[Patch, fortran] PR26822 - Scalarization of non-elemental intrinsic: __logical_4_l4

[Paul Thomas <paulthomas2 at wanadoo dot fr>]

:ADDPATCH fortran:

This took a depressingly long time to get to the bottom of - 
particularly since the patch changes but one character!  Fortunately, it 
was a few minutes here and there.

The problem came about because LOGICAL was not marked as elemental in 
intrinsic.c(add_functions), which put it in the wrong place in 
trans_intrinsic.c and caused the observed problem.  The error message 
should have made it obvious what was happening.

I do not see any point in a testcase to verify the correction of a typo 
and so have not included one.  The principle of this kind of intrinsic 
is more than adequately tested.

Regtested on FC3/Athlon - will commit under the obvious rule, when doing 
other commits.

Paul


2006-04-13  Paul Thomas  <pault@gcc.gnu.org>

   PR fortran/26822
   * intrinsic.c (add_functions): Mark LOGICAL as elemental.

Index: gcc/fortran/intrinsic.c
===================================================================
*** gcc/fortran/intrinsic.c    (revision 112712)
--- gcc/fortran/intrinsic.c    (working copy)
*************** add_functions (void)
*** 1670,1676 ****

   make_generic ("log10", GFC_ISYM_LOG10, GFC_STD_F77);

!   add_sym_2 ("logical", 0, 1, BT_LOGICAL, dl, GFC_STD_F95,
          gfc_check_logical, gfc_simplify_logical, gfc_resolve_logical,
          l, BT_LOGICAL, dl, REQUIRED, kind, BT_INTEGER, di, OPTIONAL);

--- 1670,1676 ----

   make_generic ("log10", GFC_ISYM_LOG10, GFC_STD_F77);

!   add_sym_2 ("logical", 1, 1, BT_LOGICAL, dl, GFC_STD_F95,
          gfc_check_logical, gfc_simplify_logical, gfc_resolve_logical,
          l, BT_LOGICAL, dl, REQUIRED, kind, BT_INTEGER, di, OPTIONAL);