Re: Rationale for /etc/init.d/* being conffiles?
On Sat, 20 Dec 1997, Santiago Vila wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>
> On Sat, 20 Dec 1997, Scott K. Ellis wrote:
>
> > > The policy does not explain why they should *all* be conffiles.
> >
> > I can think of a reason to modify almost any /etc/init.d/* script, on the
> > grounds that they effect the startup behavior of the system.
>
> /sbin/init does also effect the startup behavior of the system and does
> not mean it have to be a conffile.
>
> > The object here is to make things consistant.
>
> There are several ways to be consistent. One of them is making all init.d
> script conffiles even when they do not need to be. Another of them
> is make conffiles the scripts that need to be conffiles and not make them
> when they do not need it. I don't see the first way is more consistent
> than the second way.
Then give me a concrete example of an init.d script that doesn't need to
be a conffile. I'm certain I can find a way to change it. You haven't
proven to me that there is a useful, obvious test for when a file should
or shouldn't be a conffile.
> > Why are you so opposed to them being conffiles anyway,
>
> I am NOT opposed to them being conffiles. I am opposed to them being
> conffiles without a rationale.
I've given the rational. If they're all conffiles, then you can't be
suprised when one of them isn't. Can I bill you for my lost time when one
of these "init.d scripts that isn't a conffile because Santiago didn't
think it needed to be" clobbers my local changes?
> > it doesn't make any difference to someone who never changes them.
>
> Let's make a conffile every script in /usr/bin, then.
There is an easy way to override those already, it's called /usr/local/bin
Reply to: