X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f
X-Recipient: djgpp-workers AT delorie DOT com
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type:content-transfer-encoding;
        bh=Pp4fPpgfVqC9a4ExirYVP0HVAGTZGKETGTMEei0+HNQ=;
        b=FdBVuFxjuHtFFjIkUpixdoWjsznApHQKQzg0HyhRXTgaIcScaCKAiRPS7FobgmqIsD
         wbkxVSKN+uJQz0FmMNI0vc0DHSJ9anM3dJsCqK1wvOmErCVHC8K2D/YcFgrLmU9MnY9t
         iXUFpL6LzoEfKtyzlDnMYgqhdWRJXvAexnI0c=
MIME-Version: 1.0
In-Reply-To: <E1QuzqW-0000sK-FT@fencepost.gnu.org>
References: <CAA2C=vCABbghMEVV2Hss+eXcmkFxuHA+dE6+Nc+-yzw-cHv1YA AT mail DOT gmail DOT com>
	<CAA2C=vBLXmFb9Pu7nmivBhiOGs=VghZ8f90CvcAk+ijV3+Cq+A AT mail DOT gmail DOT com>
	<CAA2C=vAjAgCxQqu1bJGE05Rbjd+6b7z_2xKZ=YHYfsKr5uKptQ AT mail DOT gmail DOT com>
	<83ty9c453p DOT fsf AT gnu DOT org>
	<CAA-ihx9cK=hixLEsdRO1W+1KwP0EyXa1WisanS8aPbsRkt0a1Q AT mail DOT gmail DOT com>
	<E1QuzqW-0000sK-FT AT fencepost DOT gnu DOT org>
Date: Tue, 23 Aug 2011 03:03:21 -0500
Message-ID: <CAA-ihx9Cax5LqfzT1+EjDybw2pPyYUPe9FBiqOz364URwzhs=g@mail.gmail.com>
Subject: Re: [PATCH] fix cabs and cabsf build failure with gcc-4.4, test results
From: Rugxulo <rugxulo AT gmail DOT com>
To: djgpp-workers AT delorie DOT com
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from quoted-printable to 8bit by delorie.com id p7N83TWk013879
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

Hi,

On Sat, Aug 20, 2011 at 11:41 PM, Eli Zaretskii <eliz AT gnu DOT org> wrote:
>>
>>    I'm the worst person to comment on this, but ... if __STDC__ means
>> ANSI/ISO C89/C90/C99, then lacking that, K&R C won't have _Complex
>> (C99) keyword anyways.
>
> _Complex is a GCC extension, so it's always available.

Okay, I didn't know that.

> I don't know if someone can define _Complex in another compiler

But I'm fairly?? certain that most other compilers aren't trying to
use (or esp. rebuild) DJGPP's libc. Okay, so I did encounter Detlef
Reimer's "build" of LCC 4.2 for DOS recently, and he does use DJGPP
libc (2.01, heh) and heavily modified headers. I think it's a pure
ANSI C89 compiler, so I doubt it supports _Complex. Honestly, I would
be surprised if anybody only supported _Complex but not the more
useful, common stuff from C99.

Anyways, like I said, LCC is ANSI (C89) and was only intended to use,
not rebuild, DJGPP's libc, hence it doesn't require old-style K&R nor
_Complex.

> or undef __STDC__
> with GCC (with some std= switch), but I don't like gratuitous changes
> in someone else's code.

I don't like changes either, esp. ones that aren't tested or could
break. You've got a point there. And clearly I defer to your massive
experience.

I dug up GCC 2.95.3 info files, which (to me) is oldest still relevant
GCC. I didn't test _Complex, but it did have very preliminary "c9x"
[sic] support. From what it says, __STDC__ is always defined unless
you use "-traditional".

Does newer GCC even support "-traditional"? I didn't think so, but I
can't remember without double-checking. Hmmm, latest docs online say
this:

"GCC versions 3.2 and later only support traditional mode semantics in
the preprocessor, and not in the compiler front ends."

In short, I know I'm overthinking this, and it's not that crucial,
just saying ... I'm all for compatibility and checks and being
conservative and leaving code "just in case", but I don't think it's a
realistic case here in particular.