meta.anchors.evidence_status

An implementation or test entry establishes a declared association, not proof that the rule is satisfied. Source inventories retain cfg conditions without claiming that a site was compiled or a test executed. Anchor pages distinguish test source, recorded run results, and other evidence such as traces. A passing structural check or a passing dummy test must not present a TODO or Not applicable entry as established behavioral evidence.

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

Implementation

Implemented crates/sage-meta/src/browser.rs:288–293

sage_meta::browser · lib:sage_meta

#![allow(unused)]
fn main() {
// {anchor}`meta.anchors.evidence_status`
fn source_data(
    _documents: &[BrowserDocument<'_>],
    workspace: Option<&Path>,
) -> Result<SourceData, Box<dyn std::error::Error>> {
    let Some(workspace) = workspace else {

}

See more source ↗

Implemented crates/sage-meta/src/correspondence.rs:704–709

sage_meta::correspondence · lib:sage_meta

#![allow(unused)]
fn main() {
    // {anchor}`meta.anchors.evidence_status`
    fn record(
        &mut self,
        attrs: &[syn::Attribute],
        name: &syn::Ident,
        span: proc_macro2::Span,

}

See more source ↗

Implemented crates/sage-meta/src/evidence.rs:79–84

sage_meta::evidence · lib:sage_meta

#![allow(unused)]
fn main() {
// {anchor}`meta.anchors.evidence_status`
pub(crate) fn index_source(
    source: &str,
    syntax: &syn::File,
    context: SourceContext<'_>,
    entries: &mut BTreeMap<String, Vec<ImplementationSite>>,

}

See more source ↗

Implemented crates/sage-meta/src/book/anchors.rs:129–134

sage_meta::book::anchors · bin:sage-meta

#![allow(unused)]
fn main() {
    // {anchor}`meta.anchors.evidence_status`
    // {anchor}`meta.anchors.placeholder_status`
    fn page(&self, name: &str, url: &str) -> String {
        let mut text = format!(
            "# {}\n\n<nav class=\"anchor-breadcrumbs\" aria-label=\"Anchor namespaces\"><a href=\"{}\">All anchors</a>",
            escape(name),

}

See more source ↗

Tests and review

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

Implemented sage_meta::cargo_owned_out_of_line_sources_retain_inherited_module_and_cfg

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

#![allow(unused)]
fn main() {
#[test]
#[sage::test(meta.anchors.implementation_references)]
#[sage::test(meta.anchors.evidence_status)]
fn cargo_owned_out_of_line_sources_retain_inherited_module_and_cfg() {
    let directory =
        std::env::temp_dir().join(format!("sage-evidence-owned-{}", std::process::id()));
    let _ = std::fs::remove_dir_all(&directory);
    std::fs::create_dir_all(directory.join("src")).unwrap();
    std::fs::write(directory.join("Cargo.toml"), "[package]\nname = \"evidence-fixture\"\nversion = \"0.0.0\"\nedition = \"2024\"\n[workspace]\n").unwrap();
    std::fs::write(directory.join("src/lib.rs"), "#[cfg(any())]\nmod todo;\n").unwrap();
    std::fs::write(
        directory.join("src/todo.rs"),
        "// {anchor}`example.pending`\n// TODO: Add the implementation.\n",
    )
    .unwrap();
    std::fs::write(
        directory.join("src/unowned.rs"),
        "// {anchor}`example.unowned`\n",
    )
    .unwrap();
    let rust = CorrespondenceIndex::load(&directory).unwrap();
    std::fs::remove_dir_all(&directory).unwrap();
    assert!(rust.diagnostics.is_empty(), "{:?}", rust.diagnostics);
    assert_eq!(rust.implementations.len(), 1);
    let site = &rust.implementations["example.pending"][0];
    assert_eq!(site.file, "src/todo.rs");
    assert_eq!(site.module, "evidence_fixture::todo");
    assert_eq!(site.status, EvidenceStatus::Todo);
    assert_eq!(site.cfg, ["#[cfg(any())]"]);
    assert_eq!(
        site.explanation.as_deref(),
        Some("TODO: Add the implementation.")
    );
    assert_eq!(site.excerpt, rust.files["src/todo.rs"]);
}
}

See more source ↗

Implemented sage_meta::implementation_status_uses_module_ancestry_and_preserves_cfg_and_explanations

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

#![allow(unused)]
fn main() {
#[test]
#[sage::test(meta.anchors.placeholder_status)]
#[sage::test(meta.anchors.evidence_status)]
fn implementation_status_uses_module_ancestry_and_preserves_cfg_and_explanations() {
    let source = r#"#![cfg(feature = "root")]
#[cfg(unix)]
mod todo {
    mod nested {
        // {anchor}`example.mixed`<<
        // TODO: Connect the runtime checker.
        fn checker() {}
        // >>
    }
}
mod not_applicable {
    // {anchor}`example.mixed`
    // This rule describes authoring practice, not runtime behavior.
}
mod empty { mod todo {} mod not_applicable {} }
#[cfg(feature = "operation")]
fn todo() {
    // {anchor}`example.mixed`
}
"#;
    let rust = index(source);
    assert!(rust.diagnostics.is_empty(), "{:?}", rust.diagnostics);
    let sites = &rust.implementations["example.mixed"];
    assert_eq!(sites.len(), 3, "empty modules declare no entries");
    assert_eq!(sites[0].module, "fixture::todo::nested");
    assert_eq!(sites[0].status, EvidenceStatus::Todo);
    assert_eq!(
        sites[0].cfg,
        ["#![cfg(feature = \"root\")]", "#[cfg(unix)]"]
    );
    assert_eq!(
        sites[0].explanation.as_deref(),
        Some("TODO: Connect the runtime checker.")
    );
    assert_eq!(sites[1].status, EvidenceStatus::NotApplicable);
    assert_eq!(
        sites[1].explanation.as_deref(),
        Some("This rule describes authoring practice, not runtime behavior.")
    );
    assert_eq!(
        sites[2].status,
        EvidenceStatus::Ordinary,
        "a function named todo is not a placeholder module"
    );
    assert_eq!(sites[2].module, "fixture");
    assert_eq!(
        sites[2].cfg,
        [
            "#![cfg(feature = \"root\")]",
            "#[cfg(feature = \"operation\")]"
        ]
    );
    assert!(rust.tests.is_empty());
}
}

See more source ↗

TODO sage_meta::todo::source_associations_are_distinct_from_run_evidence

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

#![allow(unused)]
fn main() {
#[test]
    #[sage::test(meta.anchors.evidence_status)]
    fn source_associations_are_distinct_from_run_evidence() {
        // TODO: Preserve cfg and separate declarations, recorded results, and
        // traces; a passing dummy test must keep its placeholder status.
    }
}

See more source ↗

Review: Accepted · Codex /root/anchor_094.

Fresh complete followup review of evidence_status, with the root provenance clarification captured through a new start using /tmp/anchor-094-evidence-instructions.txt. The claim requires declared associations rather than proof, retention of textual cfg without compilation/execution claims, provenance distinction between source and runs/traces on anchor pages, and no promotion of TODO/N/A evidence merely because checks or dummy tests pass.

All four ordinary implementation entries were inspected in their current captured files. correspondence::record copies inherited and local conditions into RustSite and records source declarations/test annotations; it does not execute tests or evaluate cfg into a compiled-status claim. evidence::index_source builds source scopes, retains inherited cfg and explanation text, and derives implementation classification from module ancestry. browser::source_data loads this correspondence inventory and copies its source locations/files and diagnostics into browser data; it supplies no fabricated execution result. book::anchors::page renders implementation and test status badges, source snippets and links, and Conditional source cfg text. Its test section explicitly says indexed tests and review imply no test run; the index introduction also disclaims recorded execution. TODO and N/A badges derive directly from stored classifications. The additionally captured review inventory fragment preserves cfg and derives test status from the owning module rather than any test outcome. Thus a passing structural check or empty test cannot convert its placeholder association into ordinary behavioral evidence. Ordinary badge wording is contextualized by the explicit declared-association/no-run explanations, not a claim of executed proof.

Both ordinary tests were read completely. cargo_owned_out_of_line_sources_retain_inherited_module_and_cfg creates a real Cargo fixture with a cfg(any()) module and unowned file, loads it, and asserts exactly one implementation site, correct file/module, TODO status, the inherited impossible cfg text, explanation, and exact source excerpt. This establishes source inventory even when compilation would exclude the module and excludes unrelated unowned files. implementation_status_uses_module_ancestry_and_preserves_cfg_and_explanations asserts nested TODO, N/A, and ordinary classification, correct module identity, file/module/function cfg strings, explanations, exactly three entries despite empty placeholder modules, and no invented tests. A function named todo remains ordinary because classification uses module ancestry. The captured index helper invokes real CorrespondenceIndex::from_sources without replacing the indexing behavior.

The sole TODO test source_associations_are_distinct_from_run_evidence is intentionally not behavioral proof. It retains the future integrated matrix for cfg, declarations, recorded results, traces, and dummy-test status preservation. In accordance with the captured root clarification, the current claim is a provenance/no-false-execution requirement, not an instruction to invent run/trace ingestion. Present source-only pages explicitly identify their provenance and imply no run; there are currently no run/trace inputs to conflate. Full mixed-evidence presentation tests require such recorded evidence fixtures/interfaces, and dummy-run integration coverage remains deferred. I do not describe that TODO as tested or implemented coverage. The two ordinary tests establish current source/status facts, while the renderer inspection establishes the current no-run distinction.

All claim aspects and every evidence entry are accounted for, with the existing limitation preserved. Additional dependencies were captured through review read (test index helper and review test-status construction); all four essential implementation files including the page renderer were initial captures. No source edits or independent suite runs were made. This was fresh source assessment and review commands, not blind refresh or a claimed test execution. No requested edits.

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.anchors.evidence_status

Implementation

Tests and review