John Fischer wrote: > Instead, JDS in Nevada uses postrun for running these programs. > Postrun generalizes the above idea by putting the complex > script into a shared script (postrun itself) and using a > single smf service for all first-boot commands.
When can consumers of postrun expect their scripts to be run at first boot? My question hints at a potential limitation of postrun for packages that expect their changes to have taken effect very early in boot (earlier than the postrun SMF service is started). It may be worthwhile to call out this limitation to set appropriate expectations. -Seb
