Gain insights into the C++ keyword 'const,' its usage in various contexts, and implications when returning pointers and references. Learn about preventing accidental data changes.

How to use const in C++ by Sándor Dargo_468 x 60 copy.png

In this course, you will learn about the C++ keyword const and its behaviour under different usage conditions, including local, global and static contexts. This course will help you to understand const and its implications when you attempt to return pointers and references in C++. By the end of the course, you will have a solid grip on const and its usage in preventing any accidental change of data.

Secure Coding Practices in C++: A Deep Dive into const

Let's explore when to use `const` parameters in the parameter list of function signatures. Let's differentiate between fundamental types (also known as primitive data types) and class types.

# Introduction
This section talks about primitive data types, such as `bool`, `int`, `float`, `char`, etc.

Should such parameters be taken as `const`?

The answer is no, because it will be inefficient. We can access these data types with one memory read if passed by value. On the other hand, if we pass them by reference or pointer, the address of the variable will be read, and then by dereferencing it, the value. That's two memory reads instead of only one.

We should not pass fundamental types by `const&`.

But should we pass them simply by `const`?

If we don't plan to modify the parameter's values, we could or we should! For better readability, for the compiler, and the future!

``` 
void setFoo(const int foo) {
   this->m_foo = foo;
}
```
It seems pedantic, but it doesn't hurt. It's explicit, and we don't know how the method will grow in the future. Maybe some additional checks will be added done, and even some exception handling.

If the parameter is not marked `const`, and the value changes, it might cause subtle errors. 

If we mark `foo` `const`, we will make this scenario impossible.

We will have to remove the `const` qualifier if we need to change the value of that parameter, but we'll do it intentionally. That change will stand out in a code review if we keep our pull request short.

On the other hand, if we have to modify the parameter, we will not mark it `const`.

# When not to use const parameters

From time to time, we can see the following pattern.

Don't make a parameter as `const` if a copy is needed.

```
void doSomething(const int foo) {
    // ...
    int foo2 = foo;
    foo2++;
    // ...
 }
```
There is no reason to take a `const` value if we plan to modify it. One more variable on the stack in vain, on more assignments without any reason. Simply take it by value.

# Parameters non-`const` for modification

Take value parameters non-`const` if we plan to modify them.

```
void doSomething(int foo) {
    // ...
    foo++;
    // ...
}
```
In short, we don't take fundamental types by `const&` or `const T*`. We only take them by `const T` when we don't plan to modify them.

Learn about const primitive data type parameters.

Introduction

Arguments using const

Local Variables

Local Members

Functions

Return Types

const Parameters

const and Smart Pointers

Rvalue References

Summary

const Primitive Data Type

Introduction