James Carlson wrote: > I still think that getting the translation between file-based and > SMF-based configuration right is going to be *really* hard. Hard > enough that perhaps we ought not do it.
I wasn't thinking of translation per se. More that cron needs to have internal structures that represent the jobs, and it would load those internal structures from two sources: legacy crontabs and new SMF-based entries. inetconv-like translation seems possible, but I don't think it'd be desirable. and in a later message: > If it's the same mechanism, then it's a read-only display for the > legacy bits with no controls. > > That's a bit different from the inetd model. (I don't think it's > wrong; just different.) Right. Consume the legacy crontab data, display it, offer little or no control. The better analogy is to /etc/rc*.d scripts.
