Reporting

report renders report.json in the formats other tools already read. -mutants is required (it carries the source locations), the shards of a run are passed together, and -srcs names the files or directories holding the mutated sources.

# Stryker mutation-testing-report-schema v2, which the Stryker dashboard accepts.
mutrim report -mutants mutants.json -srcs ./pkg report.json > mutation-report.json

# The same JSON inside the single-file mutation-test-report-app viewer.
mutrim report -format html -mutants mutants.json -srcs ./pkg -o mutation-report.html report.json

# One ::warning per surviving mutant, which GitHub Actions shows on the diff;
# -max-per-line 1 surfaces one mutant per line, as Google's code review does.
mutrim report -format github -max-per-line 1 -mutants mutants.json report.json

KILLED, LIVED, TIMEOUT and NO_COVERAGE map onto the schema’s Killed, Survived, Timeout and NoCoverage; RUN_ERROR is a RuntimeError, which the schema keeps out of the score; NOT_VIABLE is a CompileError, since the mutant never built, and IGNORED, SKIPPED, EQUIVALENT and SUSPECT_EQUIVALENT are Ignored, with the directive or filter that suppressed it, not in the diff, or the equivalence, as its statusReason (a suspect is Survived under -count-suspect). Each mutant carries the tests that reach it (coveredBy) and the ones that killed it (killedBy), so the viewer shows the kill matrix per mutant. A mutant no report mentions is left out, so one shard’s report renders that shard. A mutant’s diff from mutants.json follows its description in the Stryker report and its message in a GitHub annotation.

-format html and -format github write their own text rather than JSON: GitHub reads its annotations from the step’s stdout, and the HTML view loads the viewer from unpkg, so it needs a network at display time but is a single file to publish.