Why Use Sentinels Instead of End Iterators in C++?

Using sentinels instead of end iterators in C++ provides greater flexibility, especially in situations where the endpoint of a range is not known in advance or is determined dynamically. Sentinels can define stopping conditions that are more complex than simply reaching the end of a container.

For example, if you want to iterate through a range until a specific condition is met, such as encountering a negative number, a sentinel can be used to stop the iteration when that condition is fulfilled. Here’s a practical example:

1#include <iostream>
2#include <vector>
3
4struct Sentinel {
5  bool operator==(
6    std::vector<int>::const_iterator iter) const {
7    return *iter < 0;
8  }
9};
10
11int main() {
12  std::vector<int> numbers{1, 4, 3, 8, -2, 5};
13  Sentinel s;
14
15  for (auto it = numbers.begin(); it != s; ++it) {
16    std::cout << *it << ", ";
17  }
18}

11, 4, 3, 8,

In this case, the iteration stops as soon as the first negative number is encountered, thanks to the custom sentinel.

Using sentinels can simplify code, making it more readable and maintainable by clearly expressing the termination condition. They are also more versatile, allowing for a wider range of termination criteria beyond just reaching the end of a container. This makes sentinels a powerful tool for controlling the flow of algorithms in C++.