Subject: Re: should not ask for filesets first
From: David Laight <firstname.lastname@example.org>
Date: 07/03/2003 18:23:58
> Really? Is there some new way of determining, without accessing the
> network and downloading and unpacking the file, what the resulting
> size of the set will be? And a need to tightly fit a partition to
> the size of a particular set, without room for future expansion?
> Or are you just ignoring the recommendations you sought over in the
> current-users list about the partition layout?
> Folks seemed pretty sure that what they wanted is usage-based, not
> set-based, partition organization. That's why all of the responses
> said there was no need for the base+X menu choice.
I moved the 'set' question before the disk partitioning, so that the
disk partioning could depend on the installed sets.
I have not done anything about working out the sizes/availability
of the sets themselves. This could be done by determining the sizes
of the sets on the system where sysinst was built (under the assumption
that they won't change much), or (and probably better) by moving the
questions about where to locate the sets right to the front?
> > Maybe it would make sense to locate the sets right at the start.
> With no network connection?
> What I'd *LIKE* to see, is the preparation phase finished, and then
> ask where to locate the sets, and then list the sets available at
> that site, and select from that list.
Then you get a chiken/egg situation, you can't size the filesystems
until you know which sets to install. ....
The expectation is that all the known sets are available on the install
> > > (And doesn't come close to matching INSTALL.txt.)
> > The documentation will be way out of date!
> > and getting further away :-)
> Woah, Nelly! Are you saying you are writing code without updating
> the documentation? How does anybody know what's going on? Heck,
> how do you know whether what you are doing is easy to explain to
> everybody else, until you've carefully described the process?
I haven't finished the code yet!
(I'm getting around to the final testing of the MBR table editor,
however I'm suffering from a caffiene overdose....)
> > > Where in the code should I look for these changes?
> > Somewhere under my rototiller, but mind your fingers.
> That's a very unfriendly remark. I was looking for a file reference.
All of distrib/utils/sysinst! particularly label.c, bsddisklabel.c
and (soon) mbr.c
David Laight: email@example.com