Unit tests

[stopalopa]

https://issues.jenkins-ci.org/browse/JENKINS-57864

[stopalopa]

Couple of questions/issues

1. Is there a way to add a compareTo method and pass findBugs without also overriding the equals and hashCode methods?

2. Haven't figured out yet why the second listPlugins unit test isn't working...I think I implemented it the same way I tested getting console output for unit testing the CLI.

3. For checking the listPlugins input, I remember \n caused some issues on Windows and I ended up trimming the ends for the previous unit tests. I'm not sure if that's going to work for this one since there's so many new lines, unless I make each new line it's own string. Should I strip out all white space or would something like converting the \n to %n be a better solution?

[timja]

Couple of questions/issues

1. Is there a way to add a compareTo method and pass findBugs without also overriding the equals and hashCode methods?

You should always override equals and hashCode if you're comparing objects, your IDE should be able to generate sensible ones.

[timja]

please no see my comment in the main thread

[stopalopa]

Ok, good to know. I thought in particular generating the hashCode method was going to be really complicated, but that was fairly easy.

[timja]

you could probably extract this to a method as it's called 3 times and then use something like:

Suggested change

	.forEach(p -> System.out.println(p.getKey() + " " + p.getValue()));
	.forEach(this::logPlugin);

[timja]

?

[stopalopa]

It's currently commented out because it's not working. For some reason the value returned/output to the console is empty.

[kwhetstone]

Looks like this was hit with the pom version issues. The tests are not on the CI server.

[timja]

what do you mean @kwhetstone ?
https://ci.jenkins.io/blue/organizations/jenkins/Tools%2Fplugin-installation-manager-tool/detail/PR-58/1/tests

[kwhetstone]

Random thought: the windows failure in the test could be line endings. You might have to do more of line by line comparison to see if the output is being printed correctly,

[stopalopa]

Random thought: the windows failure in the test could be line endings. You might have to do more of line by line comparison to see if the output is being printed correctly,

Yeah, I knew this was potentially going to be a problem. See my question #3. Is there an easier way for me to test without having to push a new commit each time?

[timja]

Random thought: the windows failure in the test could be line endings. You might have to do more of line by line comparison to see if the output is being printed correctly,

Yeah, I knew this was potentially going to be a problem. See my question #3. Is there an easier way for me to test without having to push a new commit each time?

Windows VM?

[kwhetstone]

I think what you want to do is a simple string replace for the generated string like content.replaceAll("\\r\\n?", "\n") We really don't care about what the invisible characters look like, so hard replacing them shouldn't break the display testing. I can't find anything about libraries testing for this particular thing.