Crash-restart blocks UI thread for up to 60 seconds

[kKaskak]

server.rs:187-199 calls recv_timeout(60s) from start(), which is called from process_event on the GUI message-pump thread during crash restart. The entire NWG UI is frozen until the new server emits readiness or 60s elapses. Should move the wait off the UI thread and navigate the WebView via a Notice callback.

[kKaskak]

Verdict: valid.

Current start() is even more blocking than the issue describes: it calls rx.recv().unwrap() with no timeout. Because the crash notice handler calls self.start() on the NWG event thread, the UI can freeze indefinitely until the new runtime emits the readiness line. Restart should happen off the UI thread and notify the WebView after readiness.