------- Comment #2 from janus at gcc dot gnu dot org 2010-06-08 02:32 -------
Ok, this one took me a while to figure out, but in the end it turns out the fix
is actually pretty simple, once you found out where the problem is:
Index: gcc/fortran/resolve.c
===================================================================
--- gcc/fortran/resolve.c (revision 160351)
+++ gcc/fortran/resolve.c (working copy)
@@ -5415,6 +5415,9 @@ resolve_typebound_function (gfc_expr* e)
if (st == NULL)
return resolve_compcall (e, NULL);
+ if (resolve_ref (e) == FAILURE)
+ return FAILURE;
+
/* Get the CLASS declared type. */
declared = get_declared_from_expr (&class_ref, &new_ref, e);
@@ -5487,6 +5490,9 @@ resolve_typebound_subroutine (gfc_code *code)
if (st == NULL)
return resolve_typebound_call (code, NULL);
+ if (resolve_ref (code->expr1) == FAILURE)
+ return FAILURE;
+
/* Get the CLASS declared type. */
declared = get_declared_from_expr (&class_ref, &new_ref, code->expr1);
One can construct a test case analogous to comment #0 using a type-bound
function, which would also fail. The patch above fixes both flavors.
--
janus at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
AssignedTo|unassigned at gcc dot gnu |janus at gcc dot gnu dot org
|dot org |
Status|NEW |ASSIGNED
Last reconfirmed|2010-05-20 19:01:47 |2010-06-08 02:32:39
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44211
