in updateDevices there was a race condition with the sand callbacks
when the last host entry was not successfully returned - then the
callback was not called and so the timeout was not set for the next
cycle … means that the adapter did not checked anything anymore
made sure by using a pollingTimer variable that only one timeout for
the next cycle is running. there were cases where multiple timeouts
were running (high load)
change some small coding style things on jslint advice
commented out dead code (return before code, so was never called at
all)
added some more error logging to get infos on errors to show error transparent in the log because they should be fixed, or ?!
also start very first „updateAll“ call by a timeout because I had
cases where it was started multiple times initially