claude-mpm 4.2.38__tar.gz → 4.2.40__tar.gz

claude_mpm-4.2.40/BUILD_NUMBER

@@ -0,0 +1 @@
+362

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: claude-mpm
-Version: 4.2.38
+Version: 4.2.40
 Summary: Claude Multi-Agent Project Manager - Orchestrate Claude with agent delegation and ticket tracking
 Author-email: Bob Matsuoka <bob@matsuoka.com>
 Maintainer: Claude MPM Team

claude_mpm-4.2.40/VERSION

@@ -0,0 +1 @@
+4.2.40

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "claude-mpm"
-version = "4.2.38"
+version = "4.2.40"
 description = "Claude Multi-Agent Project Manager - Orchestrate Claude with agent delegation and ticket tracking"
 readme = "README.md"
 requires-python = ">=3.8"
@@ -78,7 +78,7 @@ ignore-imports = false
 
 [tool.commitizen]
 name = "cz_conventional_commits"
-version = "4.2.38"
+version = "4.2.40"
 version_files = [ "VERSION", "src/claude_mpm/VERSION", "package.json:version",]
 tag_format = "v$version"
 update_changelog_on_bump = true

claude_mpm-4.2.40/src/claude_mpm/VERSION

@@ -0,0 +1 @@
+4.2.40

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm/cli/__init__.py

@@ -106,7 +106,8 @@ def main(argv: Optional[list] = None):
         # Check for MCP auto-configuration (pipx installations)
         _check_mcp_auto_configuration()
 
-    # Verify MCP Gateway configuration on startup (non-blocking)
+    # Re-enabled: MCP pre-warming is safe with subprocess daemon (v4.2.40)
+    # The subprocess approach avoids fork() issues entirely
     _verify_mcp_gateway_startup()
 
     # Set up logging

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm/services/cli/unified_dashboard_manager.py

@@ -133,7 +133,7 @@ class UnifiedDashboardManager(IUnifiedDashboardManager):
             if background:
                 # The daemon.start() method will handle cleanup when force_restart=True
                 # We don't need pre-emptive cleanup here as it causes race conditions
-                
+
                 # Try to start daemon with retry on port conflicts
                 max_retries = 3
                 retry_count = 0

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm/services/monitor/daemon.py

@@ -157,35 +157,67 @@ class UnifiedMonitorDaemon:
             )
             return False
 
-        # Wait for any pre-warming threads to complete before forking
-        self._wait_for_prewarm_completion()
-
-        # Use daemon manager's daemonize which includes cleanup
-        # DO NOT reset startup_status_file - it's needed for parent-child communication!
-        # self.daemon_manager.startup_status_file = None  # BUG: This breaks communication
-        success = self.daemon_manager.daemonize()
-        if not success:
-            return False
+        # Use subprocess approach for clean daemon startup (v4.2.40)
+        # This avoids all fork() + threading issues by starting in a fresh process
+        # The daemon_manager.use_subprocess_daemon() now checks for CLAUDE_MPM_SUBPROCESS_DAEMON
+        # environment variable to prevent infinite recursion
+        if self.daemon_manager.use_subprocess_daemon():
+            # Start using subprocess - this returns immediately in parent
+            success = self.daemon_manager.start_daemon_subprocess()
+            return success
+
+        # Check if we're in subprocess mode (environment variable set)
+        if os.environ.get("CLAUDE_MPM_SUBPROCESS_DAEMON") == "1":
+            # We're in a subprocess started by start_daemon_subprocess
+            # We need to write the PID file ourselves since parent didn't
+            self.logger.info("Running in subprocess daemon mode, writing PID file")
+            self.daemon_manager.write_pid_file()
+
+            # Setup signal handlers for graceful shutdown
+            self._setup_signal_handlers()
+
+            # Start the server (this will run until shutdown)
+            try:
+                result = self._run_server()
+                if not result:
+                    self.logger.error("Failed to start server in subprocess mode")
+                return result
+            except Exception as e:
+                self.logger.error(f"Server startup exception in subprocess: {e}")
+                raise
+        else:
+            # Legacy fork approach (kept for compatibility but not used by default)
+            # Wait for any pre-warming threads to complete before forking
+            self._wait_for_prewarm_completion()
+
+            # Use daemon manager's daemonize which includes cleanup
+            # DO NOT reset startup_status_file - it's needed for parent-child communication!
+            # self.daemon_manager.startup_status_file = None  # BUG: This breaks communication
+            success = self.daemon_manager.daemonize()
+            if not success:
+                return False
 
-        # We're now in the daemon process
-        # Update our PID references and status file
-        self.lifecycle.pid_file = self.daemon_manager.pid_file
-        self.lifecycle.startup_status_file = self.daemon_manager.startup_status_file
+            # We're now in the daemon process
+            # Update our PID references and status file
+            self.lifecycle.pid_file = self.daemon_manager.pid_file
+            self.lifecycle.startup_status_file = self.daemon_manager.startup_status_file
 
-        # Start the server in daemon mode
-        try:
-            result = self._run_server()
-            if not result:
-                # Report failure before exiting
-                self.daemon_manager._report_startup_error("Failed to start server")
-            else:
-                # Report success
-                self.daemon_manager._report_startup_success()
-            return result
-        except Exception as e:
-            # Report any exceptions during startup
-            self.daemon_manager._report_startup_error(f"Server startup exception: {e}")
-            raise
+            # Start the server in daemon mode
+            try:
+                result = self._run_server()
+                if not result:
+                    # Report failure before exiting
+                    self.daemon_manager._report_startup_error("Failed to start server")
+                else:
+                    # Report success
+                    self.daemon_manager._report_startup_success()
+                return result
+            except Exception as e:
+                # Report any exceptions during startup
+                self.daemon_manager._report_startup_error(
+                    f"Server startup exception: {e}"
+                )
+                raise
 
     def _start_foreground(self, force_restart: bool = False) -> bool:
         """Start in foreground mode.
@@ -602,29 +634,30 @@ class UnifiedMonitorDaemon:
                 self.logger.info(
                     f"Waiting for {len(active_threads)} background threads to complete before forking"
                 )
-                
+
                 # List thread names for debugging
                 thread_names = [t.name for t in active_threads]
                 self.logger.debug(f"Active threads: {thread_names}")
 
                 # Wait for threads to complete or timeout
                 while time.time() - start_time < timeout:
-                    remaining_threads = [
-                        t for t in active_threads if t.is_alive()
-                    ]
+                    remaining_threads = [t for t in active_threads if t.is_alive()]
                     if not remaining_threads:
                         self.logger.debug("All threads completed")
                         break
-                    
+
                     # Log remaining threads periodically
                     if int(time.time() - start_time) % 1 == 0:
-                        self.logger.debug(f"{len(remaining_threads)} threads still active")
-                    
+                        self.logger.debug(
+                            f"{len(remaining_threads)} threads still active"
+                        )
+
                     time.sleep(0.1)
-                
+
                 # Final check
                 final_threads = [
-                    t for t in threading.enumerate() 
+                    t
+                    for t in threading.enumerate()
                     if t.is_alive() and t != threading.current_thread()
                 ]
                 if final_threads:

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm/services/monitor/daemon_manager.py

@@ -149,7 +149,7 @@ class DaemonManager:
             # Try IPv4 first (most common)
             test_sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
             test_sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
-            
+
             # Use 127.0.0.1 for localhost to match what the server does
             bind_host = "127.0.0.1" if self.host == "localhost" else self.host
             test_sock.bind((bind_host, self.port))
@@ -201,7 +201,10 @@ class DaemonManager:
         try:
             # Find processes using the port
             result = subprocess.run(
-                ["lsof", "-ti", f":{self.port}"], capture_output=True, text=True, check=False
+                ["lsof", "-ti", f":{self.port}"],
+                capture_output=True,
+                text=True,
+                check=False,
             )
 
             if result.returncode != 0 or not result.stdout.strip():
@@ -220,35 +223,41 @@ class DaemonManager:
                     process_info = subprocess.run(
                         ["ps", "-p", str(pid), "-o", "comm="],
                         capture_output=True,
-                        text=True, check=False,
+                        text=True,
+                        check=False,
                     )
 
                     # Get full command to check if it's our monitor process
                     cmd_info = subprocess.run(
                         ["ps", "-p", str(pid), "-o", "command="],
                         capture_output=True,
-                        text=True, check=False,
+                        text=True,
+                        check=False,
                     )
-                    
+
                     if cmd_info.returncode != 0:
                         continue
-                        
+
                     full_command = cmd_info.stdout.strip().lower()
                     process_name = process_info.stdout.strip().lower()
-                    
+
                     # Check if this is our monitor/socketio process specifically
                     # Look for monitor, socketio, dashboard, or our specific port
-                    is_monitor = any([
-                        "monitor" in full_command,
-                        "socketio" in full_command,
-                        "dashboard" in full_command,
-                        f"port={self.port}" in full_command,
-                        f":{self.port}" in full_command,
-                        "unified_monitor" in full_command,
-                    ])
-                    
+                    is_monitor = any(
+                        [
+                            "monitor" in full_command,
+                            "socketio" in full_command,
+                            "dashboard" in full_command,
+                            f"port={self.port}" in full_command,
+                            f":{self.port}" in full_command,
+                            "unified_monitor" in full_command,
+                        ]
+                    )
+
                     if is_monitor and "python" in process_name:
-                        self.logger.info(f"Killing monitor process {pid}: {full_command[:100]}")
+                        self.logger.info(
+                            f"Killing monitor process {pid}: {full_command[:100]}"
+                        )
                         os.kill(pid, signal.SIGTERM)
 
                         # Wait briefly for graceful shutdown
@@ -329,7 +338,7 @@ class DaemonManager:
 
     def _kill_claude_mpm_processes(self) -> bool:
         """Kill any claude-mpm monitor processes specifically.
-        
+
         This targets monitor/dashboard/socketio processes only,
         NOT general Claude instances.
 
@@ -338,7 +347,9 @@ class DaemonManager:
         """
         try:
             # Look for monitor-specific processes
-            result = subprocess.run(["ps", "aux"], capture_output=True, text=True, check=False)
+            result = subprocess.run(
+                ["ps", "aux"], capture_output=True, text=True, check=False
+            )
 
             if result.returncode != 0:
                 return False
@@ -349,10 +360,14 @@ class DaemonManager:
             for line in lines:
                 line_lower = line.lower()
                 # Only target monitor/dashboard/socketio processes
-                if any(["monitor" in line_lower and "claude" in line_lower,
+                if any(
+                    [
+                        "monitor" in line_lower and "claude" in line_lower,
                         "dashboard" in line_lower and "claude" in line_lower,
                         "socketio" in line_lower,
-                        f":{self.port}" in line_lower and "python" in line_lower]):
+                        f":{self.port}" in line_lower and "python" in line_lower,
+                    ]
+                ):
                     parts = line.split()
                     if len(parts) > 1:
                         try:
@@ -395,7 +410,8 @@ class DaemonManager:
                     process_info = subprocess.run(
                         ["ps", "-p", str(pid), "-o", "comm="],
                         capture_output=True,
-                        text=True, check=False,
+                        text=True,
+                        check=False,
                     )
 
                     if "python" in process_info.stdout.lower():
@@ -441,7 +457,10 @@ class DaemonManager:
         """
         try:
             result = subprocess.run(
-                ["lsof", "-ti", f":{self.port}"], capture_output=True, text=True, check=False
+                ["lsof", "-ti", f":{self.port}"],
+                capture_output=True,
+                text=True,
+                check=False,
             )
 
             if result.returncode == 0 and result.stdout.strip():
@@ -485,9 +504,145 @@ class DaemonManager:
                 self.logger.error(f"Cannot start daemon - port {self.port} is in use")
                 return False
 
-            # Daemonize the process
+            # Use subprocess for clean daemon startup (v4.2.40)
+            # This avoids fork() issues with Python threading
+            if self.use_subprocess_daemon():
+                return self.start_daemon_subprocess()
+            # Fallback to traditional fork (kept for compatibility)
             return self.daemonize()
 
+    def use_subprocess_daemon(self) -> bool:
+        """Check if we should use subprocess instead of fork for daemonization.
+
+        Returns:
+            True to use subprocess (safer), False to use traditional fork
+        """
+        # Check if we're already in a subprocess to prevent infinite recursion
+        if os.environ.get("CLAUDE_MPM_SUBPROCESS_DAEMON") == "1":
+            # We're already in a subprocess, use traditional fork
+            return False
+
+        # Otherwise, use subprocess for monitor daemon to avoid threading issues
+        return True
+
+    def start_daemon_subprocess(self) -> bool:
+        """Start daemon using subprocess.Popen for clean process isolation.
+
+        This avoids all the fork() + threading issues by starting the monitor
+        in a completely fresh process with no inherited threads or locks.
+
+        Returns:
+            True if daemon started successfully
+        """
+        try:
+            # Build command to run monitor in foreground mode in subprocess
+            import sys
+
+            python_exe = sys.executable
+
+            # Run 'claude-mpm monitor start' in subprocess with environment variable
+            # to indicate we're already in a subprocess (prevents infinite recursion)
+            cmd = [
+                python_exe,
+                "-m",
+                "claude_mpm.cli",
+                "monitor",
+                "start",
+                "--background",  # Run as daemon
+                "--port",
+                str(self.port),
+                "--host",
+                self.host,
+            ]
+
+            # Set environment variable to prevent recursive subprocess creation
+            env = os.environ.copy()
+            env["CLAUDE_MPM_SUBPROCESS_DAEMON"] = "1"
+
+            self.logger.info(f"Starting monitor daemon via subprocess: {' '.join(cmd)}")
+
+            # Open log file for output redirection
+            log_file_handle = None
+            if self.log_file:
+                log_file_handle = open(self.log_file, "a")
+                log_file = log_file_handle
+            else:
+                log_file = subprocess.DEVNULL
+
+            try:
+                # Start the subprocess detached from parent
+                # Redirect stdout/stderr to log file to capture output
+                process = subprocess.Popen(
+                    cmd,
+                    stdin=subprocess.DEVNULL,
+                    stdout=log_file,
+                    stderr=subprocess.STDOUT if self.log_file else subprocess.DEVNULL,
+                    start_new_session=True,  # Create new process group
+                    close_fds=(
+                        False if self.log_file else True
+                    ),  # Keep log file open if redirecting
+                    env=env,  # Pass modified environment
+                )
+
+                # Close the log file handle now that subprocess has it
+                if log_file_handle:
+                    log_file_handle.close()
+
+                # Get the process PID
+                pid = process.pid
+                self.logger.info(f"Monitor subprocess started with PID {pid}")
+
+                # Wait for the subprocess to write its PID file
+                # The subprocess will write the PID file after it starts successfully
+                max_wait = 10  # seconds
+                start_time = time.time()
+
+                while time.time() - start_time < max_wait:
+                    # Check if process is still running
+                    if process.poll() is not None:
+                        # Process exited
+                        self.logger.error(
+                            f"Monitor daemon exited with code {process.returncode}"
+                        )
+                        return False
+
+                    # Check if PID file was written
+                    if self.pid_file.exists():
+                        try:
+                            with open(self.pid_file) as f:
+                                written_pid = int(f.read().strip())
+                            if written_pid == pid:
+                                # PID file written correctly, check port
+                                if (
+                                    not self._is_port_available()
+                                ):  # Port NOT available means it's in use (good!)
+                                    self.logger.info(
+                                        f"Monitor daemon successfully started on port {self.port}"
+                                    )
+                                    return True
+                        except:
+                            pass  # PID file not ready yet
+
+                    time.sleep(0.5)
+
+                # Timeout waiting for daemon to start
+                self.logger.error("Timeout waiting for monitor daemon to start")
+                # Try to kill the process if it's still running
+                if process.poll() is None:
+                    process.terminate()
+                    time.sleep(1)
+                    if process.poll() is None:
+                        process.kill()
+                return False
+            finally:
+                # Clean up log file handle if still open
+                if log_file_handle and not log_file_handle.closed:
+                    log_file_handle.close()
+
+        except Exception as e:
+            self.logger.error(f"Failed to start daemon via subprocess: {e}")
+            return False
+
     def daemonize(self) -> bool:
         """Daemonize the current process.
 
@@ -495,12 +650,14 @@ class DaemonManager:
             True if successful (in parent), doesn't return in child
         """
         # Guard against re-entrant execution after fork
-        if hasattr(self, '_forking_in_progress'):
-            self.logger.error("CRITICAL: Detected re-entrant daemonize call after fork!")
+        if hasattr(self, "_forking_in_progress"):
+            self.logger.error(
+                "CRITICAL: Detected re-entrant daemonize call after fork!"
+            )
             return False
-            
+
         self._forking_in_progress = True
-        
+
         try:
             # Clean up asyncio event loops before forking
             self._cleanup_event_loops()
@@ -787,7 +944,7 @@ class DaemonManager:
                     f.write("success")
                     f.flush()  # Ensure it's written immediately
                     os.fsync(f.fileno())  # Force write to disk
-            except Exception as e:
+            except Exception:
                 # Logging might not work in daemon process after fork
                 pass
 

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm/services/monitor/server.py

@@ -239,7 +239,11 @@ class UnifiedMonitorServer:
             except OSError as e:
                 # Port binding error - make sure it's reported clearly
                 # Check for common port binding errors
-                if "Address already in use" in str(e) or "[Errno 48]" in str(e) or "[Errno 98]" in str(e):
+                if (
+                    "Address already in use" in str(e)
+                    or "[Errno 48]" in str(e)
+                    or "[Errno 98]" in str(e)
+                ):
                     error_msg = f"Port {self.port} is already in use. Another process may be using this port."
                     self.logger.error(error_msg)
                     self.startup_error = error_msg

{claude_mpm-4.2.38 → claude_mpm-4.2.40}/src/claude_mpm.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: claude-mpm
-Version: 4.2.38
+Version: 4.2.40
 Summary: Claude Multi-Agent Project Manager - Orchestrate Claude with agent delegation and ticket tracking
 Author-email: Bob Matsuoka <bob@matsuoka.com>
 Maintainer: Claude MPM Team

claude_mpm-4.2.38/BUILD_NUMBER

@@ -1 +0,0 @@
-360

claude_mpm-4.2.38/VERSION

@@ -1 +0,0 @@
-4.2.38

claude_mpm-4.2.38/src/claude_mpm/VERSION

@@ -1 +0,0 @@
-4.2.38