fix(openai): also emit response.failed in ensureForwardErrorResponse after Writer.Written

Case B: when a slot wait flushes SSE ping comments first (Writer.Written
becomes true), the previous ensureForwardErrorResponse short-circuited
on `c.Writer.Written()` and returned false without notifying the client.
Subsequent upstream errors (http2 timeout, stream INTERNAL_ERROR, etc.)
produced silent EOF; Codex CLI reported "stream closed before
response.completed" just like the user-slot timeout case.

Remove the Written() early return; coerce streamStarted to true when
Writer has already been written to, and let handleStreamingAwareError
walk the existing logic — which now (thanks to the previous commits)
emits a protocol-compliant response.failed for /responses paths and the
legacy `event: error` for others.

Update tests that previously asserted "do not override written response":
the new contract is to *append* an SSE terminal frame so the client sees
a clean close instead of EOF. recoverResponsesPanic inherits this fix.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

backend/internal/handler/openai_gateway_handler.go

@@ -1719,9 +1719,17 @@ func (h *OpenAIGatewayHandler) handleStreamingAwareError(c *gin.Context, status
 
 // ensureForwardErrorResponse 在 Forward 返回错误但尚未写响应时补写统一错误响应。
 func (h *OpenAIGatewayHandler) ensureForwardErrorResponse(c *gin.Context, streamStarted bool) bool {
-	if c == nil || c.Writer == nil || c.Writer.Written() {
+	if c == nil || c.Writer == nil {
 		return false
 	}
+	// 旧实现在 Writer.Written 时直接 return false，导致 ping 已 flush 之后的
+	// 上游错误（http2 timeout、连接中断等）完全无法把错误传给客户端——
+	// HTTP 200 已锁死，TCP 直接 EOF，Codex CLI 报 "stream closed before response.completed"。
+	// 这里改成：Writer 已写过时强制走 streamStarted 分支，让
+	// handleStreamingAwareError 通过 SSE 发协议合规的 response.failed。
+	if c.Writer.Written() {
+		streamStarted = true
+	}
 	h.handleStreamingAwareError(c, http.StatusBadGateway, "upstream_error", "Upstream request failed", streamStarted)
 	return true
 }