Fix for is_intxx_type() overloaded methods

[Michael Goffioul]

The patch below fixes a missing "const" modifier in overloaded 
is_intxx_type()
functions of integer octave types (matrix and scalar). In MSVC, the 
result is that
the predicate declared in ov-intx.h does not overload the one of 
octave_base-value,
but defines another one. Calling is_int32_type() on octave_int32_scalar 
object then
returns false (I guess the compiler prefer to use the const version by 
default).

Michael.

Index: src/ov-intx.h
===================================================================
RCS file: /cvs/octave/src/ov-intx.h,v
retrieving revision 1.23
diff -c -r1.23 ov-intx.h
*** src/ov-intx.h       18 Aug 2006 21:27:04 -0000      1.23
--- src/ov-intx.h       22 Oct 2006 16:04:45 -0000
***************
*** 58,64 ****
   octave_base_value *empty_clone (void) const
     { return new OCTAVE_VALUE_INT_MATRIX_T (); }

!   bool OCTAVE_TYPE_PREDICATE_FUNCTION (void) { return true; }

   int8NDArray
   int8_array_value (void) const { return int8NDArray (matrix); }
--- 58,64 ----
   octave_base_value *empty_clone (void) const
     { return new OCTAVE_VALUE_INT_MATRIX_T (); }

!   bool OCTAVE_TYPE_PREDICATE_FUNCTION (void) const { return true; }

   int8NDArray
   int8_array_value (void) const { return int8NDArray (matrix); }
***************
*** 237,243 ****
       return retval;
     }

!   bool OCTAVE_TYPE_PREDICATE_FUNCTION (void) { return true; }

   octave_int8
   int8_scalar_value (void) const { return octave_int8 (scalar); }
--- 237,243 ----
       return retval;
     }

!   bool OCTAVE_TYPE_PREDICATE_FUNCTION (void) const { return true; }

   octave_int8
   int8_scalar_value (void) const { return octave_int8 (scalar); }