This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Patch for `LL' warnings in insn-*.c
- From: Zack Weinberg <zack at codesourcery dot com>
- To: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 4 Aug 2002 23:19:59 -0700
- Subject: Re: Patch for `LL' warnings in insn-*.c
- References: <20020724154505.GA7224@codesourcery.com> <Pine.LNX.4.33.0208050003530.15986-100000@kern.srcf.societies.cam.ac.uk>
On Mon, Aug 05, 2002 at 12:06:40AM +0100, Joseph S. Myers wrote:
> On Wed, 24 Jul 2002, Zack Weinberg wrote:
>
> > That particular warning comes from cpplib, which is not currently
> > aware of __extension__ since it treats each token in isolation.
>
> That is something that needs fixing somehow, at least for the warning for
> imaginary constants, since at present the following (compiled with
> -std=c99 -pedantic on a glibc system) shows a regression between 3.1 and
> mainline (since <complex.h> uses __extension__ in the definition of I to
> avoid warnings).
This should be covered by the rule that warnings do not issue from the
use of a macro defined in a system header. If that isn't currently
true, it will be much easier to fix the regression thus than by
teaching cpplib about __extension__, which would entail cpp knowing
the extent of "the expression which follows" - only the parser knows
that.
zw