frysk.event.EventLoop has been modified so that it extends Thread, this makes
eventLoop = new EventLoop ();
(the async bugs are hopefully also fixed)
begs the question, who should catch a failing event loop?
perhaphs the thing to do is extend EventLoop (possible?) with a wrapper that
starts the thread, but then catches any throw notifying the GUI which can then
decide if restarting the event loop is worthwhile.
I tried this last night, and for some reason if I use the code as below I get no
updates from the Manager.host.requestRefreshXXX(). Using the old way
(Manager.eventloop.(...) and the updates come in fine.