This section covers all the operators in C and C++.

All operators produce a value from their operands. This value is produced without modifying the operands, except with the assignment, increment, and decrement operators. Modifying an operand is called a side effect. The most common use for operators that modify their operands is to generate the side effect, but you should keep in mind that the value produced is available for your use just as in operators without side effects.

Assignment is performed with the operator =. It means “Take the right-hand side (often called the rvalue) and copy it into the left-hand side (often called the lvalue).” An rvalue is any constant, variable, or expression that can produce a value, but an lvalue must be a distinct, named variable (that is, there must be a physical space in which to store data). For instance, you can assign a constant value to a variable (A = 4;), but you cannot assign anything to constant value – it cannot be an lvalue (you can’t say 4 = A;).

The basic mathematical operators are the same as the ones available in most programming languages: addition (+), subtraction (-), division (/), multiplication (*), and modulus (%; this produces the remainder from integer division). Integer division truncates the result (it doesn’t round). The modulus operator cannot be used with floating-point numbers.

C and C++ also use a shorthand notation to perform an operation and an assignment at the same time. This is denoted by an operator followed by an equal sign, and is consistent with all the operators in the language (whenever it makes sense). For example, to add 4 to the variable x and assign x to the result, you say: x += 4;.

This example shows the use of the mathematical operators:

The rvalues of all the assignments can, of course, be much more complex.

Notice the use of the macro PRINT( ) to save typing (and typing errors!). Preprocessor macros are traditionally named with all uppercase letters so they stand out – you’ll learn later that macros can quickly become dangerous (and they can also be very useful).

The arguments in the parenthesized list following the macro name are substituted in all the code following the closing parenthesis. The preprocessor removes the name PRINT and substitutes the code wherever the macro is called, so the compiler cannot generate any error messages using the macro name, and it doesn’t do any type checking on the arguments (the latter can be beneficial, as shown in the debugging macros at the end of the chapter).

Relational operators establish a relationship between the values of the operands. They produce a Boolean (specified with the bool keyword in C++) true if the relationship is true, and false if the relationship is false. The relational operators are: less than (<), greater than (>), less than or equal to (<=), greater than or equal to (>=), equivalent (==), and not equivalent (!=). They may be used with all built-in data types in C and C++. They may be given special definitions for user-defined data types in C++ (you’ll learn about this in Chapter 12, which covers operator overloading).

The logical operators and (&&) and or (||) produce a true or false based on the logical relationship of its arguments. Remember that in C and C++, a statement is true if it has a non-zero value, and false if it has a value of zero. If you print a bool, you’ll typically see a ‘1’ for true and ‘0’ for false.

This example uses the relational and logical operators:

You can replace the definition for int with float or double in the program above. Be aware, however, that the comparison of a floating-point number with the value of zero is strict; a number that is the tiniest fraction different from another number is still “not equal.” A floating-point number that is the tiniest bit above zero is still true.

The bitwise operators allow you to manipulate individual bits in a number (since floating point values use a special internal format, the bitwise operators work only with integral types: char, int and long). Bitwise operators perform Boolean algebra on the corresponding bits in the arguments to produce the result.

The bitwise and operator (&) produces a one in the output bit if both input bits are one; otherwise it produces a zero. The bitwise or operator (|) produces a one in the output bit if either input bit is a one and produces a zero only if both input bits are zero. The bitwise exclusive or, or xor (^) produces a one in the output bit if one or the other input bit is a one, but not both. The bitwise not (~, also called the ones complement operator) is a unary operator – it only takes one argument (all other bitwise operators are binary operators). Bitwise not produces the opposite of the input bit – a one if the input bit is zero, a zero if the input bit is one.

Bitwise operators can be combined with the = sign to unite the operation and assignment: &=, |=, and ^= are all legitimate operations (since ~ is a unary operator it cannot be combined with the = sign).

The shift operators also manipulate bits. The left-shift operator (<<) produces the operand to the left of the operator shifted to the left by the number of bits specified after the operator. The right-shift operator (>>) produces the operand to the left of the operator shifted to the right by the number of bits specified after the operator. If the value after the shift operator is greater than the number of bits in the left-hand operand, the result is undefined. If the left-hand operand is unsigned, the right shift is a logical shift so the upper bits will be filled with zeros. If the left-hand operand is signed, the right shift may or may not be a logical shift (that is, the behavior is undefined).

Shifts can be combined with the equal sign (<<= and >>=). The lvalue is replaced by the lvalue shifted by the rvalue.

What follows is an example that demonstrates the use of all the operators involving bits. First, here’s a general-purpose function that prints a byte in binary format, created separately so that it may be easily reused. The header file declares the function:

Here’s the implementation of the function:

The printBinary( ) function takes a single byte and displays it bit-by-bit. The expression

produces a one in each successive bit position; in binary: 00000001, 00000010, etc. If this bit is bitwise anded with val and the result is nonzero, it means there was a one in that position in val.

Finally, the function is used in the example that shows the bit-manipulation operators:

Once again, a preprocessor macro is used to save typing. It prints the string of your choice, then the binary representation of an expression, then a newline.

In main( ), the variables are unsigned. This is because, in general, you don't want signs when you are working with bytes. An int must be used instead of a char for getval because the “cin >>” statement will otherwise treat the first digit as a character. By assigning getval to a and b, the value is converted to a single byte (by truncating it).

The << and >> provide bit-shifting behavior, but when they shift bits off the end of the number, those bits are lost (it’s commonly said that they fall into the mythical bit bucket, a place where discarded bits end up, presumably so they can be reused...). When manipulating bits you can also perform rotation, which means that the bits that fall off one end are inserted back at the other end, as if they’re being rotated around a loop. Even though most computer processors provide a machine-level rotate command (so you’ll see it in the assembly language for that processor), there is no direct support for “rotate” in C or C++. Presumably the designers of C felt justified in leaving “rotate” off (aiming, as they said, for a minimal language) because you can build your own rotate command. For example, here are functions to perform left and right rotations:

Try using these functions in Bitwise.cpp. Notice the definitions (or at least declarations) of rol( ) and ror( ) must be seen by the compiler in Bitwise.cpp before the functions are used.

The bitwise functions are generally extremely efficient to use because they translate directly into assembly language statements. Sometimes a single C or C++ statement will generate a single line of assembly code.

Bitwise not isn’t the only operator that takes a single argument. Its companion, the logical not (!), will take a true value and produce a false value. The unary minus (-) and unary plus (+) are the same operators as binary minus and plus; the compiler figures out which usage is intended by the way you write the expression. For instance, the statement

has an obvious meaning. The compiler can figure out:

but the reader might get confused, so it is safer to say:

The unary minus produces the negative of the value. Unary plus provides symmetry with unary minus, although it doesn’t actually do anything.

The increment and decrement operators (++ and --) were introduced earlier in this chapter. These are the only operators other than those involving assignment that have side effects. These operators increase or decrease the variable by one unit, although “unit” can have different meanings according to the data type – this is especially true with pointers.

The last unary operators are the address-of (&), dereference (* and ->), and cast operators in C and C++, and new and delete in C++. Address-of and dereference are used with pointers, described in this chapter. Casting is described later in this chapter, and new and delete are introduced in Chapter 4.

The ternary if-else is unusual because it has three operands. It is truly an operator because it produces a value, unlike the ordinary if-else statement. It consists of three expressions: if the first expression (followed by a ?) evaluates to true, the expression following the ? is evaluated and its result becomes the value produced by the operator. If the first expression is false, the third expression (following a :) is executed and its result becomes the value produced by the operator.

The conditional operator can be used for its side effects or for the value it produces. Here’s a code fragment that demonstrates both:

Here, the conditional produces the rvalue. a is assigned to the value of b if the result of decrementing b is nonzero. If b became zero, a and b are both assigned to -99. b is always decremented, but it is assigned to -99 only if the decrement causes b to become 0. A similar statement can be used without the “a =” just for its side effects:

Here the second B is superfluous, since the value produced by the operator is unused. An expression is required between the ? and :. In this case, the expression could simply be a constant that might make the code run a bit faster.

The comma is not restricted to separating variable names in multiple definitions, such as

Of course, it’s also used in function argument lists. However, it can also be used as an operator to separate expressions – in this case it produces only the value of the last expression. All the rest of the expressions in the comma-separated list are evaluated only for their side effects. This example increments a list of variables and uses the last one as the rvalue:

In general, it’s best to avoid using the comma as anything other than a separator, since people are not used to seeing it as an operator.

As illustrated above, one of the pitfalls when using operators is trying to get away without parentheses when you are even the least bit uncertain about how an expression will evaluate (consult your local C manual for the order of expression evaluation).

Another extremely common error looks like this:

The statement a = b will always evaluate to true when b is non-zero. The variable a is assigned to the value of b, and the value of b is also produced by the operator =. In general, you want to use the equivalence operator == inside a conditional statement, not assignment. This one bites a lot of programmers (however, some compilers will point out the problem to you, which is helpful).

A similar problem is using bitwise and and or instead of their logical counterparts. Bitwise and and or use one of the characters (& or |), while logical and and or use two (&& and ||). Just as with = and ==, it’s easy to just type one character instead of two. A useful mnemonic device is to observe that “Bits are smaller, so they don’t need as many characters in their operators.”

The word cast is used in the sense of “casting into a mold.” The compiler will automatically change one type of data into another if it makes sense. For instance, if you assign an integral value to a floating-point variable, the compiler will secretly call a function (or more probably, insert code) to convert the int to a float. Casting allows you to make this type conversion explicit, or to force it when it wouldn’t normally happen.

To perform a cast, put the desired data type (including all modifiers) inside parentheses to the left of the value. This value can be a variable, a constant, the value produced by an expression, or the return value of a function. Here’s an example:

Casting is powerful, but it can cause headaches because in some situations it forces the compiler to treat data as if it were (for instance) larger than it really is, so it will occupy more space in memory; this can trample over other data. This usually occurs when casting pointers, not when making simple casts like the one shown above.

C++ has an additional casting syntax, which follows the function call syntax. This syntax puts the parentheses around the argument, like a function call, rather than around the data type:

Of course in the case above you wouldn’t really need a cast; you could just say 200f (in effect, that’s typically what the compiler will do for the above expression). Casts are generally used instead with variables, rather than constants.

Casts should be used carefully, because what you are actually doing is saying to the compiler “Forget type checking – treat it as this other type instead.” That is, you’re introducing a hole in the C++ type system and preventing the compiler from telling you that you’re doing something wrong with a type. What’s worse, the compiler believes you implicitly and doesn’t perform any other checking to catch errors. Once you start casting, you open yourself up for all kinds of problems. In fact, any program that uses a lot of casts should be viewed with suspicion, no matter how much you are told it simply “must” be done that way. In general, casts should be few and isolated to the solution of very specific problems.

Once you understand this and are presented with a buggy program, your first inclination may be to look for casts as culprits. But how do you locate C-style casts? They are simply type names inside of parentheses, and if you start hunting for such things you’ll discover that it’s often hard to distinguish them from the rest of your code.

Standard C++ includes an explicit cast syntax that can be used to completely replace the old C-style casts (of course, C-style casts cannot be outlawed without breaking code, but compiler writers could easily flag old-style casts for you). The explicit cast syntax is such that you can easily find them, as you can see by their names:

The first three explicit casts will be described more completely in the following sections, while the last one can be demonstrated only after you’ve learned more, in Chapter 15.

A static_cast is used for all conversions that are well-defined. These include “safe” conversions that the compiler would allow you to do without a cast and less-safe conversions that are nonetheless well-defined. The types of conversions covered by static_cast include typical castless conversions, narrowing (information-losing) conversions, forcing a conversion from a void*, implicit type conversions, and static navigation of class hierarchies (since you haven’t seen classes and inheritance yet, this last topic will be delayed until Chapter 15):

In Section (1), you see the kinds of conversions you’re used to doing in C, with or without a cast. Promoting from an int to a long or float is not a problem because the latter can always hold every value that an int can contain. Although it’s unnecessary, you can use static_cast to highlight these promotions.

Converting back the other way is shown in (2). Here, you can lose data because an int is not as “wide” as a long or a float; it won’t hold numbers of the same size. Thus these are called narrowing conversions. The compiler will still perform these, but will often give you a warning. You can eliminate this warning and indicate that you really did mean it using a cast.

Assigning from a void* is not allowed without a cast in C++ (unlike C), as seen in (3). This is dangerous and requires that programmers know what they’re doing. The static_cast, at least, is easier to locate than the old standard cast when you’re hunting for bugs.

Section (4) of the program shows the kinds of implicit type conversions that are normally performed automatically by the compiler. These are automatic and require no casting, but again static_cast highlights the action in case you want to make it clear what’s happening or hunt for it later.

If you want to convert from a const to a nonconst or from a volatile to a nonvolatile, you use const_cast. This is the only conversion allowed with const_cast; if any other conversion is involved it must be done using a separate expression or you’ll get a compile-time error.

If you take the address of a const object, you produce a pointer to a const, and this cannot be assigned to a nonconst pointer without a cast. The old-style cast will accomplish this, but the const_cast is the appropriate one to use. The same holds true for volatile.

This is the least safe of the casting mechanisms, and the one most likely to produce bugs. A reinterpret_cast pretends that an object is just a bit pattern that can be treated (for some dark purpose) as if it were an entirely different type of object. This is the low-level bit twiddling that C is notorious for. You’ll virtually always need to reinterpret_cast back to the original type (or otherwise treat the variable as its original type) before doing anything else with it.

In this simple example, struct X just contains an array of int, but when you create one on the stack as in X x, the values of each of the ints are garbage (this is shown using the print( ) function to display the contents of the struct). To initialize them, the address of the X is taken and cast to an int pointer, which is then walked through the array to set each int to zero. Notice how the upper bound for i is calculated by “adding” sz to xp; the compiler knows that you actually want sz pointer locations greater than xp and it does the correct pointer arithmetic for you.

The idea of reinterpret_cast is that when you use it, what you get is so foreign that it cannot be used for the type’s original purpose unless you cast it back. Here, we see the cast back to an X* in the call to print, but of course since you still have the original identifier you can also use that. But the xp is only useful as an int*, which is truly a “reinterpretation” of the original X.

A reinterpret_cast often indicates inadvisable and/or nonportable programming, but it’s available when you decide you have to use it.

The sizeof operator stands alone because it satisfies an unusual need. sizeof gives you information about the amount of memory allocated for data items. As described earlier in this chapter, sizeof tells you the number of bytes used by any particular variable. It can also give the size of a data type (with no variable name):

By definition, the sizeof any type of char (signed, unsigned or plain) is always one, regardless of whether the underlying storage for a char is actually one byte. For all other types, the result is the size in bytes.

Note that sizeof is an operator, not a function. If you apply it to a type, it must be used with the parenthesized form shown above, but if you apply it to a variable you can use it without parentheses:

sizeof can also give you the sizes of user-defined data types. This is used later in the book.

This is an escape mechanism that allows you to write assembly code for your hardware within a C++ program. Often you’re able to reference C++ variables within the assembly code, which means you can easily communicate with your C++ code and limit the assembly code to that necessary for efficiency tuning or to use special processor instructions. The exact syntax that you must use when writing the assembly language is compiler-dependent and can be discovered in your compiler’s documentation.

These are keywords for bitwise and logical operators. Non-U.S. programmers without keyboard characters like &, |, ^, and so on, were forced to use C’s horrible trigraphs, which were not only annoying to type, but obscure when reading. This is repaired in C++ with additional keywords:

If your compiler complies with Standard C++, it will support these keywords.