Question

I don’t know why the destruction of an object in vector is called at the following point of time.

class Something
{
public:
    Something() {}
    ~Something()    { cout << "destruction called" << endl; }
};

int main()
{
    std::vector<Something> vec;
    Something sth1 = Something();   
    Something sth2 = Something();
    vec.push_back(sth1);
    vec.push_back(sth2);
    vec.clear();
}

After I push sth2, destruction for sth1 is called. Why? Shouldn’t sth1 be kept in vec[0]?

Answer 1

Because the vector has to resize its capacity to be able to store two elements instead of one. It allocates a new buffer, it copies the old buffer to the new one, removes the old buffer, and then adds the new object.