[1491] in NetBSD-Development
An experiment
daemon@ATHENA.MIT.EDU (Greg Hudson)
Tue Nov 11 15:43:22 1997
Date: Tue, 11 Nov 1997 15:41:19 -0500
From: Greg Hudson <ghudson@MIT.EDU>
To: sipb-athena@MIT.EDU, source-developers@MIT.EDU
(Some of the background material here may be a little unfamiliar to
non-SIPB people reading this mail. Bear with me.)
SIPB's NetBSD-Athena port has been rotting, to say the least, in large
part because I have the real Athena source tree to take care of and
because I did a bad job making sure that other people could take care
of its release engineering process.
My experience in Athena rel-eng has also exposed some weaknesses in
SIPB's processes for supporting Athena ports. Chief among them are
lack of code review and lack of good experience in keeping a divergent
source tree up to date with respect to the real Athena source tree.
In an attempt to make things generally better, I would like to see the
NetBSD 1.3-based Athena port involve some informal cooperation between
SIPB and IS. The steps I would like to see IS take are fairly modest:
* Make the athena hierarchy of the source tree work on NetBSD
and Linux, except for one or two of the very highly
platform-dependent packages (machtype and xdm come to mind).
There aren't a whole lot of changes to make, and I've
already done about two thirds of the work (or so I
estimate).
* Spend some time over the next month making the mainline
Athena tree a bit more ready for release. Nothing we
wouldn't be doing for 8.2 anyway.
* Possibly provide space in the dev cell for SIPB use. When
you have a 100MB+ source tree, a proper job of release
engineering involves a lot of disk space, and the sipb cell
is traditionally a bit tight while the dev cell, with many
fewer consumers, typically has vast reaches of unexplored
space.
* Possibly provide a bit of hardware. SIPB's office Intel
heads do not always constitute a big enough set of machines
to support release engineering on two platforms.
The incremental cost is small. The payoff for IS is that SIPB will
find bugs in the mainline release during the first quarter of 1998,
making the Athena 8.2 rleease more reliable.
SIPB, meanwhile, needs to decide what we want to do for the NetBSD
1.3-based release, who's going to do it, and how it ties into Linux.
There are a bunch of issues, which I don't have organized in my head
right now. I'd like to have at least one meeting to help hammer
things out. If you'd like to attend, please tell me whether you're
free tomorrow, Thursday, or Friday evening, and I will pick a time and
place.
