected nw_connection nw_connection_copy_connected_local_endpoint_block_invoke [C16] Client called nw_connection_copy_connected_local_endpoint on unconnected nw_connection nw_connection_copy_connected_remote_endpoint_block_invoke [C16] Client called nw_connection_copy_connected_remote_endpoint on unconnected nw_connection [C16 192.168.0.1:443 tcp, tls, attribution: developer] is already cancelled, ignoring cancel [C16 192.168.0.1:443 tcp, tls, attribution: developer] is already cancelled, ignoring cancel 🔴 ERROR InterfaceDisconnectedViewModel.connect():51 - Sequence timeout.

When I run this in a playground: var meDate = Calendar.current.date(from: DateComponents(year: 2024, hour: 7, weekday: 3, weekdayOrdinal: 2))! print(meDate) I see: 2024-01-09 15:00:00 +0000 This seems correct to me. jan 9th is the second Tuesday in 2024 I'm in the pacific TZ, 07:00 PDT matches 15:00GMT But then I do this: meDate = Calendar.current.date(bySetting: .weekday, value: 4, of: meDate)! print(meDate) and I see: 2024-01-10 08:00:00 +0000 I would have expected my hour value (7PST/15GMT) to have been preserved. Is there a way I can update weekday, but not lose my hour?