[822] in NetBSD-Development

home help back first fref pref prev next nref lref last post

Updated installation distribution

daemon@ATHENA.MIT.EDU (Greg Hudson)
Tue Jun 27 02:51:35 1995

Date: Tue, 27 Jun 1995 02:50:22 -0400
From: Greg Hudson <ghudson@MIT.EDU>
To: netbsd-dev@MIT.EDU


I have updated the -current installation distribution to match our
source tree.  Users who install now should get a libafs.o, ps, and w
that match the kernel, and should get the make and .mk changes to
allow building out of our source tree (specifically "make objdir").

To my knowledge, the route bug on startup still exists.  Also, Tom is
still having a problem where the script complains about being unable
to write to /tmp (although he could after he got a prompt back); I
don't know what would be causing that, and might have to go to Fenway
to figure it out.

I will do a limekiller test tomorrow.

Things left to do:

	* Fix the route bug.
	* Fix things so sendmail runs periodically
	* Make the Athena login use /etc/athena/shellcvt.
	* Come up with a good upgrade path for people who installed
	  1.0, which we can also use in the future.
	* Try to make X configuration easier, in the common case
	* Move dev/current-source/build to a non-backed-up volume
	* Think about how to set up a local printer at installation
	  time.  (This is getting into the realm of things people will
	  want to do after installation, which we don't have covered
	  yet.)
	* Modify the kernel so that you can access DOS partitions on
	  other disks.
	* Write a tool (to go in /sbin) to perform reverse-arp queries,
	  and use it during installation to try to find a default for
	  the machine IP address.

We are starting to get enough users, and particularly users of a
non-hacker nature, that I'm beginning to think we should instate a
real release cycle, where we would assign version numbers to revisions
of the source tree and Athena utilities, provide real update scripts,
etc..  This would consume a lot of disk space, so it will have to wait
until we get more space in the SIPB cell, but I know Yoav would prefer
this approach, and it would solve some of the headaches we've had.


home help back first fref pref prev next nref lref last post