In TFS 2015 I have the following structure for a Project:
ProjectA
Scripts
SystemA
SystemB
SystemC
Now let's say I have a build definition for SystemA with the following mappings:
$/ProjectA/SystemA/trunk/src
$/ProjectA/SystemA/trunk/deploy
When I build SystemA the source branch will be:
$/ProjectA/SystemA/trunk
I will be able to see the changeset for the build only for SystemA (what I expect).
Now let's say I want to add a powershell script to my build definition in the Scripts folder. Here are the new mappings for SystemA's build definition:
$/ProjectA/SystemA/trunk/src
$/ProjectA/SystemA/trunk/deploy
$/ProjectA/Scripts
When I build again SystemA, the following source branch will be:
$/ProjectA
The problem now is when I click a build done and troubleshoot the errors why the build failed, I have all the changesets from SystemA, SystemB, SystemC and Scripts.
I would like to manually set the Source Branch for the changesets to $/ProjectA/SystemA/trunk/, even if Scripts is outside of that scope and being used. Is there any way to do that?
We are not able to change the Source Branch, it capture the same part in all mappings. You could consider create the Scripts folder under $/ProjectA/SystemA.
When creating your mapping, instead of
$/ProjectA/Scripts, type
$/ProjectA/SystemA/trunk/../../Scripts
and it will trick it to use $/ProjectA/SystemA/trunk as the branch.
Unfortunately, I think this is more a bug than a feature since once saved, the path is shown as $/ProjectA/Scripts and that if you save again without manually setting the backward path again, the original situation comes back. Anyhow, it works as long as you think of setting the path each time you need to modify it.
Related
Recently, my company started to focus Extension_v2 development for Dynamics NAV BC. We store our code in an internal Git-Server. So far, so good.
But startig a new project is still a very fiddly task. You have to create a repository, clone it, execute the AL Code-Task, move the files to the fitting location push the repository to the correct upstream etc. And all this does not include the first initial Steps (README, CHANGELOG and all other fundamental files...).
So I wanted to write a small PowerShell-Script, to do all these initial steps before being able to start working on the Project.
The Problem: I could not find a way to execute the "AL-GO!" task via script.
I have already searched the Internet and some forums for an answer... but it seems like microsoft did not consider the possibility to execute tasks from the AL-Language-Extension via script.
I also played around with the New-NAVAppManifest and the New-NAVAppManifestFile command for the old Extension_V1 development, but that did not do the trick.
I am looking for a fair and easy way to combine the creation of the app.json file and the launch.json file with other commands to easily initialize a new Project without haveing to write all commands manually. Maybe I did not recognize the easy solution. Or maybe this is just the way we have to do it in Extension_v2.
Anyway, thanks for all your help nevertheless.
Greetings.
Stay away from Ext V1. It's highly deprecated at this point.
First of all, why do you need to execute the "AL-Go!" via script? The "AL-Go!" command should already include all necessary steps to create an empty project including the launch.json and app.json. (minimal adjuments required dependent on your BC environment)
There is an extension/plugin for Git in Visual Studio Code which will handle all the repository stuff for you. You don't need to change file locations if everything is set up for Git. I rarely use it yet, but saw a demo for it on the Directions EMEA last year and I'm pretty sure it works at its current state (someone correct me if I'm wrong)
A way to implement the "AL-GO!" command for a script or for setting up additional steps in your project setup might be to write your own visual studio code extension/plugin which requires some additional know-how for that.
OR
You just change the settings/files of the default project, I bet there is at least a file for creating the initial AL project. Just change that to your requirements
I have a folder in TFS which has SQL Scripts. At the moment I am manually adding a comment and updating a version number inside the comment every time i make a change and check it back it. This works however was hoping there might be a better way. Is there a way to automate this in TFS?
I have read the following article
Version control project files
do i have to go through such a process for simple .sql files? Are there any other simple ways.
There are a few ways you can do this:
Create an automated build in TFS and write a custom build step / PowerShell script to parse the appropriate SQL scripts, read the version, increment it, and store the new version by either checking in the updated file or a local store
Use a database project (part of SQL Server Data Tools) which will output a DACPAC. Inside the database project, you can set the version as specified here. This stores the version in the project file. If you update your TFS build number to be digits only, you can then update the project file to set that value to match the build using a custom build task. For example, if your build number was yyyy.m.d.R where R is the number of times that build was run today (TFS manages that - it's the revision variable). Or, you could set the the <DacVersion> tag to something like 2.1.0.0 and your build replaces the last digit with yyyymmddr.
I'd recommend using a database project. It's pretty easy to create a new database project off an existing database.
The first way mentioned by Jacob above can achieve that if you just want to incremental the version number of the script/folder, just create a CI build definition.
Actually you can just enable Label sources and set the Label format with predefined environment variables such as $(build.buildNumber), and set without publish any artifacts during build process.
Thus, it will automatically trigger the CI build when you check in files, and the source (SQL Script /folder) will be labeled with the incremental number.
Then you can find the specific versions with the label.
We are using the "new, scriptable build system" of TFS 2015 and all of a sudden one of our build definitions is gone. It seems that it deleted itself.
Is there a way to undelete/recover a build definition?
Where are the build definitions (those JSON files) acutally stored?
If you are on prem and you just chose delete from the GUI... run this in the TFS database for your specific id:
update Build.tbl_Definition
set Deleted = 0 --was 1
where DefinitionId = <your build ID goes here>
You will loose your build history without some extra work i did not take the time to dig into, but that might also be do-able.
NOTE: You will also loose credentials stored as build parameters you will need to recreate. Might even be best for you to recreate a new build using this as a template to avoid other unknowns.
TFS build definitions are not version controlled items. So, It's not possible to restore the deleted build definitions for now.
There has been a feature request in uservoice: provide a way to version-control build definitions According to the response from PM, this is still in process.
In the new build system coming with TFS 2015 you can see the full
history of the changes to your build definition. The feature that is
currently missing is the ability to undo or rollback to a previous
revision.
We expect to get the rollback deployed to our service in the next few
months.
Chris Patterson
Program Manager
Moreover, the build definitions are stored in the TFS database.
I have the following variables defined:
Now once a build is complete (the last step in the build process), I want to update the VersionRevision variable, basically increment it.
So I'm looking for an API I can call from C# and create a console application or a powershell script to edit the build definition (if I have to do this)?
You can use VSTS Rest API to update the variable value in Build Definition. Both Console Application and Powershell Script is OK for this.
If I understand correctly, you want to get these build variables and them assignment them as your version number.
After the build completes, update and increment the VersionRevision. It's not a good way and seems not available to achieve it.
In TFS build there is a $(Rev:.r) which means
Use $(Rev:.rr) to ensure that every completed build has a unique name.
When a build is completed, if nothing else in the build number has
changed, the Rev integer value is incremented by one.
Source: Specify general build definition settings
To version your assemblies you could just add an powershell script in your build definition, detail ways to achieve please follow this link from MSDN: Version your assemblies
And usually we only define and assignment variables with the Major and Minor version. If you want to change the value of them. You may need manually edit the build definition.
More related link about how to manage version numbers as part of your vNext builds.
vNext Build Awesomeness – Managing Version Numbers
Generate custom build numbers in TFS Build vNext
We have a particular file, say X.zip that is only modified by 1 or 2 people. Hence we don't want the build to trigger on every check-in, as the other files are mostly untouched.
I need to check for a condition prior to building, whether the checked-in item is "X.zip" or not.. if yes, then trigger a build, else don't. We use only CI builds.
Any idea on how to trigger the build only when this particular file is checked-in? Any other approaches would be greatly appreciated as i am a newbie in TFS...
Tara.
I don't know of any OOTB feature which can do this, what you would need to do is write your own custom MSBuild task which is executed prior to the build running (pre-build action).
The task will then need to use the TFS API to check the current check in for the file you want and if it's not found you'll have to set the task to failed.
This isn't really ideal as it'll indicate to Team Build a build failure, which, depending on whether you're using check in policies, may be unhelpful. It'd also be harder to at-a-glance work out which builds failed because of the task and which failed because of a real problem.
You can change the build to occur less frequently rather than every check in, which will reduce load on your build server.
Otherwise you may want to dig into Cruise Control .NET, it may support better conditional builds.
If you could move X.zip into it's own folder, then you could set up a CI build with a workspace that only looked at the folder containing X.zip.
You would then need to add an explicit call to tf get to download the rest of the code as Team Build only downloads what the workspace is looking at.
But this might be simpler than the custom task approach?