Merge pull request #2283 from HUQIANTAO/fix/turn-stall-watchdog

fix(engine): recover from stalled in-progress turns

crates/tui/src/tui/ui.rs

@@ -149,6 +149,11 @@ const UI_IDLE_POLL_MS: u64 = 48;
 const UI_ACTIVE_POLL_MS: u64 = 24;
 const WEB_CONFIG_POLL_MS: u64 = 16;
 const DISPATCH_WATCHDOG_TIMEOUT: Duration = Duration::from_secs(30);
+/// Maximum wall-clock time a turn may stay in `"in_progress"` before the UI
+/// assumes the engine stalled (e.g. sub-agent hang, lost completion event,
+/// engine panic).  Matched to [`DEFAULT_STREAM_IDLE_TIMEOUT`] so legitimate
+/// long-running tool chains are not interrupted prematurely.
+const TURN_STALL_WATCHDOG_TIMEOUT: Duration = Duration::from_secs(300);
 // Forced repaint cadence while a turn is live (model loading, compacting,
 // sub-agents running). Drives the footer water-spout animation as well as
 // the per-tool spinner pulse — keep this fast enough that the spout reads as
@@ -4008,6 +4013,40 @@ fn reconcile_turn_liveness(app: &mut App, now: Instant, has_running_agents: bool
         return true;
     }
 
+    // Branch 3: turn started but never completed — engine may have
+    // panicked, sub-agent may be stuck, or the completion event was lost.
+    if app.is_loading
+        && matches!(app.runtime_turn_status.as_deref(), Some("in_progress"))
+        && !has_running_agents
+        && !app.is_compacting
+        && app.turn_started_at.is_some_and(|started| {
+            now.saturating_duration_since(started) > TURN_STALL_WATCHDOG_TIMEOUT
+        })
+    {
+        // Finalize in-flight thinking / assistant / tool cells so the
+        // transcript doesn't show permanent spinners after recovery.
+        streaming_thinking::finalize_current(app);
+        app.finalize_streaming_assistant_as_interrupted();
+        app.finalize_active_cell_as_interrupted();
+        app.streaming_state.reset();
+        app.streaming_message_index = None;
+        app.streaming_thinking_active_entry = None;
+
+        app.is_loading = false;
+        app.turn_started_at = None;
+        app.runtime_turn_status = None;
+        app.runtime_turn_id = None;
+        app.dispatch_started_at = None;
+        // Per-turn scroll lock — clear so the next turn auto-scrolls.
+        app.user_scrolled_during_stream = false;
+        app.push_status_toast(
+            "Turn stalled — no completion signal received. Please try again.",
+            StatusToastLevel::Error,
+            None,
+        );
+        return true;
+    }
+
     false
 }
 

crates/tui/src/tui/ui/tests.rs

@@ -2111,17 +2111,43 @@ fn turn_liveness_leaves_active_turn_running() {
     let mut app = create_test_app();
     app.is_loading = true;
     app.runtime_turn_status = Some("in_progress".to_string());
-    app.dispatch_started_at =
-        Some(Instant::now() - DISPATCH_WATCHDOG_TIMEOUT - Duration::from_secs(10));
+    app.turn_started_at = Some(Instant::now() - Duration::from_secs(60));
 
     let recovered = reconcile_turn_liveness(&mut app, Instant::now(), false);
 
     assert!(!recovered);
     assert!(app.is_loading);
-    assert!(app.dispatch_started_at.is_some());
+    assert!(app.turn_started_at.is_some());
     assert!(app.status_toasts.is_empty());
 }
 
+#[test]
+fn turn_liveness_recovers_stalled_in_progress_turn() {
+    let mut app = create_test_app();
+    app.is_loading = true;
+    app.runtime_turn_status = Some("in_progress".to_string());
+    app.runtime_turn_id = Some("stale-turn-id".to_string());
+    app.turn_started_at =
+        Some(Instant::now() - TURN_STALL_WATCHDOG_TIMEOUT - Duration::from_millis(1));
+    app.streaming_message_index = Some(0);
+    app.user_scrolled_during_stream = true;
+
+    let recovered = reconcile_turn_liveness(&mut app, Instant::now(), false);
+
+    assert!(recovered);
+    assert!(!app.is_loading);
+    assert!(app.turn_started_at.is_none());
+    assert!(app.runtime_turn_status.is_none());
+    assert!(app.runtime_turn_id.is_none());
+    assert!(app.dispatch_started_at.is_none());
+    assert!(app.streaming_message_index.is_none());
+    assert!(app.streaming_thinking_active_entry.is_none());
+    assert!(!app.user_scrolled_during_stream);
+    let toast = app.status_toasts.back().expect("stall toast");
+    assert_eq!(toast.level, StatusToastLevel::Error);
+    assert!(toast.text.contains("Turn stalled"));
+}
+
 #[test]
 fn fixed_model_auto_thinking_skips_auto_model_router() {
     let mut app = create_test_app();