Subject: Re: NetBSD/Alpha in 1.1 release?
To: John Birrell <cimaxp1!>
From: Chris G Demetriou <Chris_G_Demetriou@BALVENIE.PDL.CS.CMU.EDU>
List: port-alpha
Date: 09/27/1995 09:20:20
> What NetBSD/Alpha stuff should we expect to see in the 1.1 release?

There will be no NetBSD/Alpha binaries _at all_ for the 1.1 release.

The alpha port isn't currently at the stage where it's sufficiently
integrated into the master source tree to even consider building the
1.1 sources into NetBSD/Alpha binaries.

I'll probably make a snapshot of the 1.1 release sources, integrated
with my own, however.  It won't be called "1.1" though.

> Any new source for the devo tree?

Yes, i've received a new set of these sources, but have been hesitant
to release them to the world until i iron out, or at leastlook deeply
into, a few problems.

One of them was my own fault.  (i'd biffed the return value in the
mem{cpy,move}() wrappers i'd put around Trevor Blackwell's assembly
bcopy()...)  That's fixed now.

It appears that the optimizer isn't _quite_ safe in the new gcc i've
got.  It appears to miscompile some X11 client code.  (i saw one
case where i printed one variable, another, and (the first < the
second).  both variables had the value "7," both were integer
variables, and it said that the first _was_ less than the second!  8-)
I've not yet looked into this.  It _does_ appear to compile the rest
of the sources reasonably, though.

some news:

(1) i went out on a limb, and decided to make g++ work.  It appears to
    work well enough to compile groff.

(2) groff compiles and (dig this!) RUNS!  8-)

(3) I've been whacking some more on X.  (I changed my source base to
    be the XF86 3.1.2 sources, which are R6pl12 + more fixes.  I did
    this mostly because the XF86 folks actually call NetBSD a
    "supported" platform, which means that stuff like the sys_errlist
    gunk has been fixed...)  I whacked things to the point of making
    the Xvfb (virtual frame buffer) server work, so it should be
    possible to get an X server going once the appropriate code is in
    the kernel to deal with mapping video memory, dealing with
    keyboard/mouse events, etc.  I'm working on this now.

I hope to get a snapshot out sometime in the next week or so.
(yeah, i know i've been saying that for a while, but now that i've
fixed my memcpy/memmove bug, i think i'll actually do it...  8-)