To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Cc: Andris Pavenis <pavenis AT lanet DOT lv>,
        "Gurunandan R. Bhat" <grbhat AT unigoa DOT ernet DOT in>,
        Alexey Yakovlev <jack AT catalysis DOT nsk DOT su>, djgpp-workers AT delorie DOT com,
        fortran AT gnu DOT org
Subject: Re: Inconsistencies between g77 v0.5.23 and v0.5.19
References: <Pine DOT SUN DOT 3 DOT 91 DOT 980714100626 DOT 6831A-100000 AT is>
From: Dave Love <d DOT love AT dl DOT ac DOT uk>
Date: 14 Jul 1998 19:18:46 +0100
In-Reply-To: Eli Zaretskii's message of "Tue, 14 Jul 1998 10:06:55 +0300 (IDT)"
Message-ID: <rzqoguse1o9.fsf@djlvig.dl.ac.uk>
Lines: 41
Precedence: bulk

>>>>> "Eli" == Eli Zaretskii <eliz AT is DOT elta DOT co DOT il> writes:

 >> I'm not sure that's wise.  I'd rather special-case DJGPP and
 >> similar systems.

 Eli> I would be glad to know why do you feel that way.

Because they are the special cases.

 Eli> If you mean the non-ANSI libc on these systems, then they are
 Eli> the minority these days.  So if you need at all to test for
 Eli> system-specific definition (as opposed to something tested by
 Eli> the configure script), I'd suggest to test for those special
 Eli> systems which cannot live with "b" as ANSI requires.

I wouldn't want to start breaking Unix systems where it's worked fine
for years that I don't even know about (though I've actually been
interested personally in more than one where this was an issue).  It's
not ultimately my descision but I did originate this stuff, do know a
fair amount about the issues and will argue with Burley if necessary.
I'm surprised to be arguing at all :-(.

If you get to build it native with configure && make (like the
windows32 systems) I just don't see the problem.

 >> I certainly failed with the only autoconfed (non-GNU) program I
 >> can remember trying under DJGPP.  I'd expect various things to
 >> break because of filenames, if nothing else.

 Eli> Only for files like config.h.in etc., and even that can be taken
 Eli> care of by using the output:input feature.

I think I even changed AC_OUTPUT once (perhaps when that feature first
appeared in autoconf), but I don't think that's the only relevant file
name, and we might as well sort things out properly for a complete
distribution.  This will presumably need help from interested DJGPP
people to make it happen in a reasonable way.

I did actually try unpacking and configuring the egcs core
distribution (not even including g77) with the current DJGPP, and that
certainly doesn't just work.