On Wed, Jul 12, 2006 at 03:07:59PM +0800, Darren Reed wrote: > If that were to happen, which way should the compatibility > interface fall here? My guess would be to have it cause a > new contract for each successive fork, as does sshd.
Ironically, that would largely limit daemon()'s usefulness to those consumers we are trying to exterminate in Solaris; i.e. random commands that start persistent daemon processes. For things started from SMF, you typically either want the daemon to remain in the same contract so SMF can track it, or you want explicit knowledge of the creation and identity of contracts so that you yourself can monitor their health. Dave
