Re: fortran/10188: gcc test.c or g77 test.f -malign-double

Re: fortran/10188: gcc test.c or g77 test.f -malign-double

[toon]

Synopsis: gcc test.c  or g77 test.f -malign-double

Responsible-Changed-From-To: unassigned->toon
Responsible-Changed-By: toon
Responsible-Changed-When: Sat Mar 22 13:40:49 2003
Responsible-Changed-Why:
    Fortran Maintainer.
State-Changed-From-To: open->closed
State-Changed-By: toon
State-Changed-When: Sat Mar 22 13:40:49 2003
State-Changed-Why:
    Not a bug.  A floating point constant in Fortran is a
    REAL unless indicated otherwise.
    
    If you need a DOUBLE PRECISION rendition of 5/9, you have
    to write 0.555555555555555555D0.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10188

RE: fortran/10188: gcc test.c or g77 test.f -malign-double

[Billinghurst, David (CRTS)]

The following reply was made to PR fortran/10188; it has been noted by GNATS.

From: "Billinghurst, David (CRTS)" <David.Billinghurst@riotinto.com>
To: <bording@bnl.gov>,
	<gcc-gnats@gcc.gnu.org>
Cc:  
Subject: RE: fortran/10188: gcc test.c  or g77 test.f -malign-double
Date: Sat, 22 Mar 2003 12:06:50 +1100

 The behaviour you see is expected (and standard conforming)
 
 Try =20
       b =3D 0.555555555555555555d0
 
 
 I will confirm this when I get near a computer with g77 installed.
 
 
 -----Original Message-----
 From: bording@bnl.gov [mailto:bording@bnl.gov]
 Sent: Saturday, 22 March 2003 11:54 AM
 To: gcc-gnats@gcc.gnu.org
 Subject: fortran/10188: gcc test.c or g77 test.f -malign-double
 
 
 
 >Number:         10188
 >Category:       fortran
 >Synopsis:       gcc test.c  or g77 test.f -malign-double
 >Confidential:   no
 >Severity:       serious
 >Priority:       medium
 >Responsible:    unassigned
 >State:          open
 >Class:          sw-bug
 >Submitter-Id:   net
 >Arrival-Date:   Sat Mar 22 00:56:00 UTC 2003
 >Closed-Date:
 >Last-Modified:
 >Originator:     bording@bnl.gov
 >Release:        gcc-3.2.7
 >Organization:
 >Environment:
 Red-Hat 8.0 on Pentium Pro
 >Description:
 It seems double precision is not used:
 ----fortran code----
       real a
       double precision b
 
       a =3D 0.555555555555555555
       b =3D 0.555555555555555555
       write(6,1) a
       write(6,2) b
 1     format('real   ',f30.28)
 2     format('double ',f30.28)
       end
 
 gives :=20
 real   0.5555555820465087890625000000
 double 0.5555555820465087890625000000
 
 ------C code -----
 #include <stdio.h>
 float a;
 double b;
 main()
 {
   a =3D b =3D 0.555555555555555555;
   printf("real   %.28f\n",a);
   printf("double %.28f\n",b);
 }
 
 gives:
 real   0.5555555820465087890625000000
 double 0.5555555555555555802271783250
 
 -----
 For C double has higher resolution, but for fortran=20
 there is no differece. I have tried various
 compiler flags without luck !
 >How-To-Repeat:
 
 >Fix:
 
 >Release-Note:
 >Audit-Trail:
 >Unformatted:

fortran/10188: gcc test.c or g77 test.f -malign-double

[bording]

>Number:         10188
>Category:       fortran
>Synopsis:       gcc test.c  or g77 test.f -malign-double
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Mar 22 00:56:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     bording@bnl.gov
>Release:        gcc-3.2.7
>Organization:
>Environment:
Red-Hat 8.0 on Pentium Pro
>Description:
It seems double precision is not used:
----fortran code----
      real a
      double precision b

      a = 0.555555555555555555
      b = 0.555555555555555555
      write(6,1) a
      write(6,2) b
1     format('real   ',f30.28)
2     format('double ',f30.28)
      end

gives : 
real   0.5555555820465087890625000000
double 0.5555555820465087890625000000

------C code -----
#include <stdio.h>
float a;
double b;
main()
{
  a = b = 0.555555555555555555;
  printf("real   %.28f\n",a);
  printf("double %.28f\n",b);
}

gives:
real   0.5555555820465087890625000000
double 0.5555555555555555802271783250

-----
For C double has higher resolution, but for fortran 
there is no differece. I have tried various
compiler flags without luck !
>How-To-Repeat:

>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted: