Why is second initialization allowed in C++11

Question

Under C++11 it is possible to initialize class members directly upon declaration. But it is also ok to initialize them once more in the initialization list of the constructor... why?

#include <iostream>



struct MyStr
{

    MyStr()
    :j(0)
    {
        std::cout << "j is " << j << std::endl; // prints "j is 0"
    }

    const int j = 1;

};

int main()
{
    const int i = 0;

    MyStr mstr; 
}

Because doing something like this is an error, understandably:

MyStr()
:j(0),
j(1)
{
}

What is different about the first example, where the data member gets initialized upon declaration and then again in the constructor's init list?

Answer 1

Only one initialization actually happens. It's just that you're allowed to write a "default" one in the form of the brace-or-equals initializer, but if your constructor initializer list specifies an initializer, too, that one is the only one that's used.

As a side note, as of C++14, a brace-or-equals initializer can be provided for a non-static data member of an aggregate (which cannot have constructors).

Answer 2

So that an individual constructor may override it.

From the original feature proposal:

It may happen that a data member will usually have a particular value, but a few specialized constructors will need to be cognizant of that value. If a constructor initializes a particular member explicitly, the constructor initialization overrides the member initializations as shown below: [..]

Remember, you can have more than one constructor.