Why we can use the declared variable in its constructor in C++

Question

I recently ran into some weird bugs due to this kind of code.

vector<int> a(a);

Why the code above is accepted? When is this necessary? And how to ask the compiler to forbid this kind of usage? Thank you.

Answer 1

The most common use case for that is actually in C, not C++ (you should not use malloc in C++, although you can), but in this case C++ is backwards compatible:

mytype *p = malloc(sizeof *p);

By allowing the use of p in the initializer expression, you can pass sizeof *p to malloc ensuring that whatever the size of mytype is, the right size will be allocated. If you were not allowed to do *p, then the expression above would have to have the type mytype twice, and later maintenance in the code might update one of the types and fail to update the other.