Could somebody please clarify my thoughts on this.
With arrays, you can either declare a constant value of them such as [10] which is hard coded into the program. If I wanted variable amount of memory at runtime, I would define the array as 'new int[n]' for example, and then define n at run time. Is that correct?
Thanks.
yes, you but you need to declare a variable of type int*
like
Code:int* myArray = NULL; int n = 10; myArray = new int[n];
Thought for the day:FLTK: "The most fun you can have with your clothes on.""Are you sure your sanity chip is fully screwed in sir?" (Kryten)
Stroustrup:
"If I had thought of it and had some marketing sense every computer and just about any gadget would have had a little 'C++ Inside' sticker on it'"
But at run time, I could update the value of n before the array is created?Originally Posted by rogster001
You could do that. Also remember to use nullptr and not NULL.
If you only really need dynamic memory (ie an array whose size will expand if you need to store more data in it), std::vector is fine.
And if you don't use them, don't forget smart pointers!
Thanks. Yeah, I would definately almost always use vectors. I'm just practising memory management with new/delete.
I've not got as far as smart pointers just yet. Do they perform similar tasks to what vectors do? Ie, knowing array size, etc to help with out of bounds?
They release memory for you automatically.
No more memory leaks.Code:{ std::shared_ptr<int> p(new int); } // Memory is released
Okay, thats handy. A little bit similar to garbage collection that i've read about in other languages?
Should I always try to use smart pointers as above, in place of the standard type?
Thanks.
Somewhat. The fundamental difference is deterministic destruction. We know that the memory will be released when the last pointer to the memory is destroyed. Hence we will only be using as much memory as we need at any point in time. Garbage collection typically runs now and then and gathers unused memory and frees it. That means at any given point in time, your application may be using more memory than you need / should be using.
Yes, you should always strive to use smart pointers with raw pointers. This becomes fundamentally important later when when you start using exceptions and multiple exit paths in your functions. When you use smart pointers, you can be sure that your memory will be freed, no matter how you exit your functions.
Thanks. Will have a read up on them.
