Date:	Wed, 19 May 1999 10:18:28 +0300 (IDT)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	"Mark E." <snowball3 AT bigfoot DOT com>
cc:	djgpp-workers AT delorie DOT com
Subject:	Re: wctype.h and ctype.h
In-Reply-To:	<199905181753.RAA20958@out5.ibm.net>
Message-ID:	<Pine.SUN.3.91.990519101753.14875A-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Tue, 18 May 1999, Mark E. wrote:

> > The question is: why did whoever created wctype.h need to include
> > ctype.ha in it?  Does Addendum 1 say something that implies that
> > including wctype.h pulls in the definitions of is* functions?  
> 
> I'm the guilty one. I had that include in so later I could use it
> for the isw* equivalents of the is* functions in ctype.h.

Why couldn't you simply include ctype.h in the same place where you
use its macros?

> But the include can be removed since none of this is implemented yet.

Will do.

Btw, I'm a bit uneasy about including incomplete headers in the
distribution.  The problem is that some packages which auto-configure
themselves test for a header and expect ALL of its standard functions
to be present if the header is available.  Granted, that's wrong (they
should test for the functions instead), but that's life.  We already
had such problems with the Unix part of sys/ioctl.h, and lately with
gnu-regex.c in the GDB distribution.

So, in general, I think we should have at least a trivial
implementation of all the functions described in the Addendum to be
added ASAP to wctype.h and wchar.h.  I don't think this *must* be done
before v2.03 is released, but I'd surely feel better if someone would
come up with some minimal code.

- Raw text -