[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

question about initscripts



Hi,

please redirect me if this is the wrong place for this discussion.
I've been reading

http://www.debian.org/doc/debian-policy/ch-opersys.html#s-sysvinit

and involved in a bug discussion

https://launchpad.net/distros/ubuntu/+source/avahi/+bug/56426

about avahi-daemon's init script. It seems from a recent comment that
avahi-daemon might actually be a special case (in which case perhaps
it should not have an entry in /etc/init.d) so I'll phrase this as a
general question.

The document above says that "start" should "start the service" but it
seems there are quite a few debian packages where this does not work
because their initscripts contain things like

---
test -f /etc/default/fetchmail || exit 0
. /etc/default/fetchmail
if [ ! "x$START_DAEMON" = "xyes" ]; then
       echo "Edit /etc/default/fetchmail to start fetchmail"
       exit 0
fi
---

I can see no point to this, it is a redundant second knob that needs
to be turned in order to start the service. It leads to confusion (see
the many duplicate ubuntu bugs about avahi daemon), it causes users to
have to go read the initscripts, it prevents GUI tools like
services-admin from "just working" and it seems to be against the
spirit of the policy document - "start" starts the daemon and
"update-rc.d" enables or disables the service.

Do you consider what happens inside the initscript to be beyond the
scope of debian policy or was it your intention that "start the
service" would mean that the initscript should definitely attempt to
start the service?

Thanks,

Fergal



Reply to: