fix: eliminate 5+ second delays from stale daemon.lock files (bd-htwx)

Root cause: When a bd daemon crashes, its daemon.lock file remains with
the old PID. If that PID gets reused by an unrelated process, the code
would wait 5 seconds for a socket that will never appear.

Fix: Use flock-based check as authoritative source for daemon liveness.
The OS releases flocks when a process dies, so this is immune to PID reuse.

Changes:
- handleExistingSocket: Check daemon flock before waiting for socket
- acquireStartLock: Verify daemon lock is held before waiting
- handleStaleLock: Use flock check to detect stale startlocks
- lockfile/process_*.go: Add pid <= 0 check to prevent false positives
  (PID 0 signals process group on Unix, not a specific process)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

cmd/bd/daemon_autostart.go

@@ -11,6 +11,7 @@ import (
 
 	"github.com/steveyegge/beads/internal/config"
 	"github.com/steveyegge/beads/internal/debug"
+	"github.com/steveyegge/beads/internal/lockfile"
 	"github.com/steveyegge/beads/internal/rpc"
 	"github.com/steveyegge/beads/internal/ui"
 )
@@ -217,7 +218,7 @@ func acquireStartLock(lockPath, socketPath string) bool {
 	// nolint:gosec // G304: lockPath is derived from secure beads directory
 	lockFile, err := os.OpenFile(lockPath, os.O_CREATE|os.O_EXCL|os.O_WRONLY, 0600)
 	if err != nil {
-		// Lock file exists - check if it's from a dead process (stale) or alive daemon
+		// Lock file exists - check if daemon is actually starting
 		lockPID, pidErr := readPIDFromFile(lockPath)
 		if pidErr != nil || !isPIDAlive(lockPID) {
 			// Stale lock from crashed process - clean up immediately (avoids 5s wait)
@@ -227,7 +228,17 @@ func acquireStartLock(lockPath, socketPath string) bool {
 			return acquireStartLock(lockPath, socketPath)
 		}
 
-		// PID is alive - daemon is legitimately starting, wait for socket to be ready
+		// PID is alive - but is daemon actually running/starting?
+		// Use flock-based check as authoritative source (immune to PID reuse)
+		beadsDir := filepath.Dir(socketPath)
+		if running, _ := lockfile.TryDaemonLock(beadsDir); !running {
+			// Daemon lock not held - the start attempt failed or process was reused
+			debugLog("startlock PID %d alive but daemon lock not held, cleaning up", lockPID)
+			_ = os.Remove(lockPath)
+			return acquireStartLock(lockPath, socketPath)
+		}
+
+		// Daemon lock is held - daemon is legitimately starting, wait for socket
 		debugLog("another process (PID %d) is starting daemon, waiting for readiness", lockPID)
 		if waitForSocketReadiness(socketPath, 5*time.Second) {
 			return true
@@ -242,11 +253,24 @@ func acquireStartLock(lockPath, socketPath string) bool {
 
 func handleStaleLock(lockPath, socketPath string) bool {
 	lockPID, err := readPIDFromFile(lockPath)
-	if err == nil && !isPIDAlive(lockPID) {
-		debugLog("lock is stale (PID %d dead), removing and retrying", lockPID)
+
+	// Check if PID is dead
+	if err != nil || !isPIDAlive(lockPID) {
+		debugLog("lock is stale (PID %d dead or unreadable), removing and retrying", lockPID)
 		_ = os.Remove(lockPath)
 		return tryAutoStartDaemon(socketPath)
 	}
+
+	// PID is alive - but check daemon lock as authoritative source (immune to PID reuse)
+	beadsDir := filepath.Dir(socketPath)
+	if running, _ := lockfile.TryDaemonLock(beadsDir); !running {
+		debugLog("lock PID %d alive but daemon lock not held, removing and retrying", lockPID)
+		_ = os.Remove(lockPath)
+		return tryAutoStartDaemon(socketPath)
+	}
+
+	// Daemon lock is held - daemon is genuinely running but socket isn't ready
+	// This shouldn't happen normally, but don't clean up a legitimate lock
 	return false
 }
 
@@ -260,19 +284,24 @@ func handleExistingSocket(socketPath string) bool {
 		return true
 	}
 
-	pidFile := getPIDFileForSocket(socketPath)
-	if pidFile != "" {
-		if pid, err := readPIDFromFile(pidFile); err == nil && isPIDAlive(pid) {
-			debugLog("daemon PID %d alive, waiting for socket", pid)
-			return waitForSocketReadiness(socketPath, 5*time.Second)
-		}
+	// Use flock-based check as authoritative source (immune to PID reuse)
+	// If daemon lock is not held, daemon is definitely dead regardless of PID file
+	beadsDir := filepath.Dir(socketPath)
+	if running, pid := lockfile.TryDaemonLock(beadsDir); running {
+		debugLog("daemon lock held (PID %d), waiting for socket", pid)
+		return waitForSocketReadiness(socketPath, 5*time.Second)
 	}
 
-	debugLog("socket is stale, cleaning up")
+	// Lock not held - daemon is dead, clean up stale artifacts
+	debugLog("socket is stale (daemon lock not held), cleaning up")
 	_ = os.Remove(socketPath) // Best-effort cleanup, file may not exist
+	pidFile := getPIDFileForSocket(socketPath)
 	if pidFile != "" {
 		_ = os.Remove(pidFile) // Best-effort cleanup, file may not exist
 	}
+	// Also clean up daemon.lock file (contains stale metadata)
+	lockFile := filepath.Join(beadsDir, "daemon.lock")
+	_ = os.Remove(lockFile) // Best-effort cleanup
 	return false
 }