🚀 go-pugleaf

In article <MC#^-?@uzi-9mm.fulcrum.bt.co.uk>,
    igb@fulcrum.bt.co.uk (Ian G Batten) writes:

>I've just ordered a 386 machine from Dell to run Unix V.4.

I have an ICL DRS6000 here, running V.4 on a SPARC architecture.  I'm
expecting a 486-based machine running the same, RSN.

>working on Emacs, gcc or gdb for such beasts?  I've ported gdb and emacs
>before, so I'll start on it, but I suspect other people will have done
>some work.  I assume that for emacs there should be almost total
>commonality with the Commodore offering.  What about gdb?  gcc?

After some hacking, I have emacs running on the beast, but I haven't
attempted to produce a real xemacs (I defined CANNOT_UNEXEC and CANNOT_DUMP),
and I haven't finished testing it yet, but it seems to be going well.  It's
a hell of a lot faster than the CCI 6/32 I was using before, so I can live
with not having a dumped emacs (I never got the 6/32 to dump either).

I'm not sure how to go about modifying unexec.c and crt0.c to work with
V.4.  First of all, COFF has been replaced by ELF, secondly it might be
better to use some form of shared object as supported by ELF, rather than
just dumping the whole thing back out, and thirdly, I'm not sure I really
understand how unexec works, anyway.

I can at least put up a working s-usg5-4.h, and a list of changes I've made
to other files (mostly to #include statements).  Beware the cpp on the
machine.  On mine it had the distressing habit of inserting tabs in the
xmakefile in front of every token (/usr/ucbinclude became /usr /ucbinclude,
and it broke horribly when evaluating UNEXEC as a target).

>ian

Good Luck,

Mike
-- 
Mike Benson			+---------------+	   mikb@stcns3.stc.OZ.AU
Alcatel-STC Pty Ltd		|    CAVEAT	|
11th Floor, 5 Blue St		|     LECTOR	|   Any clod can have the facts,
North Sydney NSW 2060 AUSTRALIA +---------------+ but having opinions is an art.