[Please wrap your lines at 72 characters!]
> Given the comments on my proposal, I'm rewriting it to the following:
>
> 1. to provide support for separation of local initialization
> scripts, allow update-rc.d to handle subdirectories. So, a
> /etc/init.d/local (or whatever) dir could be created and a script
> within this dir could be linked to rc?.d with 'update-rc.d
> local/99ascript defaults'.
Just a small question: have you actually tried this method and
compared it to the alternative of just using the /etc/init.d and
/etc/rc?.d directories? What significant improvements does your
proposal provide? It's not good to create new policy without having
some experience of the impact or problems caused (as recent events
have shown).
> 2. to let regular users have init scripts, allow ~/.rc.d directories
> (to be run after all the system initialization for the active users
> in the passwd database)
No way. This could potentially be a huge security problem (and this
is why some sites disallow user cron jobs). Why should a user need
personal stuff run at init time? They should wait until they log on.
And if it's a system user, such as ftp or the like, they can have a
job in /etc/init.d.
Julian
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, QMW, Univ. of London. [EMAIL PROTECTED]
Debian GNU/Linux Developer, see http://www.debian.org/~jdg
