Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 17 Sep 1996 16:01:55 -0600
From:      Warner Losh <imp@village.org>
To:        graphix@iastate.edu
Cc:        hackers@freebsd.org
Subject:   Re: OpenBSD <-> FreeBSD 
Message-ID:  <199609172201.QAA12496@rover.village.org>
In-Reply-To: Your message of Wed, 11 Sep 1996 19:51:02 CDT

next in thread | raw e-mail | index | archive | help
:   Is there anyone tracking the changes that are made to the 4.4 source
: in NetBSD/OpenBSD and incorporating the changes into the FreeBSD tree on
: a regular basis?  If not, what is the common feeling of if this would be

Having looked into this problem before I upgraded the size of my disk
drive collection....

The kernel sources generally would be hard to do.  There are some
parts of the kernel that would be easy to do this for, and other parts
that would be hard (the vm system comes to mind for the latter).  This
would be complex and would require more effort than it currently is
worth to make it happen.  Better to tilt at smaller windmills :-).

A smaller windmill might be userland.  It should be basically trivial,
but it likely isn't at this point.  It is a *LOT* of work.  There are
hacks on all three *BSD systems that need to be reconciled with the
4.4 Lite-2 changes and eachother.  OpenBSD, for example, has gone
"paranoid" in the use of sprintf due to potential buffer overflow
security holes.  These would be great to merge into FreeBSD, for
example.  I know that OpenBSD merges a lot of FreeBSD stuff into it
(the ports system, for example) and it would be good to narrow the gap
between the *BSDs.

A good first step would be to merge in as much of the differences in
the user level code as possible.  This would take someone a lot of
time, and then be a bear to keep in sync between the three.  OpenBSD
keeps in sync with NetBSD by a brute force approach that involves
running diffs, lots of patches and the watchful eyes of the core
team.  Its a lot of work for them (judging from the timestamps of
various messages and such).

However, let us say that FreeBSD merges every last change from NetBSD
and OpenBSD into its tree.  I would maintain that this would be a
wasted effort because NetBSD and OpenBSD might not pick them up and
we'd still have version skew.

Maybe what would work best would be for people in each of the commiter
groups to work towards ironing out the differences one at a time.
Take /usr/src/bin or something small and make all of those files the
same (or as close to it as is sane) and slowly integrated the changes
from the other trees.  Once there is convergence, then it would be
easier to keep things in sync.  Once there is a large part of the
world that is nearly identical on all the platforms, then it might not
be a bad idea to talk of having a Grand Unified Source Tree (GUST, you
heard it hear first :-).  Until then, I think it will remain a project
too large to ever make good headway.

In conclusion: Merging from other sources is a good idea.  Trying to
make all three share parts of their source tree is an idea whose time
has not yet arrived.

Warner



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199609172201.QAA12496>