POD members release moment in Destructor

[oseri]

I'm curious when are POD member variables released when calling a destructor.

class L2
{
public:
	int eCount;
        int GetCount() {return eCount;}
        ~L2()
        { std::cout << GetCount(); }
}

Are the POD members released at the beginning of the destructor call or at the end of it ? Do I risk printing garbage?

If they were released at the beginning, then writing something like that :

        ~L2()
        {  int a = 40;
           std::cout << GetCount(); }

would mean that I risk "a" will be pun on the released space of eCount and thus GetCount would print value of "a" - thats mostly what concerns me.

[Peter87]

They will be "released" after. You can always use data members of the class in the destructor, no matter if they are POD or not.

[Disch]

When an object is destroyed:

1) Your destructor body is run
2) Destructors are run for any member objects of the class (like if you have any strings/vectors/etc)
3) Destructor bodies for the parent class(es) are run
4) Destructors for the parent(s) members are run
5) Once all destructors are run, memory for the object is released.


So:

Are the POD members released at the beginning of the destructor call or at the end of it ?

The end. Having a destructor would be useless if you couldn't use your members in it.

Do I risk printing garbage?

No.