This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [patch, fortran] Fix PR29458 - spurious warning for implied do-loop counter
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Daniel Franke <franke dot daniel at gmail dot com>
- Cc: fortran at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- Date: Sun, 5 Apr 2009 09:47:50 -0700
- Subject: Re: [patch, fortran] Fix PR29458 - spurious warning for implied do-loop counter
- References: <200904051516.58984.franke.daniel@gmail.com>
On Sun, Apr 05, 2009 at 03:16:58PM +0200, Daniel Franke wrote:
>
> This code gives a warning about 'i' being used uninitialized:
>
> integer :: n, i
> n = 5
> n = SUM((/(i,i=1,n)/))
> end
>
> Attached patch fixes this by shadowing 'i' with a temporary variable. Further,
> it wraps the implied-do in its own block instead of placing it in the parent
> block directly. The relevant part now looks like this, 'i' is never touched:
>
> {
> shadow_loopvar.4 = 1;
> D.1515 = n;
> [...]
> while (1)
> {
> if (shadow_loopvar.4 > D.1515) goto L.1;
> (*(integer(kind=4)[0] *) atmp.2.data)[offset.3] = shadow_loopvar.4;
> offset.3 = offset.3 + 1;
> shadow_loopvar.4 = shadow_loopvar.4 + 1;
> }
> L.1:;
> }
What happens if someone used shadow_loopvar as a variable
in their code? Do we need to use _shadow_loopvar (note
the leading underscore)?
> 2009-04-05 Daniel Franke <franke.daniel@gmail.com>
>
> PR fortran/29458
> * trans-array.c (gfc_trans_array_constructor_value): Shadow implied
> do-loop variable to avoid spurious middle-end warnings.
>
> 2009-04-05 Daniel Franke <franke.daniel@gmail.com>
>
> PR fortran/29458
> * gfortran.dg/implied_do_1.f90: New.
Can you change the testcase to
! { dg-do "run" }
! PR fortran/29458 - spurious warning for implied do-loop counter
integer :: n, i
n = 5
i = 10
n = SUM((/(2*i,i=1,n,2)/))
if (i /= 10) call abort
end
This will test that future code changes doesn't break the
preservation of the value of i.
The code looks ok to me. Please commit to trunk.
--
Steve