On Friday 22 January 2010 16:10:08 Jamie Lokier wrote:
> Mike Frysinger wrote:
> > On Thursday 21 January 2010 20:21:31 Jamie Lokier wrote:
> > > Mike Frysinger wrote:
> > > > GCC-2.x has a bug with empty arg expansion in macros.
> > > >
> > > > - fprintf(stderr, "%s:%i: " fmt, __func__, __LINE__, ## args); \
> > > > + fprintf(stderr, "%s:%i: " fmt, __func__, __LINE__ , ## args); \
> > >
> > > Assuming it's what it looks like, it's not a GCC-2.x bug, it's the old
> > > GCC varargs behavior, which was documented and behaved like that since
> > > very old versions of GCC.
> >
> > i'm not really familiar with gcc-2 ... i didnt get into messing with
> > toolchain stuff until like gcc-3.4 era.
>
> Wow. I haven't messed much with toolchain stuff since the 2.7.2 era ;-)
>
> > it looks like a bug to me as gcc-3+ works fine.
> > if you wanted to elaborate ...
>
> GCC 2.95 documentation:
>
> This is a special feature of the GNU C preprocessor: @samp{##}
> before a rest argument that is empty discards the preceding
> sequence of non-whitespace characters from the macro definition.
> (If another macro argument precedes, none of it is discarded.)
>
> It might be better to discard the last preprocessor token instead
> of the last preceding sequence of non-whitespace characters; in
> fact, we may someday change this feature to do so. We advise you
> to write the macro definition so that the preceding sequence of
> non-whitespace characters is just a single token, so that the
> meaning will not change if we change the definition of this
> feature.
>
> It's behaved like that since the macro varargs feature was introduced
> in GCC 2.2 (released June 1992), and was consciously changed to
> in GCC 3.0 (released June 2001).
>
> > this one fails:
> > #define debug(fmt, args...) x(fmt, moo, ## args)
> > main(){debug("a");}
> >
> > these work:
> > #define debug(fmt, args...) x(fmt, moo , ## args)
> > main(){debug("a");}
> >
> > #define debug(fmt, args...) x(fmt, ## args)
> > main(){debug("a");}
>
> The last one compiles for a surprising reason: fmt is preexpanded to
> ' "a" ' (complete with spaces), so the word-removal rule goes back as
> far as the comma and stops at the last space of the preexpanded fmt.thanks for the info > Anyway, your patch is fine! :-) > > Only the commit message is misleading; this would be appropriate imho: > > GCC-2.x's variadic macro behaviour requires an isolated comma > > The behaviour of GCC before 3.0 was to remove the whole > non-whitespace word prior to ## (after argument expansion), > not just the comma. i'd agree, but it's already been committed -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ uClinux-dev mailing list uClinux-dev@uclinux.org http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by uclinux-dev@uclinux.org To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
