[core] Fix Kotest based tests

[adangel]

Describe the PR

The changes are in pmd-java only:

• Renamed BranchingExprsTestCases to BranchingExprsTests so that it is picked up by surefire
• Fixed BranchingExprsTests (the test code sample wouldn't actually be compileable, so I guess, expected a wildcard for the type is ok)
• Added a check in pmd-java's ParserTestSpec, so that test sets ("container" in kotest speak) that don't have any test cases, are failing. This makes sure, that the tests are reported correctly via surefire (only for single test cases the succeed counter is incremented). Now such tests don't end up with ("Tests run: 0...") anymore.
• Fixed tests in pmd-java, that didn't define test cases yet. I simply wrapped the test code inside a "doTest", which registers a test case. Some tests did this already, but many didn't. This created many whitespace changes due to indentation/reformatting.
• Update build-tools to pick up Improve surefire reporter build-tools#31 - note: this is only cosmetical, so that junit suites are displayed as indented cases.

Related issues

• Fixes [core] Kotest tests aren't picked up by surefire #5009
• Refs Improve surefire reporter build-tools#31

Ready?

• Added unit tests for fixed bug/feature
• Passing all unit tests
• Complete build ./mvnw clean verify passes (checked automatically by github actions)
• Added (in-code) documentation (if needed)

[pmd-test]

	1 Message
📖	No regression tested rules have been changed.

Generated by 🚫 Danger

[oowekyala]

This looks good, thanks :)

[adangel]

FYI - I was now investigating the root cause for the following problem: We have a disabled test in TypeDisambiguationTest (disabled via bang) - this is listed correctly when the test is running, but it is missing in the summary. So, disabled kotest tests can be overseen very easily:

$ ./mvnw surefire:test -pl pmd-java -Dtest=TypeDisambiguationTest,CtorInvocMirrorTest,ReportTest
...
[INFO] -------------------------------------------------------
[INFO]  T E S T S
[INFO] -------------------------------------------------------
[INFO] Running net.sourceforge.pmd.lang.java.ReportTest
[INFO] Tests run: 6, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 1.136 s -- in net.sourceforge.pmd.lang.java.ReportTest
[INFO] Running net.sourceforge.pmd.lang.java.ast.TypeDisambiguationTest
[INFO]     └─ ↷  TODO Import on demand of class defined in same compilation unit that has an extends clause
[INFO] Tests run: 14, Failures: 0, Errors: 0, Skipped: 1, Time elapsed: 2.856 s -- in net.sourceforge.pmd.lang.java.ast.TypeDisambiguationTest
[INFO] Running net.sourceforge.pmd.lang.java.types.internal.infer.ast.CtorInvocMirrorTest
[INFO] Tests run: 2, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 0.206 s -- in net.sourceforge.pmd.lang.java.types.internal.infer.ast.CtorInvocMirrorTest
[INFO] 
[INFO] Results:
[INFO] 
[INFO] Tests run: 21, Failures: 0, Errors: 0, Skipped: 0

It turns out, that the surefire/junit-platform integration is tailored against methods based test frameworks, where each test is based on a method. For kotest, when it reports the tests up to junit-platform, it only provides a class source (and not a method source): https://github.com/kotest/kotest/blob/6f371fbd0af65577f9a5092d3abab017613afa98/kotest-runner/kotest-runner-junit5/src/jvmMain/kotlin/io/kotest/runner/junit/platform/descriptors.kt#L34
When surefire receives such a TestDescriptor, it tries to convert it here https://github.com/apache/maven-surefire/blob/790e332790b1d601cb58e566c29eef3e47353f7d/surefire-providers/surefire-junit-platform/src/main/java/org/apache/maven/surefire/junitplatform/RunListenerAdapter.java#L366 into classname and methodname, but the method name will always be null for kotest tests. This is used to create a ReportEntry here: https://github.com/apache/maven-surefire/blob/790e332790b1d601cb58e566c29eef3e47353f7d/surefire-providers/surefire-junit-platform/src/main/java/org/apache/maven/surefire/junitplatform/RunListenerAdapter.java#L218 - so the report entry for the test has a classname, but no method name. These ReportEntries are collected and converted into TestMethodStats instances (https://github.com/apache/maven-surefire/blob/790e332790b1d601cb58e566c29eef3e47353f7d/maven-surefire-common/src/main/java/org/apache/maven/plugin/surefire/report/TestSetRunListener.java#L288). The "classMethodName" will be something like "TestClassName.null" - for all test cases in the same kotest class. These TestMethodStats are later on accumulated: https://github.com/apache/maven-surefire/blob/790e332790b1d601cb58e566c29eef3e47353f7d/maven-surefire-common/src/main/java/org/apache/maven/plugin/surefire/report/DefaultReporterFactory.java#L243 - the grouping happens by "ClassMethodName" (Map<String, List<TestMethodStats>> mergedTestHistoryResult). We still have the one skipped TestMethodStat and a couple of other, but they are dealt here as multiple executions of the same test (probably for the feature to automatically repeat failing tests x times until they succeed - flaky tests). But the tests should actually be dealt as separate test cases...
For now, I don't see how this could be fixed without a change to either surefire or kotest. I think, this problem is also the cause for kotest/kotest#4024 ...