[JENKINS-58102] GlobalAnnotator does not annotate lines that are part of a multi-line color sequence

[basil]

See JENKINS-58102. Fixes a bug introduced in #25 that is only observable when timestamper is used with Pipeline and the multi-line escape sequence support introduced in jenkinsci/ansicolor-plugin#137. The problem is that GlobalAnnotator doesn't account for the <span> elements introduced by ansicolor when looking for timestamps to annotate. The fix is similar in approach to the existing code that accounts for the <span> elements introduced by Pipeline.

I wrote a new test that exercises this use case. When the test is executed against Jenkins 2.145 and ansicolor 0.6.2 (with the src/main changes from this PR reverted), it fails with the following stack trace:

15:23:35 [2019-06-23T22:23:35.328Z] Finished: SUCCESS expected:<21> but was:<17>
	at org.junit.Assert.fail(Assert.java:88)
	at org.junit.Assert.failNotEquals(Assert.java:834)
	at org.junit.Assert.assertEquals(Assert.java:645)
	at hudson.plugins.timestamper.pipeline.PipelineTest.globalDecoratorAnnotator(PipelineTest.java:101)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50)
	at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12)
	at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:47)
	at org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17)
	at org.junit.internal.runners.statements.RunBefores.evaluate(RunBefores.java:26)
	at org.junit.internal.runners.statements.RunAfters.evaluate(RunAfters.java:27)
	at org.jvnet.hudson.test.JenkinsRule$1.evaluate(JenkinsRule.java:552)
	at org.junit.internal.runners.statements.FailOnTimeout$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)

When the test is executed against Jenkins 2.145 and ansicolor 0.6.2 (with the src/main changes in this PR), the test passes.

Unfortunately, I couldn't use ansicolor 0.6.2 in this PR because it requires Jenkins 2.145 or later, and only 73% of timestamper users are on Jenkins 2.145 or later. Rather than drop support for 27% of the user base for the sake of one test, I compromised by using ansicolor 0.5.3, whose minimum Jenkins version is 2.121.2. When run against ansicolor 0.5.3, the new test isn't exercising the changes in src/main from this PR. However, it is exercising the bulk of the new logic introduced in #25, which is still a net improvement from the state of affairs before this PR.

[jglick]

only 73% of timestamper users are on Jenkins 2.145 or later

But virtually all of those who have already updated to 1.9 are. Just bump the jenkins.version, say to 2.150.3, and be done with it.

[basil]

Just bump the jenkins.version, say to 2.150.3, and be done with it.

Done.

[jglick]

A tested fix, good. The broader issue is that GlobalAnnotator should not have to be doing these hacks to begin with, which suggests a logic error somewhere in console handling.