I have a repository in VS Team Services which is setup as :
-root
|
| -Dev (.sln)
|
| -Releases
|
| - Test (.sln)
| - Live (.sln)
Each build is working fine, however its pulling the entire repo and I just want it to pull a specific branch (Dev, Test or Live)
So I'm selecting the Solution.sln file in the dev branch...
On the Repository tab you can specify your workspace mappings, the default mapping grabs your whole repository excluding the $/project/drops folder.
By changing the first mapping to not take $/Project, but $/project/dev you'll limit the scope of the get operation. If you no longer get the root of your team project, you can remove the cloak of the drop folder, that will only raise a warning otherwise:
The next thing you should do, if you're setting this build up to be a CI build, it to specify the folders that will trigger this build on the Tiggers tab, because, again by default, CI builds will trigger on any change in the whole team project.
To allow you to keep your build definition as a template, I suggest to use the $(Build.SourcesDirectory)\MySolution.sln instead of $/project/dev/Mysolution.sln, that way you only need to update the mappings of a build definition after cloning it to another branch root or when saving it as a template.
Related
One of my colleagues has made some changes to a project, named "DB Conversion". There are 2 TFS builds associated with the PharmacyWarehouse project. Here's a picture so you can see what I'm talking about.
A previous TFS Administrator set this configuration up. I inherited it when he left. As you can see PharmacyWarehouse is just a folder. But the 2 builds have a Visual Studio Build step which only looks at the PharmacyWarehouse.sln file in the PharmacyWarehouse folder. I've looked at that .sln file, it only references packages, PharmacyWarehouse and UnitTests folders. I've looked at the .csproj files, they only reference those folders/VS projects.
I've looked at the DBConversion.sln and DBConversion.csproj files in the Projects | "DB Conversion" folder. It only references that folder/VS project. The files my colleague was editing are in the Projects | "DB Conversion" | Models folder.
Why would checking in changes in Projects | "DB Conversion" | Models folder cause the TFS trigger watching the PharmacyWarehouse.sln to fire? Is it because their folders and not TFS Projects?
We're working with TFS 2015
According to the screenshot, you are using TFVC version control.
Open build definition, click the tab Trigger and check the trigger field Path specification and ensure the path is $/Pharmacy Warehouse /DB Conversion
Open build definition, click the tab Get sources and check the field Workspace mappings, ensure the path is $/Pharmacy Warehouse /DB Conversion.
After configuration, if we push the changes to the folder or project DB Conversion, it will trigger this build and only build DB Conversion.
My TFS repo has the following structure:
Project
- Dev
- 1.0.0_Branch1
- 1.1.0_Branch2
- N.0.0_BranchN
The branches are actual branches in TFS.
I have an automated CI build set up for this project on a TFS server. The issue I am having is that I need the build name in TFS to use the name of the branch.
I have tried a number of the build definition variables listed here , for examle $(SourceBranchName) and and $(Build.SourceBranchName) but all of these are instead using the name of the Project (Project in this example).
I would like the build to be named along the lines of 1.0.0_Branch1.1 with the last number being the revision.
The reason I would like to do this is so that when I generate Nuget packages at the end of the build, they can be versioned using the version of the branch that is being built. Currently they are being versioned using the date which means that the highest version is only ever the most recently built, which may not be the case in practice. I don't want to have to manually set the versions each time.
All assemblies that are being built are correctly versioned to match the branch that they are within.
Is there a way to achieve this?
For build definitions, $(SourceBranchName) can be used in the build number format:
example build:
The environment variable is BUILD_SOURCEBRANCHNAME so it will be available as $(BUILD_SOURCEBRANCHNAME) in msbuild.
The list of available variables is found at: https://www.visualstudio.com/en-us/docs/build/define/variables#predefined-variables
This could be caused by the "Mappings" setting you configured under "Repository" tab in your build definition. When you build with TFVC repository, the BUILD_SOURCEBRANCHNAME variable is filled with the last path segment in the root server path of the workspace. So I'm wondering that the mapped server path in your build definition is "$/Project" rather than "$/Project/Dev/1.0.0_Branch1".
TFVC repo branch: The last path segment in the root server path for
the workspace. For example in $/teamproject/main this value is main.
Refer to this link for details: Predefined Variables.
I have setup a CI Workflow in Jenkins to build project from bitbucket server.
I followed this Excluded Regions in Jenkins with Git
I made configuration such that ,when ever any changes are pushed into the repository, Jenkins will trigger a build.
I felt Repository is having too many checkins and each project is having separate workflows in Jenkins. So I used the "Polling ignores commits in certain places" option
But it seems not working.
I think this may be a bug as well.
Please see screenshot attached.
Scenario
I have 3 projects named Project1,Project2 and Project3
I want to look only changes in Project1 and then build.
For that i gave "Project1/.*" in the Included Regions. But even when changes are done to Project2 / Project 3, Jenkins triggers a build.
Doubt:
The Repository structure is as below
Project : TestForJenkins
Repository : JenkinsRepo
Project1,Project2,Project3,Packages,.gitignore
Since Project1 is at toplevel in repository, please let me know if i am giving wrong lookup path by providing "Project1/.*" in included region section.
I have Microsoft TFS build process to deploy web project to azure web role, this occurs automated way every day. I have followed Azure article https://azure.microsoft.com/en-in/documentation/articles/cloud-services-dotnet-continuous-delivery/
I have following MSBuild arguments in my build process, as above article suggested
/t:Publish /p:PublishDir=C:\MSCD\
When i add database project to my project solution, build is keep failing. Getting error message
The "SqlPublishTask" task was not given a value for the required
parameter "SqlPublishProfilePath"
When i publish my web project, i don't want database project to be published. How to drop DACPAC file to drop folder ? so i can use powershell to update my database in azure.
I am using TFS 2012 On-Premise. Could someone give suggestion on how to solve this problem?
You need to create Master-child build definitions and in master build definition configure solution with build and deploy disabled in configuration manager .and in child build configure database project alone .
Create Master and child build definition such that they share common drop .
#Siva palla's answer solved this exact same issue for me. Here are the complete set of changes we made to get this working.
Changes in Visual Studio
Originally I was using a single configuration called 'Deployment' that was set to deploy both our WinForms project and our SQL project (VisionShellWin is the WinForms project, the two Vision.SQLMigration projects are the SSDT projects) so everything below is based on changes to Deployment. For most people you'll have Release instead of Deployment, all of the below should still work fine, just substitute in Release for Deployment:
To separate that single configuration in two I created a new configuration:
I set that new deployment to copy its settings from the old Deployment configuration and not to create new configurations for each of the projects (the existing project level Deployment ones are fine):
In that new deployment I then unticked the boxes to Build and Deploy the database projects:
I then did the exact same thing again to create a Database specific deployment:
And in that one I unticked everything except the database files:
That's everything needed in Visual Studio, so I committed all of that and synced it back to DevOps.
Changes in Azure DevOps
In DevOps I cloned my existing Visual Studio Build stage (called 'Build Winforms solution') and renamed the clone to Build Databases:
I added two new build variables in the Variables tab, named ClickOnceBuildStageConfiguration and DatabasesBuildStageConfiguration with their values set to the names of the new configurations we just created in VS:
And then I changed the WinForms build stage to use the new ClickOnceBuildStageConfiguration variable - note that we still have the /T:"VisionShellWin" /target:Publish MSBuild Arguments needed for ClickOnce set:
And the new Build Databases stage to use the databases variable - note that we don't have any MSBuild Arguments:
Finally, in addition to the copy stage I already had for copying the ClickOnce application files into the drop artifact:
I also added a Copy Files stage (called 'Copy Dacpacs') to copy the DacPac into the drop too:
Once you've done all of that, you should end up with a build that works and an artifact that contains both the ClickOnce files and the DacPacs
If you follow some of MS's recommended branching strategies you can easily end up with a project structure such as:
$PROJECT\
DEV\
MyProject
STAGE\
MyProject
PROD\
MyProject
Now let's say I have three different build definitions. One each for DEV, STAGE, and PROD. This should be common considering that the build definition will define the exact solutions to build.
If I turn on CI for each of them, STAGE will be built even though the checkin occurred in DEV...
Now my question. How can I limit the build definition to execute only when a check in occurs in either a path or a solution that is part of the build definition?
When defining the working folder configuration screen - only have it start at the root of the branch you want to build.
For example, your DEV branch would be configured so that $/TEAMPROJECT/DEV/MyProject was mapped to $(SourceDir) rather than the default mapping which would have been set to just $/TEAMROJECT.
FYI - Personally, I only have CI Builds on Dev branches and queue manual builds for a push to QA. I also normally don't do a re-build for production but just push the build binaries that were QA'd. I also keep by build configuration folder inside the branch i.e. $/TEAMPROJECT/DEV/TeamBuild rather than the default $/TEAMPROJECT/TeamBuildTypes and therefore changes to the build configuration are also pushed up through the branches. That said, you have to stick with the default if you wanted the build configuration to be visible to VS2005 clients.
Hope that helps,
Martin.