// - File 1 -
#include "a.h"
#include "b.h"
B b;
A::A(){
  b.Use();
}

// - File 2 -
#include "a.h"
A a;

// - File 3 -
#include "a.h"
#include "b.h"
extern A a;
extern B b;

int main() {
  a.Use();
  b.Use();
}

For this example, The interpretation for the deferred dynamic initialization may exist something wrong. The comment is:

If, however, a is initialized at some point after the first statement of main, b will be initialized prior to its use in A​::​A.

In other words, the sentence interprets that when the dynamic initialization is deferred, how to initialize these deferred variable. Although, the non-inline function A::A() is defined at the same TU as the variable b, however the invocation of A::A() is caused by the initialization of a, in other words, the use of A::A() shouldn't be considered as non-initialization odr-use, whose term definition is:

A non-initialization odr-use is an odr-use ([basic.def.odr]) not caused directly or indirectly by the initialization of a non-local static or thread storage duration variable.

Hence, in this situation, the use of A::A() shouldn't cause the initialization of the variable b. The use of b in expression b.Use() is caused indirectly by the initialization of the variable a(as aforementioned), which shouldn't trigger the initialization of the variable b. So, why the comment says that b will be initialized prior to its use in A​::​A?

Such an issues has been discussed at here.