G3: F02 — inject reconnect callback for send_appstart after reconnect

ConnectionManager._attempt_reconnect called self.connection.connect()
directly, bypassing MeshCore.connect() which runs send_appstart().
Firmware requires CMD_APP_START after every transport-level connection
to initialize the session.  Without it, the reconnected transport has
no active session — sends go unanswered, tcp_no_response fires after
5 attempts, handle_disconnect re-enters _attempt_reconnect, and the
reconnect storm begins.

Fix: add an optional reconnect_callback parameter to
ConnectionManager.__init__.  MeshCore passes self._on_reconnect which
calls send_appstart() after the transport reconnects.  The callback
is invoked inside _attempt_reconnect immediately after a successful
connect(), before the CONNECTED event is emitted.  Callback failures
are logged as warnings but do not break the reconnect — the transport
is up regardless.  Default None keeps the API backwards-compatible
for direct ConnectionManager users.

Refs: Forensics report finding F02

src/meshcore/connection_manager.py

@@ -48,11 +48,13 @@ class ConnectionManager:
         event_dispatcher=None,
         auto_reconnect: bool = False,
         max_reconnect_attempts: int = 3,
+        reconnect_callback: Optional[Callable[[], Awaitable[None]]] = None,
     ):
         self.connection = connection
         self.event_dispatcher = event_dispatcher
         self.auto_reconnect = auto_reconnect
         self.max_reconnect_attempts = max_reconnect_attempts
+        self._reconnect_callback = reconnect_callback
 
         self._reconnect_attempts = 0
         self._is_connected = False
@@ -139,6 +141,16 @@ class ConnectionManager:
                 if result is not None:
                     self._is_connected = True
                     self._reconnect_attempts = 0
+
+                    # Invoke reconnect callback (e.g. send_appstart) if provided
+                    if self._reconnect_callback is not None:
+                        try:
+                            await self._reconnect_callback()
+                        except Exception as cb_err:
+                            logger.warning(
+                                f"Reconnect callback failed: {cb_err}"
+                            )
+
                     await self._emit_event(
                         EventType.CONNECTED,
                         {"connection_info": result, "reconnected": True},

src/meshcore/meshcore.py

@@ -28,10 +28,17 @@ class MeshCore:
         auto_reconnect: bool = False,
         max_reconnect_attempts: int = 3,
     ):
-        # Wrap connection with ConnectionManager
+        # Wrap connection with ConnectionManager.
+        # The reconnect callback ensures send_appstart() runs after every
+        # transport-level reconnect, which is required by firmware to
+        # initialize the session (F02).
         self.dispatcher = EventDispatcher()
         self.connection_manager = ConnectionManager(
-            cx, self.dispatcher, auto_reconnect, max_reconnect_attempts
+            cx,
+            self.dispatcher,
+            auto_reconnect,
+            max_reconnect_attempts,
+            reconnect_callback=self._on_reconnect,
         )
         self.cx = self.connection_manager  # For backward compatibility
 
@@ -174,6 +181,15 @@ class MeshCore:
             return None
         return mc
 
+    async def _on_reconnect(self):
+        """Callback invoked by ConnectionManager after a successful reconnect.
+
+        Firmware requires CMD_APP_START after every transport-level connection
+        to initialize the session.  MeshCore.connect() does this on the initial
+        connection; this callback ensures it also happens on reconnects (F02).
+        """
+        await self.commands.send_appstart()
+
     async def connect(self):
         await self.dispatcher.start()
         result = await self.connection_manager.connect()