jenkins_jobs 0.1.4

Listing with Jenkins CLI

My first attempt was to use the official Jenkins CLI to extract that information.

If you are curious, you can use Vagrant with the following Vagrantfile and download the Vagrant VirtualBox box I created with a sample Jenkins server and some examples jobs over there:

Vagrant.configure("2") do |config|
  config.vm.box = "arfreitas/jenkins-centos7"
  config.vm.box_version = "0.0.1"
  config.vm.network 'forwarded_port', guest: 8080, host: 8080, id: 'jenkins'

  config.vm.provider 'virtualbox' do |vb|
    vb.gui = false
    vb.memory = '2048'
    vb.cpus = '2'
    vb.name = 'jenkins-sample'
end

These are the credentials already setup:

• user: admin

• password: admin

• token: 116f3e55f677416a7c054faa20fbbcf0be

Finally, fire up the VM:

$ vagrant up

After the setup is complete, fire up the commands below:

$ vagrant ssh
$ java -jar jenkins-cli.jar -s http://localhost:8080/ -webSocket -auth admin:116f3e55f677416a7c054faa20fbbcf0be list-jobs
freestyle-sample
pipeline-sample

Not a very exciting output. You will get the job names and that’s it.

Listing with the REST API

So then I tried the Jenkins REST API with the sample.py Python 3 program:

$ cd /vagrant
$ ./sample.py
{'_class': 'hudson.model.FreeStyleProject', 'name': 'freestyle-sample', 'url': 'http://localhost:8080/job/freestyle-sample/', 'color': 'notbuilt', 'fullname': 'freestyle-sample'}
XML information:
OrderedDict([('project',
              OrderedDict([('keepDependencies', 'false'),
                           ('properties', None),
                           ('scm',
                            OrderedDict([('@class', 'hudson.scm.NullSCM')])),
                           ('canRoam', 'false'),
                           ('disabled', 'false'),
                           ('blockBuildWhenDownstreamBuilding', 'false'),
                           ('blockBuildWhenUpstreamBuilding', 'false'),
                           ('triggers', None),
                           ('concurrentBuild', 'false'),
                           ('builders', None),
                           ('publishers', None),
                           ('buildWrappers', None)]))])
{'_class': 'org.jenkinsci.plugins.workflow.job.WorkflowJob', 'name': 'pipeline-sample', 'url': 'http://localhost:8080/job/pipeline-sample/', 'color': 'notbuilt', 'fullname': 'pipeline-sample'}
XML information:
OrderedDict([('flow-definition',
              OrderedDict([('@plugin', 'workflow-job@2.40'),
                           ('keepDependencies', 'false'),
                           ('properties', None),
                           ('triggers', None),
                           ('disabled', 'false')]))])

If you take in consideration this is almost raw output, it’s an improvement because of the additional details, but far from enough.

First because part of the output is XML. Second, it is not exactly easy to understand the output.

Worst, the XML specification depends on the job underline mechanism: if the job is based on a plugin, the XML format will depend on that plugin, not on Jenkins.

Getting documentation about the XML format is another challenge.

Solution

What the jenkins-jobs project tries to do is to map desired information from the XML format based on the Python classes under jenkins_jobs.jobs module.