Fix flakiness by synchronising the State

[Pldi23]

Description:

Cloudbees CI reported a flake in
org.jenkinsci.plugins.workflow.support.steps.StageStepTest.serializability

Error Log:
stderr.txt

How to reproduce:
I wasn't managed to reproduce it locally without code changes, but if we add TimeUnit.MILLISECONDS.sleep(200); before this line the StageStepTest.serializability started to flake regularly.

Assumption:
My assumption that there is a deadlock possibility.

Solution:
As a solution I want to suggest to synchronize the State local variable (the same synchronisation mechanism is used here)

• Make sure you are opening from a topic/feature/bugfix branch (right side) and not your main branch!
• Ensure that the pull request title represents the desired changelog entry
• Please describe what you did
• Link to relevant issues in GitHub or Jira
• Link to relevant pull requests, esp. upstream and downstream changes
• Ensure you have provided tests - that demonstrates feature works or fixes the issue

[Pldi23]

I did not notice some tests that failed because of the race condition here, but since we are synchronizing the success method, then we need to synchronize here too.

[Pldi23]

Waiting for a review from maintainers

[Pldi23]

@jglick could you please review

[jglick]

IIUC to match

workflow-support-plugin/src/test/java/org/jenkinsci/plugins/workflow/test/steps/SemaphoreStep.java

Lines 141 to 148 in a1c7106

	synchronized (s) {
	while (!s.started.contains(k)) {
	if (b != null && !b.isBuilding()) {
	throw new AssertionError(JenkinsRule.getLog(b));
	}
	s.wait(1000);
	}
	}