[Bug fortran/52162] Bogus -fcheck=bounds with realloc on assignment to unallocated LHS

["pault at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=52162

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |pault at gcc dot gnu.org

--- Comment #8 from Paul Thomas <pault at gcc dot gnu.org> ---
Changing the line mentioned in comment #2 to:
  if (gfc_is_reallocatable_lhs (expr1)
        && !(expr2->expr_type == EXPR_FUNCTION
             && expr2->value.function.isym != NULL
             && !(expr2->value.function.isym->elemental
                  || expr2->value.function.isym->conversion)))
    lss->is_alloc_lhs = 1;

fixes the problem. This testcase now runs correctly:

! { dg-do run }
! { dg-options "-fbounds-check" }
!
! Test the fix for PR52162 in which the elemental and conversion
! intrinsics in lines 14 and 19 would cause the bounds check to fail.
!
! Contributed by Dominique d'Humieres  <dominiq@lps.ens.fr>
!
    integer(4), allocatable :: a(:)
    integer(8), allocatable :: b(:)
    real, allocatable :: c(:)
    allocate (b(7:11), source = [7_8,8_8,9_8,10_8,11_8])

    a = b ! Implicit conversion

    if (lbound (a, 1) .ne. lbound(b, 1)) call abort
    if (ubound (a, 1) .ne. ubound(b, 1)) call abort

    c = sin(real(b(9:11))/100_8) ! Elemental intrinsic

    if ((ubound(c, 1) - lbound(c, 1)) .ne. 2) call abort
    if (any (int(asin(c)*100.0) .ne. b(9:11))) call abort
    deallocate (a, b, c)
  end

I will submit as soon as the patch for PR34640 clears my tree.

Paul