Provide official Maven aggregator plugin that enforces thresholds on aggregated PIT results

[verhasi]

Problem

The existing report-aggregate goal produces a combined report but does not enforce thresholds.
I attempted to use an aggregator plugin, but only org.pitest:pitest-aggregator exists and it is a library jar (no plugin.xml), so Maven cannot execute it as a plugin.

Expected

An official Maven plugin (or goal) that can:

• aggregate PIT results across modules, and
• enforce thresholds (mutation, coverage, test strength) on the aggregated data.

Use Case

Multi-module build with production code with unit tests + integration tests split, where only the overall quality gate should decide build status.

Suggested solution

A new aggregate goal in pitest-maven that can enforce thresholds

Environment

• Maven: 3.9.8
• Java: 21.0.6
• PIT: 1.22.0

[verhasi]

In the meantime I have figured out that at least partially it is available. Even it is not documented, you can use the following configuration items:

                    <aggregatedTestStrengthThreshold>85</aggregatedTestStrengthThreshold>
                    <aggregatedMutationThreshold>80</aggregatedMutationThreshold>
                    <aggregatedMaxSurviving>80</aggregatedMaxSurviving>

Actually the coverage threshold is missing. Comparing the PitMojo.java as

      if (result.isPresent()) {
        throwErrorIfTestStrengthBelowThreshold(result.get().getMutationStatistics());
        throwErrorIfScoreBelowThreshold(result.get().getMutationStatistics());
        throwErrorIfMoreThanMaximumSurvivors(result.get().getMutationStatistics());
        throwErrorIfCoverageBelowThreshold(result.get().getCoverageSummary());
      }

to the AbstractPitAggregationReportMojo.java as

      AggregationResult result = reportAggregator.aggregateReport();

      throwErrorIfTestStrengthBelowThreshold(result.getTestStrength());
      throwErrorIfScoreBelowThreshold(result.getMutationCoverage());
      throwErrorIfMoreThanMaximumSurvivors(result.getMutationsSurvived());

the difference is clear.
I was looking for the source code as the documentation missed these options. At first I tried to look at the tests of aggregation but unfortunately there is no test for this area!!! I plan to create a PR to add tests at least.

[verhasi]

One more interesting issue with the aggregated results. If my understanding is good the mutation coverage and the mutation score are aliases with the same meaning. The result of the aggregation shows 85% on the html report but 46% as the enforcement. I accept reading the documentation that in case of overlapping of the tested class in the sources of aggregation the result is not reliable but even this case they should be equal on the different report outputs.