From: Ian W. C. <ico...@re...> - 2006-11-06 07:37:00
|
On Nov 5, 2006, at 12:18 PM, Nigel Kersten wrote: > On 02/11/2006, at 8:54 AM, Jeremy Reichman wrote: >> On 10/31/06 10:15 PM, "Andrew Mortensen" <adm...@um...> wrote: >>> Unfortunately, ra.sh explicitly changes directory to the root of =20 >>> the boot disk at line 315 or so ("cd /"). Would you be willing to =20= >>> write a patch that would allow the administrator to specify which =20= >>> directory ra.sh should change to prior to running fsdiff? >> >> If it's not otherwise specified, I'm all in favor of leaving the =20 >> default behavior as-is, fwiw. > > Am I the only one who wants to be able to do this sort of thing? I don't have an immediate use for this functionality, but I'd support =20= its inclusion. Why not make ra.sh as flexible as the tools =20 themselves are? > Say we're booting from a NetInstall image, and want to radmind /=20 > Volumes/System. > > Should we use /Volumes/System for everything in such a situation? > > ie, say we've passed -D /Volumes/System for DIRECTORY, should we have: > > KFILE=3D${DIRECTORY}/var/radmind/client/command.K > DEFAULTS=3D${DIRECTORY}/etc/defaults/radmind > PREAPPLY=3D${DIRECTORY}/var/radmind/pretapply > POSTAPPLY=3D${DIRECTORY}/var/radmind/postapply > TMPDIR=3D${DIRECTORY}/tmp/.ra.$$ > > etc? I would vote no=97or rather, not by default. ra.sh, like the tools, =20 should by default always get its configuration files from /var/=20 radmind, no matter the comparison path. However it would be nice to =20 have a flag (say -D, to match radmind(8) and ktcheck(1)) that asked =20 the script to get its configuration from somewhere else. So to get =20 the behavior you describe, you could do: ra.sh -C /Volumes/System -D /Volumes/System/var/radmind (Taking, for the moment, the -C flag to provide the path from which =20 fsdiff is run.) =97IWC |