Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!zds-ux!gerry
From: gerry@zds-ux.UUCP (Gerry Gleason)
Newsgroups: comp.unix.i386
Subject: Re: Gnu Emacs on 386/ix Ver 2.0.2
Keywords: HOW ?
Message-ID: <82@zds-ux.UUCP>
Date: 5 Jan 90 17:58:52 GMT
References: <1445@fcs280s.ncifcrf.gov>
Reply-To: gerry@zds-ux.UUCP (Gerry Gleason)
Organization: Zenith Data Systems
Lines: 40

In article <1445@fcs280s.ncifcrf.gov> adam@ncifcrf.gov writes:

>I tried building emacs version 18.55 on 386/ix v 2.0.2, using
>s-usg5-3.h and m-intel386.h with no luck (compiler bombs in sysdep.c,
>not to mention that it can't seem to handle x11term.c at all). I know
>people have done it, so my question is...HOW?? Am I missing something ??

>Internet: adam@ncifcrf.gov			Adam W. Feigin

I've just been doing the same thing.  Same system and machine includes,
and found one major problem and another minor one.  First, it seems that
in XENIX there is a ioctl defined (termio.h I think) that conflicts with
one of theirs (TIOCGETC).  There is a #ifdef XENIX that undefines it, but
now these definitions are included with UNIX also.  The minor problem was
that cpp didn't seem to be able to handle a define name with a dollar
sign ($) in it.

This is the diff for the second minor problem:
74c74,75
< #ifndef RAB$C_BID
---
> /*#ifndef RAB$C_BID*/
> #ifndef RABC_BID /* for some reason, the $ confuses cpp */

And this for the first one which is a stopper:
130a132,134
> /* we need this because termio.h has this defined for Xenix compatibility */
> #undef TIOCGETC  /* Avoid confusing some conditionals that test this.  */
> 
453c457

And throughout I had to do this type of thing so cpp wouldn't complain:
< #endif HPUX
---
> #endif /* HPUX */

At present it's dumping core all over the place, so I'm about to recompile
fns.c without the optimizer as someone elso suggested.

Gerry Gleason