------- Comment #3 from fxcoudert at gcc dot gnu dot org 2008-04-23 18:27
-------
> BTW: The testcase doesn't fail for x86_64-linux. Can you please regression
> test the patch (and write a testcase ;) on *-mingw target?
I'm not set up for regtesting on x86_64-mingw32 (nor on 32-bit mingw32, in any
case), but I checked that your patch does fix both the reduced and the original
testcase (sent to me by private mail). I propose to add the following
testcases:
$ cat z1.f90
! { dg-do compile }
!
subroutine test4
integer, parameter :: wp = 4
complex(wp), parameter :: i = (0._wp, 1._wp)
complex(wp) :: c(12)
integer :: m, N
N = 12
c = (/(exp(i*m),m=1,N)/)
print *, c(1)
end
subroutine test8
integer, parameter :: wp = 8
complex(wp), parameter :: i = (0._wp, 1._wp)
complex(wp) :: c(12)
integer :: m, N
N = 12
c = (/(exp(i*m),m=1,N)/)
print *, c(1)
end
$ cat z2.f90
! { dg-do compile }
! { dg-require-effective-target fortran_large_real }
!
subroutine test_large
integer, parameter :: wp = selected_real_kind (precision (0.0_8) + 1)
complex(wp), parameter :: i = (0._wp, 1._wp)
complex(wp) :: c(12)
integer :: m, N
N = 12
c = (/(exp(i*m),m=1,N)/)
print *, c(1)
end
(I think covering all real types is a good idea, in that case.)
--
fxcoudert at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |patch
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36006
