What causes "Codewhale responses request should succeed" in Hmbown/CodeWhale?

Confirm the mock route matches the Responses endpoint (`{base}/responses`), not the chat default

Codewhale responses request should succeed

Error message

Codewhale responses request should succeed

What it means

This `expect` panics when `client.create_message(...)` returns `Err` for a Responses-wire-format request, aborting with 'Codewhale responses request should succeed' at crates/tui/src/client.rs:7911. The client is expected to stream against the mock server and aggregate usage from the terminal `response.completed` event; any transport, HTTP, or stream-parse failure surfaces here.

Solutions

Confirm the mock route matches the Responses endpoint (`{base}/responses`), not the chat default
Log the returned error to see whether it is HTTP status or stream parsing
Ensure the mock SSE body ends with a `response.completed` event carrying usage
Re-enable retry or fix the fixture stream if the failure is transient

Example fix

// before
let response = client.create_message(minimal_zen_request("openai/gpt-5.6")).await
    .expect("Codewhale responses request should succeed");
// after
let response = client.create_message(minimal_zen_request("openai/gpt-5.6")).await
    .unwrap_or_else(|e| panic!("Codewhale responses request should succeed: {e:?}"));

Defensive patterns

Strategy: try-catch

Validate before calling

assert_eq!(client.wire_format, WireFormat::Responses);
assert!(!client.retry.enabled); // deterministic failure surfacing

Try / catch

let response = client.create_message(req).await
    .unwrap_or_else(|e| panic!("Codewhale responses request should succeed: {e:?}"));

Prevention

Register the mock on the /responses path, not the chat default
Ensure fixture SSE streams include a terminal response.completed event with usage
Keep retry disabled only when the mock is deterministic

When it happens

Trigger: Calling `create_message` with `wire_format == WireFormat::Responses` and `retry.enabled = false` when the mock returns a non-2xx status, the SSE stream is malformed, or `response.completed` never arrives so usage cannot be built.

Common situations: Mock route registered for `/chat/completions` instead of `/responses`; the fixture SSE events missing the terminal `response.completed` frame; retries disabled so a transient mock hiccup fails immediately.

Understand the failure class

Background: 'Something went wrong' / 'Request failed (500)' / 'HTTP error! status: 404' — what failed HTTP requests actually mean and how to find the real cause — this error's family across 28 libraries.

Related errors

AI-assisted analysis of Hmbown/CodeWhale@73e0f67d83 (2026-09-22). Data as JSON: /api/errors/fbb993a902919f65. Report an issue: GitHub.

Appendix: source

Thrown at crates/tui/src/client.rs:7911

            "codewhale",
            &fingerprint,
            now_unix(),
        )
        .expect("fixture catalog parses");
        crate::provider_catalog_live::record_success(ProviderCatalogDelta {
            provider: "codewhale".to_string(),
            base_url_fingerprint: fingerprint,
            fetched_at: now_unix(),
            offerings,
        });

        let mut client = codewhale_client(&server, "openai/gpt-5.6");
        assert_eq!(client.wire_format, WireFormat::Responses);
        client.retry.enabled = false;
        let response = client
            .create_message(minimal_zen_request("openai/gpt-5.6"))
            .await
            .expect("Codewhale responses request should succeed");

        // Responses usage arrives on the terminal `response.completed` event —
        // the dialect's equivalent of chat's `stream_options.include_usage`.
        assert_eq!(response.usage.input_tokens, 3);
        assert_eq!(response.usage.output_tokens, 1);
        let requests = server.received_requests().await.expect("recorded request");
        assert_eq!(requests.len(), 1);
        assert_codewhale_bearer(&requests[0]);
        let body: Value = serde_json::from_slice(&requests[0].body).expect("responses JSON body");
        assert_eq!(
            body.get("model").and_then(Value::as_str),
            Some("openai/gpt-5.6")
        );
        assert!(body.get("input").is_some(), "Responses body: {body}");
        assert!(body.get("messages").is_none(), "Responses body: {body}");

        crate::provider_catalog_live::reset_cache_for_test();
        crate::provider_lake::clear_live_snapshot();

View on GitHub (pinned to 73e0f67d83)