Page 2 of 2
Posted: Tue Aug 28, 2007 10:12 am
by eboyd
oh shiza.
i addd the copy ctor for list like you pretty much had osdev64 and then all friggin hell broke loose.
if i let the compiler supply it, it seems fine. im lost.
which brings me to another question:
Do i want object declared like this:
Code: Select all
dlist<int> list;
dlist<int> list2 = list;
dlist<int> list3(list2);
to behave the SAME AS this:
Code: Select all
dlist<int> *list = new dlist<int>();
dlist<int> *list2 = list;
dlist<int> *list3(list2);
???????
It seems that the copy ctor is only envoked for objects like so:
Code: Select all
object myObject;
object another = myObject;
but not for:
Code: Select all
object *myObject = new object();
ojbject *another = myObject;
Somebody pleez explain what the frick is up with all this!?!
Posted: Tue Aug 28, 2007 6:55 pm
by frank
Code: Select all
object myObject;
object another = myObject;
is assigning the value of an object to an object so a copy would be necessary.
Code: Select all
object *myObject = new object();
ojbject *another = myObject;
is assigning the value of a pointer to a pointer. IE the only thing that should be copied is the address.
Posted: Thu Aug 30, 2007 10:43 am
by eboyd
I'm sorry. I guess I should have clarified my question.
Is the whole point of a copy constructor to STOP copied objects from referencing the same portion of memory?
So if myObj is a list from 1 - 10, and myObj2 is a copy of myObj, if myObj reverses its order, should myObj2 reverse as well?
Posted: Thu Aug 30, 2007 1:53 pm
by eboyd
If what I'm assuming in my previous reply is true, than this is what I've come up with:
Code: Select all
dlist(const dlist& src) : head(new type(T())), tail(new type(T())) {
clear();
for(type* temp = (src.head)->next; temp != src.tail; temp =temp->next){
pushb(temp->data);
}
}
Look good?
Posted: Sun Sep 02, 2007 2:48 am
by os64dev
Ok, this will problably not hlp much
but if you want to know what happens or need to happen then take a pencil and a piece of paper and draw the objects. trust me it will become very clear then.
Posted: Tue Sep 04, 2007 6:54 am
by Solar
eboyd wrote:Is the whole point of a copy constructor to STOP copied objects from referencing the same portion of memory?
The point of a copy constructor is to do what is "right" for the class in question. If the "right" thing goes beyond a shallow copy, you have to define a copy constructor of your own.
So if myObj is a list from 1 - 10, and myObj2 is a copy of myObj, if myObj reverses its order, should myObj2 reverse as well?
Up to you.
I'd say it would be confusing, but your design requirements might be different. (You'd better document it in your headers, though.)