Kill lib.rs CLI escapes via function and mutator slices (no coverage)

[cursor]

Parent

Part of #40.

What to build

Raise kill rate on command-entry mutants in src/lib.rs using only memory-safe batches: one function and one mutator group at a time, no --coverage, narrow --test cli --lib. Add kill tests through messrust::run for every escaped mutant that is not proven equivalent. Delete production code that no command path can reach.

Acceptance criteria

• Every CLI entry function is measured with function-scoped + mutator-group batches (--workers 1, CARGO_BUILD_JOBS=1).
• Escaped mutants have a kill test via messrust::run, or a short note that the mutant is equivalent / dead code removed.
• No --coverage and no full-suite mutation run.
• Free RAM abort stays in force; temps under $TMPDIR/mutarust-* are cleared after each batch.
• cargo test passes. No new clippy warnings. No mutator-disable / baseline / blacklist.

Blocked by

• Land CLI messrust::run tests (no mutation score gate) — the sibling ticket created first under Mutation score 75/80 for the command entry and exit codes #40.

[jonbaldie]

This was generated by AI during triage.

Agent Brief

Category: enhancement
Summary: Increase mutation kill rate on command-entry functions in src/lib.rs using memory-safe, function-scoped batches without coverage

Current behavior:
Basic CLI tests call messrust::run for standard flag combinations (landed in #68), but several mutation escapes remain in src/lib.rs across CLI parsing and execution functions (e.g. parse_args, run, apply_flag_option, apply_value_option, parse_priority, suffix_list, source_suffixes). Previous full-file mutation runs exhausted host memory.

Desired behavior:
Every command-entry function in src/lib.rs is evaluated using memory-safe, single-function and mutator-group batches without --coverage. Escaped mutants are killed by adding integration tests that exercise messrust::run and assert exit codes or text output. Equivalent mutants or unreachable dead code are documented or removed.

Key interfaces:

• Entry point messrust::run(args, stdout, stderr): tests must invoke this function directly rather than introducing unit seams
• Command line arguments parsing and validation in src/lib.rs (parse_args, apply_flag_option, apply_value_option, parse_priority, run)
• Integration test suite in tests/cli.rs

Acceptance criteria:

• Every CLI entry function is measured with function-scoped + mutator-group batches (--workers 1, CARGO_BUILD_JOBS=1)
• Escaped mutants have a kill test via messrust::run, or a note explaining equivalence or dead code removal
• No --coverage and no full-suite mutation run in this issue
• Free RAM abort guards remain in force and temporary directories under $TMPDIR/mutarust-* are cleaned up after each batch
• cargo test passes and cargo clippy --all-targets introduces no new warnings
• No mutator-disable annotations, baseline files, or blacklist files are added

Out of scope:

• Running --coverage batches (deferred to Covered-MSI probe per CLI function (coverage, narrow tests) #70)
• Running the full 75/80 acceptance gate across src/lib.rs and src/main.rs (deferred to Final 75/80 mutarust gate for command entry (lib.rs + main.rs) #71)
• Changing ruleset behavior or adding new CLI features

[jonbaldie]

Completion Summary

All CLI entry functions in src/lib.rs have been measured with function-scoped and mutator-group batches without --coverage (--workers 1, CARGO_BUILD_JOBS=1, --test-flags "--test cli --lib").

Measurement Results

Function	Mutants	Killed	Escaped	Errored	Skipped	MSI
handle_info_flags	4	1	0	0	3	100.00%
print_usage	0	0	0	0	0	N/A
parse_priority	6	6	0	0	0	100.00%
source_suffixes	2	0	0	0	2	100.00%
suffix_list	2	0	0	0	2	100.00%
split_list	0	0	0	0	0	N/A
required_value	0	0	0	0	0	N/A
apply_flag_option	13	13	0	0	0	100.00%
apply_value_option	11	11	0	0	0	100.00%
prepare_analysis	1	1	0	0	0	100.00%
run_analysis	10	6	0	0	4	100.00%
parse_args	32	29	0	2 (timeouts)	1	100.00%
run	32	26	2 (equivalent)	0	4	93.75%
default (Options)	4	2	1 (equivalent)	0	1	75.00%

Escaped Mutants and Equivalence Notes

Across all CLI entry functions, 3 escaped mutants remain, and all 3 are proven behaviorally equivalent:

1. src/lib.rs:90 in run:
   return EXIT_SUCCESS; -> return 0; (statement/return)
   Equivalence note: EXIT_SUCCESS is pub const EXIT_SUCCESS: i32 = 0;. Returning literal 0 produces identical behavior.

2. src/lib.rs:95 in run:
   return code; -> return 0; (statement/return)
   Equivalence note: handle_info_flags returns Some(EXIT_SUCCESS) (which is 0) on info flags (--version, --help, -h), and None otherwise. When code is returned, its value is always 0. Returning literal 0 produces identical behavior.

3. src/lib.rs:72 in default:
   max_priority: 1 -> max_priority: 0; (composite/field-clear)
   Equivalence note: Rule priorities are integers in range 1..=5. Rules are filtered by max priority when opts.max_priority > 0 && priority < opts.max_priority. If max_priority == 1, priority < 1 is false for all rules (since priority >= 1). If max_priority == 0, opts.max_priority > 0 is false. Both conditions filter zero rules. Behavior is identical.

The fourth mutant, max_priority: 1 -> max_priority: 2 (numbers/incrementer), escaped previously. It is killed by the new test default_maximumpriority_includes_priority_one_rules in tests/cli.rs (testing cleancode ruleset containing BooleanArgumentFlag with priority 1).

Resolution

Landed on main via PR #151 (merge commit 04ff941ac236d4626804de21fcbb0da01615ea2f).