Why isn't Groovy (Jenkins pipeline) not recognizing class (static) variables?

Question

I'm following an example in Getting To Know Groovy Variables And Their Functions and have the following. EnvironmentMgr is a singleton class I've defined somewhere else, and it works.

@Singleton
class EnvironmentMgr {
  String ecsCluster = "sandbox"
  ...
}

abstract class EcsService {
  def setup() {
    if (envMgr == null) {
      println "envMgr is null"
      envMgr = EnvironmentMgr.instance
    }
  }

  protected static def envMgr = null
}

class Service1 extends EcsService {
  def setup() {
    super.setup()
    // Do something for Service1
  }
}

class Service2 extends EcsService {
  def setup() {
    super.setup()
    // Do something for Service2
  }
}

def svc1 = new Service1()
svc1.setup()
def svc2 = new Service2()
svc2.setup()

I expect the println statement in EcsService:setup() to be printed once, but it's printed both calls. Why is that? How can I make it work? Thanks.

EDIT: The above code works as expected as straight Groovy code on my Mac, but when run in a Jenkins pipeline, the static-ness isn't recognized.

Answer 1

It seems like the issue is due to the envMgr property is not behaving as a static property in the Jenkins environment. To ensure that the envMgr property is a shared/static property across all instances of EcsService in a Jenkins pipeline, you can explicitly declare it as a static property using the static keyword:

abstract class EcsService {
  static def envMgr = null  // Declare envMgr as a static property

  def setup() {
    if (envMgr == null) {
      println "envMgr is null"
      envMgr = EnvironmentMgr.instance
    }
  }
}