Introduction to Functions

In C++, a function that is called MyFunction that returns void (ie, nothing) and does nothing looks like this:

1void MyFunction() {}

Here are examples of functions that return data. Note that we specify the return type in the function heading:

1float ReturnFloat() {
2  return 5.5f;
3}
4
5bool ReturnBool() {
6  return true;
7}

Similar to variables, return statements can be implicitly cast to the return type of the function:

1// This returns the integer 5
2int ReturnInt() {
3  return 5.5f;
4}

We invoke functions using their name, and the () operator:

1ReturnFloat();

Our program needs to define an entry point - the function that will be run when our program is executed.

This is typically a function that returns an int and is called main:

1int main() {
2  return 0;
3}

The return value of the main function is an exit code, which describes why our program stopped. Returning 0 indicates the program ran as expected. Any other number is an error code.

If the main function returns nothing, it is assumed to have returned 0. This is a behavior that is specific just to the main function. Every other function must return something of the type specified in its heading, or something that can be converted to that type.

1// This is valid
2int main() {}
3
4// This is not valid
5int SomeFunction() {}

We define parameters between the () in the function definition. We specify their type and name, and separate multiple parameters using commas. Below, we declare our function has two parameters, both of type int, which we're calling x and y:

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

We provide arguments for these parameters between the ( and ) when calling the function. Below, we pass 1 to x and 2 to y:

1int main() {
2  Add(1, 2); // 3
3}

We can define default values for parameters, making them optional. We do this using the = token within the function definition:

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

We can pass {} as an argument to use the default value in that position, or omit it entirely if it's at the end of our argument list:

1int Add(int x = 1, int y = 2) {
2  return x + y;
3}
4
5int main() {
6  // Use no defaults
7  Add(3, 4); // 7
8
9  // Use default for 1st parameter
10  Add({}, 5); // 6
11
12  // Use default for 2nd parameter
13  Add(2, {}); // 4
14  Add(2); // 4
15
16  // Use default for both parameters
17  Add({}, {}); // 3
18  Add({}); // 3
19  Add(); // 3
20}

Function return types can be set to auto which will ask the compiler to infer what it should be based on the return statements:

1// Compiler deduces it will return an int
2auto Add(int x, int y) {
3  return x + y;
4}

Like with variables, the compiler will attempt to implicitly convert function arguments and return values into the appropriate type. It will only present an error if it cannot do the conversion.

1// Will return 2
2int GetInt() {
3  return 2.5f;
4}
5
6int Add(int x, int y) {
7  return x + y;
8}

1// All expressions will return 2
2// they are equivalent to Add(2, 2)
3Add(2, 2);
4Add(2.5f, 2.5f);
5Add(2, 2.5f);
6Add(2.5, 2.5);

It's possible to have two functions with the same name in the same scope.

The way the compiler determines which function we want to call is by comparing the argument types to the parameter types:

1// Integer Overload
2int Add(int x, int y) {
3  return x + y;
4}
5
6// Float Overload
7float Add(float x, float y) {
8  return x + y;
9}
10
11Add(2, 2); // Calling Integer Overload
12Add(2.5f, 2.5f); // Calling Float Overload

If it is ambiguous what function we want to call, the compiler will throw an error. The following is ambiguous because the compiler doesn't know which variation of Add to use:

1Add(2, 2.5f);

It could convert the first argument to a float, and call the overloads that accept a (float, float) argument list. Or, it could convert the second argument to an int, and call the (int, int) overload.

It doesn't know what we want, so it throws an error.

1error: 'Add': overloaded functions have similar conversions
2could be 'float Add(float,float)'
3or       'int Add(int,int)'
4trying to match argument list '(int, float)'

The following also throws an error, as we provide two double arguments. A double can be converted to either a float or an int so again, we're being ambiguous:

1Add(2.5, 2.5);

1error: 'Add': ambiguous call to overloaded function
2could be 'float Add(float,float)'
3or       'int Add(int,int)'
4trying to match argument list '(double, double)'

We can remove the ambiguity by explicitly casting our arguments to the desired type.

1// Calling (int, int) Overload
2Add(2, static_cast<int>(2.5f));
3
4// Calling (float, float) Overload
5Add(
6  static_cast<float>(2.5),
7  static_cast<float>(2.5)
8);

Functions can be declared and defined as separate statements. A function declaration looks the same as a definition, but without the body:

1int Add(int x, int y);

A function's return type, name, and parameter types are often called the function prototype or function signature.

Function prototypes technically don't need to name their parameters:

1int Add(int, int);

But it's common to include the names, to remind ourselves (or other developers reading our code) what the parameters are for.

Function declarations allow our functions to be called even if they are not yet defined at the place where they are called.

This is referred to as a forward declaration:

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}

The difference between a declaration and a definition will become more important as we advance through this course.

We don't necessarily have access to all of our functions and variables from anywhere in our program.

Scopes define which variables are accessible, and from where. Scopes are typically created any time we have a set of braces, { and }

For example, a function body creates a scope.

Expressions can access their own scope, and any ancestor (parent, grandparent, etc) scope:

1// Global scope
2int GlobalInt{0};
3
4void MyFunction() {
5  // Local scope
6  int LocalInt{0};
7
8  // This is fine - we're accessing
9  // a variable in the parent scope
10  GlobalInt++;
11}

But they cannot access any descendant (child, grandchild, etc) scopes:

1void MyFunction() {
2  int LocalInt{0};
3
4  {
5    // Fine - LocalInt is in a parent scope
6    ++LocalInt;
7    int ChildInt{1};
8  }
9
10  // Error - we don't have access to child scope
11  ChildInt++;
12}

In this lesson, we've covered the fundamentals of functions. We've seen how to define functions, call them, and control their visibility using scopes. Key takeaways:

• Functions are defined with a return type, name, and parameter list
• Functions can be called using their name and providing arguments
• Functions can be overloaded by providing different parameter types
• Forward declarations allow functions to be used before they are defined
• Variables can be global (accessible everywhere) or local (accessible only within a specific scope)