2007-02-14 Steven G. Kargl <kargl@gcc.gnu.org>
+ PR fortran/30799
+ * primary.c (match_logical_constant): Return MATCH_ERROR on invalid
+ kind.
+
+2007-02-14 Steven G. Kargl <kargl@gcc.gnu.org>
+
* misc.c (gfc_typename): Fix potential buffer overflow.
2007-02-13 Paul Thomas <pault@gcc.gnu.org>
kind = gfc_default_logical_kind;
if (gfc_validate_kind (BT_LOGICAL, kind, true) < 0)
- gfc_error ("Bad kind for logical constant at %C");
+ {
+ gfc_error ("Bad kind for logical constant at %C");
+ return MATCH_ERROR;
+ }
e = gfc_get_expr ();
+2007-02-14 Steven G. Kargl <kargl@gcc.gnu.org>
+
+ PR fortran/30799
+ * gfortran.dg/logical_2.f90: New test.
+
2007-02-14 Joseph Myers <joseph@codesourcery.com>
* gcc.dg/torture/complex-alias-1.c: New test.
--- /dev/null
+! { dg-do compile }
+! PR fortran/30799
+! Inconsistent handling of bad (invalid) LOGICAL kinds
+! Reporter: Harald Anlauf <anlauf@gmx.de>
+! Testcase altered by Steven G. Kargl
+program gfcbug57
+ implicit none
+ !
+ ! These are logical kinds known by gfortran and many other compilers:
+ !
+ print *, kind (.true._1) ! This prints "1"
+ print *, kind (.true._2) ! This prints "2"
+ print *, kind (.true._4) ! This prints "4"
+ print *, kind (.true._8) ! This prints "8"
+ !
+ ! These are very strange (read: bad (invalid?)) logical kinds,
+ ! handled inconsistently by gfortran (there's no logical(kind=0) etc.)
+ !
+ print *, kind (.true._0) ! { dg-error "kind for logical constant" }
+ print *, kind (.true._3) ! { dg-error "kind for logical constant" }
+ print *, kind (.true._123) ! { dg-error "kind for logical constant" }
+ !
+ ! Here gfortran bails out with a runtime error:
+ !
+ print *, .true._3 ! { dg-error "kind for logical constant" }
+end program gfcbug57