Am 18.07.2010 22:25, schrieb Petter Reinholdtsen:
[Ralph Ulrich]
A trigger (like update-initramfs) could have alternative
implementations for all different init systems also. This way poor
Debian maintainers soul should be bothered less to handle init
scripts: only need to provide a valid LSB-header. Users would get
caught by less bugs then?
This can be triggered using update-rc.d. For example a call like this
will cause a reordering based on dependencies:
update-rc.d /etc/init.d/bootlogs defaults
Is update-rc.d really usable as a trigger like kernel related
"update-initramfs" is used:
- deferrable
- all of the previous configuration obsoleted and removed
Why did I saw this debootstrap problem: Debootstrap failed because of a
service couldn't install without a yet installed Required-Start. The use
of a deferred trigger would have prevented this bug. The solution was of
manual kind.
Why do I see this many not fitting LSB-header bugs: If every
init.d/SCRIPT is NOT a conffile it would possible to remove the script
also when upgrading. A deferred trigger would issue insserv to run for a
full removal of all remaining confs. Following a fresh install of the
new version of the package. This would solve and prevent all bugs where
the solution found just was to cleanly purge the package before
reinstall of the upgrade version. For example dkms is used like this as
a trigger of not standard modules of the kernel.
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
