Question about linked-list copy ctor

What is the typical way of handling list copying?

Do you normally clear() and then swap(const List& rhs)?

What I don't understand about that is clear() should effectively delete all the nodes in the list, but what happens with swap() after clear()? Does the rhs list receive a list that doesn't exist?

My first thought then is to handle it in this manner:

Code:

---

template <typename T>
List<T>::List(const List& rhs_l) : head(0), tail(0), sizeOfList(0){

    clear();

    for(Iter i(rhs_l.head); i != end(); i++){

        push_back(*i);

    }
}

---

This is the copy constructor. This creates a completely fresh object - what is there to clear?

Edit: you are probably confusing copy assignment (the assignment operator) and copy constructing.

Thanks anon,

so this is the correct copy ctor, but I should get rid of clear()?

Your copy routine is not correct (or at least it looks suspicious). Use a bog-standard iterator loop for that.

And yes, get rid of the clear.

Errr, I'm getting a seg fault when I call the copy ctor like this:

Code:

---

List<int> m_list2( m_list );

---

That's with or without the use of clear() in the copy ctor:

Code:

---

template <typename T>

List<T>::List(const List& rhs_l) : head(0), tail(0), sizeOfList(0){

   

    //clear();

   

    for(Iter i(rhs_l.head); i != end(); i++){

       

        push_back(*i);

    }

}

---

should head and tail be initialized in some other way?

lf I change the constructor to this:

Code:

---

template <typename T>

List<T>::List(const List& rhs_l) : sizeOfList(0){

       

        head = tail = new Node(NULL);

   

    for(Iter i(rhs_l.head); i != end(); i++){

       

        push_back(*i);

    }

}

---

I don't get the seg fault but of course now I have an erroneous node in the list that I don't want.

What's gong on here?

Quote:

---

Originally Posted by CornedBee

Use a bog-standard iterator loop for that.

---

a bog-standard iterator?

Perhaps this?

Code:

---

  for(Iter i(rhs_l.head); i != rhs_l.end(); i++){

---

--
Mats

As I said, your loop looks suspicious.

But it could also be that push_back tries to set the prev pointer on the tail node.


Bog-standard iterator loop.

Code:

---

for(whatever::iterator it = rhs.begin(); it != rhs.end(); ++it)

---

Quote:

---

Originally Posted by matsp

Perhaps this?

Code:

---

  for(Iter i(rhs_l.head); i != rhs_l.end(); i++){

---

--
Mats

---

Yea, that fixed it. I had previously changed end() to return null (not tail->next) so I didn't need to prefix it with rhs, but that was all a stupid mistake.