I use Azure DevOps for long time, and I always use GUI when crafting Build Pipeline so far, though I knew Azure DevOps supports YAML and its benefits.

So this time, I tried YAML instead of GUI and I learnt many things. In this article, I talk about "variables".

Variables in GUI

Let's start from GUI. There is a place to store variables for pipeline scope.

And global scope.

Use $(xxx) format to reference it.

Variable in YAML

First of all, I can reference variable same as GUI. In addition, I can directly write parameter in YAML.

variables:
  myval: 1

I can also use all the out-of-box variables as well as my own defined variable groups.

Secret value

As YAML is simply a text file, I shouldn't store any secret. I can use "Variable" feature of YAML editor to store both non-secure and secure variables.

1. Click "Variables".
   

2. Click "New Variable".
   

3. Enter info. Check "Keep this value secret" if it's secret :)
   

4. Keep adding as necessary.
   

But where this variables are stored?

1. Go to triggers first.
   

2. Then you see variables tab like GUI pipeline.
   

It's easy to store, but I needed to know the trick to use it, as I couldn't simply reference it. It was obvious if I read the document carefully, but I didn't :D. Basically, I need to map the variable to environment variable to reference it.

steps:

- powershell: |
    # Using an input-macro:
    Write-Host "This works: $(mySecret)"

    # Using the env var directly:
    Write-Host "This does not work: $env:MYSECRET"

    # Using the mapped env var:
    Write-Host "This works: $env:MY_MAPPED_ENV_VAR"    # Recommended
  env:
    MY_MAPPED_ENV_VAR: $(mySecret)

Variable Template

The reason why I use variable group is to grouping variables. Then, I realized that YAML also has similar feature, called variable template.

Template types & usage

I just need to create YAML file contains variables I want to group.

# File: vars.yml
variables:
  favoriteVeggie: 'brussels sprouts'

Then reference it in another YAML.

variables:
- template: vars.yml  # Template reference

steps:
- script: echo My favorite vegetable is ${{ variables.favoriteVeggie }}.

Templates are not only for variables but for more powerful stuff. I will find sometime later to talk about what I found in template if I can.

Service Connection as variable

One of the variable I couldn't figure out at first was service connection.
I tried to build and push docker container to Azure Container Registry and refer to the documentation below.

Build and push to Azure Container Registry

Inside the article, there is YAML example.

- stage: Build
  displayName: Build and push stage
  jobs:  
  - job: Build
    displayName: Build job
    pool:
      vmImage: $(vmImageName)
    steps:
    - task: Docker@2
      displayName: Build and push an image to container registry
      inputs:
        command: buildAndPush
        repository: $(imageRepository)
        dockerfile: $(dockerfilePath)
        containerRegistry: $(dockerRegistryServiceConnection)
        tags: |
          $(tag)

I couldn't find where this dockerRegistryServiceConnection variable comes in, and I later realized I simply pass the service connection name as parameter.

When I use template, it gave me GUID but I actually can simply use string name of the service connection. In the example below, I set variable in YAML and use it.

trigger:
- master

pool:
  vmImage: 'ubuntu-latest'

variables:
  acrConnection: 'MyAcrServiceConnection'
  imageName: 'sample-image'

steps:
- task: Docker@2
  displayName: BuildAndPushImageToACR
  inputs:
    repository: $(imageName)
    command: 'buildAndPush'
    containerRegistry: $(acrConnection)
    Dockerfile: '**/Dockerfile'
    tags: $(Build.BuildId)

Summary

In this article, I talk about variables in YAML. I am glad to find that it can use all the variables from Library and pipeline scoped ones as GUI does. In addition I can write it to YAML so it's easy to migrate to other DevOps environment.