fix: use dedicated _shutting_down flag for cancel-time checkpoint restore

The previous guard used self._running, which is False both before run() starts and during shutdown. Tests and process_direct() dispatch messages with _running=False, so cancelled turns there silently skipped the checkpoint restore and lost partial context. Introduce _shutting_down, set only by stop(), and gate the restore on it.
fix: prevent duplicate user message on session recovery re-publish
2026-07-31 15:49:31 +08:00 · 2026-07-31 15:37:41 +08:00 · 2026-07-31 15:36:51 +08:00 · 2026-07-31 15:36:51 +08:00 · 2026-07-31 15:36:51 +08:00 · 2026-07-31 15:36:51 +08:00
1 changed files with 11 additions and 3 deletions
@@ -378,6 +378,7 @@ class AgentLoop:
        )
        self._unified_session = unified_session
        self._running = False
+        self._shutting_down = False
        self._mcp_servers = mcp_servers or {}
        self._mcp_stacks: dict[str, MCPConnection] = {}
        self._mcp_connecting = False
@@ -1224,11 +1225,11 @@ class AgentLoop:
                    # it into session history now makes it visible in the
                    # next conversation turn.
                    #
-                    # During gateway shutdown (self._running is False), skip
-                    # this so the pending markers survive and
+                    # During gateway shutdown (self._shutting_down is True),
+                    # skip this so the pending markers survive and
                    # recover_stale_sessions() can notify the user and
                    # re-trigger the interrupted turn on next startup.
-                    if self._running:
+                    if not self._shutting_down:
                        try:
                            key = self._effective_session_key(msg)
                            session = self.sessions.get_or_create(key)
@@ -1318,6 +1319,7 @@ class AgentLoop:
    def stop(self) -> None:
        """Stop the agent loop."""
        self._running = False
+        self._shutting_down = True
        logger.info("Agent loop stopping")

    async def _process_message(
@@ -2151,6 +2153,12 @@ class AgentLoop:
                            sender_id=chat_id,
                            chat_id=chat_id,
                            content=last_user_msg["content"],
+                            # The message already exists in session history
+                            # (persisted before the crash); skip persisting it
+                            # again so recovery does not duplicate it.
+                            metadata={
+                                turn_continuation.SKIP_USER_PERSIST_META: True
+                            },
                        )
                    )
                    logger.info(
Author	SHA1	Message	Date
nanobot	298b7ac6f6	fix: use dedicated _shutting_down flag for cancel-time checkpoint restore Test Suite / Detect changes (push) Has been cancelled Details Test Suite / webui (push) Has been cancelled Details Test Suite / docker (push) Has been cancelled Details Test Suite / Python (Windows, 3.14) (push) Has been cancelled Details Test Suite / Python (minimum, 3.11) (push) Has been cancelled Details Test Suite / Python (latest, 3.14 + coverage) (push) Has been cancelled Details The previous guard used self._running, which is False both before run() starts and during shutdown. Tests and process_direct() dispatch messages with _running=False, so cancelled turns there silently skipped the checkpoint restore and lost partial context. Introduce _shutting_down, set only by stop(), and gate the restore on it.	2026-07-31 15:49:31 +08:00
nanobot	c7270b8d81	fix: prevent duplicate user message on session recovery re-publish The re-published last user message was persisted into session history a second time when the agent loop processed it, duplicating the user's input. Mark the re-published InboundMessage with SKIP_USER_PERSIST_META so should_persist_user_message() and _save_skip_for_turn() treat it as already-persisted.	2026-07-31 15:37:41 +08:00
nanobot	66b8743024	docs: NixOS deployment guide from Camellia	2026-07-31 15:36:51 +08:00
nanobot	617a0129b7	Revert "docs: deployment guide from Camellia experience" This reverts commit `ded5d4d8ed`.	2026-07-31 15:36:51 +08:00
nanobot	3b4eb40f1e	docs: deployment guide from Camellia experience	2026-07-31 15:36:51 +08:00
nanobot	6b7f3889ea	fix: raise error when Telegram bot not running instead of silently dropping messages When the outbound dispatcher processes recovery notifications before the Telegram bot is fully started, the send() method silently returned, causing the notification to be lost. Now it raises RuntimeError so _send_with_retry will retry until the bot is ready.	2026-07-31 15:36:51 +08:00
nanobot	9240000e75	fix: capture pending_user_turn before _restore_runtime_checkpoint clears it _restore_runtime_checkpoint() clears both the runtime_checkpoint and pending_user_turn flags (lines 1966-1967). recover_stale_sessions() checked had_pending AFTER calling it, so had_pending was always False when a checkpoint existed. This meant the re-publish of the last user message never happened — the user got a notification but the interrupted turn was never re-triggered.	2026-07-31 15:36:51 +08:00
nanobot	5248513ad1	feat: notify user and re-trigger interrupted turns on session recovery recover_stale_sessions now publishes a restart notification to the user channel and re-publishes the last user message for sessions with a pending turn, so interrupted work continues automatically after a crash or restart. Also skip checkpoint restore during shutdown so pending markers survive for startup recovery.	2026-07-31 15:36:47 +08:00