Thread safety fixes for session and message counts

NousResearch/hermes-agent · cluster 2130 · agent

Data as of

Nothing open on GitHub. Fix merged in #2466. Check it covers your case.

The short answer

0open issues
0open PRs
6items grouped
Known?
Yes. 6 related items grouped here, 1 issue already closed.
Fix event
A fix was merged to main in PR #2466. Check it covers your case. No release data is tracked.
Tracker
The maintainers' tracker marks this unresolved. That is separate from the fix event.
Still open
Nothing open on GitHub.

Multiple issues and pull requests addressing missing thread locks in session_count() and message_count() methods, with related concurrent subagent thread safety fixes. The group centers on a specific thread-safety bug and its proposed fixes.

Open issues0

No open issues.

Open PRs0

No open pull requests.

Fixes on record

Merged to main. A stamp counts the issues a fix is linked to.

  • #2466 fix: batch of 5 small contributor fixes — PortAudio, SafeWriter, IMAP, thread lock, prefill

    merged to main

    files: gateway/platforms/email.py, hermes_state.py, run_agent.py, tools/voice_mode.py

    Show the diff
    • gateway/platforms/email.py
      @@ -230,7 +230,7 @@ async def connect(self) -> bool:
                   # Mark all existing messages as seen so we only process new ones
                   imap.select("INBOX")
                   status, data = imap.uid("search", None, "ALL")
      -            if status == "OK" and data[0]:
      +            if status == "OK" and data and data[0]:
                       for uid in data[0].split():
                           self._seen_uids.add(uid)
                   imap.logout()
      @@ -295,7 +295,7 @@ def _fetch_new_messages(self) -> List[Dict[str, Any]]:
                   imap.select("INBOX")
       
                   status, data = imap.uid("search", None, "UNSEEN")
      -            if status != "OK" or not data[0]:
      +            if status != "OK" or not data or not data[0]:
                       imap.logout()
                       return results
       
    • hermes_state.py
      @@ -855,23 +855,25 @@ def search_sessions(
       
           def session_count(self, source: str = None) -> int:
               """Count sessions, optionally filtered by source."""
      -        if source:
      -            cursor = self._conn.execute(
      -                "SELECT COUNT(*) FROM sessions WHERE source = ?", (source,)
      -            )
      -        else:
      -            cursor = self._conn.execute("SELECT COUNT(*) FROM sessions")
      -        return cursor.fetchone()[0]
      +        with self._lock:
      +            if source:
      +                cursor = self._conn.execute(
      +                    "SELECT COUNT(*) FROM sessions WHERE source = ?", (source,)
      +                )
      +            else:
      +                cursor = self._conn.execute("SELECT COUNT(*) FROM sessions")
      +            return cursor.fetchone()[0]
       
           def message_count(self, session_id: str = None) -> int:
               """Count messages, optionally for a specific session."""
      -        if session_id:
      -            cursor = self._conn.execute(
      -                "SELECT COUNT(*) FROM messages WHERE session_id = ?", (session_id,)
      -            )
      -        else:
      -            cursor = self._conn.execute("SELECT COUNT(*) FROM messages")
      -        return cursor.fetchone()[0]
      +        with self._lock:
      +            if session_id:
      +                cursor = self._conn.execute(
      +                    "SELECT COUNT(*) FROM messages WHERE session_id = ?", (session_id,)
      +                )
      +            else:
      +                cursor = self._conn.execute("SELECT COUNT(*) FROM messages")
      +            return cursor.fetchone()[0]
       
           # =========================================================================
           # Export and cleanup
    • run_agent.py
      @@ -108,7 +108,7 @@
       
       
       class _SafeWriter:
      -    """Transparent stdio wrapper that catches OSError from broken pipes.
      +    """Transparent stdio wrapper that catches OSError/ValueError from broken pipes.
       
           When hermes-agent runs as a systemd service, Docker container, or headless
           daemon, the stdout/stderr pipe can become unavailable (idle timeout, buffer
      @@ -117,8 +117,13 @@ class _SafeWriter:
           run_conversation() — especially via double-fault when an except handler
           also tries to print.
       
      +    Additionally, when subagents run in ThreadPoolExecutor threads, the shared
      +    stdout handle can close between thread teardown and cleanup, raising
      +    ``ValueError: I/O operation on closed file`` instead of OSError.
      +
           This wrapper delegates all writes to the underlying stream and silently
      -    catches OSError. It is transparent when the wrapped stream is healthy.
      +    catches both OSError and ValueError. It is transparent when the wrapped
      +    stream is healthy.
           """
       
           __slots__ = ("_inner",)
      @@ -129,13 +134,13 @@ def __init__(self, inner):
           def write(self, data):
               try:
                   return self._inner.write(data)
      -        except OSError:
      +        except (OSError, ValueError):
                   return len(data) if isinstance(data, str) else 0
       
           def flush(self):
               try:
                   self._inner.flush()
      -        except OSError:
      +        except (OSError, ValueError):
                   pass
       
           def fileno(self):
      @@ -144,7 +149,7 @@ def fileno(self):
           def isatty(self):
               try:
                   return self._inner.isatty()
      -        except OSError:
      +        except (OSError, ValueError):
                   return False
       
           def __getattr__(self, name):
      @@ -2438,7 +2443,18 @@ def _sanitize_api_messages(messages: List[Dict[str, Any]]) -> List[Dict[str, Any
                       "Pre-call sanitizer: added %d stub tool result(s)",
                       len(missing_results),
                   )
      -
      +        # 3. Strip trailing empty assistant messages to prevent prefill rejection.
      +        # These can leak from Responses API reasoning-only turns (Codex/MiniMax)
      +        # where an empty assistant message is required by the Responses API but
      +        # must NOT be sent to Chat Completions or Anthropic Messages API providers.
      +        while (
      +            messages
      +            and messages[-1].get("role") == "assistant"
      +            and not (messages[-1].get("content") or "").strip()
      +            and not messages[-1].get("tool_calls")
      +        ):
      +            logger.debug("Pre-call sanitizer: removed trailing empty assistant message")
      +            messages = messages[:-1]
               return messages
       
           @staticmethod
    • tools/voice_mode.py
      @@ -81,8 +81,15 @@ def detect_audio_environment() -> dict:
                       warnings.append("No audio input/output devices detected")
               except Exception:
                   warnings.append("Audio subsystem error (PortAudio cannot query devices)")
      -    except (ImportError, OSError):
      +    except ImportError:
               warnings.append("Audio libraries not installed (pip install sounddevice numpy)")
      +    except OSError:
      +        warnings.append(
      +            "PortAudio system library not found -- install it first:\n"
      +            "  Linux:  sudo apt-get install libportaudio2\n"
      +            "  macOS:  brew install portaudio\n"
      +            "Then retry /voice on."
      +        )
       
           return {
               "available": len(warnings) == 0,

4 PRs closed without merging

From the close comments: duplicate 2 · merged via another change 1 · other 1

  • #2131 fix(state): add missing thread lock to session_count() and message_count() · duplicate
  • #2133 fix(state): add thread lock to session_count() and message_count() · duplicate
  • #2251 fix(state): add missing thread lock to session_count() and message_count() · other
  • #2288 fix(state): add missing thread lock to session_count() and message_count() · merged via another change → #2466
Closed issues · 1
  • Closed#2130Bug: session_count() and message_count() missing thread lock
How this cluster was grouped
  • Link evidence (structural links throughout): every member is connected by structural evidence: shared changed files, or a closing/superseded reference.
  • Pairwise link density 0.33: the share of item pairs that carry a link, counting a structural link in full and a relevance-only link at 0.36. 1.00 means every pair is linked.

Pick up a Lab task

Check whether a fix covers the open issues here, with your own agent. Bring back evidence for maintainers.

Open the Lab

At a glance

Subsystem
agent
Merged PRs
#2466
Closed issues
1
Attempts
4 PRs closed without merging
Fix link checked
22 Mar 2026. This confirmation is old and may be out of date (a revert or reopen would not have invalidated it here); re-verify.

For your agent

The same facts, typed.