Mutation score 75/80 for the command entry and exit codes

[jonbaldie]

Parent

Part of #33.

What to build

Strong tests for the command entry: argument parsing, option validation, exit
code selection, and the usage text. This includes --ignore-tests,
--minimumpriority, --strict, --reportfile, --only,
--ignore-errors-on-exit, and --ignore-violations-on-exit.

Exit code selection is the contract in the README table: 0 for a clean run, 1
for an error, 2 for findings, and errors take precedence. Today a mutation of a
precedence branch or an option name can escape. After this ticket the tests
cover each option, each rejection message, and each exit code path.

Acceptance criteria

• The gate command scoped to src/lib.rs and src/main.rs exits 0. It prints a total MSI of 75
  percent or more, and a covered-code MSI of 80 percent or more.
• The run uses --coverage and the full Cargo test suite, less the
  packaging smoke test.
• Every new test calls messrust::run. It asserts the exit code, the
  standard output text, or the standard error text. No new public item
  exists only to make a test possible.
• Production code that no test can reach through the command is deleted. A
  #[cfg(test)] unit test does not replace deletion.
• No rule behaviour changes, unless a mutant proves the code is dead.
• No mutator-disable comment is added under src. No baseline file
  and no blacklist file is added.
• cargo test passes. cargo clippy --all-targets adds no new warning.
• A comment on this ticket records the two scores.

Blocked by

• Mutation measurement harness: mutarust.yml, README, per-module command #34 — the measurement harness gives the command and the start scores.

[jonbaldie]

Handoff: work in progress on branch mutation/issue-40-cli-entry (pushed to origin, not yet a PR).

Done so far:

• One commit: "Cover CLI arg-parsing edge cases for issue Mutation score 75/80 for the command entry and exit codes #40 mutation gate", adding tests/cli.rs (125 lines) — covers bare 'help' word, unknown single-dash flags, lone '-' positional, missing option value, --minimumpriority/--maximumpriority integer/range validation, unprefixed --suffixes normalization.

Not done:

• Mutation score (MSI / covered-MSI) not yet measured against the 75/80 gate.
• No PR opened yet.

Next agent: check out mutation/issue-40-cli-entry, run mutarust with coverage against the in-scope CLI entry file(s), close remaining gaps, then open a PR with body "Closes #40".

[jonbaldie]

Process note: function-scoped mutarust batches only

On 16 GB hosts, a full-file mutarust --coverage run with the full suite (even at --workers 1 and CARGO_BUILD_JOBS=1) can push multi‑tens of GB of memory pressure and freeze the machine. Root-module targets such as src/lib.rs are worst case: every mutant rebuilds the whole crate, then runs every integration test.

Until a module’s scores look safe, agents must measure only in small function-scoped batches:

1. Prefer --match <function_name> on one function at a time (or a tiny related set).
2. Prefer a narrow test filter while iterating, e.g.
   --test-flags "--test <relevant_test_file> --lib"
   not the full suite.
3. Keep --workers 1 and CARGO_BUILD_JOBS=1 (or lower).
4. On macOS, set TMPDIR to a real path under $HOME before any --coverage run (see README). Default /var/folders breaks LCOV path restore.
5. Do not start a full-file acceptance gate (--coverage, full suite less pack_install_smoke, scoped only by file) until batch scores are clearly near or above 75 / 80.
6. Kill runaway mutarust / rustc / cargo processes and clear mutarust-* temp dirs if memory spikes.

Final acceptance for each child still needs the issue’s documented gate command when ready — but only after batches show the gate is likely to pass without thrashing the host.

[jonbaldie]

Working on this now on branch issue-40-cli-entry. Function-scoped mutarust batches only; host free RAM is low so measurement waits until free memory recovers.

[jonbaldie]

Handoff / progress (halted)

Branch: issue-40-cli-entry (pushed)
Commit: strengthens tests/cli.rs (59 tests via messrust::run) and adds scripts/mutarust-fn-batch.sh (memory-gated single-function batches).

Done

• Edge-case coverage for help/version, unknown flags, lone -, missing option values, priority validation, suffix normalization, usage text, comma-separated paths/rulesets/excludes, --only vs --enable precedence, ignore-errors-only exit 0, silent non-verbose load, unknown format available-list, reportfile parent missing, etc.
• cargo test --test cli: 59 passed
• Function-scoped baseline (no coverage, --test cli --lib):
  • handle_info_flags: 100% MSI (1 killed, 3 skipped compile-fail, 0 escaped)
  • print_usage: 0 mutants under --match (string body; needs different approach)
  • Mutant counts (--no-exec): parse_args 32, run 36, apply_flag_option 13, apply_value_option 11, parse_priority 6, suffix_list 2, source_suffixes 2

Not done

• Remaining function batches (parse_args, run, flags/values, analysis helpers)
• Covered MSI / full acceptance gate on src/lib.rs + src/main.rs
• PR

Halt reason

Host free RAM collapsed during parse_args batch (workers=1 still heavy because crate-root lib.rs rebuilds). Stopped on request; temps and rust processes cleared.

Next agent

1. Wait until free RAM is comfortably high (several GB free).
2. Use only scripts/mutarust-fn-batch.sh <fn> (or equivalent): --match, --workers 1, CARGO_BUILD_JOBS=1, --test-flags "--test cli --lib", macOS TMPDIR under $HOME.
3. Clean $TMPDIR/mutarust-* after every batch; kill leftover test binaries under that tree.
4. Do not run full-file --coverage acceptance until batch scores look near 75/80.
5. Open PR with closes Mutation score 75/80 for the command entry and exit codes #40 only after gate passes and CI is green.

[jonbaldie]

Command-entry gate recorded (child #71). Final acceptance run on the gate command — --coverage, full suite less pack_install_smoke, scoped to src/lib.rs and src/main.rs, --workers 1, CARGO_BUILD_JOBS=1, thresholds 75/80:

• Total MSI: 92.06% (≥ 75 required)
• Covered MSI: 98.31% (≥ 80 required)
• Mutants: 126 total — 98 killed, 1 errored (infinite-loop timeout, counts as killed), 17 skipped (compile-fail), 2 escaped, 8 not covered
• Command exited 0; reproduced twice with identical scores.

The 2 escapes are behaviorally equivalent mutants in run (return EXIT_SUCCESS → return 0, return code → return 0); no test can kill them while handle_info_flags only returns EXIT_SUCCESS (0). The 8 not-covered lines lower only total MSI and are accepted at the recorded score. Full numbers on #71; probe table on #70 (closed).