Since this morning, visionOS builds in XCode Cloud are failing.
Did change anything in my configuration, the error on XCode cloud looks like visionOS is not installed:
{ platform:visionOS, id:dvtdevice-DVTiOSDevicePlaceholder-xros:placeholder, name:Any visionOS Device, error:visionOS is not installed. To use with Xcode, first download and install the platform }
I attached a screenshot from the logs.
Anyone know what is happening here and how to fix?
Thanks,
Patrick
Xcode Cloud
RSS for tagAutomate workflows to test, analyze, build, and distribute your app, and integrate them with other developer tools, such as TestFlight and App Store Connect.
Post
Replies
Boosts
Views
Activity
Five days ago everything was working as intended. Nothing has been changed on Workflows. All of a sudden, new builds are failing. The issue is with these two tasks during the Archiving:
Export archive for ad-hoc distribution
Export archive for app-store distribution
I downloaded the Artifacts and checked the logs. If we look at the IDEDistribution.verbose.log then we see the following errors:
"Error Domain=IDEProfileLocatorErrorDomain Code=1 \"No profiles for 'uz.uzum.avto' were found\" UserInfo={IDEDistributionIssueSeverity=3, NSLocalizedDescription=No profiles for 'uz.uzum.avto' were found, NSLocalizedRecoverySuggestion=Xcode couldn't find any iOS App Store provisioning profiles matching 'uz.uzum.avto'.}",
"Error Domain=DeveloperAPIServiceErrorDomain Code=5 \"There is a problem with the request entity\" UserInfo={IDEDistributionIssueSeverity=3, NSLocalizedDescription=There is a problem with the request entity, NSLocalizedRecoverySuggestion=You already have a current Distribution Managed certificate or a pending certificate request.}",
"Error Domain=IDECodesignResolverErrorDomain Code=1 \"No signing certificate \"iOS Distribution\" found\" UserInfo={IDEDistributionIssueSeverity=3, NSLocalizedRecoverySuggestion=No \"iOS Distribution\" signing certificate matching team ID \"x\" with a private key was found., IDEProvisioningError_UserInfoKey_IDEProvisioningErrorSpecifier=iOS Distribution, NSLocalizedDescription=No signing certificate \"iOS Distribution\" found, IDEProvisioningError_UserInfoKey_IDEProvisioningErrorPlatform=com.apple.platform.iphoneos, IDEProvisioningError_UserInfoKey_IDEProvisioningErrorAction=5, IDEProvisioningError_UserInfoKey_IDEProvisioningErrorTeam=<IDEProvisioningBasicTeam: 0x600001866280; teamID='x', teamName='(null)'>}"
)}
Issues seems to be related to certificates and etc. However, I checked the member center, and everything is fine with the certificates and profiles. Moreover, I can build, archive and distribute from my local machine with no issues at all.
One particular thing that got my interest is the following: I opened the IDEDistribution.critical.log file, and saw the following issue:
It seems Xcode Cloud is having a hard time to connect to App Store Connect. Please, help me to resolve the issue. I tried to change the Environment, to add new provisioning profile, but nothing works.
FYI:
I tried different environments to build, but the latest was:
Xcode 15.0.1
macOS: Ventura 13.6.3
platform: iOS
Deployment Preparation: Testflight (Internal Testing Only)
After months of the same script working properly, all of sudden, as of last night, our ci_post_clone.sh started failing with following error:
The ci_post_clone.sh is not executable and was run using zsh (default shell on macOS). To make sure your script runs correctly, make the file executable using `chmod +x` and add an appropriate shebang line.
Here's what I've done:
Re-applied: git update-index --chmod=+x ./ci_scripts/ci_post_clone.sh; still failed to execute returning same error
I was using older flutter version so changed the flutter branch to stable - still same error.
App code builds fine in local Xcode!
APPRECIATE ANY HELP!
Here is the script.
#!/bin/sh
# ci_post_clone.sh
# Runner
#
#
set -e
echo "Running post clone script"
# The default execution directory of this script is the ci_scripts directory.
cd $CI_WORKSPACE # change working directory to the root of your cloned repo.
# Install Flutter using git.
git clone https://github.com/flutter/flutter.git --depth 1 -b stable $HOME/flutter
export PATH="$PATH:$HOME/flutter/bin"
# Install Flutter artifacts for iOS (--ios), or macOS (--macos) platforms.
echo "Doing precache"
flutter precache --ios
echo "Installing cocoa pods"
# Install CocoaPods using Homebrew.
HOMEBREW_NO_AUTO_UPDATE=1 # disable homebrew's automatic updates.
brew install cocoapods
echo "calling pub get"
# Install Flutter dependencies.
flutter pub get
echo "executing pod install"
# Install CocoaPods dependencies.
cd ios && pod install # run `pod install` in the `ios` directory.
exit 0
Hello everyone!
I've noticed that Xcode 15.2 is still not selectable as a build environment on Xcode Cloud, even though the official version of 15.2 has been out for several days. Is this part of the plan? Or is it a system bug? 🤔 I need 15.2 as the release environment, including support for building visionOS apps. I have already submitted FB13531029.
Hi there,
In my custom script in ci-scripts, I had to call my git repository rest api to fetch registries (but also had to create tags, releases and more...)
I added an ad-hoc secret environment variable to my workflow with a fresh access token as explained in this WWDC Video.
But I wonder :
Is there a way to get the git repository Access Token we provide to Xcode Cloud when we associate it to our git repository following the Source code management setup page as an Environment Variable of the project Workflows ?
If I get it right, before I had the had-oc access token, I was able to clone the other git repositories of my project... So IMHO, the access token provided during the setup should be available somewhere (at least for the git commands).
Just a quick tip from what I discovered in the build times of our app in Xcode Cloud.
Beware of selecting another simulator device instead of one of the recommended iPhones.
Prepare Simulator
Custom device: ~3m
Recommended device: ~3s
I assume Xcode Cloud comes preconfigured with the recommended iPhone. Any other device will take ~3m to setup in the testing step. (At least this is my assumption, please correct me if you have different findings.)
You don't have to use all recommended iPhone. It will work if you just select one of them.
Xcode Cloud Builds seem flaky to me. Sometimes the test step fails (not necessarily the tests itself, but the step as a whole). Xcode Cloud seems to re-attempt this step. In my most extreme example, took 3 attempts to run them. For a build which usually takes 22m, it took 1h9m, but billed just 29m. Does anyone else experience this? PR-builds with a duration of >1h are not really acceptable.
Here is the protocol. Testing usually takes <3 minutes.
Hi, we're trying to reduce the Xcode Cloud usage in our project, and in the logs of the "Test" action I can see that there's a "Save artifacts" step that takes quite a while (more than 10 minutes), and it looks like a "fixed cost" that's impossible to reduce by improving the build and the tests themselves.
In the "Artifacts" section of the Xcode Cloud build I can see that there a "Test Products" zip that's almost 5GB, which seems to be the cause of this slowness (I could verify that in other projects the same zip is smaller, and the "Save artifacts" step is much faster, as a result), but this seems a waste: we never need to download that "Test Products" zip, and its upload takes literally 10 fixed minutes from every single build with a "Test" action.
Is it possible to either stop uploading the "Test Products" zip? Or maybe to skip the "Save artifacts" step entirely?
Thanks
I have been trying for a couple of days now to get Xcode cloud running.
But after some struggle with cocoapods the ci_post_clone.sh is working.
But now it fails at the archive section with no real explenation what went worng.
I have my profisioning profile connected to the right bundle-id.
Run command: 'xcodebuild archive -workspace /Volumes/workspace/repository/wevize/ios/Runner.xcworkspace -scheme Runner -destination generic/platform=iOS -archivePath /Volumes/workspace/build.xcarchive -derivedDataPath /Volumes/workspace/DerivedData -resultBundleVersion 3 -resultBundlePath /Volumes/workspace/resultbundle.xcresult -resultStreamPath /Volumes/workspace/tmp/resultBundleStream34ff2ba9-cd2a-4e3b-befb-3243d688c7f5.json -IDEPostProgressNotifications=YES CODE_SIGN_IDENTITY=- AD_HOC_CODE_SIGNING_ALLOWED=YES CODE_SIGN_STYLE=Automatic DEVELOPMENT_TEAM=KTJT289M8K COMPILER_INDEX_STORE_ENABLE=NO -hideShellScriptEnvironment'
All the steps within our Archive action complete as expected with green checkmarks (including all Prepare Build for App Store Connect steps), but the Archive action itself never completes. It keeps spinning and never passes it off to the post-action to release the build to TestFlight.
The system status page says Xcode Cloud had issue for multiple days this week, but they resolved 14 hours ago. I've re-run our build several times today and keep running into this problem. Is this something server-side?
I see from the WWDC video that it's possible to use secrets, like an API_KEY variable, when running Xcode Cloud builds.
How are these variables accessed from code? Is it something like this?
ProcessInfo.processInfo.environment["API_KEY"]
If so, how are we supposed to run the tests locally? Is the idea that every developer creates their own non-shared schemes and adds all secret environment variables to that scheme?
Hi there,
I'm trying to find this information but I don't see it anywhere. What happens when I reach the maximum of 1000 hours/month used on Xcode Cloud? Can I buy additional hours or can I get them reset if I pay again?
I'm not near close to using that but I think we should have a public answer for this scenario.
Thanks!