This is the mail archive of the
mailing list for the GNU Fortran project.
Re: -2147483648 (again)
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Scott Robert Ladd <coyote at coyotegulch dot com>
- Cc: fortran at gcc dot gnu dot org
- Date: Tue, 6 Jan 2004 09:50:16 -0800
- Subject: Re: -2147483648 (again)
- References: <3FFA08A5.email@example.com>
On Mon, Jan 05, 2004 at 08:00:21PM -0500, Scott Robert Ladd wrote:
> I've received more opinions, both in e-mail and in comp.lang.fortran,
> regarding gfortran's rejection of the constant -2147483648. The
> overwhelming concensus: the constant is valid, real world code usues it,
> and "any commercially-viable" compiler (as one person put it) should
> accept it. Some of these opinions come from members of J3.
First, -2147483648 is not a valid constant because 2147483648
is not valid (on a 32-bit processor). Richard Maine clearly stated
that -2147483648 is a UNARY MINUS and positive literal constant.
Richard suggested that -2147483648 is valid integer value.
Second, "real world code uses it" means that those "real world codes"
are not portable.
Third, only 1 member of J3 (Richard Maine) commented on your
> Given that my proposed switch to allow such constants was rejected, I
> noew suggest that the compiler accept such constants by default,
> flagging them when the -fpedantic switch is used.
Unless Paul or Steven rejected your patch in private email, then
I don't think it has been rejected. If anything, it sparked
discussion about what to do with the intrinsic functions and it
was left at that point. For the record, I support your original