This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the EGCS project.
Re: va_list: headers
- To: Richard Henderson <rth@cygnus.com>
- Subject: Re: va_list: headers
- From: Zack Weinberg <zack@bitmover.com>
- Date: Mon, 26 Jul 1999 09:34:30 -0700
- cc: gcc-patches@gcc.gnu.org
Richard Henderson wrote:
> On Sun, Jul 25, 1999 at 11:15:42PM -0700, Zack Weinberg wrote:
> > * Can we have something along the lines of
> >
> > #if defined __STDC_VERSION__ && __STDC_VERSION__ >= 199909L
> > #define va_copy(dest, src) __va_copy(dest, src)
> > #endif
> >
> > please? That'll eliminate a deviation from C9x.
>
> va_copy is defined unconditionally in <stdarg.h>; c9x
> really defines va_copy at all for <varargs.h>?
C9x says nothing about <varargs.h>. I should have been clearer: I'd
like the above in <stdarg.h> only.
I think we should still have the ability to compile code in strict C89
mode, so we shouldn't unconditionally define the no-underscores
version.
[...]
> * Does any port use the second argument to stdarg.h va_start? If not,
> > why don't we use the same builtin for both versions of va_start, and
> > ignore the second argument to the stdarg.h version?
>
> The final arg isn't _used_, as such, but we do verify and warn
> when the argument is not the last named argument.
I realized we did that five minutes after sending the message. Never
mind.
zw