Resolve Pitest Issues - CustomImportOrderCheck (1)

[rnveach]

Child issue of #7797 ,

checkstyle/.ci/pitest.sh

Line 77 in abf829f

"CustomImportOrderCheck.java.html:<td class='covered'><pre><span class='survived'> &#38;&#38; matcher.start() &#60; betterMatchCandidate.matchPosition) {</span></pre></td></tr>"

[wltan]

I am on it

[wltan]

Pitest report:
https://wltan.github.io/checkstyle-reports/2020-03-11/customimport-pitest1/com.puppycrawl.tools.checkstyle.checks.imports/CustomImportOrderCheck.java.html#org.pitest.mutationtest.report.html.SourceFile@4d2b49be_816

Surviving mutation:
removed conditional - replaced comparison check with false → SURVIVED

Hardcoded mutation:
wltan@52eab4d

Regression:
https://wltan.github.io/checkstyle-reports/2020-03-11/pitest1-regression/diff/

[wltan]

Regression raised no discrepancies at all. However, I may have found the cause of the surviving mutation.

We already have unit tests in CustomImportOrderCheckTest.java that covers the different cases of regex matching, but the problem is that the input file is currently designed such that even if the test were to "fail" with the mutation, the output (violations) does not change, and so the tests incorrectly pass.

[rnveach]

Pitest report looks good.

Hardcoded report looks good.

Regression report looks done correctly. This is a complex check with a lot of properties. If regression didn't help in this case, I would be looking to see if I turned on/off all the right options. Since a PR is started with a test case this probably isn't necessary.

See #7801 (comment) on brief discussion what would happen if no case could be created.

[rnveach]

Fix was merged