Question

Consider the code below in a single translation unit:

class C {
private:

    struct Init {

        Init() {
            /* compute data once here */
        }
    };
    static const Init& i;
    static int data[];
public:
    /* interface for reading data */
};

const C::Init& C::i = Init();
int C::data[200];

1. Is C::i always initialized after C::data no matter the order of the definition of both?
2. Is this solution the most elegant one for computing static data once?

Answer 1

int C::data[200] is zero-initialized, which means that it is statically initialized. Static initialization comes before dynamic initialization. Since C::Init::Init() is not a constant expression, C::i is dynamically initialized, necessarily after C::data.

See 3.6.2 for details.

A bootleg quote:

Variables with static storage duration […] shall be zero-initialized before any other initialization takes place. […] Together, zero-initialization and constant initialization are called static initialization; all other initialization is dynamic initialization. Static initialization shall be performed before any dynamic initialization takes place.