http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47448
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |rejects-valid
Summary|Accepts invalid |Invalid check for
|ASSIGNMENT(=) which |ASSIGNMENT(=)
|overrides intrinsic |
|assignment |
--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-01-25
08:30:41 UTC ---
Draft patch. The typo seems to go back to the commit for PR 37425, committed
2009-08-10.
However, it seems only to be an accepts-invalid regression. Before the valid
and the invalid form were rejected, currently only the valid form is rejected -
the invalid is accepted.
--- a/gcc/fortran/interface.c
+++ b/gcc/fortran/interface.c
@@ -654,11 +654,12 @@ gfc_check_operator_interface (gfc_symbol *sym,
gfc_intrinsic_op op,
/* Allowed are (per F2003, 12.3.2.1.2 Defined assignments):
- First argument an array with different rank than second,
- - Types and kinds do not conform, and
+ - First argument is a scalar and second an array,
+ - Types and kinds do not conform, or
- First argument is of derived type. */
if (sym->formal->sym->ts.type != BT_DERIVED
&& sym->formal->sym->ts.type != BT_CLASS
- && (r1 == 0 || r1 == r2)
+ && (r2 == 0 || r1 == r2)
&& (sym->formal->sym->ts.type == sym->formal->next->sym->ts.type
|| (gfc_numeric_ts (&sym->formal->sym->ts)
&& gfc_numeric_ts (&sym->formal->next->sym->ts))))
