Jenkins Pipeline Choices attributes - jenkins

Question

I was under the impression that the Jenkins pipeline choices syntax supported only a static newline-separated list of options and does not currently support dynamic population as the result of a groovy function (returning a map, say).
However, doing some reading on .net core I found this post.
I was particularly interested in a section of one of the pipeline code extracts:
parameters {
choice(choices: services, description: 'Name of the ECS service to deploy', name: 'serviceName')
choice(choices: services, description: 'Name of Docker image to update', name: 'imageName')
string(defaultValue: 'Tag to deploy', description: 'Docker image tag', name: 'tag')
}
Where and how does the services variable get set / populated?
What is services type? Is it a map of some sort (value and description)?
Thanks in advance.

Answer 1

The services variable here is most probably either a string with newlines (services = "choice a/nchoice b") or a list (services = ["choice a", "choice b"]. Both work these days.
I think you can set the variable in the environment directive and then use it in parameters [in deed the environment version does not work] or (probably evil) outside of the pipeline:
List<String> CHOICES = ["a", "b", "c"]
pipeline {
agent any
options {
timestamps()
}
parameters {
choice(name: 'CHOICE', choices: CHOICES, description: 'Who to greet?')
}
stages {
stage("test") {
steps {
echo params.CHOICE
}
}
}
}
I once used this to get all labels of a Docker images on our registry, to choose which one should be deployed. There are multiple issues with this:
This is AFAIK completely undocumented, so it might just eventually stop working.
The code is executed every when the Jenkinsfile is read. It configures the job, so the parameters are changed for the next build. So in my case, I only got the labels which were existing during the last build, not the currently existing ones.