[Bug fortran/99852] New: Missing error "Arithmetic overflow" for some cases

[Bug fortran/99852] New: Missing error "Arithmetic overflow" for some cases

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99852

            Bug ID: 99852
           Summary: Missing error "Arithmetic overflow" for some cases
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Following examples produce an error for c and d, but not for a and b :
(similar for wik=2,4,8,16)


$ cat z1.f90
program p
   implicit none
   integer, parameter :: wik = 1
   integer(wik), parameter :: a = huge(1_wik) + 1_wik
   integer(wik) :: b = huge(1_wik) + 1_wik
   integer(wik) :: c
   c = huge(1_wik) + 1_wik
   associate (d => huge(1_wik) + 1_wik)
      print *, a, b, c, d
   end associate
end


$ cat z2.f90
program p
   implicit none
   integer, parameter :: wik = 1
   integer(wik), parameter :: a = -huge(1_wik) - 2_wik
   integer(wik) :: b = -huge(1_wik) - 2_wik
   integer(wik) :: c
   c = -huge(1_wik) - 2_wik
   associate (d => -huge(1_wik) - 2_wik)
      print *, a, b, c, d
   end associate
end


$ gfortran-11-20210328 -c z1.f90
z1.f90:7:7:

    7 |    c = huge(1_wik) + 1_wik
      |       1
Error: Arithmetic overflow at (1)
z1.f90:8:19:

    8 |    associate (d => huge(1_wik) + 1_wik)
      |                   1
Error: Arithmetic overflow at (1)

[Bug fortran/99852] Missing error "Arithmetic overflow" for some cases

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99852

--- Comment #1 from G. Steinmetz <gscfq@t-online.de> ---

In addition a side mark, the following gives for all a-d :


$ cat z3.f90
program p
   implicit none
   integer, parameter :: wik = 1
   integer(wik), parameter :: a = -huge(1_wik) - 1_wik
   integer(wik) :: b = -huge(1_wik) - 1_wik
   integer(wik) :: c
   c = -huge(1_wik) - 1_wik
   associate (d => -huge(1_wik) - 1_wik)
      print *, a, b, c, d
   end associate
end


$ gfortran-11-20210328 -c z3.f90 -pedantic
z3.f90:4:33:

    4 |    integer(wik), parameter :: a = -huge(1_wik) - 1_wik
      |                                 1
Warning: Integer outside symmetric range implied by Standard Fortran at (1)
z3.f90:5:22:

    5 |    integer(wik) :: b = -huge(1_wik) - 1_wik
      |                      1
Warning: Integer outside symmetric range implied by Standard Fortran at (1)
z3.f90:7:7:

    7 |    c = -huge(1_wik) - 1_wik
      |       1
Warning: Integer outside symmetric range implied by Standard Fortran at (1)
z3.f90:8:19:

    8 |    associate (d => -huge(1_wik) - 1_wik)
      |                   1
Warning: Integer outside symmetric range implied by Standard Fortran at (1)

[Bug fortran/99852] Missing error "Arithmetic overflow" for some cases

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99852

--- Comment #2 from anlauf at gcc dot gnu.org ---
*** Bug 104848 has been marked as a duplicate of this bug. ***

[Bug fortran/99852] Missing error "Arithmetic overflow" for some cases

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99852

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
                 CC|                            |anlauf at gcc dot gnu.org
   Last reconfirmed|                            |2024-04-02

--- Comment #3 from anlauf at gcc dot gnu.org ---
Confirming at last.

Another testcase from pr104848#c1:

program p
  integer, parameter :: a(2, 0) = 0
  integer, parameter :: y(*) = minval(a, dim=2)
  print *, y   ! no ICE
  print *, y-1 ! no ICE
  print *, y+1 ! ICE
end