Operator Overloading

FreeBASIC

Operator Overloading
 
Changing the way user defined types work with built-in operators.

Overview
Global Operators
Member Operators

Overview

Simply, operators are procedures, and their arguments are called operands. Operators that take one operand (Operator Not) are called unary operators, operators that take two operands (Operator +) are called binary operators and operators taking three operands (Operator Iif) are called ternary operators.

Most operators are not called like procedures. Instead, their operator symbol is placed next to their operands. For unary operators, their sole operand is placed to the right of the symbol. For binary operators, their operands - referred to as the left and right-hand side operands - are placed to the left and right of the operator symbol. FreeBASIC has one ternary operator, Operator Iif, and it is called like a procedure, with its operands comma-separated surrounded by parenthesis. For example, the following code calls Operator Iif to determine if a pointer is valid. If it is, Operator * (Value Of) is called to dereference the pointer, and if not, Operator / (Divide) is called to find the value of twenty divided by four.

Dim i As Integer = 420
Dim p As Integer Ptr = @i

Dim result As Integer = IIf( p, *p, CInt( 20 / 4 ) )

Notice the call to Operator Iif is similar to a procedure call, while the calls to Operator * (Value Of) and Operator / (Divide) are not. In the example, p is the operand to Operator * (Value Of), and 20 and 4 are the left and right-hand side operands of Operator / (Divide), respectively.

All operators in FreeBASIC are predefined to take operands of standard data types, like Integer and Single, but they may also be overloaded for user-defined types; that is, they can be defined to accept operands that are objects as well. There are two types of operators that can be overloaded, global operators and member operators.

Global Operators

Global operators are those that are declared in module-level scope (globally). These are the operators - (Negate), Not (Bitwise Not), -> (Pointer To Member Access), * (Value Of), + (Add), - (Subtract), * (Multiply), / (Divide), \ (Integer Divide), & (Concatenate), Mod (Modulus), Shl (Shift Left), Shr (Shift Right), And (Bitwise And), Or (Bitwise Or), Xor (Bitwise Xor), Imp (Bitwise Imp), Eqv (Bitwise Eqv), ^ (Exponentiate), = (Equal), <> (Not Equal), < (Less Than), > (Greater Than), <= (Less Than Or Equal) and >= (Greater Than Or Equal).

Declaring a custom global operator is similar to declaring a procedure. The Declare keyword is used with the Operator keyword. The operator symbol is placed next followed by the comma-separated list of parameters surrounded in parenthesis that will represent the operands passed to the operator. Unlike procedures, operators can be overloaded by default, so the Overload keyword is not necessary when declaring custom operators. At least one of the operator's parameters must be of user-defined type (after all, operators with built-in type parameters are already defined).

The following example declares the global operators - (Negate) and + (Multiply) to accept operands of a user-defined type.

Type Rational
    As Integer numerator, denominator
End Type

Operator - (ByRef rhs As Rational) As Rational
    Return Type(-rhs.numerator, rhs.denominator)
End Operator

Operator * (ByRef lhs As Rational, ByRef rhs As Rational) As Rational
    Return Type(lhs.numerator * rhs.numerator, _
        lhs.denominator * rhs.denominator)
End Operator

Dim As Rational r1 = (2, 3), r2 = (3, 4)
Dim As Rational r3 = -(r1 * r2)
Print r3.numerator & "/" & r3.denominator

Here the global operators are defined for type Rational, and are used in the initialization expression for r3. The output is -6/12.

Member Operators

Member operators are declared inside a Type or Class definition, like member procedures, and they are the cast and assignment operators Let (Assign), Cast (Cast), += (Add And Assign), -= (Subtract And Assign), *= (Multiply And Assign), /= (Divide And Assign), \= (Integer Divide And Assign), ^= (Exponentiate And Assign), &= (Concat And Assign), Mod= (Modulus And Assign), Shl= (Shift Left And Assign), Shr= (Shift Right And Assign), And= (Conjunction And Assign), Or= (Inclusive Disjunction And Assign), Xor= (Exclusive Disjunction And Assign), Imp= (Implication And Assign) and Eqv= (Equivalence And Assign).

When declaring member operators, the Declare and Operator keywords are used followed by the operator symbol and its parameter list. Like member procedures, member operators are defined outside the Type or Class definition, and the symbol name is prefixed with the name of the Type or Class name.

The following example overloads the member operators Cast (Cast) and *= (Multiply And Assign) for objects of a user-defined type.

Type Rational
    As Integer numerator, denominator
    
    Declare Operator Cast () As Double
    Declare Operator Cast () As String
    Declare Operator *= (ByRef rhs As Rational)
End Type

Operator Rational.cast () As Double
    Return numerator / denominator
End Operator

Operator Rational.cast () As String
    Return numerator & "/" & denominator
End Operator

Operator Rational.*= (ByRef rhs As Rational)
    numerator *= rhs.numerator
    denominator *= rhs.denominator
End Operator

Dim As Rational r1 = (2, 3), r2 = (3, 4)
r1 *= r2
Dim As Double d = r1
Print r1, d

Notice that the member operator Cast (Cast) is declared twice, once for the conversion to Double and once for the conversion to String. This is the only operator (or procedure) that can be declared multiple times when only the return type differs. The compiler decides which cast overload to call based on how the object is used (in the initialization of the Double d, Rational.Cast as double is called, and in the Print statement, Rational.Cast as string is used instead).