I’ve got some code that creates an AVPlayerItem from a URL, the creates an AVQueuePlayer from it. If I check the player item's status after that, it's still unknown. According to the docs, it'll remain unknown until it is associated with an AVPlayer, and then it "immediately begins enqueuing the item’s media and preparing it for playback." But checking the status right after that, I still get unknown, which tells me it’s not quite immediate. Is there any way to test if the player item will work immediately after creation? In this case, the problem is that my app doesn't have permission, due to it being a bookmark saved in a sandboxed app.

I've filed FB13497581 on this topic. But Ever since macOS 14 (or perhaps one of the releases that came after), my machine (M1 MacBook Pro) frequently locks up the UI. Sometimes it's so bad the mouse won't move. It only lasts a few seconds, but at its worst it happens every few seconds. It usually manifests as no text appearing as I type (which sucks, since I write code all day). I'm fairly certain configd is to blame. When this occurs, configd is using nearly 100% CPU, and spams logs with thousands of messages per second. Killing it does nothing. Rebooting might fix it for a bit, but it eventually comes back. I'm wondering if there’s some on-disk state I can blow away to start SystemConfiguration from scratch. I hate to do that in case someone from Apple asks me for files to analyze, but I feel like my FB is just going to languish. Meanwhile, this is rendering my machine unusable. Or if anyone knows how to interpret the log messages.

I wrote a StoreKit unit test and set the renewal rate to .oneRenewalEveryTwoSeconds for the test session. But now my App expires and renews every two seconds when running normally, even though the StoreKit Configuration file is set to "Real Time." Changing it to anything else doesn't change the renewal rate. I've tried cleaning the build folder. If, however, I set my SKTestSession explicitly to .realTime and run the test again, then my app behaves. However, again, it doesn't matter what I set the Configuration Settings to. Anyone know where this information is stored?

Because I can't find a way to share StoreKit purchases between two apps (macOS), I was forced to make my LSUIElement application a single app. This is turning out to work reasonably well, except for one issue that I hadn't noticed before (maybe it's new in 14.2?). The normal flow is this: User double-clicks app icon App calls SMAppService.agent(plistName: "com.myagent.plist").register(). The referenced plist points to the same application binary (see attached project) that the user launched in step 1. System launches a second instance of the application binary What I want the call to do instead is see that the app is already running, and just adopt it (that is, if it is killed or crashes, relaunch it). I see two inelegant workarounds: Wait until the user is done configuring the app after first launch, register the app and then quit. Thing is, I can't be sure when they're done, e.g. if they don't close the window. Allow both instances of the app to run, but quit if they close the window and it wasn't launched by launchd, or if there are other instances running. This feels fragile (e.g. if the behavior of register() is ever improved as I'd like). The best solution, of course, would be to share StoreKit context between apps, so I could separate this into the UI app and background agent. (To Apple: I've filed FB13574819 with my project attached.)

How do I respond to a SpatialTapGesture in my RealityView when the tap is on no entity whatsoever? I tried just doing RealityView {} .gesture(SpatialTapGesture().onEnded { print("foo") }) but that doesn't get called. All I can find searching is advice to add Collision and Input components to entities, but I don't want this on an entity; I want it when the user is not looking at any specific entity.

My app is a login item that gets registered as a background agent using SMAppService. On launch it checks to see that the user has an active subscription, and when I run it from Xcode, this works fine. But it seems that if I restart my machine and it is run from launchd right at login, StoreKit doesn't return any purchases. This results in my app throwing up a window to tell the user to subscribe. It’s hard to be sure this is the behavior, because I can't really use a debugger, and for whatever reason, OSLog is not persisting messages (despite being configured to do so). Does anyone know if it takes a while after restart or login for Store Kit to reliably return purchases?

I'm having an issue with logging on my system (Sonoma 14.3.1), in that log messages disappear after 60-90 seconds, despite logging being configured to persist. This seems to happen across all subsystems. The symptoms are: • Watching the log stream in Console.app, messages older than 60-90 seconds disappear, even if no new messages are coming in. • log show --last 2d only returns messages from the last couple of minutes. I've filed FB13616761, but wondered if anyone had any other insights or suggestions. TIA

It seems cross-device passkey authentication isn't supported by the Simulator, is that right? I can't tell if it's the simulator not supporting it, or my app/server failing. My app, running in the Simulator, presents a QR code during sign-in flow. I scan that with my phone, and it just spins "connecting…". Can anyone confirm that the Simulator just doesn't support this? What are my options for testing this flow? I guess I need another device not sign in to my AppleID.

I just started a little macOS app using SwiftUI and SwiftData. It seems to insist on placing the name of the file (along with the extension) in the title bar. I would prefer to put the name of the selection there instead (i.e. instead of "Untitled.paper", "New Article").. Along with that, I'd like to control the appearance of "Edited" as well. I have the suspicion that this simply isn't possible.

I'm trying to communicate with an RS-485 device using a USB-to-RS-485 adapter based on a CP21012N. It's capable of speeds up to 3 Mbaud. However, when I call tcsetattr() with speed set to anything other than one of the predefined constants (using cfsetspeed()), I get an "Invalid argument" error back from the call. Googling, I found https://github.com/avrdudes/avrdude/issues/771 Seems like macOS really can't accept baud rates outside the predefined set without resorting to ioctl?

_IOW and friends, defined in <sys/ioccom.h>, help build ioctl() calls. Is there an equivalent function in Swift?

I don't know why we’re up to Xcode 16 and this stuff is still so damn difficult. First of all, I don't know why I can't just send a .app I built for my M1 MacBook Pro to my friend who also has an M1 MacBook Pro. But even after going through the quarantine steps, he gets an alert saying the app can't be opened. So I'm trying to do direct distribution of an archive. But that gives me two errors: There is a problem with the request entity You already have a current Developer ID Application Managed (With Kext) certificate or a pending certificate request. No profiles for 'com.latencyzero.VideoBox' were found Xcode couldn't find any Developer ID provisioning profiles matching 'com.latencyzero.VideoBox'. The signing is managed by Xcode. CloudKit access works.

Bit of a stream of consciousness below. I discovered after writing all that that I needed to switch width & height when setting values, and print landscape. That gets things to print as expected. But I really need to be able to set the Brother's notion of tape width in the Basic Options portion of the print panel. I've got this Brother QL-800 USB label printer that takes spools of either die-cut labels, or continuous paper of a given width. It can cut between labels or the continuous paper at any point. My app is intended to print small labels for small parts drawers. The app has a DB of parts, and I want to print dozens of unique labels of a specific height and width. Ideally, you'd set things up once, and then print one or more specific part labels without further UI interaction. The NSPrintPanel is far too complex, and allows the user to screw things up, so I'd like to forego it altogether. I've written some simple code that does succeed in printing via the NSPrintPanel, but it’s positioning the text ultimately incorrectly. I can't figure out how to properly size a view to print a label. I can get an NSPrintInfo object that knows the label printer is selected (e.g. it shows the right name for NSPrintInfo.printer.name). But it states the paperSize is (612.0, 792.0), which if I assume 72 ppi, comes out to 8.5" x 11", which this paper clearly is not. func printLabel() { Task { let printInfo = NSPrintInfo() let printer = printInfo.printer print("Paper size: \(printInfo.paperSize)") // printInfo.paperSize = CGSize(width: 47 / 25.4 * 72.0, height: 12.0 / 25.4 * 72.0) printInfo.topMargin = 0.0 printInfo.bottomMargin = 0.0 printInfo.leftMargin = 0.0 printInfo.rightMargin = 0.0 printInfo.orientation = .landscape printInfo.isHorizontallyCentered = false printInfo.isVerticallyCentered = false printInfo.scalingFactor = 1.0 let view = NSHostingView(rootView: createLabelPrintView()) view.frame.size = CGSize(width: 25 / 25.4 * 72.0, height: 12.0 / 25.4 * 72.0) let op = NSPrintOperation(view: view, printInfo: printInfo) op.showsPrintPanel = true op.showsProgressPanel = true op.printPanel.options.insert(.showsPaperSize) op.printPanel.options.insert(.showsOrientation) op.run() } } @ViewBuilder func createLabelPrintView() -> some View { VStack(spacing: 0) { Text("Line 1") Text("Line 2") } .padding(0) } If I set the pageSize explicitly (uncommenting the line above), The printer offers a long list of available sizes (although none I've found corresponds to the 0.47"/12 mm x continuous spool). By the way, it never remembers the size I selected, and always reverts to 8.5 x 11, which it will then refuse to print, hanging in Print Center with an error about mismatched page sizes. I don’t know why the available page sizes have multiple sizes that don’t seem to match. E.g. 0.47" In the printer’s “Basic Options,” it shows the notion of a Width that's "not found," if I use a custom size: I don't really understand where those printer options can be found and set. Hmm, maybe I found a clue: I ran an NSPrintPanel by itself and examined the NSPrintInfo afterward: let panel = NSPrintPanel() panel.options.insert(.showsPaperSize) panel.options.insert(.showsOrientation) let result = await panel.beginSheet(using: printInfo, on: NSApplication.shared.windows.first!) print("Result: \(result)") print("Paper size: \(printInfo.paperSize)") print("Settings: \(printInfo.printSettings)") Paper size: (283.44000244140625, 82.08000183105469) Settings: { "AP_D_InputSlot" = ""; BrAutoTapeCut = ON; BrBiDiPrint = ON; BrBrightness = 0; BrContrast = 0; BrCutAtEnd = ON; BrCutLabel = 1; BrHalftonePattern = BrErrorDiffusion; BrMargin = "3.0"; BrMultiColor = BrMultiColorMonochrome; BrRedLevel = 0; BrRemoveBlkSpace = OFF; BrResolution = BrSpeed300x300dpi; BrTapeLength = "69.1"; ColorModel = Gray; Duplex = None; PaperInfoIsSuggested = 0; Resolution = 300x300dpi; "com_apple_print_DialogDismissedBy" = Print; "com_apple_print_DocumentTicket_PMSpoolFormat" = "application/pdf"; "com_apple_print_PDEsUsed" = "Printer Options"; "com_apple_print_PageToPaperMappingMediaName" = 29mm; "com_apple_print_PageToPaperMappingType" = 1; "com_apple_print_PrintSettings_PMCopies" = 1; "com_apple_print_PrintSettings_PMCopyCollate" = 1; "com_apple_print_PrintSettings_PMDestinationType" = 1; "com_apple_print_PrintSettings_PMDuplexing" = 1; "com_apple_print_PrintSettings_PMFirstPage" = 1; "com_apple_print_PrintSettings_PMLastPage" = 2147483647; "com_apple_print_PrintSettings_PMPageRange" = ( 1, 2147483647 ); "com_apple_print_ticket_type" = "com.apple.print.PrintSettingsTicket"; "com_apple_print_totalPages" = 2147483647; } It has added a bunch of Br-prefixed keys that clearly correspond to some of the options in the UI. But I'm not sure if I can set those programmatically and get things to print the way I expect. And I don't see anything for tape width. How do I get an NSPrintPanel to show the preview without an NSPrintOperation? I want to react to any changes the user might make before things go to print.

Every now and again, I see that Python 3.8 crashes. Now, I tried to get rid of all Python 3.8 interpreters on my system. But the crash log for this one gives a Path of /Users/USER/*/python3.8. What does that mean? That's not a real path on my filesystem. The crash is due to Python not finding a dylb it needs: Process: python3.8 [2946] Path: /Users/USER/*/python3.8 Identifier: python3.8 Version: ??? Code Type: X86-64 (Translated) Parent Process: launchd [1] User ID: 501 Date/Time: 2024-09-20 17:03:49.4230 -0700 OS Version: macOS 14.6.1 (23G93) Report Version: 12 Anonymous UUID: <redacted> Sleep/Wake UUID: <redacted> Time Awake Since Boot: 520000 seconds Time Since Wake: 43861 seconds System Integrity Protection: enabled Notes: PC register does not match crashing frame (0x0 vs 0x202E45A78) Crashed Thread: 0 Exception Type: EXC_CRASH (SIGABRT) Exception Codes: 0x0000000000000000, 0x0000000000000000 Termination Reason: Namespace DYLD, Code 1 Library missing Library not loaded: /usr/local/opt/gettext/lib/libintl.8.dylib Referenced from: <redacted> /Users/USER/*/python3.8 Reason: tried: '/usr/local/opt/gettext/lib/libintl.8.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/usr/local/opt/gettext/lib/libintl.8.dylib' (no such file), '/usr/local/opt/gettext/lib/libintl.8.dylib' (no such file), '/usr/local/lib/libintl.8.dylib' (no such file), '/usr/lib/libintl.8.dylib' (no such file, not in dyld cache) (terminated at launch; ignore backtrace) Now, this only happens (I think) when I'm running VS Code, which I normally hate doing, but I’m trying to develop an Autodesk Fusion plug-in in Python, and that’s how it’s done. I don’t know if this Python was installed by Apple, or if I got via one of the dozens of SDKs I've installed over the years. VS Code has a way to search for Python interpreters, and I used that to delete all the old ones I could find. And I'm pretty sure macOS is on a more recent version.

There are a couple of questions in here. I want to do a basic master-detail split view, as so many SwiftUI apps do. But I want to let the user select an arbitrary subset of items in the list, select them all, and delete them (or do other operations). Select All, by default, is excruciatingly slow. I have a List view with a selection binding to a Set<PersistentIdentifier>. A ForEach inside renders the items from a SwiftData @Query. I have a couple hundred items, each with a String and Date property. The list item renders the String and nothing else. I click on an item, hit Command-A, and the app locks up for several seconds while it messes with the selection. It never highlights all the items, but sometimes highlights one (usually different from the one I clicked on).I have an .onChange(of: self.selection) in there to debug the selection. It is called many, many times (2-3 times per second, very slowly), and print the selection count, which starts at 135, and goes down by one, to about 103. If I scroll the list, the selection onChange gets called a bunch more times. Sometimes you see the selection highlights change. My SwiftUI view hierarchy looks like this: NavigationSplitView OrdersList struct OrdersList : View { var body: some View { List(selection: self.$selection) { ForEach(self.orders) { order in NavigationLink { Text("Order ID: \(order.vendorOrderID ?? "<none>")") } label: { Text("Order ID: \(order.vendorOrderID ?? "<none>")") } } .onDelete(perform: self.deleteOrders) .onChange(of: self.selection) { print("Selection changed: \(self.selection.count)") } } } func deleteOrders(offsets: IndexSet) { withAnimation { for index in offsets { self.modelContext.delete(self.orders[index]) } } } @State var selection = Set<PersistentIdentifier>() @Query var orders : [Order] @Environment(\.modelContext) var modelContext } If I use the mouse to select a single item, it behaves as expected. The item is selected, the onChange gets called, and if I choose Edit->Delete, the onDelete handler is called. If I use the mouse to shift-select multiple items, the behavior is similar to Select All above, but the selection Set count starts at the number of items I selected, and then is slowly whittled down to 0 or 1, with some random item being selected. I’d like for the Edit->Delete command to work with a keystroke (delete). How can I set the shortcut on an existing menu item without having to reimplement its behavior?