pitest: increase mutation coverage for pitest-checks-misc profile to 100%

[romani]

We should increase coverage for pitest-checks-misc profile up to 100%.
This issue is a subtask of #3708

Current threshold of pitest-checks-misc profile: XX

[romani]

@Nimfadora , FYI, missed set of classes. I created new profile.

[romani]

fix is merged

[romani]

problematic test is reverted to unlock release process. See issue #5141
https://github.com/checkstyle/checkstyle/blob/master/src/main/java/com/puppycrawl/tools/checkstyle/checks/TranslationCheck.java
Now uncovered mutations are:

[rnveach]

Latest report: http://rveach.no-ip.org/checkstyle/regression/pitest-reports/13/

Report shows TrailingCommentCheck is not at 100% code coverage, and TranslationCheck is missing mutation coverage.

[rnveach]

@romani Missing mutation coverage for TranslationCheck that was reverted might be brought back if Issue #5141 is fixed. It has a PR that has been sitting.

[romani]

the only missed coverage we have is

removed call to com/puppycrawl/tools/checkstyle/api/MessageDispatcher::fireFileFinished → SURVIVED

@Nimfadora , please help us to finish this issue.

[Nimfadora]

@romani, ok, I will do it today.

[Nimfadora]

@romani xml output varies only in order of lines when I comment out line with missed coverage. And comment in testLogOutput says:

            // order is not always maintained here for an unknown reason.
            // File names can appear in different orders depending on the OS and VM.
            // This ensures we pick up the correct file based on its name and the
            // number of children it has.

So, I don't really understand how to test it, if it influences only order, which we can't check by known reasons.

XmlLogger#fileMessages are rewritten two times, when we call XmlLogger#fileFinished() and XmlLogger#auditFinished() (called from Checker#process). This causes situation when selected line doesn't really change anything special, just rewrites messages in different order.

[romani]

@Nimfadora , if line of code does not affect anything, and generated xml report is still correct (all violations reported on proper files, no false-postives).
Such line should be removed from code completely - please do.

[rnveach]

@romani
Please look over #5057 in terms of this fire call.
Just removing this line will have weird affects on any listener as we are sending violations without triggering a file has been started or a previous file is done.
Other issue might be the reason this one is so hard to kill.

See

checkstyle/src/main/java/com/puppycrawl/tools/checkstyle/XMLLogger.java

Line 141 in 3b531fe

public void fileFinished(AuditEvent event) {

Probably why this isn't the case for this is because auditFinished finishes the job, but it shouldn't have to worry about it as fileFinished should be handling it.

Edit: I can confirm that fixing that issue will make pitest 100%.

[romani]

Final fix is merged