This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: __va_copy - for compatibility
> Mail-Copies-To: never
> Cc: gcc-patches@gcc.gnu.org, Kaoru Fukui <k_fukui@highway.ne.jp>
> From: Andreas Jaeger <aj@suse.de>
> Date: 26 Nov 1999 08:23:03 +0100
> User-Agent: Gnus/5.070099 (Pterodactyl Gnus v0.99) XEmacs/21.1 (Biscayne)
> >> #define va_end __builtin_va_end
> >> #define va_arg __builtin_va_arg
> >> #define va_copy(d,s) __builtin_va_copy(&(d),(s))
> >> +#define __va_copy(d,s) __builtin_va_copy(&(d),(s))
>
> Geoff> Should we really be defining va_copy (the one without the underscores)
> Geoff> when compiling with -ansi? It would seem to be a violation of the old
> Geoff> C standard.
>
> That's a policy decision, since va_copy part of the new standard.
Well, if the user asks to us to support the new standard, we should
define it; if we are supporting C89, we should not define it; for GNU
C, I think we do define it because GNU C will be based on C99.
There's a whole bunch of flags that people can use to pick exactly
what they want.
[None of this affects the original patch, because things with
underscores are internal to gcc and glibc, so we can do whatever
works.]
--
- Geoffrey Keating <geoffk@cygnus.com>