[MNG-6268] When a reactor build fails Maven should include -f (if used) in command line suggestion

[michael-o]

I don't know whether this should be included, but it is trivial to provide. I will let the community decide. If accepted someone should write an IT for that.

[rfscholte]

Writing a unittest is much easier in this case and sufficient enough.

[eolivelli]

LGTM

I agree that a unit test is enough here

[michael-o]

Just scratching my head how the full boostrap with a unit test would work here since I'd need to analyze stdout/err.

[rfscholte]

IIRC @mthmulders had a similar usecase recently

[mthmulders]

IIRC @mthmulders had a similar usecase recently

Not sure if I understand correctly what you refer to. I recently had a contribution to Plexus Archiver where I wrote a CapturingLog. But given that this is SLF4J, it might be easier to go with something like SLF4J Test.

[michael-o]

The SFL4J test looks promising, but wouldn't it be easier to do just an IT instead of going through a new component?

[rfscholte]

Another option is to increase the visibility of logBuildResumeHint( String resumeBuildHint ) a bit, in the unittest making a new MavenCli with an override of this method. To me an IT is overkill for this check, as it is simple input/output validation, only the difficulty here is that it is writing directly the the logging framework. Better adjust the class to make it testable

[michael-o]

Another option is to have no test at all since this change is straight forward. Pipe input to output unmodified as you can see.

[rfscholte]

I'm fine with no test in this case.

[michael-o]

The issue has been closed as wontfix. So closing this, but leaving the branch for a while.