Sign in to follow this  

How does C# handle pointers internally?

This topic is 1861 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hey everyone,

I was refreshing on my pointer skills by reading about linked lists and it got me wondering how C# handles it's lists. Lets say I have a global list of type Foo
[code]
List<Foo> globalList = new List<Foo>();
[/code]
Now lets say I create an instantiate a Foo object within a local scope of a function.

[code]
void LocalFunction()
{
Foo myLocalObject = new myLocalObject();
//Do stuff with myLocalObject
globalList.Add(myLocalObject);
}
[/code]

From what I understand with c++, globalList.Add(myLocalObject); is passing the object by address (which is how c# passes classes anyway) so even though myLocalObject is no longer in scope, I can still access it and it's information by using globalList[whateverIndexItsat].whateverMethod();

I'm able to do this because C# has a pointer internally somewhere in which it is pointing to the memory created by that localObject. Is my understanding of this correct? Also, does the garbage collector manage the pointers by reference counting?

I'm just curious as to how everything is working here and the linked list datastructure in C++ is helping me understand where to use pointers a little better.

Thanks, hopefully I'm understanding this correctly.

Share this post


Link to post
Share on other sites
[url=http://msdn.microsoft.com/en-us/library/0xy59wtx.aspx].NET Garbage Collection[/url] is significantly more sophisticated than just reference counting.

Share this post


Link to post
Share on other sites
Pointers internally [url="http://msdn.microsoft.com/en-us/magazine/cc163791.aspx"]are pretty complicated too[/url].

The linked article is rather dated, it covers .NET 1.1 after all, but the basics of memory management are all there.

Share this post


Link to post
Share on other sites
Here is some more good reading on the subject of references and pointers in C#. Not necessarily directly related to your question, but good reading on the topics, nonetheless.

[url="http://blogs.msdn.com/b/ericlippert/archive/2009/02/17/references-are-not-addresses.aspx"]http://blogs.msdn.com/b/ericlippert/archive/2009/02/17/references-are-not-addresses.aspx[/url]
[url="http://blogs.msdn.com/b/ericlippert/archive/2011/03/07/references-and-pointers-part-one.aspx"]http://blogs.msdn.com/b/ericlippert/archive/2011/03/07/references-and-pointers-part-one.aspx[/url]
[url="http://blogs.msdn.com/b/ericlippert/archive/2011/03/10/references-and-pointers-part-two.aspx"]http://blogs.msdn.com/b/ericlippert/archive/2011/03/10/references-and-pointers-part-two.aspx[/url]

A slight aside, List<T> in C# is more analogous to std::vector in C++. It's not a linked list; it's just an indexable collection of objects. Generally speaking a List<T> in C# is just implemented as a raw T[] under the hood, handling resizing the array as necessary. If you wanted a linked list there is the aptly named [url="http://msdn.microsoft.com/en-us/library/he2s3bh7.aspx"]LinkedList<T>[/url] class.

Share this post


Link to post
Share on other sites
They aren't "pointers" in C#, they are references. This sounds like a non-important distinction, but it is important because C# has [i]actual pointers[/i] as well.

.NET does not use reference counting, it uses a generational garbage collector where objects are moved along the generations as they are no longer rooted to another non-collected object. You'll hear the terms "anchored" or "rooted" a lot when dealing with this kind of GC. You'll also hear terms like "gen1 object" or "gen2 object", and this just has to do with where in the GC process the object is. Note that this is all true for the reference implementation of the .NET framework. It's quite possible that projects like Mono have their own implementation, though I am almost positive that Mono uses a very similar generational GC.

That said, your understanding of what's going on from a high level is correct. When you put a reference to your object in to that list, the list now holds it's "root" (or one of them anyway, you may have references in other places). The fact that the variable the reference was assigned to is local means nothing. That local variable will fall out of scope, but the reference it was attached to will not be swept by the GC because you have it anchored in that list. Edited by GnomeTank

Share this post


Link to post
Share on other sites
Sign in to follow this