back to listing index


[web search]
Original source (
Tags: jenkins groovy workflow pipeline
Clipped on: 2016-03-11

Skip to content
Open this file in GitHub Desktop
201 lines (152 sloc) 6.27 KB

Pipeline Global Library

When you have multiple Pipeline jobs, you often want to share some parts of the Pipeline scripts between them to keep Pipeline scripts DRY. A very common use case is that you have many projects that are built in the similar way.

This plugin adds that functionality by creating a "shared library script" Git repository inside Jenkins. Every Pipeline script in your Jenkins see these shared library scripts in their classpath.

Directory structure

The directory structure of the shared library repository is as follows:

 +- src                     # groovy source files
 |   +- org
 |       +- foo
 |           +- Bar.groovy  # for class
 +- vars
     +- foo.groovy          # for global 'foo' variable/function
     +- foo.txt             # help for 'foo' variable/function

The  src  directory should look like standard Java source directory structure. This directory is added to the classpath when executing Pipelines.

The  vars  directory hosts scripts that define global variables accessible from Pipeline scripts. The basename of each  *.groovy  file should be a Groovy (~ Java) identifier, conventionally  camelCased . The matching  *.txt , if present, can contain documentation, processed through the system’s configured markup formatter (so may really be HTML, Markdown, etc., though the  txt  extension is required).

The groovy source files in these directories get the same sandbox / CPS transformation just like your Pipeline scripts.

Other directories under the root are reserved for future enhancements.

Accessing repository

This directory is managed by Git, and you'll deploy new changes through  git push . The repository is exposed in two endpoints:

  •  ssh://USERNAME@server:PORT/workflowLibs.git  through Jenkins SSH
  •  http://LOCATION/workflowLibs.git  (when your Jenkins app is located on the url  http://LOCATION/ ). As noted in JENKINS-26537, this mode will not currently work in an authenticated Jenkins instance.

Having the shared library script in Git allows you to track changes, perform tested deployments, and reuse the same scripts across a large number of instances.

Note that the repository is initially empty of any commits, so it is possible to push an existing repository here. Normally you would instead  clone  it to get started, in which case Git will note

normalwarning: remote HEAD refers to nonexistent ref, unable to checkout.

To set things up after cloning, start with:

normalgit checkout -b master

Now you may add and commit files normally. For your first push to Jenkins you will need to set up a tracking branch:

normalgit push --set-upstream origin master

Thereafter it should suffice to run:

normalgit push

Writing shared code

At the base level, any valid Groovy code is OK. So you can define data structures, utility functions, and etc., like this:

// src/org/foo/Point.groovy

// point in 3D space
class Point {
  float x,y,z;

However classes written like this cannot call step functions like  sh  or  git . More often than not, what you want to define is a series of functions that in turn invoke other Pipeline step functions. You can do this by not explicitly defining the enclosing class, just like your main Pipeline script itself:

// src/org/foo/Zot.groovy

def checkOutFrom(repo) {
  git url: "${repo}"

You can then call such function from your main Pipeline script like this:

def z = new

Defining global functions

You can define your own functions that looks and feels like built-in step functions like  sh  or  git . For example, to define  helloWorld  step of your own, create a file named  vars/helloWorld.groovy  and define the  call  method:

// vars/helloWorld.groovy
def call(name) {
    // you can call any valid step functions from your code, just like you can from Pipeline scripts
    echo "Hello world, ${name}"

Then your Pipeline can call this function like this:

helloWorld "Joe"

If called with a block, the  call  method will receive a  Closure  object. You can define that explicitly as the type to clarify your intent, like the following:

// vars/windows.groovy
def call(Closure body) {
    node('windows') {

Your Pipeline can call this function like this:

windows {
    bat "cmd /?"

See the closure chapter of Groovy language reference for more details about the block syntax in Groovy.

Defining global variables

Internally, scripts in the  vars  directory are instantiated as a singleton on-demand, when used first. So it is possible to define more methods, properties on a single file that interact with each other:

// vars/acme.groovy
def setFoo(v) { = v;
def getFoo() {
def say(name) {
    echo "Hello world, ${name}"

Then your Pipeline can call these functions like this: = "5";
echo; // print 5
acme.say "Joe" // print "Hello world, Joe"

Define more structured DSL

If you have a lot of Pipeline jobs that are mostly similar, the global function/variable mechanism gives you a handy tool to build a higher-level DSL that captures the similarity. For example, all Jenkins plugins are built and tested in the same way, so we might write a global function named  jenkinsPlugin  like this:

// vars/jenkinsPlugin.groovy
def call(body) {
    // evaluate the body block, and collect configuration into the object
    def config = [:]
    body.resolveStrategy = Closure.DELEGATE_FIRST
    body.delegate = config

    // now build, based on the configuration provided
    node {
        git url: "${}-plugin.git"
        sh "mvn install"
        mail to: "...", subject: "${} plugin build", body: "..."

With this, the Pipeline script will look a whole lot simpler, to the point that people who don't know anything about Groovy can write it:

jenkinsPlugin {
    name = 'git'