meta.reviews.complete_test_set

One review covers one anchor’s claim and its complete set of associated tests. The input records every eligible test associated with that anchor in a deterministic order. Several tests may cover different parts of a claim; each test need not establish the whole claim independently.

Read in specification context · design/meta/reviews.md

Implementation

Implemented crates/sage-meta/src/review.rs:256–261

sage_meta::review · lib:sage_meta

#![allow(unused)]
fn main() {
        // {anchor}`meta.reviews.complete_test_set`
        tests.sort();
        tests.dedup();
        for ((path, crate_name, target, rust_path), (_, _, contents)) in test_fragments {
            fragments.push(InputFragment {
                path,

}

See more source ↗

Tests and review

Indexed tests and the implementation-completeness and test-adequacy review; no test run is implied.

Implemented sage_meta::added_or_removed_associations_invalidate_unchanged_saved_fragments

crates/sage-meta/tests/review.rs · test:review

#![allow(unused)]
fn main() {
#[test]
#[sage::test(meta.reviews.complete_test_set)]
#[sage::test(meta.reviews.fragment_associations)]
#[sage::test(meta.reviews.input_format)]
#[sage::test(meta.reviews.review_diagnostics)]
#[sage::test(meta.reviews.staleness)]
fn added_or_removed_associations_invalidate_unchanged_saved_fragments() {
    let fixture = Fixture::new();
    fixture.approve();
    let original = fixture.input();
    let second = "#[test] #[sage::test(demo.rule)] fn second() { assert_eq!(1 + 3, 4); }";
    fixture.write("src/tests.rs", &format!("{TEST}\n{second}"));
    let error = sage_meta::check(&fixture.0)
        .ensure_valid()
        .unwrap_err()
        .to_string();
    assert!(error.contains("demo.rule"), "{error}");
    assert!(error.contains("test associations"), "{error}");
    assert!(error.contains("sage-review-checklist/SKILL.md"), "{error}");
    let current = fixture.start();
    assert_eq!(current.tests.len(), 2);
    assert_eq!(current.tests, fixture.start().tests);
    assert_eq!(
        original.tests[0],
        *current
            .tests
            .iter()
            .find(|t| t.rust_path.ends_with("verifies_result"))
            .unwrap()
    );
    review::record(&fixture.0, "demo.rule", &Fixture::report()).unwrap();
    fixture.write("src/tests.rs", TEST);
    let error = sage_meta::check(&fixture.0)
        .ensure_valid()
        .unwrap_err()
        .to_string();
    assert!(error.contains("demo.rule"), "{error}");
    assert!(error.contains("test associations"), "{error}");
    assert!(error.contains("sage-review-checklist/SKILL.md"), "{error}");
    fixture.approve();
    fixture.write("src/lib.rs", "#[sage::item(crate)] mod model {}\nmod behavior;\n#[cfg(test)] mod tests;\n// {anchor}`demo.rule`\nfn another_implementation() {}\n");
    let error = sage_meta::check(&fixture.0)
        .ensure_valid()
        .unwrap_err()
        .to_string();
    assert!(error.contains("demo.rule"), "{error}");
    assert!(error.contains("implementation associations"), "{error}");
    assert!(error.contains("sage-review-checklist/SKILL.md"), "{error}");
    assert_eq!(fixture.start().implementations.len(), 2);
}
}

See more source ↗

Review: Accepted · final_quality.

Reviewed the full meta.reviews.complete_test_set claim, the associated review.rs implementation, and the complete added_or_removed_associations_invalidate_unchanged_saved_fragments test. This is source inspection; no behavior tests were rerun, and central execution is separate.

inventory iterates one canonical anchor at a time and creates one ReviewInput containing that claim and every site in rust.tests for that name. Each eligible site's full function source is sliced from its recorded span, metadata is added to tests, and a semantic test fragment is keyed by file/crate/target/Rust path. Conflicting spans for one semantic identity cause an ambiguity error rather than silently selecting a test. tests.sort/dedup gives deterministic metadata order, while the BTreeMap gives deterministic fragment order and deduplicates repeated associations to the same test. Distinct tests and targets remain distinct. Captured CorrespondenceIndex::record_tests is the eligibility producer: module-level built-in test functions in enabled standard harness targets, excluding ignored/conditionally ignored tests; accepted references append to the per-anchor vector. Captured evidence::placeholder_status preserves ordinary/TODO/N/A classification without filtering tests out of the review.

One review record is stored for the anchor, not one required proof per individual test. The prompt asks the reviewer to assess all requirements across the supplied implementation/test fragments, and inventory imposes no requirement that each test independently establish the complete claim. input_errors compares the complete saved/current tests metadata and initial fragments, so association-set changes invalidate the review even if an existing test body remains unchanged. Captured shared check wiring reports these review failures through ensure_valid.

The associated test starts with an approved one-test fixture, adds a second independently named test, and requires an anchor-specific test-associations diagnostic with the review-skill guidance. It then asserts the fresh input contains two tests, repeated starts yield identical ordered metadata, and the original test record remains exactly present. After recording the new set, removing the second test must again fail with the same association reason. The final implementation-association portion separately confirms adding a second implementation also invalidates the saved review and yields two current implementation entries; this supports the broader association regression without substituting for the explicit test-set assertions. Captured Fixture/TEST helpers create a real Cargo and Design workspace, call review::start/record and deserialize saved input; the key assertions are visible in the annotated body.

The implementation and test account for per-anchor scope, inclusion of the complete eligible set, deterministic ordering, association changes and collective rather than per-test claim coverage. There are no associated TODO or N/A entries. No concrete defect or additional necessary edit was found. No source changes are requested.

Browse meta anchors

116 anchors

Anchor Implementation Tests

meta (116) 102 Implemented24 TODO 103 Implemented36 TODO

meta.anchors (22) 22 Implemented 21 Implemented12 TODO

anchor_pages Implemented ImplementedTODO

Every specification anchor has a stable page identifying the rule and showing its definition, context, implementation entries, and tests/evidence. The page includes Rust excerpts and access to surrounding source, and exposes each entry’s placeholder status and explanation. Multiple sites remain individually inspectable rather than being reduced to a single coverage badge.

Symposiun AGentic Environment (SAGE)

meta.reviews.complete_test_set

Implementation

Tests and review