On Mon, 2006-06-05 at 12:48 -0500, Michael D Schleif wrote:
> Did you really want to correspond OUTSIDE of the bugreport?
My reply was CC'ed to the BTS. Yours was not. Something weird in your
email client, perhaps?
> What do you think?
I haven't been actively involved in the NTP packaging for a while, now,
and a lot has apparently changed since I last contributed. Either I'm
missing something, or I actually see what I think are several problems
in the current version of the maintainer scripts.
1) I don't see that the daemon is explicitly stopped during a package
remove or purge. That means we could still have a running daemon when
you try to reinstall the package, which could perhaps lead to two copies
of the daemon running. I don't think that's your current problem, but
it feels broken regardless.
2) The ntp-simple postinst always tries to use 'restart', which will
fail on a new install because...
3) The debian/ntp-server.init.d script's restart clause (which is
installed by the ntp-simple package) tries to do
start-stop-daemon --stop --quiet --pidfile /var/run/ntpd.pid
which will fail if there is no running process. At the minimum, this
needs an --oknodo if 'restart' is always going to be used even when
'start' would be more appropriate. However, combined with the lack of
an explicit stop in any of the pre/postrm scripts, I worry that this
could lead to the multiple copies of the daemon running problem if the
pid file is out of sync for some reason.
Perhaps one of the folks listed in the changelog since my last
contribution would care to comment on why the current maintainer script
structure makes sense to them... since it doesn't to me.
Bdale
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
