[JENKINS-50476] offer a way to assert that Jenkins won't start #96
Conversation
| if(entry.getValue()) { | ||
| throw e; | ||
| } | ||
| // Failure ignored as requested |
There was a problem hiding this comment.
Might be worth logging something to indicate we saw expected failure, so we know rule ran okay.
| try { | ||
| j.apply(step, description).evaluate(); | ||
| if (!entry.getValue()) { | ||
| Assert.fail("The current JenkinsRule should have failed to start Jenkins."); |
There was a problem hiding this comment.
Super-🐜 - I like to phrase this as "X happened but we saw Y" because that makes it clearer.
| /** | ||
| * List of {@link Statement}. For each one, the boolean value says if Jenkins is expected to start or not. | ||
| */ | ||
| private final Map<Statement, Boolean> steps = new LinkedHashMap<>(); |
There was a problem hiding this comment.
🐜 I would declare the type here as explicitly LinkedHashmap since no other type could be legally used since we need ordering preserved.
What we're doing here is a bit dodgy (using a LinkedHashmap to simulate a List of Tuple types, with no intent to really use mapping)... but it's also much less verbose than the "proper" alternative and doesn't require adding Apache Commons Lang with a modern version to get tuple types.
|
@svanoort thanks a lot for the review. I will address your comments in a followup PR. I'm merging and releasing to make the associated core PR ready (instead of using a JTH SNAPSHOT). |
JENKINS-50476
Used in jenkinsci/jenkins#3364
@reviewbybees