On enums in C++
Enums are tremendously useful to group names in your namespace. In C++ an enum is basically a dictionary where the name that you declare in the enum is the key for an index value. By default this is a zero-index incrementing by one. Hence:
enum gem_stones {
Quartz,
Ruby,
Diamond,
Jade,
};
Here Quartz=0, Ruby=1, Diamond=2, Jade=3.
But this default behaviour also leads to some logical problems. Say we define another set of enums later in the code:
enum baby_names {
Sarah,
Leila,
Jane,
Annabelle,
};
We will then have this ridiculous situation where:
if ( Quartz == Sarah ) {
std::cout << “ Quartz == Sarah is True “ << std::endl;
} else {
std::cout << “ Quartz == Sarah is False “ << std::endl;
}
Would compile and produce1:
Quartz == Sarah is True
A work around from C++11 is to define the start of the index. Hence:
enum gem_stones {
Quartz = 20, // 20
Ruby, // 21
Diamond, // 22
Jade, // 23
};
enum baby_names {
Sarah = 30, // 30
Leila, // 31
Jane, // 32
Annabelle, // 33
};
That seems to solve the problem. But you have to track the indexing that you have used and if you use many enums in your code, there is still a risk of collision in the index number.
C++11 also introduced enum Classes which solves that problem. If you think about it, a class is a collection of related names, functions etc. so it is perfectly logical that enums are a class.
So the above examples would be written as:
enum class gem_stones {
Quartz,
Ruby,
Diamond,
Jade,
};
enum class baby_names {
Sarah,
Leila,
Jane,
Annabelle,
};
Now any gem_stones::Quartz == baby_names::Sarah comparisons will throw a compile error. It is far safer to catch these kinds of problems at compile time.
-
In a modern
g++, you will get a “comparison of different enumeration types” warning, and if you use an LSP interface, you will get a real-time warning, but the code will compile. ↩︎