[JENKINS-62014] Add support for build step environment filters

[daniel-beck]

See JENKINS-62014.

This adds support for global build step environment filters to durable task based pipeline steps.

Downstream from jenkinsci/jenkins#4683.
Upstream from jenkinsci/workflow-durable-task-step-plugin#135.

[daniel-beck]

Wasn't me:

Failed to build image (1): 32edfdd58111
Stacktrace
java.lang.Error: Failed to build image (1): 32edfdd58111
	at org.jenkinsci.test.acceptance.docker.Docker.build(Docker.java:131)
	at org.jenkinsci.test.acceptance.docker.Docker.build(Docker.java:174)
	at org.jenkinsci.test.acceptance.docker.Docker.build(Docker.java:160)
	at org.jenkinsci.test.acceptance.docker.Docker.build(Docker.java:140)
	at org.jenkinsci.test.acceptance.docker.DockerRule.build(DockerRule.java:67)
	at org.jenkinsci.test.acceptance.docker.DockerClassRule$1.evaluate(DockerClassRule.java:63)
	at org.junit.rules.ExternalResource$1.evaluate(ExternalResource.java:48)
	at org.jvnet.hudson.test.JenkinsRule$1.evaluate(JenkinsRule.java:598)
	at org.junit.internal.runners.statements.FailOnTimeout$CallableStatement.call(FailOnTimeout.java:298)
	at org.junit.internal.runners.statements.FailOnTimeout$CallableStatement.call(FailOnTimeout.java:292)
	at java.util.concurrent.FutureTask.run(FutureTask.java:266)
	at java.lang.Thread.run(Thread.java:748)
Standard Output
=== Starting org.jenkinsci.plugins.durabletask.EncodingTest
Building Docker image `docker build -t jenkins/sshd:32edfdd58111 C:\Jenkins\workspace\ugins_durable-task-plugin_PR-124\target\tmp\Dockerfile8705698837630806719dir`: logfile is at C:\Jenkins\workspace\ugins_durable-task-plugin_PR-124\target\tmp\docker-JavaContainer-build8578491675782530461.log
---%<--- docker-JavaContainer-build8578491675782530461.log
Sending build context to Docker daemon  11.78kB
Step 1/7 : FROM ubuntu:bionic
bionic: Pulling from library/ubuntu
no matching manifest for windows/amd64 10.0.17763 in the manifest list entries
--->%---

[dwnusbaum]

Hmm, I recall @car-roll fixing those issues in #121 by changing some of the assumptions in the tests, so I'm not immediately sure what is happening.

[car-roll]

That's really strange this issue is showing up again. The issue I ran into was because of the loading order of ClassRule and BeforeClass.
You can see the changes I made here

[daniel-beck]

Weird, these commits are in my history -- this PR is based off of latest master.

[daniel-beck]

Re-ran and it fixed a bunch of tests, but the quoted one remains.

[car-roll]

I wonder if the unix() and assumeDocker() functions should just be replicated within EncodingTest.setUp()?

[jeffret-b]

This looks good, though I'm not an expert in this area.

I'm not able to get the tests to complete locally. It seems to just hang on BourneShellScriptTest.

[jeffret-b]

I'm not sure about the best approach here. It doesn't seem wise to use this flag. It's a big switch and can hide other potential issues.

I guess this can't be merged until after all supported versions have the new API anyway. We've got enough pieces involved here and demonstrations of behavior that the Beta restriction doesn't help. The purpose is to make these API changes available and then use them as soon as possible. Marking them as Beta, waiting for propagation, and then removing the restriction, and waiting for further propagation makes little sense.

[jeffret-b]

I tested basic operations with a few different configurations. Everything worked as expected.

The console log messages are a little weird, but it looks like they don't originate here.

This PR could use an improved description of what it adds and how it needs to be tested.

[dwnusbaum]

Latest failures look like flaky infra issues, rebuilding:

EC2 (aws) - Windows 2019 (i-054e26c30a9572cef) was marked offline: Connection was broken: java.io.EOFException

[dwnusbaum]

Looks like #125 should fix the CI build.

[dwnusbaum]

Rebuilding to pick up #125.

[daniel-beck]

On-holding due to feedback in downstream PR.