What does void mean in C, C++, and C#?

Question

Looking to get the fundamentals on where the term VOID comes from and why it would be called void. The intention of the question is to assist someone who has no C experience and is suddenly looking at a C-based codebase.

Answer 1

It means "no value". You use void to indicate that a function doesn't return a value or that it has no parameters or both. Pretty much consistent with typical uses of word void in English.

Answer 2

In c# you'd use the void keyword to indicate that a method does not return a value:

public void DoSomeWork()
{
//some work
}

Answer 3

Void is used only in method signatures. For return types it means method will not return anything to the calling code. For parameters it means, no parameters are passed to the method

e.g.

void MethodThatReturnsAndTakesVoid(void)
{
// Method body
}

In C# we can omit the void for parameters and can write the above code as:

void MethodThatReturnsAndTakesVoid()
{
// Method body
}

Void should not be confused with null. Null means for the variable whose address is on stack, the value on the heap for that address is empty.

Answer 4

It indicates the absence of a return value in a function.

Some languages have two sorts of subroutines: procedures and functions. Procedures are just a sequence of operations, whereas a function is a sequence of operations that return a result.

In C and its derivatives, the difference between the two is not explicit. Everything is basically a function. the void keyword indicates that it's not an "actual" function, since it doesn't return a value.

Answer 5

void mean that you won't be returning any value form the function or method

Answer 6

Basically it means "nothing" or "no type"

There are 3 basic ways that void is used:

1) Function argument: int myFunc(void) -- the function takes nothing.

2) Function return value: void myFunc(int) -- the function returns nothing

3) Generic data pointer: void* data; -- 'data' is a pointer to data of unknown type, and cannot be dereferenced

Note: the void in a function argument is optional in C++; int myFunc() is exactly the same as int myFunc(void). It is required for a return value.

Answer 7

Void means no value required in return type from a function in all of three language.

Answer 8

There are two ways to use void:

void foo(void);

or

void *bar(void*);

The first indicates that no argument is being passed or that no argument is being returned.

The second tells the compiler that there is no type associated with the data effectively meaning that the you can't make use of the data pointed to until it is cast to a known type.

For example you will see void* used allot when you have an interface which calls a function whose parameters can't be known ahead of time.

For example, in the Linux Kernel when deferring work you will setup a function to be run at a latter time by giving it a pointer to the function to be run and a pointer to the data to be passed to the function:

struct _deferred_work {
sruct list_head mylist;
.worker_func = bar;
.data        = somedata;
} deferred_work;

Then a kernel thread goes over a list of deferred work and when it get's to this node it effectively executes:

bar(somedata);

Then in bar you have:

void bar(void* mydata) {
    int *data = mydata;
    /* do something with data */;
}

Answer 9

Three usage cases for void:

1. Function signatures. void foo(int bar) does not return a value. int bar(void) does not take any parameters but this is usually expressed with empty argument list: int bar(). Usage of the void keyword here corresponds to its meaning in English.

2. Generic top-type pointer void * that points to unspecified data and cannot be dereferenced. Here the meaning of void is different from other meanings of void: universal type vs. no type.

3. In casts such as (void) new Foo(this) to signify that the return value is deliberately thrown away. Here the keyword usage also matches its meaning in English.

Cases 1 and 2 were already covered by @Gerald but case 3 has not been addressed yet.

Answer 10

Void is the equivalent of Visual Basic's Sub.

Answer 11

If you're explaining the concept to a beginner, it might be helpful to use an analogy. The use of void in all these cases is analogous in meaning to a page in a book which has the following words, "This page left intentionally blank." It is to differentiate to the compiler between something which should be flagged as an error, versus a type which is intentionally to be left blank because that is the behavior you want.

It always appears in code where normally you would expect to see a type appear, such as a return type or a pointer type. This is why in C#, void maps to an actual CLR type, System.Void because it is a type in itself.

Some programming languages never developed the concept of void, just like some human cultures never invented the concept of the number zero. Void represents the same advancement in a programming language as the concept of zero represents to human language.

Answer 12

I have always taken it to mean absent. Here are four cases in the C language that matches to this use of absent

• R f(void) - Function parameters are absent
• void f(P) - Return value is absent
• void *p - Type of what is pointed to is absent
• (void) p - Usage of value is absent

Other C descendants use it for other things. The D programming language uses it for cases where an initializer is absent

• T t = void; - initializing value is absent

Answer 13

Think of void as the "empty structure". Let me explain.

Every function takes a sequence of parameters, where each parameter has a type. In fact, we could package up the parameters into a structure, with the structure slots corresponding to the parameters. This makes every function have exactly one argument. Similarly, functions produce a result, which has a type. It could be a boolean, or it could be float, or it could be a structure, containing an arbitrary set of other typed values. If we want a languge that has multiple return values, it is easy to just insist they be packaged into a structure. In fact, we could always insist that a function returned a structure. Now every function takes exactly one argument, and produces exactly one value.

Now, what happens when I need a function that produces "no" value? Well, consider what I get when I form a struct with 3 slots: it holds 3 values. When I have 2 slots, it holds two values. When it has one slot, one value. And when it has zero slots, it holds... uh, zero values, or "no" value". So, I can think of a function returning void as returning a struct containing no values. You can even decide that "void" is just a synonym for the type represented by the empty structure, rather than a keyword in the language (maybe its just a predefined type :)

Similarly, I can think of a function requiring no values as accepting an empty structure, e.g., "void".

I can even implement my programming language this way. Passing a void value takes up zero bytes, so passing void values is just a special case of passing other values of arbitrary size. This makes it easy for the compiler to treat the "void" result or argument. You probably want a langauge feature that can throw a function result away; in C, if you call the non-void result function foo in the following statement: foo(...); the compiler knows that foo produces a result and simply ignores it. If void is a value, this works perfectly and now "procedures" (which are just an adjective for a function with void result) are just trivial special cases of general functions.

Void* is a bit funnier. I don't think the C designers thought of void in the above way; they just created a keyword. That keyword was available when somebody needed a point to an arbitrary type, thus void* as the idiom in C. It actually works pretty well if you interpret void as an empty structure. A void* pointer is the address of a place where that empty structure has been put.

Casts from void* to T* for other types T, also work out with this perspective. Pointer casts are a complete cheat that work on most common architectures to take advantage of the fact that if a compound type T has an element with subtype S placed physically at the beginning of T in its storage layout, then casting S* to T* and vice versa using the same physical machine address tends to work out, since most machine pointers have a single representation. Replacing the type S by the type void gives exactly the same effect, and thus casting to/from void* works out.

The PARLANSE programming language implements the above ideas pretty closely. We goofed in its design, and didn't pay close attention to "void" as a return type and thus have langauge keywords for procedure. Its mostly just a simple syntax change but its one of things you don't get around to once you get a large body working code in a language.