Evidence
The full locked Rust suite on 2026-08-31 produced 969 passed, 23 ignored, and 3 failed. The failures are deterministic byte-for-byte fixture comparisons and predate the external-tool collector slice:
- commands::business_rule_archaeology::qualification_comparison::tests::checked_report_is_exact_and_reproducible
- commands::deterministic_review::tests::recorded_benchmark_never_emits_an_invalid_position_after_qualification
- commands::business_rule_archaeology::refresh_command::correctness_qualification::real_pipeline_correctness_measurement_is_exact_and_reproducible
Focused collector tests, CLI tests, and release cargo check pass. The golden outputs were not refreshed in the tooling/security change because that would hide an unrelated semantic change without review.
Acceptance
- Identify the source change that moved each generated output.
- Review the semantic delta rather than replacing bytes blindly.
- Refresh only outputs whose new semantics are correct.
- Prove each focused test and the full locked Rust suite pass.
- Preserve the exact/reproducible assertion; do not downgrade it.
Evidence
The full locked Rust suite on 2026-08-31 produced 969 passed, 23 ignored, and 3 failed. The failures are deterministic byte-for-byte fixture comparisons and predate the external-tool collector slice:
Focused collector tests, CLI tests, and release cargo check pass. The golden outputs were not refreshed in the tooling/security change because that would hide an unrelated semantic change without review.
Acceptance