I’m seeing what looks like a change or possible bug in Xcode when creating a new Swift Playground. When I create a Swift Playground using Xcode (File → New → Playground), the project is generated without an Assets folder. I’ve reproduced this multiple times with fresh Playgrounds to rule out user error and the folder is consistently missing. Previously, new Playgrounds included an Assets catalog by default, which makes this behavior surprising.

Hi, I have a recurrent problem with an Xcode workspace that I cannot figure out how to solve: When I command-click a symbol (or press ctrl-cmd-J) to find its definition, and when that definition is in a header file belonging to a custom static library, Xcode opens a header located in the derivedData folder rather than in the project's source codes. The exact location of this file is not in the main 'Build' folder but rather in a folder that seems dedicated to indexing: derivedData/myProject/Index.noIndex/Build/Products/Debug/Include/theHeader.h Consequently the file opened by Xcode is uneditable, I have to close it and reopen the source using another method. Even when the correct header is already opened, and I try again, Xcode reopens the wrong file so that it's opened twice with different paths. In the config file of the main application project, I reference the sources of these static libraries like this: HEADER_SEARCH_PATHS = ${SRCROOT}/myLibrary1/Sources ${SRCROOT}/myLibrary2/Sources ${SRCROOT}/myLibrary3/Sources However, removing this doesn't fix the issue and I don't see any other related-settings. Any idea to fix this annoying issue would be greatly appreciated, thanks!

Hi; Here is a simple program that reads a file and counts number of 1s bits in it. The file it reads, simple.txt, is a static text file, resides in the file system, and nothing is touching it during the program execution. The file's length is 436 bytes. When the program is run in the terminal it produces consistent results. Each run counts 1329 1s bits. It behaves differently when run in Xcode. A simple command-line tool project having a single file. Running the program multiple times, via Command-R, produces something strange. That is each execution randomly produces one of two results. One is expected 1329 1s bits, another one shows 1333 1s bits. That is it counts four bits more. Again, each run randomly produces one of those two results. And it happens only when I run the program in Xcode. I tried a different sample text file and experienced the same behavior, but the difference in 1s bits count was five bits. Any idea of how this behavior can be explained? func countOnes(in byte: UInt8) -> Int { var nOnes = 0 var mask: UInt8 = 0x1 while mask != 0 { if (byte & mask) > 1 { nOnes += 1 } mask <<= 1 } return nOnes } func countOnes(in buffer: UnsafeMutableRawBufferPointer, length: Int) -> Int { var nOnes = 0 for byte in buffer[0...length] { nOnes += countOnes(in: byte) } return nOnes } var fd = try FileDescriptor.open("sample.txt", .readOnly) var buffer = UnsafeMutableRawBufferPointer.allocate(byteCount: 4096, alignment: 4) var nBytes = try fd.read(into: buffer) var nOnes = 0 while nBytes > 0 { nOnes += countOnes(in: buffer, length: nBytes) nBytes = try fd.read(into: buffer) } buffer.deallocate() try fd.close() print(nOnes) My machine Darwin MacBookPro 24.6.0 Darwin Kernel Version 24.6.0: Wed Oct 15 21:08:19 PDT 2025; root:xnu-11417.140.69.703.14~1/RELEASE_ARM64_T8103 arm64 macOS Sequoia 15.7.2 Xcode 26.1.1 (17B100) Thanks, Igor

Hi, I'm interested in trying out Xcode Assist to help with things like complicated refactors or writing tests cases. The ChatGPT and Claude options both share your code with third parties, which is not acceptable for my use case. Has anyone used a fully local model for Xcode Assist? I see that you can select one in the Apple Intelligence section of Xcode's Preferences screen, but don't really know where to start. Are there local models that work well with Xcode Assist and that truly keep your source code private?

Our apps, such as iCMTGIS PRO, provides the ability for users to open a job file received as an email attachment. Our apps use application:openurl:option to get the URL for the job file (*.pmp). We use [URL path] to convert it to NSString format. Then we use fopen to read the .pmp job file in "read only" mode. On iPad Air (5th generation) running iOS Version 18.6.2, our apps are able to read the .pmp job file. However, on iPhone 11 running iOS: 26.2, using fopen to read the .pmp file, we get a NULL value and are not able to correctly read the .pmp file.

Apple made them both, and for some reason they are acting like each is speaking a foreign language. Once you plug the iPad in, shouldn't the "trust this computer" kick in and allow the iMac to turn on developer mode? I'm the dumbest person in the world when it comes to Code, but this is a no brainer! And the fact that the iPad is in Developer mode, is the most frustrating part. This needs a fix.

Xcode 26.2 RC has been out for over 24 hours and it's still not possible to submit apps.

Has anyone been able to use Google Gemini with Intelligence in Xcode? I got the Claude models working in there but as much as I try I can't see any Gemini models.

Gitlab runner on M4 Mac Mini 2024, XCode 26.2, supported platforms: iOS, native UIKit swift project. Cocoapods. Trying to run tests via fastlane: run_tests(workspace: 'Project.xcworkspace', destination: "platform=macOS,id=#{deviceId}", scheme: "Release", clean: true) And having an error: $ xcodebuild -showBuildSettings -workspace Project.xcworkspace -scheme Release -destination platform\=macOS,id\=01234567-0123456789B9001C 2>&1 [12:23:53]: Could not read the "SUPPORTED_PLATFORMS" build setting, assuming that the project supports iOS only. [12:23:54]: Found simulator "iPhone 16 Pro (18.5)" The tests runs just fine in xcode on "My Mac (Designed for iPad)" device on the very same machine. Compared the output of xcodebuild -showdestinations in terminal: $ xcodebuild -showdestinations -workspace Project.xcworkspace -scheme Release Available destinations for the "Release" scheme: { platform:macOS, arch:arm64, variant:Designed for [iPad,iPhone], id:01234567-0123456789B9001C, name:My Mac } { platform:iOS, id:dvtdevice-DVTiPhonePlaceholder-iphoneos:placeholder, name:Any iOS Device } { platform:iOS Simulator, id:dvtdevice-DVTiOSDeviceSimulatorPlaceholder-iphonesimulator:placeholder, name:Any iOS Simulator Device } { platform:iOS Simulator, arch:arm64, id:A7E98A79-B29B-4E9A-8103-ECBA55ABC0C6, OS:26.2, name:iPhone 17 Pro } But when I launch this command in gitlab runner, I get this: Available destinations for the "Release" scheme: { platform:iOS, id:dvtdevice-DVTiPhonePlaceholder-iphoneos:placeholder, name:Any iOS Device } { platform:iOS Simulator, id:dvtdevice-DVTiOSDeviceSimulatorPlaceholder-iphonesimulator:placeholder, name:Any iOS Simulator Device } { platform:iOS Simulator, arch:arm64, id:A7E98A79-B29B-4E9A-8103-ECBA55ABC0C6, OS:26.2, name:iPhone 17 Pro } I suspect gitlab runner to ruin some ENV variable, probably. So I googled the most common issues on the internet and tried everything I found: LANG: "en_US.UTF-8" LC_ALL: "en_US.UTF-8" unset CFProcessPath And none is worked. I also compared the printenv output in terminal and gitlab runner, didn't found anything suspicious. Ran out of ideas. How can I troubleshoot this?

The Send to Device in Reality Composer Pro no longer works for me. The progress spinner keeps spinning... Happens with Wifi and Developer Strap (first gen). Any known workarounds / ways to recover?

Hi all, Has anyone else run into this issue in Xcode 26? I’m logged into my paid ChatGPT Plus account, but the Xcode integration doesn’t seem to recognize the subscription. After a short period of use, I get the following error: “Over daily limit. ChatGPT in Xcode will be unavailable for up to 24 hours. For higher limits, sign in with a paid ChatGPT account.” Since I’m already signed in with a paid account, this looks like either a bug or a limitation specific to Xcode. Is this expected behavior, or has anyone found a workaround to make Xcode properly recognize Plus accounts? Thanks in advance for any guidance.

MacBook Air M2(2022) macOS Tahoe 26.2 Xcode Version 26.1.1 (17B100) 上記環境でXcodeのGPT-5を使用すると表題の不具合が発生します。 GPTの回答に短いサンプルコードが表示されている状況下でスクロールすると、 カーソルが待機状態になり操作が全くできなくなります。 双眼鏡のアイコンをオフにしてみても症状は解消されませんでした。

The developer downloads page now lists an Xcode 26.1 which was released on 11th Dec (the original Xcode 26.1 was posted on 3rd Nov). Strangely, this new Xcode 26.1 has a CFBundleShortVersionString of 26.1.1, and a DTXcodeBuild of 17B55 % ls -ln total 4413136 -rw-r--r--@ 1 503 20 2259523057 16 Dec 19:01 Xcode_26.1_Apple_silicon.xip % xip --expand Xcode_26.1_Apple_silicon.xip xip: signing certificate was "Software Update" (validation not attempted) xip: expanded items from "/Users/me/Downloads/temp/Xcode_26.1_Apple_silicon.xip" % plutil -p Xcode.app/Contents/Info.plist | grep CFBundleShort "CFBundleShortVersionString" => "26.1.1" % plutil -p Xcode.app/Contents/Info.plist | grep DTXcodeBuild "DTXcodeBuild" => "17B55" 17B55 does correspond to the original Xcode 26.1 final release. The Xcode 26.1.1 release that was previously posted had a DTXcodeBuild of 17B100, though. The pairing of 26.1.1 and 17B55 looks new and probably a packaging error?

I have Mac mini intel, and updated top Sequia 15.7.2 and updated Xcode to 26 and now it crashes on launch. I uninstalled and reinstalled and still crashes. I uninstalled and installed Xcode 16.4 and that crashes on launch also. I've looked all over for help but cannot find anything. Can anyone please help me with this?

Hi When attempting to upload a React Native app (version 0.77) we encountered the following error: ITMS-90426: Invalid Swift Support - The SwiftSupport folder is missing. Rebuild your app using the current public (GM) version of Xcode and resubmit it. If to check ipa folder we can see that content includes only 2 folders: Payload Symbols Could you please tell us why it does not include Swift Support folder? We tried to use XCode - 16.2 and 16.3 Thank you

It seems Xcode's predictive code completion model is censored. Specifically, when typing the word "torrent," the model stops working completely. It doesn't matter whether the word is written directly in the code or in a comment. It could also be part of another word, such as "qBittorrent." In either case, the model stops working. Reproducing this issue is fairly simple. Create a Swift file and type the word "torrent." The model will stop generating code. Xcode Version 26.2 (17C52) Predictive Code Completion Model: [com.apple.fm.code.generate_small_v2.base: 700.0.81600.13.202379,0] [com.apple.fm.code.generate_safety_guardrail.base: 1.6.81619.13.202072,0] [com.apple.gm.safety_deny.input.code_intelligence.base: 32025010.20251009.91600.100.1651,0] [com.apple.gm.safety_deny.output.code_intelligence.base: 32025010.20251009.91600.100.1651,0] (Installed)

I am building a data logging app that logs the iphone's movement data at 120 hertz and 200 but for some reason its always capped at 100. does anyone know if there a hardware issue or is there a code problem that could be linked to this. made in Xcode.

Hi everyone, We run our CI builds on a Mac mini. On my local MacBook, incremental builds work properly and build times are fast, but on CI it looks like Xcode rebuilds everything from scratch every time, and the build is about 6 minutes slower than local. In Xcode 26, “compilation caching” became available. My understanding was that if DerivedData is preserved between CI runs, compilation caching / incremental builds should reduce build time. So I tried specifying the DerivedData path as an absolute path in our xcodebuild command to reuse it across builds. However, it still seems to do a full rebuild every time and the build time didn’t improve. Has anyone seen a similar issue with incremental builds on CI (Mac mini) with Xcode 26? Any advice on what to check or how to configure CI so incremental builds / caching actually work would be greatly appreciated. Thanks!

We’re facing an issue with Xcode 26.1 where code coverage is not being generated. All our test cases run and pass successfully, but the .xccovreport / .xccovarchive files are never produced. Code coverage is enabled in the scheme, and this setup used to work correctly in earlier Xcode versions. We are trying to determine whether this is a configuration issue on our end or a possible Xcode 26.1 bug. Is anyone else experiencing the same problem with code coverage in Xcode 26.1? Any insights or workarounds would be appreciated.

I have been using Xcode Version 26.1.1 on macOS Tahoe Version 26.1. In VIM mode, whenever I quickly tap d twice, to delete the complete line, most of the time Xcode crashes. Is anyone else facing this issue? Should I move to any older version of Xcode or any other solve for this?