Date: Thu, 11 Jul 1996 09:04:05 +1000 From: Bruce Evans <bde@zeta.org.au> To: dgy@rtd.com, freebsd-hackers@freefall.freebsd.org, freebsd-ports@freefall.freebsd.org Subject: Re: Q: macro expansion Message-ID: <199607102304.JAA31461@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
> #define X 1
> #define Y 2
> X+Y
>I was surprised to see that ``X+Y'' was expanded to ``1 +2''. So,
>I started digging through ANSI and couldn't seem to locate something
>to clearly define this behaviour.
> - why the inserted whitespace?
See the ISO C standard section 6.8.3 (Macro Replacement). Macros are
expanded as if they were tokenized before expansion (so a `+' at the
end of X doesn't get joined with the `+' in X+ to form a `++' token).
This is usually implemented by inserting whitespace so that tokenizing
after expansion gives the same result.
> - why no whitespace after `+'?
This is a bug in gcc-2.6.3. It actually gives `` 1 +2 '', while
the version in gcc-2.7.2 gives ``1 + 2 ''. The
difference is important for
#define Y +
int z;
main() { printf("%d\n", +Y z); }
For gcc-2.6.3, ``+Y z'' expands to ``++ z'' so there is a bogus `++'
token and the result is 1. For gcc-2.7.2, it expands to ``+ + z''
and the result is 0.
Anyway, don't use `gcc -E' as a general purpose macro expander.
`gcc -E -traditional' and /usr/bin/cpp work better.
Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199607102304.JAA31461>