Liane Praza wrote: > It'd be good to have someone who's been in svccfg more recently than I > look over this too. (Bustos? Tom?) > > Liane Praza wrote: >> Alan Maguire wrote: >>> http://cr.opensolaris.org/~amaguire/svccfg_refresh/ >> >> I'll look at it this afternoon. > > (Alan updated this to do the svcadm refresh as we discussed too.) > > svccfg_help.c: > > nit. Not really an either/or. Maybe: "Update the Running snapshot > for the instance with the values from the current configuration. > If svccfg is not operating on an alternate repository, also inform > the restarter for the instance that the configuration has been > updated." > > Or, less accurate/less pedantic, but more helpful?: "Update the > Running snapshot for the instance with the values from the current > configuration. If svccfg is not operating on an alternate > repository, inform the restarter to invoke the instance's refresh > method." > i went with the latter, as it seems more user-centric. > svccfg_libscf.c > > l12335: I'd think calling lscf_prep_hndl() would be more appropriate > than checking g_hndl. Is there a reason you didn't? > no reason, i've changed this. > I found the switch working for both the refresh_entity() and > take_snap() returns a little weird. At the least, it makes the > bad_error() a lie. > i figured sharing the switch might make sense since the functions share most of the same return values, but i've split it into two separate switch statements now.
thanks Liane (and Tom!) for taking a look. i've refreshed the webrev, so let me know if there's anything else amiss. http://cr.opensolaris.org/~amaguire/svccfg_refresh/ alan
