[Patch, libfortran] PR 55539 Regression with -fno-sign-zero

[Patch, libfortran] PR 55539 Regression with -fno-sign-zero

[Janne Blomqvist]

[-- Attachment #1: Type: text/plain, Size: 411 bytes --]

Hello,

the attached patch fixes PR 55539 (4.8 regression). Regtested on
x86_64-unknown-linux-gnu, Ok for trunk?


2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/55539
	* io/write_float.def (output_float): Take into account decimal dot.


testsuite ChangeLog:

2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>

	PR fortran/55539
	* gfortran.dg/nosigned_zero_3.f90: New testcase.



-- 
Janne Blomqvist

[-- Attachment #2: pr55539.diff --]
[-- Type: application/octet-stream, Size: 1041 bytes --]

diff --git a/libgfortran/io/write_float.def b/libgfortran/io/write_float.def
index 6521f3c..1e30dde 100644
--- a/libgfortran/io/write_float.def
+++ b/libgfortran/io/write_float.def
@@ -483,16 +483,19 @@ output_float (st_parameter_dt *dtp, const fnode *f, char *buffer, size_t size,
   /* Scan the digits string and count the number of zeros.  If we make it
      all the way through the loop, we know the value is zero after the
      rounding completed above.  */
-  for (i = 0; i < ndigits; i++)
+  int hasdot = 0;
+  for (i = 0; i < ndigits + hasdot; i++)
     {
-      if (digits[i] != '0' && digits[i] != '.')
+      if (digits[i] == '.')
+	hasdot = 1;
+      else if (digits[i] != '0')
 	break;
     }
 
   /* To format properly, we need to know if the rounded result is zero and if
      so, we set the zero_flag which may have been already set for
      actual zero.  */
-  if (i == ndigits)
+  if (i == ndigits + hasdot)
     {
       zero_flag = true;
       /* The output is zero, so set the sign according to the sign bit unless

[-- Attachment #3: nosigned_zero_3.f90 --]
[-- Type: application/octet-stream, Size: 359 bytes --]

! { dg-do run }
! { dg-options "-fno-sign-zero" }
!
! PR fortran/55539
!
program nosigned_zero_3
  implicit none
  character(len=20) :: s
  real(4) :: x = -1.2e-3
  real(8) :: y = -1.2e-3
  write(s,'(7f10.3)') x
  if (trim(adjustl(s)) /= "-0.001") call abort
  write(s, '(7f10.3)') y
  if (trim(adjustl(s)) /= "-0.001") call abort
end program nosigned_zero_3

Re: [Patch, libfortran] PR 55539 Regression with -fno-sign-zero

[Janne Blomqvist]

Ping!

On Sat, Dec 15, 2012 at 6:32 PM, Janne Blomqvist
<blomqvist.janne@gmail.com> wrote:
> Hello,
>
> the attached patch fixes PR 55539 (4.8 regression). Regtested on
> x86_64-unknown-linux-gnu, Ok for trunk?
>
>
> 2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>
>
>         PR fortran/55539
>         * io/write_float.def (output_float): Take into account decimal dot.
>
>
> testsuite ChangeLog:
>
> 2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>
>
>         PR fortran/55539
>         * gfortran.dg/nosigned_zero_3.f90: New testcase.
>
>
>
> --
> Janne Blomqvist



-- 
Janne Blomqvist

Re: [Patch, libfortran] PR 55539 Regression with -fno-sign-zero

[Jerry DeLisle]

On 12/15/2012 08:32 AM, Janne Blomqvist wrote:
> Hello,
>
> the attached patch fixes PR 55539 (4.8 regression). Regtested on
> x86_64-unknown-linux-gnu, Ok for trunk?
>
>
> 2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>
>
> 	PR fortran/55539
> 	* io/write_float.def (output_float): Take into account decimal dot.
>
>
> testsuite ChangeLog:
>
> 2012-12-15  Janne Blomqvist  <jb@gcc.gnu.org>
>
> 	PR fortran/55539
> 	* gfortran.dg/nosigned_zero_3.f90: New testcase.
>

OK, thanks for patch.

Jerry