代码提交

frontend-web/src/core/threads/hooks.ts

@@ -244,6 +244,11 @@ const TERMINAL_RUN_STATUSES = new Set([
 // for a while, then reconcile it from the authoritative run/state endpoints.
 const STREAM_RECONCILIATION_IDLE_MS = 12_000;
 const STREAM_RECONCILIATION_POLL_MS = 5_000;
+// A run can legitimately be quiet while a long tool call is in progress, so
+// reconnect more conservatively than the terminal-status check above.  The
+// reconnection only replaces the local SSE reader; the backend run keeps going.
+const STREAM_RECONNECT_IDLE_MS = 30_000;
+const STREAM_RECONNECT_MAX_ATTEMPTS = 3;
 
 async function fetchThreadTitleRefreshState(
   threadId: string,
@@ -466,9 +471,15 @@ export function useThreadStream({
   const startedRef = useRef(false);
   const lastStreamActivityAtRef = useRef(0);
   const streamReconciliationInFlightRef = useRef(false);
+  const streamReconnectAttemptsRef = useRef(0);
+  const streamReconnectToastShownRef = useRef(false);
+  const streamReconnectFailureToastShownRef = useRef(false);
 
   const markStreamActivity = useCallback(() => {
     lastStreamActivityAtRef.current = Date.now();
+    streamReconnectAttemptsRef.current = 0;
+    streamReconnectToastShownRef.current = false;
+    streamReconnectFailureToastShownRef.current = false;
   }, []);
 
   const listeners = useRef({
@@ -761,13 +772,6 @@ export function useThreadStream({
   threadStreamRef.current = thread;
 
   useEffect(() => {
-    if (!thread.isLoading) {
-      // A normal terminal event reached the browser.  A later run may use the
-      // same thread, so it must be eligible for recovery again.
-      streamReconciliationInFlightRef.current = false;
-      return;
-    }
-
     const activeThreadId = threadIdRef.current ?? onStreamThreadId ?? null;
     if (!activeThreadId || typeof window === "undefined") {
       return;
@@ -776,7 +780,7 @@ export function useThreadStream({
     let disposed = false;
     let refreshTimer: number | null = null;
 
-    const reconcileCompletedRun = async () => {
+    const reconcileRun = async () => {
       if (
         disposed ||
         streamReconciliationInFlightRef.current ||
@@ -793,7 +797,51 @@ export function useThreadStream({
       try {
         const client = getAPIClient(isMock);
         const run = await client.runs.get(activeThreadId, runId);
-        if (disposed || !TERMINAL_RUN_STATUSES.has(run.status)) {
+        if (disposed) {
+          return;
+        }
+
+        if (!TERMINAL_RUN_STATUSES.has(run.status)) {
+          // ``useStream`` retries one failed connection on its own.  If that
+          // reader quietly stalls, however, its internal queue never gets a
+          // terminal error and the page remains frozen despite the backend
+          // continuing the run.  Abort only the local reader, then join the
+          // same run so no work or tokens are lost server-side.
+          if (Date.now() - lastStreamActivityAtRef.current < STREAM_RECONNECT_IDLE_MS) {
+            return;
+          }
+
+          if (streamReconnectAttemptsRef.current >= STREAM_RECONNECT_MAX_ATTEMPTS) {
+            if (!streamReconnectFailureToastShownRef.current) {
+              streamReconnectFailureToastShownRef.current = true;
+              toast.error("回答仍在后台处理中，但页面连接暂未恢复；请稍后返回此会话查看结果。");
+            }
+            return;
+          }
+
+          streamReconnectAttemptsRef.current += 1;
+          // Avoid enqueueing another reconnect before this new reader has had
+          // enough time to receive its first event.  A genuine event resets
+          // the attempt counter through ``markStreamActivity``.
+          lastStreamActivityAtRef.current = Date.now();
+          if (!streamReconnectToastShownRef.current) {
+            streamReconnectToastShownRef.current = true;
+            toast.info("回答连接出现波动，正在自动恢复…");
+          }
+
+          const stream = threadStreamRef.current;
+          void Promise.resolve(stream.stop())
+            .catch(() => {
+              // Stopping a stale local reader is best-effort; joining is still
+              // useful if the SDK already considers the stream stopped.
+            })
+            .finally(() => {
+              if (disposed || threadIdRef.current !== activeThreadId) return;
+              void stream.joinStream(runId).catch(() => {
+                // The status poll will either reconcile a terminal run or
+                // schedule the next bounded reconnect attempt.
+              });
+            });
           return;
         }
 
@@ -801,13 +849,17 @@ export function useThreadStream({
         // receive any more useful events.  Load the checkpoint directly before
         // resetting the SDK stream; this preserves the normal onFinish side
         // effects (e.g. marking a position delivery complete) even when the
-        // browser missed the terminal event.
+        // browser missed the terminal event or already reported a transport
+        // error and dropped ``isLoading``.
         const finalState = await client.threads.getState(activeThreadId);
-        if (disposed || !threadStreamRef.current.isLoading) {
+        if (disposed) {
           return;
         }
 
         storage?.removeItem(`lg:stream:${activeThreadId}`);
+        streamReconnectAttemptsRef.current = 0;
+        streamReconnectToastShownRef.current = false;
+        streamReconnectFailureToastShownRef.current = false;
         if (run.status === "success") {
           listeners.current.onFinish?.(finalState.values as AgentThreadState);
           void queryClient.invalidateQueries({ queryKey: ["threads", "search"] });
@@ -836,7 +888,7 @@ export function useThreadStream({
     };
 
     const interval = window.setInterval(
-      () => void reconcileCompletedRun(),
+      () => void reconcileRun(),
       STREAM_RECONCILIATION_POLL_MS,
     );