fix(connection): let the library own keep_reconnecting's task — no double scheduling on protocol start/stop

custom_components/nvidia_shield/connection.py

@@ -166,7 +166,6 @@ class ShieldConnectionManager:
         self._protocol = protocol
         self._protocol_settings = protocol_settings
         self._invalid_auth_callback = invalid_auth_callback
-        self._protocol_task: asyncio.Task | None = None
         self._protocol_loop: asyncio.AbstractEventLoop | None = None
         self._listeners: list[Callable[[], None]] = []
         self._tasks: list[asyncio.Task] = []
@@ -251,10 +250,10 @@ class ShieldConnectionManager:
         client.add_is_available_updated_callback(self._on_is_available)
         client.add_current_app_updated_callback(self._on_current_app)
         client.add_volume_info_updated_callback(self._on_volume_info)
-        self._protocol_task = self._create_task(
+        # One owner per transport: the library's keep_reconnecting() creates
-            client.keep_reconnecting(self._on_invalid_auth),
+        # and owns its reconnect task internally — we never wrap or duplicate
-            "shield_protocol_keep_reconnecting",
+        # it (the stop path closes the client, which tears the task down).
-        )
+        client.keep_reconnecting(self._on_invalid_auth)
 
     def _on_is_on(self, on) -> None:
         self._protocol_push(FIELD_IS_ON, bool(on))
@@ -420,9 +419,6 @@ class ShieldConnectionManager:
         tasks, self._tasks = self._tasks, []
         for task in tasks:
             task.cancel()
-        if self._protocol_task is not None:
-            self._protocol_task.cancel()
-            self._protocol_task = None
         self._listeners.clear()
         if self._poller is not None:
             with contextlib.suppress(Exception):