Forward Declarations

1int main() {
2  // Add is not defined yet
3  Add(1, 2);
4}
5
6int Add(int x, int y) {
7  return x + y;
8}

1Error: 'Add': identifier not found

Circular dependencies occur when two or more parts of our code reference each other in a cycle.

With functions, it might look something like this:

1void HandlePositive(int x) {
2  if (x < 0) HandleNegative(x);
3}
4
5void HandleNegative(int x){
6  if (x >= 0) HandlePositive(x);
7}

Here, HandlePositive calls HandleNegative and vice versa, creating a circular dependency.

This dependency cannot be resolved by merely reordering the functions, as each function requires the other to be declared beforehand.

The primary method C++ provides to let us solve circular dependencies is the ability to separate the declaration and definition of our function. This means we can declare a function in one region of code, but define it elsewhere.

We can think of a declaration as a minimalist description of our function, whilst the definition is the complete implementation.

For code to use our function, it only needs to be declared at the point of use - not fully defined. The full definition can come later in the file or, in larger projects, in a different file entirely.

Taking advantage of this mechanism is referred to as forward declaring the function.

To forward declare a function, we first need to understand what a function prototype is. This is sometimes also called the signature.

We can think of the prototype as just the heading of the function, without its body. For example, if we have the following Add function:

1int Add(int x, int y) {
2  return x + y;
3}

Its prototype would be this:

1int Add(int x, int y);

So, the prototype is the function's return type, the function name, and the parameter types.

In this example, the return type is int, the name is Add, and the parameter types are int, int

Within the function prototype, the names of our parameters are optional. They are typically worth including anyway for the benefit of anyone reading our code.

If the names are descriptive, including them helps anyone reading our prototype get a better understanding of how our function works.

But technically, only the types are required, so the following is also a valid prototype for our function:

1int Add(int, int);

1float Calculate(int x, float y) {
2  return x * y;
3}

Now, we no longer have to define our function before we use it - we just have to declare it, using its prototype.

Forward declarations are typically put near the top of the files that use them, but as long as they're provided before the functions are used, our code will work as expected:

1int Add(int x, int y);
2
3int main() {
4  Add(1, 2);
5}
6
7int Add(int x, int y) {
8  return x + y;
9}

1int main() {
2  int x { Calculate() };
3}
4
5int Calculate() {
6  return 10 * 10;
7}

1int Calculate();
2
3int main() {
4  int x { Calculate() };
5}
6
7int Calculate() {
8  return 10 * 10;
9}

This separation between declaring a function and defining it is likely to feel quite strange, especially at this stage.

However, this is a very important concept. Most C++ code is created this way. This is because the separation can have some major workflow benefits for larger projects that span many files. We'll see examples of this soon.

In this lesson, we've explored several key concepts regarding forward declarations in C++:

• The necessity of forward declarations when functions are used before their definitions.
• Understanding circular dependencies and their resolution using forward declarations.
• The distinction between declaring and defining functions.
• How function prototypes work, including their syntax and purpose.
• Practical application of forward declarations in code, demonstrating how they enable flexible ordering of function definitions.

This knowledge not only helps in writing more organized and maintainable code but also lays the groundwork for understanding more complex patterns we'll see later in the course.