fix: macOS block index corruption, dbcache auto-sizing, import key rescan height

- Shutdown: 3-phase stop (wait for RPC stop → SIGTERM → SIGKILL) prevents
  LevelDB flush interruption on macOS/APFS that caused full re-sync on restart
- dbcache: auto-detect RAM and set -dbcache to 12.5% (clamped 450-4096 MB)
  on macOS (sysctl), Linux (sysconf), and Windows (GlobalMemoryStatusEx)
- Import key: pass user-entered start height to z_importkey and trigger
  rescanblockchain from that height for t-key imports
- Bump version to 1.1.1

src/app.cpp

@@ -2124,15 +2124,19 @@ void App::stopEmbeddedDaemon()
     }
     
     if (stop_sent) {
-        DEBUG_LOGF("Waiting for daemon to begin shutdown...\n");
-        shutdown_status_ = "Waiting for daemon to begin shutdown...";
-        std::this_thread::sleep_for(std::chrono::milliseconds(500));
+        DEBUG_LOGF("Waiting for daemon to flush block index and shut down...\n");
+        shutdown_status_ = "Waiting for daemon to flush block index...";
+        // Give the daemon time to flush LevelDB to disk before we
+        // escalate to SIGTERM.  On macOS/APFS, LevelDB compaction +
+        // fsync can take 15-20s on a large chain.  The stop() method
+        // will wait this long for a *natural* exit (via the RPC stop
+        // we already sent) before falling back to SIGTERM.
     }
     
-    // Wait for process to exit; SIGTERM/TerminateProcess as last resort.
-    // 10 seconds is generous — if the daemon hasn't exited by then it's stuck.
     shutdown_status_ = "Waiting for dragonxd process to exit...";
-    embedded_daemon_->stop(10000);
+    // 20s grace period for the RPC "stop" to complete (LevelDB flush).
+    // Only after that does stop() escalate to SIGTERM, then SIGKILL.
+    embedded_daemon_->stop(20000);
 }
 
 bool App::isEmbeddedDaemonRunning() const