How Can I Redefine a Classs Outside of A Loop?

Question

I am getting redefinition of class error for the below code. But if I put this definition in a loop there is no error. What is the difference?

using namespace std;

#include <iostream>

using namespace std;

struct MyClass
{
    int x;
};


int main()
{
    MyClass A;
    MyClass A;
    return 0;
}

I tried this but still there is same redefinition error:

#include <iostream>

using namespace std;

struct MyClass
{
    int x;
};


int main()
{
    MyClass* A;
    delete(A);
    MyClass* A;
    delete(A);
    return 0;
}

However If I put it in a loop no problem at all:

using namespace std;

struct MyClass
{
    int x;
};


int main()
{
    for (int i = 0; i < 5; i++) {
        MyClass A;
    }
    return 0;
}

So what is the difference? And how can I delete and redefine a class with the same name. I don't want to find different names. I will use it outside of a loop and push_back those in a vector.

What I want to do is actually:

#include <iostream>
#include <vector>

using namespace std;

struct MyClass
{
    int x;
    // other variables vectors etc.
};


vector <MyClass> v_A;

int main()
{
    MyClass A;
    // Use A's methods and give values to its variables and vectors.
    v_A.push_back(A);

    MyClass A2;
    // Use A's DIFFERENT methods and give values to its variables and vectors.
    v_A.push_back(A2);

    MyClass A3;
    // Use A's AGAIN DIFFERENT methods and give values to its variables and vectors.
    v_A.push_back(A3);

    // Goes on. I don't want to create different objects A A2 A3 ... A30. I don't want to keep them in the memory.
    // I don't want to write 2 3 ... 30
    // Is there a way to get rid of them from memory?

    return 0;
}

UPDATING THE CODE TO INCLUDE A SOLUTION DEPENDING ON linsock's ANSWER:

#include <iostream>
#include <vector>

using namespace std;

struct MyClass
{
    int x;
    // other variables vectors etc.
};


vector <MyClass> v_A;

int main()
{
    {
        MyClass A;
        // Use A's methods and give values to its variables and vectors.
        v_A.push_back(A);
    }

    {
        MyClass A;
        // Use A's DIFFERENT methods and give values to its variables and vectors.
        v_A.push_back(A);
    }

    {
        MyClass A;
        // Use A's AGAIN DIFFERENT methods and give values to its variables and vectors.
        v_A.push_back(A);
    }

    // Goes on. I dont want to create different classes A A2 A3 ... A30. I dont want to keep them in the memory.
    // I dont want to write 2 3 ... 30
    // Is there a way to get rid of them from memeory?

    return 0;
}

Answer 1

Well, here

    MyClass A;
    MyClass A;

you're defining the same class variable twice (which by the way should be highlighted by your compiler with a great compilation failure message)

The same you're doing here, but worse

 MyClass* A;
 delete(A);
 MyClass* A;
 delete(A);

since you're also trying to delete the pointer you defined but never initialized to point to a valid memory area.

In the last case of the for loop (which could be any block scope for what concern object creation/destruction) you're creating a stack allocated object, which internally contains only primitive types, then each cycle it gets created and destroyied with default ctor/dtor provided by the compiler.

I would suggest you to find a good tutorial to better understand how you're supposed to deal with this kind of behaviors in C++ .

EDIT from comments: delete will delete the data pointed-by the pointer you're trying to delete. If you want to define the same variable one after the other you should define the first in a block-scoped section in this way

 {
     MyClass* A;
 }
 MyClass* A;

In this way, the inner-defined MyClass* A will be lost after the exit from the scope and you can define it in the outer scope.

Even if this is actually possible, I would suggest you don't do that to enhance code readability.