rtk-ai/rtk · error

fixture listed in CORPUS

Error message

fixture listed in CORPUS

What it means

Test panic from `.expect("fixture listed in CORPUS")`: `CORPUS.iter().find(...)` returned `None` because the fixture name (e.g. `git_format_patch_cover` or `git_format_patch_sha256`) is not a key in the CORPUS list of `(name, fixture_content)` pairs. This is a test-suite wiring error: the loop iterates names that must exist in the corpus, and a missing entry is a hard error before any token-savings assertion runs.

Solutions

1. Add the missing `(name, include_str!("..."))` entry to the CORPUS constant for the listed name.
2. Fix the typo in the loop's name array so it matches an existing CORPUS key exactly.
3. Optionally replace the find/expect with a debug_assert or a compile-time check that every loop name exists in CORPUS.

Example fix

// before
.expect("fixture listed in CORPUS")
// after
let fixture = CORPUS.iter().find(|(n, _)| *n == name)
    .unwrap_or_else(|| panic!("fixture {name} missing from CORPUS")).1;

// verify all loop names exist in CORPUS before running
for name in NAMES {
    assert!(CORPUS.iter().any(|(n, _)| n == name), "{name} missing from CORPUS");
}

fn corpus_fixture<'a>(corpus: &'a [(&'a str, &'a str)], name: &str) -> Option<&'a str> {
    corpus.iter().find(|(n, _)| *n == name).map(|(_, f)| *f)
}

let fixture = corpus_fixture(&CORPUS, name)
    .unwrap_or_else(|| panic!("fixture {name} missing from CORPUS"));

When it happens

Trigger: Adding a new fixture name to the loop's array (e.g. a new format-patch fixture) without adding the corresponding `(name, include_str!(...))` entry to the CORPUS constant, or renaming/deleting a fixture file referenced by CORPUS.

Common situations: Renaming fixture files under tests/fixtures without updating CORPUS, typo'd fixture names in the test loop, or a merge that keeps one side's list.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

• a bodyless region of prose must not force raw
• a table in the message must not force raw
• f.txt entry
• --format=%B must parse
• hg diff must parse

AI-assisted analysis of rtk-ai/rtk@0924356b4c (2026-09-19). Data as JSON: /api/errors/495634c4f3c56d4c. Report an issue: GitHub.

Appendix: source

Thrown at src/cmds/git/diff_cmd.rs:7109

    #[test]
    fn metadata_heavy_fixtures_clear_the_admission_bar() {
        // Pins the CONTRIBUTING.md 20% admission bar on the metadata-heavy
        // corpus fixtures so the percentages cited above can't rot into
        // fiction. Content-heavy fidelity fixtures (single-file `diff -u`
        // shapes) are deliberately absent: their savings are input-dependent
        // and only the never-larger property above binds them.
        for name in [
            "git_log_p",
            "git_diff_multifile",
            "git_format_patch_single",
            "git_format_patch_series",
            "git_format_patch_cover",
            "git_format_patch_sha256",
        ] {
            let fixture = CORPUS
                .iter()
                .find(|(n, _)| *n == name)
                .expect("fixture listed in CORPUS")
                .1;
            let out = condense_unified_diff(fixture);
            let (input, output) = (count_tokens(fixture), count_tokens(&out));
            // output <= 80% of input  <=>  savings >= 20%.
            assert!(
                output * 5 <= input * 4,
                "{name}: savings below the 20% admission bar ({input} -> {output} tokens)"
            );
        }
    }

    #[test]
    fn test_no_truncation_large_diff() {
        // Verify compute_diff returns all changes without truncation
        let mut a = Vec::new();
        let mut b = Vec::new();
        for i in 0..500 {
            a.push(format!("line_{}", i));

View on GitHub (pinned to 0924356b4c)