Build logic based on BRANCH

[Matthew Ceroni]

Looking for some assistance on how to accomplish executing different build logic based on branch names. In my use case for branches master or PR-? I want to execute build + test (basically continuous integration). For branches named release-? I want to execute deployment steps. 

Now I know I can do this by using a single Jenkinsfile and putting crazy amounts of if statements around each stage. But after a while that just becomes overly confusing and hard to follow. What I want to accomplish is to actually run separate Jenkinsfile based on the branch name. I believe this can be accomplished using the load (or maybe apply from) functionality.

So far I have tried this:

main Jenkinsfile that lives in the repo:

@Library('shared-library@master') _

node('test') {
  stage('Load Build') {
    checkout scm
    load(BRANCH_NAME == 'master' ? 'master.groovy' : 'deploy.groovy')
  }
}

master.groovy that lives in the repo:

#!/usr/bin/env/groovy

podTemplate(label: buildId(), name: buildId, containers: []) {
  node(buildId()) {
    stage('Build') {
      checkout scm
    }
  }
}

This somewhat works but I don't know if it is the best way to accomplish this.

First, issue is that in Blue Ocean it collapses everything to one stage, that stage being 'Load Build'.

Second, I have to allocate two nodes and checkout the source code twice. The first node allocation is in the main Jenkinsfile. That is required as I need to checkout the source code due to the fact that master.groovy lives in the repository in question. The second node allocation and code checkout is in what is the main pipeline (master.groovy). 

My hope is that I could load this file (master.groovy or deploy.groovy) from the shared library, as well as pass parameters to it. 

Any assistance on how to best accomplish this would be appreciated.

Thanks

[Matthew Ceroni]

https://dev.to/jalogut/centralise-jenkins-pipelines-configuration-using-shared-libraries

Is what I just found. I will give this a try.