Can I use the same variable to initialise two class objects?

I have to create and initialise objects of a class called Device in my main program. But the number of objects created are based on the user input. The IDE is Qt so I have access to data structures like QLists.

What I have tried:

My idea was that in the header file I would declare a Device pointer like this

Code:

---

Device* temp;
QList<Device*>myList;

---

then in the source file, I would initialise this in a for loop based on the number of inputs. Like so:

Code:

---

for(int i = 0; i<inputVal; i++){
    temp = new Device;
    myList.append(temp);
}

---

I can't run the code yet as it is part of a bigger module, so I wanted to know if this is completely wrong from a theoretical perspective. I was hoping that I could reuse the variable name temp and in the end, temp would contain the pointer to the last Device object created, and to clean up the memory I would use a foreach loop on the QList.

Please advice!!

And what are going to do if some Device object is about to destruct?
Are you going to remove its pointer from the QList<Device*>? How?

Quote:

---

Originally Posted by VictorN

And what are going to do if some Device object is about to destruct?
Are you going to remove its pointer from the QList<Device*>? How?

---

I honestly hadn't considered this. I had only thought about cleaning up. Each Device object has a unique ID, which the user provides. I could save the Device pointers along with their IDs in a hash table of sorts (QHash or QMap<int,Device*>). This could allow me to delete the specific pointer. And I could add an event in the destructor, so I am informed if a particular object is deleted.
Does this sound right? Or am I overcomplicating what should be a straightforward task?

Quote:

---

Originally Posted by VBCherry

I honestly hadn't considered this. I had only thought about cleaning up. Each Device object has a unique ID, which the user provides. I could save the Device pointers along with their IDs in a hash table of sorts (QHash or QMap<int,Device*>). This could allow me to delete the specific pointer. And I could add an event in the destructor, so I am informed if a particular object is deleted.
Does this sound right? Or am I overcomplicating what should be a straightforward task?

---

I am not sure you need this global pointer

Code:

---

Device* temp;

---

at all.

Quote:

---

I have to create and initialise objects of a class called Device

---

Unless you need to store object pointers, why not just have a list of class Device?

Code:

---

QList<Device> myList;

---

or

Code:

---

QMap<int, Device> Mymap;

---

if you want easy access to a particular Device.

You then don't have to worry about memory management. If you want to remove an item just deleting the item in the container will work just fine.

Quote:

---

Originally Posted by VictorN

I am not sure you need this global pointer

---

I hadn't considered this. This would at least relieve my concern that I am reusing the same variable.

Quote:

---

Originally Posted by 2kaud

Unless you need to store object pointers, why not just have a list of class Device?

---

If I use pointers I then don't have to worry about the object going out of scope somewhere and getting deleted. I know that until I call delete, the pointer stays put. I could be totally wrong on this though.

Have you considered using smart pointers?