This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: patch for -Wno-long-long and early GNAT compilers
- From: Zack Weinberg <zack at codesourcery dot com>
- To: Richard Kenner <kenner at vlsi1 dot ultra dot nyu dot edu>
- Cc: gcc at gcc dot gnu dot org
- Date: Tue, 19 Mar 2002 12:07:20 -0800
- Subject: Re: patch for -Wno-long-long and early GNAT compilers
- References: <10203191922.AA15047@vlsi1.ultra.nyu.edu>
On Tue, Mar 19, 2002 at 02:22:38PM -0500, Richard Kenner wrote:
> I can make arguments either way, but consider that there's
> considerable time and effort been expended verifying that 2.95 and 3.0
> compile the back end correctly, on a wide variety of architectures.
> 2.8.1, on the other hand, has not received much attention at all, and
> is known to have problems on some architectures. So I think a patch
> which allows people to continue using 2.95 or 3.0 for everything but
> Ada is safer than a patch which forces everything to get built with
> 2.8.1.
>
> I disagree because you are adding extra complexity to an already complex
> process. Having one compiler run autoconf and another do the actual
> compilations is asking for trouble, for example.
Does it make a difference if I point out that none of the code being
compiled by the compiler that didn't run autoconf can be affected by
<stdbool.h>? Most of raise.c is under #ifdef IN_RTS. The only bits
relevant to the compiler are __gnat_builtin_longjmp,
__gnat_unhandled_terminate, and the dummy _Unwind_RaiseException; none
of them use bool.
I could split raise.c to make this explicit if it would make you more
comfortable with the change.
zw