Subject: The case for importing XFree86 4.4.0 RC1
To: None <tech-x11@NetBSD.org>
From: Frederick Bruckman <fredb@immanent.net>
List: tech-x11
Date: 12/13/2003 10:09:58
I suggest now that 1.6.2 is basicly in the can, and there's talk about
branching for NetBSD 2.0, we should talk about importing the latest RC
for 4.4.0 from XFree86.

4.4.0 is significant mainly because of the new IPv6 support, not
only in the Xserver and programs, but in the protocol -- XDM auth
is extended to handle 128-bit addresses. I tested the first cut from
X.org's web site early this summer, and there were only minor issues
with building, and now as of last month, 4.4.0 builds and runs fine.
There seem to be no unrelated changes of any significance. I think
we really want this in NetBSD 2.0.

There was an issue with using old xdm with new Xservers (or
vice-versa, I forget) with the X.org patches. If that's still present
in XFree86, it could be an issue for "gdm" or "kdm". In my opinion,
that would be a reason to import now, to give time to shake out the
bugs, not a reason to wait.

The danger of inaction is that our timing will be off, and no one
will want to risk XFree86 4.4.0 going in just as NetBSD 2.0 is nearing
cut-off for release. That would be a shame. If, on the other hand, the
RC goes in soon, it'll be well tested by that time. Patches from RC to
final should be small, so if 4.4.0 final happens to be released just
as NetBSD 2.0 is in the can, the pull-up to 4.4.0 final should not be
an obstacle to the 2.0 release.

Frederick