I think a better short-term fix, if we want to put one in, is to wait for the round trip reply in dbus-send. That is correct anyway even once we fix this underlying bug, and will ensure dbus-send returns a failure code if the method call fails. Havoc