Date: Mon, 25 Sep 2000 20:32:29 +0200
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: djgpp-workers AT delorie DOT com
Message-Id: <9743-Mon25Sep2000203228+0300-eliz@is.elta.co.il>
X-Mailer: Emacs 20.6 (via feedmail 8.2.emacs20_6 I) and Blat ver 1.8.5h
CC: ams AT ludd DOT luth DOT se
In-reply-to: <20000924003748.8850823139@titan.progiciels-bpi.ca> (message from
	Eli Zaretskii on Sat, 23 Sep 2000 20:37:48 -0400 (EDT))
Subject: Re: (fwd) startup-code
References: <200009231730 DOT TAA27218 AT father DOT ludd DOT luth DOT se> <20000924003748 DOT 8850823139 AT titan DOT progiciels-bpi DOT ca>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
> Date: Sat, 23 Sep 2000 20:37:48 -0400 (EDT)
> 
> > From: Martin Str|mberg <ams AT ludd DOT luth DOT se>
> > Date: Sat, 23 Sep 2000 19:30:43 +0200 (MET DST)
> > > 
> > > Please hold it.  I think Emacs uses one of those symbols, but I don't
> > > have any good way of checking it where I'm writing this.
> > 
> > Ok - no problem. If I manage to build emacs with the patch, it'll be
> > ok from emacs' perspective, right?
> 
> No ;-)
> 
> The way Emacs uses this symbol is rather tricky.  I don't remember the
> details and don't have Emacs sources handy to check (in fact, I'm not
> even sure Emacs uses _end, it might be some other related symbol)

Emacs uses the symbols (not C variables!) `_etext' and `start' (the
latter via inline asm).

The symbol `_etext' is also used in mcount.c, so removing it will
break profiling.

I'd suggest to search the mail archives of djgpp-workers and djgpp for
those symbols (I couldn't do this now because I have some strange
connectivity problems with www.delorie.com).  I think you will find
more software that uses some of them.

Frankly, the more I think about this change, the less I like it.  Why
do we need to change this?

Neither the DJGPP port of Emacs nor mcount.c invented the use of these
variables: they are evidently used by other Unix platforms, and that's
how we came about using them as well.  They were in DJGPP since day
one (well, at least since v1.11, the first version I ever used), and
AFAIR this is the first time someone complained (the mail-archives
search should tell if I'm wrong).

On balance, I think we should leave these symbols alone.