[Bug fortran/54932] New: Invalid loop code generated by Fortran FE for loops with bounds in HIGH(type)

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54932

             Bug #: 54932
           Summary: Invalid loop code generated by Fortran FE for loops
                    with bounds in HIGH(type)
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: hubicka@gcc.gnu.org


The following fortran testcase (derrived from do-1.f90):
! { dg-do run }
! Program to check corner cases for DO statements.
program do_1
  implicit none
  integer i, j

  ! limit=HUGE(i), step 1
  j = 0
  do i = HUGE(i) - 10, HUGE(i), 1
    j = j + 1
  end do
  if (j .ne. 11) call abort

end program

gets invalid estimate on number of iterations of the loop

Statement i_9 = i_8 + 1;
 is executed at most 9 (bounded by 9) + 1 times in loop 1.
Loop 1 iterates at most 9 times.

(should be 10) this is because:
  <bb 3>:
  # i_8 = PHI <2147483637(2), i_9(3)>
  # j_6 = PHI <0(2), j_7(3)>
  # DEBUG j => j_6
  # DEBUG i => i_8
  j_7 = j_6 + 1;
  # DEBUG j => j_7
  i_9 = i_8 + 1;
  # DEBUG i => i_9
  if (i_8 == 2147483647)
    goto <bb 4>;
  else
    goto <bb 3>;

the loop is produce din a way that i_9 overflows in the last iteration. It is
unused by according to discussion with Richard on IRC the program is not valid.
This is already wrong in the output from the fortran FE:
<bb 3>:
  j = j + 1;
  test (&i);
  i.1 = i;
  D.1852 = i.1 == 2147483647;
  i.1 = i;
  i.2 = i.1 + 1;
  i = i.2;
  if (D.1852 != 0)
    goto <bb 4>;
  else
    goto <bb 3>;

the increment really should happen in the latch block.
Or is the testcase bogus?

Honza