Re: Anyone working on the rc.d installer problem?
From: "H. Peter Anvin" <hpa@transmeta.com>
Date: Wed, 4 Nov 1998 20:21:25 -0800 (PST)
> True, but the way it is done isn't particularly abstract. It uses pure
> numbers for everything, instead of abstract names of the location points
> and the run levels.
... which is bad, because they're not constant. Note there needs to
be multiple things specifiable in the "start" and "stop" sections,
because the order may be different on different systems.
I think we need to standardize things, or at least use abstract enough
start and stop points so that this isn't an issue.
The basic idea here is that if you need something which requires
networking and not much else, then you might put in "start = after network"
If it requires that the remote filesystems be mounted, you should be
able to say "start = after remotefs", and so on.
What we *don't* want to force the ISV to do is to research the boot
script order for every single Linux distribution, and configure a
separate boot script start line for each different Linux distribution.
This way lies madness....
- Ted
Reply to: