Retain successful response read failures

internal/llm/openai_compatible_client.go

@@ -370,40 +370,69 @@ func tokenCount(value *int) int {
 	return *value
 }
 
+type responseErrorReader struct {
+	reader io.Reader
+	err    error
+}
+
+func (r *responseErrorReader) Read(p []byte) (int, error) {
+	n, err := r.reader.Read(p)
+	if err != nil && !errors.Is(err, io.EOF) {
+		r.err = err
+	}
+	return n, err
+}
+
 func decodeOpenAIChatResponse(body io.Reader) (openAIChatResponse, error) {
+	observed := &responseErrorReader{reader: body}
 	limited := &io.LimitedReader{
-		R: body,
+		R: observed,
 		N: maxOpenAIChatResponseBytes + 1,
 	}
 	decoder := json.NewDecoder(limited)
 
 	var response openAIChatResponse
 	if err := decoder.Decode(&response); err != nil {
-		if errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded) {
-			return openAIChatResponse{}, &requestFailedError{cause: err}
+		if errors.Is(observed.err, context.Canceled) || errors.Is(observed.err, context.DeadlineExceeded) {
+			return openAIChatResponse{}, &requestFailedError{cause: observed.err}
 		}
 		if limited.N == 0 {
 			return openAIChatResponse{}, openAIChatResponseTooLargeError()
 		}
+		if observed.err != nil {
+			return openAIChatResponse{}, &requestFailedError{cause: observed.err}
+		}
 		return openAIChatResponse{}, fmt.Errorf("%w: failed to decode response", ErrMalformedResponse)
 	}
+	if errors.Is(observed.err, context.Canceled) || errors.Is(observed.err, context.DeadlineExceeded) {
+		return openAIChatResponse{}, &requestFailedError{cause: observed.err}
+	}
 	if limited.N == 0 {
 		return openAIChatResponse{}, openAIChatResponseTooLargeError()
 	}
 
 	var trailing any
 	if err := decoder.Decode(&trailing); !errors.Is(err, io.EOF) {
-		if errors.Is(err, context.Canceled) || errors.Is(err, context.DeadlineExceeded) {
-			return openAIChatResponse{}, &requestFailedError{cause: err}
+		if errors.Is(observed.err, context.Canceled) || errors.Is(observed.err, context.DeadlineExceeded) {
+			return openAIChatResponse{}, &requestFailedError{cause: observed.err}
 		}
 		if limited.N == 0 {
 			return openAIChatResponse{}, openAIChatResponseTooLargeError()
 		}
+		if observed.err != nil {
+			return openAIChatResponse{}, &requestFailedError{cause: observed.err}
+		}
 		return openAIChatResponse{}, fmt.Errorf("%w: response contains trailing data", ErrMalformedResponse)
 	}
+	if errors.Is(observed.err, context.Canceled) || errors.Is(observed.err, context.DeadlineExceeded) {
+		return openAIChatResponse{}, &requestFailedError{cause: observed.err}
+	}
 	if limited.N == 0 {
 		return openAIChatResponse{}, openAIChatResponseTooLargeError()
 	}
+	if observed.err != nil {
+		return openAIChatResponse{}, &requestFailedError{cause: observed.err}
+	}
 
 	return response, nil
 }