Whenever I call this method on the PHPhotoLibrary class I get a warning which I've included below. I've tried looking in the Apple forums, Stack Overflow etc. and nobody seems to mention this issue that I can find. Is this warning unavoidable or am I using the API wrong? I can reproduce this trivially in a simple untitled project. Warning: Retain cycle risk: client-side managed object fetch on long-lived context for entity Asset, on context named: PHPhotoLibrary (inode=19657147) .Generic, consider using -[PHPhotoLibrary performAndWaitOnAutoResettingLibraryWithName:workBlock:] or +[PLDatabaseContext newShortLivedLibraryWithName:bundle:]...work, or using a dictionary result type, or suppressing this warning using 'clientSideLongLivedFetchWarningsSuppressed'). STEPS TO REPRODUCE Create a new untitled App project in Xcode. Set a photo library usage privacy description in the Info for the project target Replace ContentView.swift with the code included below. Make sure there is at least one photo in the photo library on the target device. A simulator is fine. Run the app Click the Get Cloud ID button, respond to the authorization dialog Click the Get Cloud ID button again Look in Xcode and see the warning import SwiftUI import Playgrounds import Photos struct ContentView: View { var body: some View { Button( action: getCloudID ) { Text("Get Cloud ID") } .padding() } private func getCloudID() { var fetchOptions = PHFetchOptions() fetchOptions.fetchLimit = 1 let assetResult = PHAsset.fetchAssets(with: .image, options: fetchOptions) var localIds = [String]() assetResult.enumerateObjects() { asset, _, _ in localIds.append(asset.localIdentifier) } let mappings = PHPhotoLibrary.shared().cloudIdentifierMappings(forLocalIdentifiers: localIds ) for mapping in mappings { debugPrint("cloud ID \(mapping.value)") } } } `

Hello, I develop an audio processing demo app that has been working fine on iOS/iPad and on Mac ("developed for iPad") until I upgraded to MacOS27. The app is demonstrating low latency / high temporal resolution audio spectral analysis. I am using AVAudioSession, I request the audio IO buffer to be 32 samples. This still works fine on iOS/iPadOS, but it seems that when running on MacOS27 the buffer size always defaults to 512 samples no matter what I request (smaller or larger). This applies to the "Designed for iPad" version running on MacOS and any version running in the iPhone and iPad simulators on MacOS. The app seems to be working as intended on physical iPhones and iPads. I made a catalyst version and used AudioObjectSetPropertyData to force the IO bugger sizes to 32 samples on the system's active hardware device. This seems to succeed but still results in incorrect behaviour in my app; I am still diagnosing but I believe it's because this leaves AVAudioSession with incorrect data, and I am using AudioKit which relies on AVAudioSession (not confirmed). Regardless, is this change in behaviour intentional in MacOS27? Is there a way around it? The whole point of my app is to demonstrate efficient audio processing with very low latency - I need small input buffer size for that. Thank you for any help with this

I’m building a personal iPhone app that organizes my music library. When I tap an album in my app, I want Apple’s Music app to open directly to that album’s page in Library, showing the tracks synced from my computer. My app can already read the albums and their tracks using MusicKit. I’ve tried normal Apple Music sharing links, but those open the catalog version of the album. I need the page I see when I find the album through Library in the Music app. Starting or changing playback is fine, but the album page must appear. Only starting the music or opening Now Playing would not be enough. Is there a supported API or link format that can do this? A documentation link or a small code example would be helpful.

Hello, I'm developing a video application for Apple platforms and would like to implement accurate Apple Log and Apple Log 2 color conversion using official Apple APIs. Final Cut Pro and Final Cut Camera support SDR Rec.709 and HDR Rec.2020 HLG monitoring. I have found Apple's official Apple Log white papers and Rec.709 LUTs, but I could not find an official HDR HLG LUT or a public API that provides equivalent color conversion. I have three questions: Does Apple provide a public API through AVFoundation, Core Image, ColorSync, or another framework to convert Apple Log / Apple Log 2 to SDR Rec.709 and HDR Rec.2020 HLG? Are the color transformations used by Final Cut Pro and Final Cut Camera available to third-party developers, either through public APIs or official LUT resources? If no such API exists, is implementing the transfer functions and gamut conversions from Apple's white papers, followed by standard SDR/HDR rendering, the recommended approach? My goal is to preserve the original Apple Log footage while allowing independent SDR/HDR monitoring without modifying the recorded data. Any guidance, documentation, or sample code from Apple engineers would be greatly appreciated. Thank you!

Hi, We are building a network audio device running embedded Linux (headless — no display, no user-facing browser UI). I'd like to clarify whether MusicKit JS can be used in this environment, both technically and from a licensing standpoint. Our setup Embedded Linux on ARM, no screen Headless Chromium running as a background service MusicKit JS loaded in a page to authenticate the user and play Apple Music Questions Is MusicKit JS supported in a headless browser context, or is it intended only for interactive, user-facing browser sessions? MusicKit JS requires a Developer Token with an origin claim, and a valid web origin. If the page is served locally on the device (e.g. http://127.0.0.1) with no public domain, what origin should be used? Is a local origin acceptable? Is embedding MusicKit JS inside device firmware — where the browser is never visible to the user and playback is not initiated through a normal browsing session — within the MusicKit terms of use? I want to confirm whether running MusicKit JS in headless Chromium would be considered an acceptable implementation。 For context: we are already MFi licensees with AirPlay 2 integrated. We simply want to be certain any additional work stays within the authorized scope. Any guidance — including a pointer to the relevant section of the MusicKit terms — would be much appreciated. Thanks in advance

Installed the latest beta releases of TVos and Home pod OS. 27.2 (24k122) The problem of audio cutting out whilst video keeps playing persists. The set up is: Mac with homeshare enabled, MP4 video in TV app on Mac encoded with ffmpeg, Apple tv 4k tvos 27, tv, Two homepods 27, in stereo pair. Problem: When playing a video from the mac on the tv via the appletv on the computers app the audio cuts out whilst the video keeps playing. Pausing and playing the video restores the video only for it to cut-out again and again. Switching audio from homepods to tv speakers there is no cut-out. Other sources play fine through the apple tv, streaming appletv, iplayer etc. The homepods work fine when watching tv with homepods using a satellite box. This only started when the TvOS was upgraded to 27, 26 was fine. It's not a problem with the Macs as I use two, one an old Mac pro and a new Mac mini, the problem is on both. I don't think it's a problem with the homepods as the play music and other sources fine. I think it's the update in TvOS from 26 to 27 that introduced the bug/problem. I'm going to submit another feedback report with the above details again.

When wrapping CABTMIDICentralViewController in a UIViewControllerRepresentable and presenting modally, I experience layout crashes since iOS 27. Does anyone have a workaround? import SwiftUI import CoreAudioKit struct BluetoothButton: View { @State var showsBluetooth = false let title = "Check for Bluetooth MIDI Devices" var body: some View { #if os(macOS) Button { let wc = CABTLEMIDIWindowController() wc.showWindow(nil) } label: { label } #else Button { showsBluetooth = true } label: { label } .sheet(isPresented: $showsBluetooth) { BluetoothCentralView() } #endif } var label: some View { Text(title) } } #if os(iOS) struct BluetoothCentralView: UIViewControllerRepresentable { func makeUIViewController(context: Context) -> CABTMIDICentralViewController { return CABTMIDICentralViewController() } func updateUIViewController(_ uiViewController: CABTMIDICentralViewController, context: Context) { } } #endif

The set bitrate is not respected when kVTCompressionPropertyKey_AverageBitRate is used. constant and variable bitrates seems to work, only average that is broken. All three modes works in iOS 26.

Hi everyone, I’m developing a native Apple TV application using MusicKit for Swift. The app provides a custom interface for browsing and playing the user’s personal Apple Music library. The application was working correctly for several months, but recently certain songs from the user’s personal library have stopped playing. I’ve been investigating this issue extensively and have reproduced it in a completely separate, minimal tvOS test application. The issue I retrieve songs using MusicLibraryRequest() and play them using ApplicationMusicPlayer. Some songs play successfully, while others consistently fail during playback preparation with: `MPMusicPlayerControllerErrorDomain Code=6 "Prepare to play failed"` The underlying errors include : NSCocoaErrorDomain Code=4099 Example of a failing song Michael Jackson — Bad (2012 Remaster) Library song ID: i.ldDOAp0UrAa62J Successfully retrieved using MusicLibraryRequest() Fails when queued with ApplicationMusicPlayer Also fails with SystemMusicPlayer Example of a working song Harry Nilsson — Everybody’s Talkin’ (1971 Mix) Library song ID: i.WmYQp9PcdZ4Y0E Retrieved using the exact same MusicLibraryRequest() Successfully plays using ApplicationMusicPlayer Both songs have library identifiers beginning with i.. What I’ve tested ApplicationMusicPlayer with prepareToPlay() followed by play(). SystemMusicPlayer with the same library song. Direct play() without explicitly calling prepareToPlay(). Retrieving songs directly using MusicLibraryRequest(), rather than accessing them through albums. A completely separate minimal tvOS application. An older version of my application’s code, recompiled using the current Xcode version. The issue persists across these tests. Interestingly, the corresponding Apple Music catalog song plays successfully when retrieved through MusicCatalogSearchRequest. However, replacing library songs with catalog equivalents is not an acceptable solution because the application needs to support users’ actual personal libraries, including imported music. Important observation The exact same failing song plays correctly on the same Apple TV and Apple Music account in: The official Apple Music application. Sleeve, another third-party Apple TV music player. This suggests that the song itself is playable and that another third-party application can successfully initiate playback. Environment Platform: tvOS Device: Physical Apple TV Framework: MusicKit for Swift Apple Music authorization: Granted MusicKit App Service: Enabled Signing: Automatic, using a paid Apple Developer account Question Is there a known issue with resolving or playing certain personal library songs through MusicKit on tvOS? Are there additional playback parameters, library synchronization requirements, or recommended queue configurations that could explain why some MusicLibraryRequest() items fail while others work? Has anyone encountered a similar issue or found a workaround that preserves playback of the user’s actual library items? Any guidance would be greatly appreciated. Thank you!

While an AVAudioPlayer/AudioUnit is rendering to AirPods over A2DP with system Spatial Audio + Head Tracking enabled, switching the shared AVAudioSession to playAndRecord + voiceChat (A2DP → HFP) blocks the main thread for tens of seconds on synchronous AVAudioSession APIs (setCategory / setMode / channel getters / setPreferred*). iOS hang detection reports a Severe Hang of 41.02 s. Stopping playback and deactivating the session before the identical switch completes in ~87 ms. Feedback: FB25324099 Environment iPhone 15 Pro (iPhone16,1), iOS 27.0.1 (24A446) AirPods worn; Spatial Audio + Head Tracking on AVFoundation-only repro (no third-party SDK): https://github.com/shinobuwz/ios27-spatial-audio-stall-repro Info.plist: LSApplicationCategoryType = public.app-category.games, LSSupportsGameMode = YES; does NOT set AVGameBypassSystemSpatialAudio (those two keys are required for system Spatial Audio on this path) Steps Wear AirPods with Spatial Audio + Head Tracking enabled Launch the repro app; allow mic (or use auto-repro) Play 440 Hz (BluetoothA2DP, spatial=1) After ~2 s, run Repro: keep playing and switch to PlayAndRecord + VoiceChat Compare with Control: stop playback + setActive:NO, then the same switch Expected: each call returns in ms; whole reconfiguration ≪ 50 ms Actual (cold launches): configuration total 20.8–28.9 s (options=13); with options=237 ≈ 36.2 s; first input buffer ~40–58 s from switch start. Control ≈ 87 ms. Evidence in the repo: docs/MEASUREMENTS-2026-10-26.md, evidence/*.log, hangs.xml, main-thread stacks in avas::client::SessionCore (sync XPC to the audio daemon). Even inputNumberOfChannels / outputNumberOfChannels getters stall for seconds. Question: Is this a known regression on iOS 27? Any workaround besides stopping all rendering AudioUnits before the HFP switch? Happy to attach sysdiagnose / Instruments / screen recording if useful.

For an iPhone recording utility with no calling functionality, what is the supported way to keep the front camera live in Picture in Picture while a ReplayKit broadcast records another foreground app? Is using AVPictureInPictureVideoCallViewController with the voip background mode supported for this purpose?

I’m developing Segue, an iPhone and Apple Watch app for discovering concerts and keeping personal concert memories. Apple Music artist pages display upcoming concerts, including dates, venues, lineups, and ticket links. I would like users to search for an artist and view those concerts directly within Segue. Is there a supported public API or Apple framework for accessing these concert listings? I reviewed MusicKit and Apple Music API documentation but couldn’t find concert-related endpoints. Example Apple Music concert: https://music.apple.com/us/concerts/ce.91c8bd7d-0669-4c33-8c2a-6809b9f65129 If this requires separate partner access or approval, could you direct me to the appropriate team?

Since iOS 27 beta we are seeing a behavior change in our video streaming app and I would like to know whether others can reproduce it. Behavior on iOS 27 beta: Fully opening the Notification Center pauses playback. Audio continues for about 5 more seconds, then the app is suspended. Closing the Notification Center leaves the player paused. On iOS 26 the same build keeps playing in this situation. Setup: AVQueuePlayer playing video with audio, not muted audiovisualBackgroundPlaybackPolicy = .automatic (default) AVAudioSession category .playback, UIBackgroundModes: audio entitlement AVPictureInPictureController attached with canStartPictureInPictureAutomaticallyFromInline = true Filed as FB23893965 Questions: Can anyone reproduce this on iOS 27 beta? Is this intentional or a regression?

Hello, We are developing LastLoop, an iOS application that continuously records the user’s screen and retains the last two hours of video locally. The user explicitly starts recording and grants permission. We would like recording to continue while the application is in the background, including after the device is locked and subsequently unlocked. Our question: Does ScreenCaptureKit on iOS 27 support this behavior? If recording stops when the device locks, is there an officially supported way to resume automatically without requesting permission again? Thank you.

I am testing SFSpeechRecognizer using the en_US locale. When the iPhone Simulator’s system language is set to Japanese, SFSpeechRecognizer.isAvailable returns false for en_US, so speech recognition is unavailable. As far as I have tested, this issue does not occur on iOS Simulator 26.2 or earlier. Is this a Simulator-specific issue, or is it a behavior change that could also occur on physical devices? Has any additional setup become necessary to use speech recognition in Simulator? I then changed the iPhone Simulator’s system language to English. After doing so, SFSpeechRecognizer.isAvailable returned true for en_US. However, starting a recognition task still failed immediately with kLSRErrorDomain Code=300, “Failed to initialize recognizer.” The following error was returned: Error Domain=kLSRErrorDomain Code=300 "Failed to initialize recognizer" UserInfo={ NSLocalizedDescription=Failed to initialize recognizer, NSUnderlyingError={ Error Domain=kLSRErrorDomain Code=300 "Failed to create recognizer from=/Users/<USERNAME>/Library/Developer/CoreSimulator/Devices/<DEVICE-UUID>/data/private/var/MobileAsset/AssetsV2/com_apple_MobileAsset_UAF_Siri_Understanding/purpose_auto/ c079bfa6b8856202dc8cb2135fef3b06229ced6e.asset/AssetData/mini.json" UserInfo={ NSLocalizedDescription=Failed to create recognizer from=/Users/<USERNAME>/Library/Developer/CoreSimulator/Devices/<DEVICE-UUID>/data/private/var/MobileAsset/AssetsV2/com_apple_MobileAsset_UAF_Siri_Understanding/purpose_auto/ c079bfa6b8856202dc8cb2135fef3b06229ced6e.asset/AssetData/mini.json } } } I have not observed either of these issues on iOS Simulator 26.2 or earlier: With the Simulator language set to Japanese, the en_US recognizer does not become unavailable. When SFSpeechRecognizer.isAvailable is true, recognition does not fail with kLSRErrorDomain Code=300. Environment Xcode: 26.6 (17F113) iOS Simulator 26.5 (23F77), iPhone 17 (A3258J/A) iOS Simulator 26.4 (23E244), iPhone 17 (A3258J/A)

I’m migrating to the new notifications described in Handling audio interruptions. On an iPhone 14 running iOS 27.0.1 (24A446), I call pause() when the system deactivates the session. On .shouldResume, I await successful session.activate() and call record() on the same recorder. It returns true, but the saved file contains only the audio recorded after Siri. To reproduce: record for five seconds, invoke and dismiss Siri, record for another five seconds, then stop and play. The earlier five seconds are missing. A legacy-notification control preserved both portions on the same phone. Is another step required to preserve the recording with the new notifications, or is this a known issue? I found this similar report, although it uses the legacy API. Minimal reproduction Built with Xcode 27.1. Create a SwiftUI iOS app targeting iOS 27, replace its generated App.swift with the code below, and add these Info.plist entries: NSMicrophoneUsageDescription: Test audio recording UIBackgroundModes: an array containing audio import SwiftUI import AVFoundation @main struct RecorderReproApp: App { var body: some Scene { WindowGroup { RecorderReproView() } } } @MainActor struct RecorderReproView: View { private let session = AVAudioSession.sharedInstance() private let file = URL.documentsDirectory.appendingPathComponent("recording.m4a") @State private var recorder: AVAudioRecorder? @State private var player: AVAudioPlayer? @State private var running = false @State private var interrupted = false @State private var status = "Ready" var body: some View { VStack(spacing: 24) { Button("Start") { Task { await start() } }.disabled(running) Button("Stop") { running = false interrupted = false recorder?.stop() status = file.path }.disabled(!running) Button("Play") { do { player = try AVAudioPlayer(contentsOf: file) status = "Saved duration: \(player!.duration) seconds" player?.play() } catch { status = error.localizedDescription } }.disabled(running) Text(status) } .padding() .task { for await notification in NotificationCenter.default.notifications( named: AVAudioSession.didBecomeInactiveNotification, object: session ) { guard running, let context = notification.userInfo?[AVAudioSession.deactivationContextKey] as? AVAudioSession.DeactivationContext, context.source == .system else { continue } recorder?.pause() interrupted = true status = "Interrupted" } } .task { for await notification in NotificationCenter.default.notifications( named: AVAudioSession.resumptionRecommendationNotification, object: session ) { guard running, interrupted, let context = notification.userInfo?[AVAudioSession.resumptionContextKey] as? AVAudioSession.ResumptionContext else { continue } guard context.recommendation == .shouldResume else { interrupted = false continue } do { guard try await session.activate(), running, interrupted else { continue } status = "Resumed: \(recorder!.record())" interrupted = false } catch { status = error.localizedDescription } } } } private func start() async { guard await AVAudioApplication.requestRecordPermission() else { return } do { try session.setCategory(.playAndRecord) guard try await session.activate() else { return } recorder = try AVAudioRecorder(url: file, settings: [ AVFormatIDKey: kAudioFormatMPEG4AAC_HE, AVSampleRateKey: 44_100, AVEncoderBitRateKey: 32_000, AVNumberOfChannelsKey: 2 ]) running = recorder!.record() status = "Recording: \(running)" } catch { status = error.localizedDescription } } }

I have not seen any issues with didActivateAudioSession not getting called by iOS in many many years with many thousands of devices. However with iOS27 beta code I have seen a few times that when making an outgoing call it never gets called. All subsequent outgoing calls fail until I dismiss and relaunch the App.

Summary On iOS build 24A446, an incoming CallKit VoIP call is reported and answered, but the system never creates or activates the CallKit audio session, so CXProviderDelegate provider:didActivateAudioSession: is never called. Both directions of call audio stay silent. The failure is inside callservicesd / audiomxd, before the app starts its audio unit. Once it starts, later calls from the same processes keep failing until the app process is killed. App and device Bundle ID: us.zoom.videomeetings OS build: 24A446 (iPhone) Log: sysdiagnose_2026.10.16_11-15-45-0700_iPhone-OS_iPhone_24A446 Sample call: 2026-10-15 13:29:06–13:29:22 PDT Processes in that window: Zoom PID 816, callservicesd PID 2474, audiomxd PID 115 How the user hits it The app is not in the foreground. A VoIP push wakes it and we call reportNewIncomingCallWithUUID:update:completion:. The user answers from the system incoming-call UI (phone locked or unlocked). The ringtone plays. After answer there is no call audio. The same failure then happens on later calls, including a call started from inside the app, until the Zoom process is killed. Signing out or rebooting only clears it temporarily. We cannot reproduce it in our lab. It became much more frequent after devices moved to this OS. What the app does We configure one CXProvider (PlayAndRecord is applied when the system activates the session). We fulfill CXAnswerCallAction when the user answers. We start the voice audio unit only from provider:didActivateAudioSession:. We do not call AVAudioSession setActive: during this answered call. Our log line "ZMCallKitManager::didActivateAudioSession" is not present for this call. System log for the sample call 13:29:06 — incoming call reported Zoom log: Provider was asked to report a new incoming call At that moment Zoom's existing primary session is sid:0x7304f, Zoom(816), 'prim', and the app is BackgroundRunning. callservicesd then fails while creating the CallKit session: audiomxd AVAudioSessionXPCServer.mm:231 -[AVAudioSessionRemoteXPCClient createProxySession:reply:] failed due to session lookup failure for SessionID 0x0 callservicesd AVAudioSession_iOS.mm:643 Creating proxy session failed, session ID = 0x0, NSOSStatusErrorDomain -50, "Session lookup failed" audiomxd ServerSessionManager.mm:192 Sibling session creation failed. Source sessionID: 0x0. SessionID is not valid. Source PID: 2474 audiomxd AVAudioSessionXPCServer.mm:1287 Session creation failed callservicesd SessionCore_Create.mm:92 AudioSession creation failed, NSOSStatusErrorDomain 2003329396, "Session creation failed in server" callservicesd Audio session not found for call. Returning nil The ringtone still plays. That audio is callservicesd's own session sid:0x732fa, category Ringtone/Default, which activates at 13:29:06.650. It is not the call session. 13:29:15 — user answers callservicesd: Performing answer request / Perform answer call callservicesd: No provider returned for descriptor callservicesd: Audio session not found for call. Returning nil callservicesd: Answer call action fulfilled callservicesd: Error retrieving AVAudioSession provider with ID 0 Sibling session creation fails again with Source sessionID 0x0, Source PID 2474. The ringtone session is stopped. From 13:29:15 to 13:29:22 there is no "action":"activate" for a CallKit call audio session, and no provider:didActivateAudioSession: notification to Zoom. Zoom's primary session 0x7304f is still inactive. Its category is updated to PlayAndRecord / VoiceChat, but it is not activated while this call is up. 13:29:22 — call ends Provider was asked to report that call ended with reason 2 callservicesd: Setting disconnected reason to remote hangup because ended reason is 2 sid:0x732fa callservicesd going inactive The call lasted about 7 seconds with no call audio session ever created. 13:29:33 — following call, same stuck state About 11 seconds later a CXStartCallAction is fulfilled. Zoom's own session then activates: sid:0x7304f, Zoom(816), 'prim' PlayAndRecord_WithBluetooth / VoiceChat, NonMixable, going active callservicesd still cannot build its session: "No existing proxy AVAudioSession interruption provider for ID 0, creating a new one" Session creation failed in server Error retrieving AVAudioSession provider with ID 0 Audio session not found for call So an app-side setActive: after the failed call does not repair the CallKit provider. The system session id remains 0x0. How often it happens in this sysdiagnose Filtering the archive for the hard failure (not the generic "Audio session not found" line, which also appears on earlier days): "Sibling session creation failed. Source sessionID: 0x0": 93 hits, grouped into 57 bursts First burst: 2026-10-15 09:52:11 PDT Last burst: 2026-10-16 11:14:30 PDT (about a minute before this sysdiagnose was taken) "Error retrieving AVAudioSession provider with ID 0": 34 hits "Session creation failed in server": 93 hits Every one of these sibling-session failures is Source PID 2474. callservicesd stays PID 2474 and audiomxd stays PID 115 across the whole window, including overnight. The bad session id is stuck in those daemons, not in one Zoom call object. There is no log line in this archive showing callservicesd successfully activating a provider audio session for these calls. Request Please confirm this is the known CallKit failure where didActivateAudioSession is not delivered because callservicesd cannot create the audio session (Session lookup failed / Session creation failed in server, session ID 0x0). Related public reports: Feedback FB19429215, "CallKit does not activate audio session after iOS 18.4.1", and the iOS 27 thread "iOS 27 CallKit didActivateAudioSession". DTS previously advised apps not to call setActive: themselves, because the CallKit session has higher priority and an app activation can fail or disturb system state. This log matches that: the session is never created, so there is nothing for the system to activate. Please also advise which of these is safe on 24A446 when provider:didActivateAudioSession: does not arrive: Call CXProvider setConfiguration: again (same configuration) before every reportNewIncomingCallWithUUID:update:completion:. On a timeout, fail the unanswered CXAnswerCallAction so the CallKit state machine resets, without invalidating or recreating the CXProvider. Anything else that clears the stuck session ID 0x0 inside callservicesd without a device reboot. I can provide another sysdiagnose if you need a logging profile installed before the next reproduction.

If a single asset in the array passed to insertOrUpdateAssets cannot be read, the returned AsyncSequence throws and delivers zero elements — including for the valid assets that appear before the offending one in the array. In my measurement a batch of 5 — two valid photographs, one zero-byte .jpg, then two more valid photographs — delivered 0 of 5 elements and reported 0 faces. The framework's own stdout log shows it had already processed the valid photographs before failing, so the work was done and then thrown away. This is the behaviour I would expect from a function returning [Result] after processing everything, not from a streaming AsyncSequence. The whole reason to expose an async sequence is to deliver results as they are produced; here the sequence produces nothing at all, which makes the streaming shape actively misleading. Why this matters at library scale. I am cataloguing 97 photo libraries, many of them archives of scanned family photographs going back to the 1970s, where a handful of damaged files is normal rather than exceptional. As the API stands, the batch size I choose for throughput is also the amount of work a single corrupt file destroys — with a batch of 100, one bad file costs 100 images. The only safe strategy is to catch the failure and re-submit the batch one asset at a time to find the culprit, which turns a rare bad file into a full re-run of that batch and makes the worst case quadratic in the number of bad files. Note also that the error gives no indication of which asset failed (see the related enhancement request on error taxonomy), so isolating the culprit by re-submission is the only option available. Feedback: FB24174733

An image file whose data is cut in half is accepted by insertOrUpdateAssets, produces no error, and is reported as an image containing zero faces. That result is indistinguishable from a photograph that genuinely has no people in it. In my measurement the intact file reported 4 faces and a copy truncated to half its bytes reported 0 faces — both without throwing. Why this is worse than an error. A client that receives "zero faces, no error" will record the asset as analysed, no people present. That is a permanent, silent, incorrect result: the file is never revisited, and the people in it are lost from the catalogue with no diagnostic anywhere. An error would have been recorded as a failure and retried later. Silence is the one outcome that cannot be recovered from. The failure mode is also inconsistent. Other kinds of damaged input do throw — a zero-byte file, a text file with a .jpg extension, and a file whose interior bytes have been destroyed all raise MediaIntelligenceError.faceGroupProcessing. Truncation is the case that silently succeeds, and truncation is precisely the damage that partial downloads, interrupted copies and failing disks produce — the most common form of corruption in a real photo archive, and the one I have to handle in 97 libraries of mixed provenance. My current workaround is to fully decode every image myself before handing it to the framework, purely to detect truncation. That means every file is decoded twice, which roughly doubles the I/O of the analysis pass. Feedback: FB24174749