make server start/stop more stable: shutdown sockets properly, try up to 20x to bind the server socket in 5 sec intervalls on fail give spooler/backends ore timme to come up - On Solaris it takes about a minute 'til a unclean sockets gets reclaimed by the OS

*ctl: fix to fullfill service contracts: don't exit with 0 before the service is running (sync via pipes) Spooler: avoid waisting cpu cycles like 'while 1: sleep(N) || pass'; don't block on a single job; use Thread.Condition to wake up the queue runner instead of fixed sleep intervalls AbstractServer: wait for signal instead of waisting cpu cycles using while 1: sleep(0.1) for the whole runtime *.manifest: use :kill method for stoping backends - more reliable than backendctl ... stop

fix design flaw in backend: don't return exit code 0, if the backend can not be started successfully (Q&D done per pipe)

remove default values for *dir properties

fix stop method

new SMF related files