Solution for long term memory for agent coding CLIs and to facilitate handoff between different agent vendors
Rust
test(llm): cover openai-oauth SSE error/truncation/empty/multi-line paths
AkitaOnRails1ea5125fb68a
diff --git a/crates/ai-memory-llm/src/openai_oauth.rs b/crates/ai-memory-llm/src/openai_oauth.rsindex 15ef8154..42d1f8dc 100644--- a/crates/ai-memory-llm/src/openai_oauth.rs+++ b/crates/ai-memory-llm/src/openai_oauth.rs@@ -829,6 +829,78 @@ mod tests {);}+ /// Audit BLOCKING fix: a `data: {"error":{...}}` event without a+ /// `type` field and without a preceding `event:` header used to+ /// fall through the kind-dispatch silently. Now any top-level+ /// `error` payload is promoted to a terminal error so the real+ /// upstream failure reaches the caller.+ #[test]+ fn parse_sse_response_promotes_top_level_error_without_event_header() {+ let body = "data: {\"error\":{\"message\":\"context length exceeded\"}}\n\n";+ let err = parse_sse_response(body)+ .expect_err("top-level error payload must surface as a terminal error");+ match err {+ LlmError::Provider { status, body } => {+ assert_eq!(status, super::SSE_TERMINAL_ERROR_STATUS);+ assert!(+ body.contains("context length exceeded"),+ "body must carry upstream error message: {body}"+ );+ }+ other => panic!("expected Provider error, got {other:?}"),+ }+ }++ /// Audit BLOCKING fix: a stream that ends mid-`data:` JSON (the+ /// upstream socket closes before the closing brace) used to surface+ /// as a generic serde parse error. The post-loop flush now catches+ /// that case when no `response.completed` event has arrived and+ /// emits the actionable "stream truncated" diagnostic.+ #[test]+ fn parse_sse_response_truncated_data_surfaces_as_truncated_stream() {+ // `data:` line carrying an unterminated JSON object, no blank+ // line after — the post-loop flush is what runs.+ let body = "data: {\"type\":\"response.output_text.delta\",\"delta\":\"hel";+ let err = parse_sse_response(body).expect_err("truncated stream must fail");+ assert!(+ err.to_string()+ .contains("stream truncated before final event closed"),+ "expected truncation-aware error, got: {err}"+ );+ }++ /// Empty body — no `data:` lines at all — falls through to the+ /// `completed.is_none()` branch with the original+ /// "stream closed before response.completed" error.+ #[test]+ fn parse_sse_response_empty_body_reports_no_completed_event() {+ let err = parse_sse_response("").expect_err("empty body must fail");+ assert!(+ err.to_string()+ .contains("stream closed before response.completed"),+ "expected canonical 'no completed' error, got: {err}"+ );+ }++ /// SSE spec allows an event to carry multiple `data:` lines; the+ /// parser joins them with `\n` before JSON-decoding. Exercise the+ /// continuation path so a future refactor doesn't drop it.+ #[test]+ fn parse_sse_response_handles_multi_line_data_continuation() {+ // Real upstreams sometimes split a wide JSON object across+ // multiple `data:` lines; the parser joins on '\n' before+ // serde_json parses, so a JSON value spanning newlines must+ // survive (here we exercise a JSON STRING that contains a+ // literal newline, which IS valid serde input).+ let body = concat!(+ "event: response.completed\n",+ "data: {\"type\":\"response.completed\",\"response\":{\"model\":\"gpt-5.5\",\"output_text\":\"line1\\nline2\",\"output\":[]}}\n\n",+ "data: [DONE]\n\n",+ );+ let resp = parse_sse_response(body).expect("multi-line content must parse");+ assert_eq!(resp.output_text.as_deref(), Some("line1\nline2"));+ }+#[test]fn structured_request_uses_responses_json_schema_format() {let request = ChatRequest::user_prompt("json please");