Philip.Tait@phxase.allied.com (Philip Tait) writes:

 >In article <199707180328.NAA14642@lux.sprc.qut.edu.au>,
 >        gmcgarry@lux.sprc.qut.edu.au (Gregory McGarry) writes:

 >> Has anyone started on an installation script for pmax?  I am
 >> willing to work on it over the weekend.  No hardware to hassle
 >> me.

 >Wouldn't the OpenBSD/pmax installation script be a good start?  I don't
 >think there are any restrictions on stealing from it.


It would, but I think starting with the ``standard'' NetBSD
installation scripts would be better still.  The only reason not to
use them as-is is that they're written to install from a miniroot
running either out of swap or out of a RAMdisk.

Neither option is viable for NetBSD/pmax for a couple of reasons:
	(1) Too many PROM versions only boot out of the 'a' partition,
	    which makes miniroot-in-swap kind of hard;

	(2)  the standard tools  to crunch binaries together into
	     one statically-linked binary don't work with mips ELF.

I think the required changes to the standard NetBSD install scripts
boil down to *not* trashing and newfs'ing the target root (since we
might be running out of it) and changing the explicit mountpoints
(e.g,. /tmp/ for the newly-installed root) to use a variable.

That variable would be /tmp/ for normal miniroot installs, and the
emtpy string or "/" when we're installing over the currently-running
root.