We are trying to modify some files on the server and then check them back in, before continuing the build.
We do msbuild /t:updateuid to update the XAML files and then I'm thinking using a tf checkin command for checking the files back in.
Problem A: TFS won't recognize file changes. We are using a "Server" workspace - is this preventing TFS from finding the changed files?
If I issue a "tf status" from command prompt, TFS won't find the changed files.
Problem B: Since the tf command will check files in. It might just trigger another build. Probably a minor issue, since the loop hopefully ends there...
Problem A:
As it is a server workspace, you need to perform a tf checkout on the files before changing them. They will then show up in the workspace as a pending change.
Problem B:
End your check in comment with ***NO_CI*** - this will stop any CI builds from triggering.
e.g. Modified by Build ***NO_CI***
If this is a gated build you may need to do other things.
Related
I am trying to diagnose an error during the build of a project with appveyor. This same project is also built with travis-ci, without any problems. I assume it is windows related.
The script produces some log files, but I have no clue on how these can be viewed after appveyor is done trying to build.
As a specific example: See the log of this build. At line 11706 it says:
Logs have been written to: C:\stack.stack-work\logs\yaml-0.8.28.log
How can I view the contents of that file?
You can push this file as artifact at on_finish stage, or simple RDP to the build worker and explore it interactively.
Side note: you can also try to debug your build in RDP, but note that environment variables from the build session are not available in the RDP session, so you need to re-create all or part of them.
We have an MSBuild task that builds our *.sqlproj file, the output of which is loaded back into TeamCity as an artifact for subsequent deployment.
Similar to this user, we are having some unpredictable output file naming happen:
Invalid file names when trying to deploy SSDT project with TeamCity 8
It appears that sometimes, it produces this output file under /bin/Release:
MyProj.sqlproj.dacpac
Then subsequent builds produce this file in the same folder:
MyProj.dacpac
We haven't done indepth testing yet - I was wondering if anyone else has seen similar or has a suggested troubleshooting path?
To be clear, it's the same task, running the same command against the same project - just run repeatedly overtime as new checkins happen.
Sounds like you have conflicting .sqlproj files being checked in.
I'd start by checking the history.
Well, after looking more closely at the build log I could see that the TeamCity MSBuild runner appears to be creating some temporary virtual project files (or something, I don't know for sure because they get deleted) with names like:
MyProj.sqlproj.teamcity
I theorized that this may be confusing MSBuild or one of the targets related to building dacpacs, so I replaced the TeamCity MSBuild build step with a Command Line build step that calls MSBuild on the original project file itself, and this appeared to solve the problem. It now produces the dacpac with the file I'm expecting.
I don't have time to dig further now, but I could believe that there's some logic in the chain somewhere that is deriving the name for the final *.dacpac from the name of the project file being used. My guess is that it just strips off everything after the last "." and attaches the ".dacpac" suffix.
I don't fully know why it would occasionally create a *.dacpac file with the correct name, but I was at times doing a manual build in the TeamCity agent work folder from the commandline on the build server itself, so this may have just been a file leftover from previous executions.
I have a plan in Bamboo that starts whenever changes are made to the attached repositories (via polling).
Now, on each build, if successful, a CHANGELOG file is updated in the repo, which in turn, triggers another build. How can I omit certain files from Bamboo's polling, so that a build isn't started if changes are found for those files? Because otherwise, I enter in infinite loop, with a change to CHANGELOG triggering another build which in turn updated CHANGELOG and so on.
If this is not possible, what other viable solutions are there? Is it possible to attach a shell script somewhere before the build starts to check whether it's desired to start a new build?
It turned out that this was simpler than I've thought. In Plan Configuration, in the Repositories tab, on each repository, under Advanced, there is an Include / exclude files input where you can Customise what files Bamboo uses to detect changes. By adding a regular expression there, I got everything solved and working as expected.
Bamboo pattern matching reference: https://confluence.atlassian.com/display/BAMBOO/Pattern+matching+reference
The Bamboo Documentation says:
Bamboo will ignore build triggers if the local working copy and the
repository copy have the same revision numbers.
This might not be the best solution, but you might add an additional task at the end of the job/build which updates the repository again to avoid triggering a new build.
I'm not sure if this would then skip builds from repository updates which occur during the current build.
I am getting a strange error on the build server, this does NOT happen locally.
Error during file generation. The target file 'C:...\AcceptanceTest\Features\HelloWorld.feature.cs'
is read-only, but different from the transformation result. This
problem can be a sign of an inconsistent source code package. Compile
and check-in the current version of the file from the development
environment or remove the read-only flag from the generation result.
To compile a solution that contains messaging project on a build
server, you can also exclude the messaging project from the
build-server solution or set the msbuild
project parameter to 'true' in the messaging project file.
I'm not sure what to do to fix this, I don't have a "messaging" project (or know what is being referred too). And the code is the latest
This is the full message (some folder names have been obfuscated):
13>C:\Builds\2\Services\PService\src\packages\SpecRun.Excel.1.1.0\tools\TechTalk.SpecFlow-buildfix.targets(47,5):
error : Error during file generation. The target file
'C:\Builds\2\Services\PService\src\PService\Test\PService.AcceptanceTest\Features\HelloWorld.feature.cs'
is read-only, but different from the transformation result. This
problem can be a sign of an inconsistent source code package. Compile
and check-in the current version of the file from the development
environment or remove the read-only flag from the generation result.
To compile a solution that contains messaging project on a build
server, you can also exclude the messaging project from the
build-server solution or set the msbuild
project parameter to 'true' in the messaging project file.
[C:\Builds\2\Services\PService\src\PService\Test\PService.AcceptanceTest\PService.AcceptanceTest.csproj]
13>C:\Builds\2\Services\PService\src\packages\SpecRun.Excel.1.1.0\tools\TechTalk.SpecFlow-buildfix.targets(47,5):
error : Could not find file
'C:\Users\tfsbuild\AppData\Local\Temp\tmpPWithData.feature.xlsx.cs'.
[C:\Builds\2\Services\PService\src\PService\Test\PService.AcceptanceTest\PService.AcceptanceTest.csproj]
13>C:\Builds\2\Services\PService\src\PService\Test\PService.AcceptanceTest\Features\PWithData.feature.xlsx
: error : Generation error: Could not find file
'C:\Users\tfsbuild\AppData\Local\Temp\tmpPWithData.feature.xlsx.cs'.
[C:\Builds\2\Services\PService\src\PService\Test\PService.AcceptanceTest\PService.AcceptanceTest.csproj]
Delete HelloWorld.feature.cs from source control.
TFS will checkout files as read-only during a build and so SpecFlow fails to re-generate the file because it already exists and SpecFlow doesn't overwrite read-only files.
I created a new request for SpecFlow to not automatically add the file to source control here: https://github.com/techtalk/SpecFlow/issues/592
I'm not sure what the messaging project refers to, that's something I haven't seen before.
However the error reads as if the problem is due to an issue with the generation of the *.feature.cs from the *.feature file, but I am confused as to why this should be happening on the build server. Normally the generation of *.feature.cs occurs in the IDE when you save the *.feature file. This gets checked in and with some version control systems (VCS) you may end up with a read-only file status.
Your error reads as the build server is trying to re-generate the *.feature.cs file and cannot overwrite it.
Some things to check;
What VCS are you running? Does it leave checked-in files as read-only?
How are you building the files on the build server? Is there a custom step involved, or is it just a simple compile?
Can you replicate this error on your machine by setting the *.feature.cs to read-only and running the same build script on your local machine?
Good luck.
I'm in a bit of a pickle. I'm trying to run some environmental scripts before I run the build in a m2 project, but it seems no matter how hard I try - the 'pre' build script are never run early enough.
Before the 'pre-build' scripts are run, the project checks to see if the correct files are in the workspace - files that won't be there until the scripts I've written are executed.
To make them 'pre-build', I'm using the M2 Extra Steps plugin - but's it's not 'pre' enough.
Has anyone got any suggestions as to how I can carry out what I want to do?
Cheers.
Have you considered breaking it up into two projects, and setting the pre-build project to be upstream of the build project?
e.g.,
Foo Pre-build
Foo Build
After Foo Pre-build runs, cause "Foo Build" to run.
I have used this, admittedly in different scenarios than yours, quite successfully. This has the added benefit (if you need it) of allowing you to manually run a build without going through the pre-build steps, if you know they aren't necessary.
You should use the free form project type and not the maven project type.
If this is a problem (ie, there are projects that are expecting to be triggered by or triggering from), consider using a custom workspace location and having a free form project execute in this workspace before the maven project runs. The free form project can be used as the trigger for the maven project.
Does adding another build step as a shell script work?
My problem stemmed from the fact I wanted to set-up my workspace before I ran anything due to an issue with Dynamic Views (ClearCase) not being accessible from the workspace - I wanted to add a symlink to fix this.
However, Andrew Bayer has made a change to the plugin that I'm currently testing that should fix this...so the question is probably invalid in it's current form.
Will edit it once we come to a conclusion.