Hmbown/CodeWhale · error

repaired result

Error message

repaired result

What it means

A panic from `prepared.messages.last().expect("repaired result")` in the dangling-tool-call repair test. The test asserts the LAST prepared message is the repaired user-role tool result; the expect fires when the prepared message list is empty — i.e. projection produced no messages at all, so the repair path emitted nothing.

Solutions

1. Dump prepared.messages (length and roles) before the assertion to see whether repair emitted anything.
2. Verify the repair matcher still expects tool_use_id "call-secret-test" and content containing "crashed_and_repaired" after refactors.
3. Check recent edits to the repair-before-adapter-projection path for early-return or filtering that drops the repaired message.
4. Confirm client_with_config_secret_sentinels still constructs a client whose model routes through the repairing adapter.

Example fix

// before
assert_eq!(prepared.messages.last().expect("repaired result").role, "user");
// after
let last = prepared.messages.last().unwrap_or_else(|| panic!("no repaired message, prepared = {prepared:#?}"));
assert_eq!(last.role, "user");

// rust: check the repair actually emitted a message before asserting role
assert!(!prepared.messages.is_empty(), "repair produced no messages");

let last = prepared.messages.last()
    .unwrap_or_else(|| panic!("empty prepared.messages: {prepared:#?}"));
assert_eq!(last.role, "user");

When it happens

Trigger: `prepared.messages.last()` returns None because the model-bound repair produced zero messages — usually after projection/repair refactors that drop the dangling tool-call's repaired user message, or when the input request's tool_use_id/content guards no longer match.

Common situations: Changes to the dangling tool-call repair that skip emission when the sentinel content 'crashed_and_repaired' or tool_use_id 'call-secret-test' matching changes; upstream projection filtering out user messages for the configured model.

Related errors

• Absolute path should not warn
• child assignment
• child-local search
• configured Web evidence tool
• expected authentication error, got

AI-assisted analysis of Hmbown/CodeWhale@73e0f67d83 (2026-09-22). Data as JSON: /api/errors/e036ebdfab00255a. Report an issue: GitHub.

Appendix: source

Thrown at crates/tui/src/client.rs:8421

        let prepared = client.prepare_model_bound_request(request);

        assert!(prepared.messages.iter().any(|message| {
            message.content.iter().any(|block| {
                matches!(
                    block,
                    ContentBlock::ToolResult {
                        tool_use_id,
                        content,
                        is_error: Some(true),
                        ..
                    } if tool_use_id == "call-secret-test"
                        && content.contains("crashed_and_repaired")
                )
            })
        }));
        assert_eq!(
            prepared.messages.last().expect("repaired result").role,
            "user"
        );
        assert!(!prepared.messages.iter().any(|message| {
            message.content.iter().any(|block| {
                matches!(
                    block,
                    ContentBlock::Text { text, .. }
                        if text.contains("[tool_history_repair]")
                )
            })
        }));
    }

    #[test]
    fn model_bound_tool_results_keep_ordinary_code_byte_exact() {
        // #5546: key-only hits in source files must reach the model unchanged
        // so exact-match edits and read-back verification keep working.
        let client = client_with_config_secret_sentinels();

View on GitHub (pinned to 73e0f67d83)