Comment 1 for bug 111834

I'm attaching a patch which skips the early test for whether the daemon is running, and creates the DISABLE_TAG before trying to stop the daemon.

Note that there is still a race here: it is possible that avahi-daemon had actually begun to start before DISABLE_TAG was created, but has not reached the point where avahi-daemon -c has indicated that it has actually started. I took the lazy man's approach of putting a 2 second sleep after creating DISABLE_TAG. If avahi-daemon's own locking mechanism is sufficiently robust, this should not be an issue.