+2018-09-19 Paul Thomas <pault@gcc.gnu.org>
+
+ PR fortran/84109
+ * trans-expr.c (gfc_trans_assignment_1): The rse.pre for the
+ assignment of deferred character intrinsic elemental function
+ results to a realocatable lhs must not be added to the exterior
+ block if they are array valued but must go to the loop body.
+
2018-09-18 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/29550
parameter available to the caller; gfortran saves it in the .mod files.
NOTE ALSO: The concatenation operation generates a temporary pointer,
whose allocation must go to the innermost loop.
- NOTE ALSO (2): A character conversion may generate a temporary, too. */
+ NOTE ALSO (2): Elemental functions may generate a temporary, too. */
if (flag_realloc_lhs
&& expr2->ts.type == BT_CHARACTER && expr1->ts.deferred
&& !(lss != gfc_ss_terminator
+ && rss != gfc_ss_terminator
&& ((expr2->expr_type == EXPR_FUNCTION
&& expr2->value.function.esym != NULL
&& expr2->value.function.esym->attr.elemental)
- || (expr2->expr_type == EXPR_OP
- && expr2->value.op.op == INTRINSIC_CONCAT)
|| (expr2->expr_type == EXPR_FUNCTION
&& expr2->value.function.isym != NULL
- && expr2->value.function.isym->id == GFC_ISYM_CONVERSION))))
+ && expr2->value.function.isym->elemental)
+ || (expr2->expr_type == EXPR_OP
+ && expr2->value.op.op == INTRINSIC_CONCAT))))
gfc_add_block_to_block (&block, &rse.pre);
/* Nullify the allocatable components corresponding to those of the lhs
+2018-09-19 Paul Thomas <pault@gcc.gnu.org>
+
+ PR fortran/84109
+ * gfortran.dg/elemental_function_3.f90 : New test.
+
2018-09-19 Marek Polacek <polacek@redhat.com>
PR c++/87357 - missing -Wconversion warning
--- /dev/null
+! { dg-do run }
+!
+! Test the fix for PR84109 in which the call to the elemental function
+! 'adjustl' was being added before the scalarization loop in the assignment.
+! Since the result temporary was being declared in the loop body, this
+! drove the gimplifier crazy. It is sufficient to compile this testcase
+! since it used to ICE. This is the intrinsic counterpart to PR87239,
+! which is tested for the absence of an ICE in elemental_function_2.f90.
+! In this fix, a further improvement was to keep scalar calls outside the
+! scalarization loop and this is tested with 'my_adjustl'.
+!
+! Contributed by Willem Vermin <wvermin@gmail.com>
+!
+program prog
+ implicit none
+ character(len=:), allocatable :: c(:)
+ integer :: cnt = 0
+
+ allocate(character(len=20) :: c(10))
+ c = " ab "
+ c = adjustl(c) ! Used to ICE
+ if (trim (c(1)) .ne. "ab") stop 1
+
+ c = my_adjustl (" abcdefg ")
+ if (trim (c(1)) .ne. "abcdefg") stop 2
+ if (cnt .ne. 1) stop 3 ! Outside the scalarization loop
+ if (size (c, 1) .ne. 10) stop 4
+ if (len (c) .ne. 20) stop 5
+
+ cnt = 0
+ c = my_adjustl ([" uv ", " xy "])
+ if (trim (c(2)) .ne. "xy") stop 6
+ if (cnt .ne. size (c, 1)) stop 7 ! Inside the scalarization loop
+ if (size (c, 1) .ne. 2) stop 8
+
+contains
+
+ impure elemental function my_adjustl(arg) result (res)
+ character(*), intent(in) :: arg
+ character(len = len (arg)) :: res
+ res = adjustl (arg)
+ cnt = cnt + 1 ! Test how many calls are made
+ end function
+end program