[Patch, Fortran] PR34901 - Make mismatched-types error message less odd

Tobias Burnus burnus@net-b.de
Tue Jan 22 01:59:00 GMT 2008


Tobias Schlüter wrote:
>> Error: Type mismatch in argument 'a' at (1); passed REAL(kind=4) to 
>> REAL(kind=8)
> Please use gfc_typename() for this purpose.
Thanks! I felt it had to be somewhere, but I could not find it last time.

Tobias

Committed revision 131700:

Index: interface.c
===================================================================
--- interface.c (revision 131699)
+++ interface.c (working copy)
@@ -1470,18 +1470,10 @@
   if ((actual->expr_type != EXPR_NULL || actual->ts.type != BT_UNKNOWN)
       && !gfc_compare_types (&formal->ts, &actual->ts))
     {
-      if (where && actual->ts.type == BT_DERIVED
-         && formal->ts.type == BT_DERIVED)
-       gfc_error ("Type mismatch in argument '%s' at %L; passed type(%s) to "
-                  "type(%s)", formal->name, &actual->where,
-                  actual->ts.derived->name, formal->ts.derived->name);
-      else if (where)
+      if (where)
        gfc_error ("Type mismatch in argument '%s' at %L; passed %s to %s",
-                  formal->name, &actual->where,
-                  actual->ts.type == BT_DERIVED ? "derived type"
-                                    : gfc_basic_typename (actual->ts.type),
-                  formal->ts.type == BT_DERIVED ? "derived type"
-                                    : gfc_basic_typename (formal->ts.type));
+                  formal->name, &actual->where, gfc_typename (&actual->ts),
+                  gfc_typename (&formal->ts));
       return 0;
     }

Index: ChangeLog
===================================================================
--- ChangeLog   (revision 131699)
+++ ChangeLog   (working copy)
@@ -1,3 +1,9 @@
+2007-01-21  Tobias Burnus  <burnus@net-b.de>
+
+       PR fortran/34901
+       * interface.c (compare_parameter): Improved error message
+       for arguments of same type and mismatched kinds.
+
 2008-01-20  Paul Thomas  <pault@gcc.gnu.org>

        PR fortran/34861





More information about the Fortran mailing list