feat: enhance WebSocket error handling and cleanup logic in simulation and ESP32 libraries

backend/app/api/routes/simulation.py

@@ -40,8 +40,14 @@ async def simulation_websocket(websocket: WebSocket, client_id: str):
             logger.info('[%s] system event: %s', client_id, data.get('event'))
         elif event_type == 'error':
             logger.error('[%s] error: %s', client_id, data.get('message'))
+        elif event_type == 'serial_output':
+            text = data.get('data', '')
+            logger.info('[%s] serial_output uart=%s len=%d: %r', client_id, data.get('uart', 0), len(text), text[:80])
         payload = json.dumps({'type': event_type, 'data': data})
-        await manager.send(client_id, payload)
+        try:
+            await manager.send(client_id, payload)
+        except Exception as _send_exc:
+            logger.debug('[%s] qemu_callback send failed (%s): %s', client_id, event_type, _send_exc)
 
     def _use_lib() -> bool:
         return esp_lib_manager.is_available()
@@ -177,13 +183,19 @@ async def simulation_websocket(websocket: WebSocket, client_id: str):
                     )
 
     except WebSocketDisconnect:
-        manager.disconnect(client_id)
-        qemu_manager.stop_instance(client_id)
-        await esp_lib_manager.stop_instance(client_id)
-        esp_qemu_manager.stop_instance(client_id)
+        # Guard: only clean up if this coroutine still owns the connection for client_id.
+        # A newer simulation_websocket may have already connected and replaced us.
+        if manager.active_connections.get(client_id) is websocket:
+            manager.disconnect(client_id)
+            qemu_manager.stop_instance(client_id)
+            await esp_lib_manager.stop_instance(client_id)
+            esp_qemu_manager.stop_instance(client_id)
+        else:
+            logger.info('[%s] old WS session ended; newer session is active — skipping cleanup', client_id)
     except Exception as exc:
         logger.error('WebSocket error for %s: %s', client_id, exc)
-        manager.disconnect(client_id)
-        qemu_manager.stop_instance(client_id)
-        await esp_lib_manager.stop_instance(client_id)
-        esp_qemu_manager.stop_instance(client_id)
+        if manager.active_connections.get(client_id) is websocket:
+            manager.disconnect(client_id)
+            qemu_manager.stop_instance(client_id)
+            await esp_lib_manager.stop_instance(client_id)
+            esp_qemu_manager.stop_instance(client_id)

backend/app/services/esp32_lib_manager.py

@@ -160,7 +160,10 @@ class EspLibManager:
 
         logger.info('Launching esp32_worker for %s (machine=%s, script=%s, python=%s)',
                     client_id, machine, _WORKER_SCRIPT, sys.executable)
-        await callback('system', {'event': 'booting'})
+        try:
+            await callback('system', {'event': 'booting'})
+        except Exception as exc:
+            logger.warning('start_instance %s: booting event delivery failed: %s', client_id, exc)
 
         try:
             proc = subprocess.Popen(
@@ -362,6 +365,17 @@ class EspLibManager:
                 raw = raw.strip()
                 if not raw:
                     continue
+                # With -nographic, serial0 is connected to the stdio mux so
+                # qemu_chr_fe_write() writes the raw UART byte to fd 1 just
+                # before picsimlab_uart_tx_event emits the JSON line.  Strip
+                # any prefix bytes before the JSON object marker.
+                idx = raw.find(b'{"type":')
+                if idx > 0:
+                    raw = raw[idx:]
+                elif idx < 0:
+                    logger.debug('[%s] ignoring non-JSON worker line: %s',
+                                 client_id, raw[:200])
+                    continue
                 try:
                     event = json.loads(raw)
                 except Exception:

frontend/src/store/useSimulatorStore.ts

@@ -307,6 +307,13 @@ export const useSimulatorStore = create<SimulatorState>((set, get) => {
         bridge.onCrash = () => {
           set({ esp32CrashBoardId: id });
         };
+        bridge.onDisconnected = () => {
+          set((s) => {
+            const boards = s.boards.map((b) => b.id === id ? { ...b, running: false } : b);
+            const isActive = s.activeBoardId === id;
+            return { boards, ...(isActive ? { running: false } : {}) };
+          });
+        };
         bridge.onLedcUpdate = (update) => {
           // Route LEDC duty cycles to PinManager as PWM (0.0–1.0).
           // If gpio is known (from GPIO out_sel sync), use the actual GPIO pin;
@@ -619,6 +626,13 @@ export const useSimulatorStore = create<SimulatorState>((set, get) => {
           if (boardPm) boardPm.triggerPinChange(gpioPin, state);
         };
         bridge.onCrash = () => { set({ esp32CrashBoardId: boardId }); };
+        bridge.onDisconnected = () => {
+          set((s) => {
+            const boards = s.boards.map((b) => b.id === boardId ? { ...b, running: false } : b);
+            const isActive = s.activeBoardId === boardId;
+            return { boards, ...(isActive ? { running: false } : {}) };
+          });
+        };
         bridge.onLedcUpdate = (update) => {
           const boardPm = pinManagerMap.get(boardId);
           if (boardPm && typeof boardPm.updatePwm === 'function') {