Accessing Variables in Azure pipeline - azure

Question

I am trying to access variables from Pipeline > Library like this
$(AndroidKeyAlias) but I can't see any output. I have tried a couple of other ways as well.
azure-pipelines.yml
trigger:
branches:
include:
- azure-develop
variables:
- group: Example Mobile App
pool:
vmImage: 'macos-latest'
stages:
- stage: initialise_environment
displayName: Initialise Environment
jobs:
- template: steps/azure-initialise-pipeline.yml
azure-initialise-pipeline.yml
jobs:
- job: initialise_node
displayName: initialise node
steps:
- checkout: self
persistCredentials: true
clean: true
- bash: |
FOO="some value"
echo "##vso[task.setvariable variable=FOO]$AndroidKeyAlias"
- bash: |
set -ex
mkdir -p packages
echo $AndroidKeyAlias > packages.json
cat packages.json
env:
CACHIX_AUTH_TOKEN: $(AndroidKeyAlias)
- bash: |
echo "$(FOO)"
As you can see I have tried getting variables values but I am not getting any output.
How do I get this value? I have tied fetching the values using the Azure CLI tool and I am getting following output
authorized: true
description: ''
id: 123
name: Example Mobile App
providerData: null
type: Vsts
variables:
AndroidKeyAlias:
isSecret: true
value: null
AndroidKeyAliasPassword:
isSecret: true
value: null
AndroidKeyStorePassword:
isSecret: true
value: null
What am I doing wrong?
Thanks in advance.

Answer 1

You're completely messing up the notation for bash environment variable substitution and the Azure Pirelines variable substitution. On top of that there are rules for when variables are available to a template and when not and also specific notation that might be needed there...
In general, when you want Azure pipelines to inline the value of a variable for you, use:
$(VariableName)
These variables will be substituted at the time that the task is invoked and will hold whatever value the variable has at that step in the job
Alternatively, use
${{ variables.Variablename }}
For 'compiletime' substitution. These will be expanded before the job even starts to run and will use the value that is available during template expansion and workflow initialisation.
If you are inside a bash script you can use any of the above mentioned syntaxis. You can also use an environment variable to pass the contents in
- bash: |
$VAR
env:
VAR: $(Variablename)
- pwsh: |
$env:VAR
env:
VAR: $(Variablename)
By passing in variables through the environment you don't have to mix the different variable substitution notations. It makes it easier to see what's going on.
When setting an environment variable in a bash step, the value won't be available in a subsequent step. The environment is cleaned up between steps for security and reasons.
For this reason, this won't work:
# Doesn't work
- bash: |
FOO="some value"
- bash: |
echo $FOO
There are special strings you can stream to the log that will instruct the agent to set a variable for the subsequent steps in the same job.
With templates things can become even harder, as you can also pass variables to a template where it's being called. You can do this with template parameters. The advantage of this syntax is that's it's very evident when you're passing in data:
jobs:
- template: templates/npm-with-params.yml # Template reference
parameters:
name: Linux
vmImage: 'ubuntu-latest'
With a template defined as:
parameters:
- name: name # defaults for any parameters that aren't specified
default: ''
- name: vmImage
default: ''
jobs:
- job: ${{ parameters.name }}
pool:
vmImage: ${{ parameters.vmImage }}
steps:
- script: npm install
- script: npm test
See:
templates
variables
stetting variables in scripts