Post not yet marked as solved
Research "fastlane" to help you automate that process more simply.
Post not yet marked as solved
@Coder Dreamer
Please note, that link is in the "Documentation Archive" meaning it is not part of the current, living documentation set.
It was last updated 2016-03-21. Moving the documentation from the live docs to the "archive" was what prompted my original post (IIRC). Everything continues to work for me, but I'm still worried about the server API.
Today I see that in Xcode 12.3 the API documentation is not available in Xcode Help (it used to be).
Post not yet marked as solved
I spoke too soon. This morning, things were moving more smoothly than before.
Tonight, I can't get Xcode Server to build an .xcarchive no matter what I try.
Fixing permissions, clearing out the caches, rebooting the server... doesn't seem to matter.
I see in the GUI "Uploading integration logs" and it's been that way for 11 minutes now.
This time, deleting the bot and rebuilding it worked... for now... it is not a guarantee, though it usually works for a while before eventually degrading.
Post not yet marked as solved
I'm getting much better stability if I do the following sudo xcrun xcscontrol --fix-permissions is run constantly by the maintenance script
It’s also run prior to doing an integration
Set the bot to Always Clean
Occasionally run Edit bot.. > Repositories > Replace Repositories… (can only be done manually it seems)
I'm going to give up, for now, in having the bot use its automatic archive generation and instead do it manually through a build script. The same bot builds its Firebase distributions flawlessly, so I guess I'll just modify that process to also do a TestFlight build. Frustrating regression...
Post not yet marked as solved
Looking at our logs, I don't think that's a 10 second timeout. I think it is a 10 minute timeout, no?
In jerimiah's log it starts at 13:28:36 and times out at 13:38:36
hour:minute:sec format, it looks like, based on relative timings for other logs.
That exactly matches mine of 10 minutes.
And where is it "uploading" to?
Post not yet marked as solved
Nov 24 04:13:47	[28884] <Info>: Uploading file /Users/server/Library/Caches/XCSBuilder/Integration-0e9d438d371a0c194d7926a703d59b19/xcodebuild_result.xcresult.zip
Nov 24 04:23:48	[28884] <Error>: Error: The request timed out.
Nov 24 04:23:48	[28884] <Error>: Error trying to upload /Users/server/Library/Caches/XCSBuilder/Integration-0e9d438d371a0c194d7926a703d59b19/xcodebuild_result.xcresult.zip: Unable to perform the request. CFNetwork error: The request timed out. (-1001 - kCFURLErrorTimedOut)
Nov 24 04:23:48	[28884] <Error>: XCSProcessTestResultsIntegrationStep finished integration with an error: Error Domain=XCSBuildServiceDomain Code=-1 "Integration results could not be uploaded. Test attachments and integration logs may be unavailable for this integration." UserInfo={NSLocalizedDescription=Integration results could not be uploaded. Test attachments and integration logs may be unavailable for this integration.}
Nov 24 04:23:48	[28884] <Info>: Executing trigger 'testflight upload'
I'm having that exact same trouble a lot. It sometimes feels like a bot will work great at first, then slowly degrade with this error occurring more and more. If a bot exhibits this issue, it is a toin-coss if I'll ever get it to build properly.
The main problem here for me is that if this occurs, the bot's .xcarchive build phase is skipped.
So I have a critical workflow which uploads the .xcarchive to TestFlight. However, if this "failed to upload integration results" error occurs, the .xcarchive is never built. Xcode Server just skips to the first post-integration script immediately.
This has broken my workflow and puts us in a tough position for QA certification of the release builds.
Bots are created fresh with each weekly release, so "recreating the bot" is not working as a solution here.
Post not yet marked as solved
This permissions problem is interesting, because today I finally saw a more explicit message than just "integration results could not be uploaded."
Build Service Warning: “com.apple.Safari.SafeBrowsing.Service-2020-10-22-030548.diag” couldn’t be copied because you don’t have permission to access “Integration-89776b80b6693d016b6ed8b7045e7606”.
Reason: You don’t have permission.
I have not found that "duplicating the offending bot" works. I have a fresh install of Xcode 12.2 beta 2 (now 3) and deleted all bots and rebuilt them from scratch. Still see the issue with fresh bots, from time to time.
Also seeing strangely long build times that seem to stall on "uploading log files"... to itself? I guess? Which eventually completes, but it adds minutes to each integration run.
Also, seeing a weird issue with bots that are instructed to build an .xcarchive after unit tests just refusing to do the work. Nothing in the log, no errors, it just ignores the request to build an archive. Re-run the integration, sometimes twice, and magically Xcode agrees to do what I specified in the bot definition. After that, Xcode 12 seems consistent with providing me with archives
Highly annoying, these intermittent failures.
(Xcode 12.0.1 and 12.2 beta 3 on macOS 10.15.6, both versions of Xcode exhibit all of the above issues)
EDIT: Just noticed OP was referring to Xcode 11. Oddly enough, I never saw OP's issue on Xcode 11 even once, but I see it very often on Xcode 12
Post not yet marked as solved
Still happening in Beta 7
Post not yet marked as solved
I want to point out that our project has 2 and only 2 .xcasset files.One is images only, the other is colors only.We have the problem; so I'm not entirely certain this diagnosis is completely accurate.However, this project was not created from scratch in Xcode 11.It is, like probably everyone else here, a pre-existing project that works in Xcode 10 and does not work with Xcode 11.
Post not yet marked as solved
Exactly the same here. Xcode 10 + iOS 11.x / 12.x works perfectly.Xcode 11 + iOS 12.x / 13.x works perfectly.Xcode 11 + iOS 11.4 == crash when calling UIColor(named:)However! Xcode 11 + iOS 11.4 + Objective-C [UIColor colorNamed:@"name"] works.
Post not yet marked as solved
Still happening in Beta 5
Post not yet marked as solved
Removing Simulator devices from the integration bot device list (i.e., test on real hardware only) "resolved" this issue for me.
Post not yet marked as solved
It seems that every time a bot integration runs, something happens in the early stages of the integration which causes the server to be removed from the team. As a result, every integration fails outright.
