The `npt-svwaitdown` program

npt-svwaitdown waits for supervised services to be brought down. It works only with npt-supervise.

Interface

     npt-svwaitdown timeout servicedir [ servicedir ...  ]

npt-svwaitdown blocks until all the supervised services defined by the servicedirs are down. Then it exits 0.
If a hard error occurs, npt-svwaitdown exits 111.
If timeout seconds elapse and npt-svwaitdown is still waiting, it exits 99. A 0 value for timeout means infinite.
If npt-supervise is not running on a servicedir, npt-svwaitdown considers that the service is down.

The goal of npt-svwaitdown is to provide a svwaitdown implementation that does not perform active waiting.

Actually, that was the starting point of all the pipe-tools business. :-)