Re: Sysinst GSOC project

[Tim Rightnour <root%garbled.net@localhost>]

Sorry I didn't reply to this earlier, I thought I was subscribed to
tech-install, but it appears I was not.

>  One thing I'm not sure of is whether to have the front end and back
> end in the same executable or not. Doing so would allow easy switching
> between the two when errors arise. Having them separate would allow
> modularity though. I think that the best option would be to have them
> be separate, and make sysinst itself just a shell script that runs
> both. That way, even the shell script is configurable. It might run
> the back end in "check" mode to see if it can fully install, and then
> run the front end it it can't.

Well.. my initial thought was something like cutting sysinst into "preinst" and
"postinst". (Or whatever we call them).  Then "sysinst" would be:

#!/bin/sh
preinst -o file
postinst -f file

So instead, you could just do:
postinst -f myconf.conf

Ideally, in the future, it would be nice if we could pull the config files from
dhcp/tftp, but I think we can worry about that later.

As for the config file format:

I agree with what others have said about using | instead of / for the option
separator.

I think your first line of action should be to do a few installs, and write
down all the questions sysinst asks you, and translate those into your config
file.  You may find that there are certain lines of reasoning that open up
different paths, where options that didn't show up before, now show up.

One such thing would be if the user does, or does not elect to use the whole
disk, or needs to edit the fdisk information.

---
Tim Rightnour <root%garbled.net@localhost>
NetBSD: Free multi-architecture OS http://www.netbsd.org/
Genecys: Open Source 3D MMORPG: http://www.genecys.org/