RustBrock/Reference Counter Smart Pointer.md

RC<T>, the Reference Counted Smart Pointer

In the majority of times ownership is clear. You always know which variable owns a value.

There are cases when a single value might have multiple owners.

An example of this includes in graph data structures, multiple edges might point to the same node, and that node is conceptually owned by all of the edges that point to it.

A node shouldn't be cleaned up unless it doesn't have any edges pointing to it and so has no owners.

You must explicitly enable multiple ownership by using the Rust type Rc<T>, this is a abbreviation for reference counting.

This type keeps track of the number of references to a value to determine whether or not the value is still in use.

If there are no references to a value, then the value can be cleaned up without any references becoming invalid.

Rc<T> can be pictured as a TV in a family room.

When one person enters to watch the TV, they turn it on.

Others can come into the room and watch the TV.

When the last person leaves the room, they turn it off because it is no longer being used.

If someone turns off the TV while others are still watching it, there would be uproar from the remaining TV watchers.

Rc<T> type is used when we want to allocate some data on the heap for multiple parts of our program to read.

We can't determine at compile time which part will finish using the data last.

If we knew which part would finished last, we could just make that part the data's owner and the normal ownership rules enforced at compile time would take effect.

Note that Rc<T> is only for use in single-threaded cases.

We discuss later how to do reference counting in multithreaded programs.

Using Rc<T> to Share Data

Lets use the cons list again to exemplify why to use a Rc<T>.

Recall that we previously defined it using Box<T>.

Now we will create two lists that both share ownership of a third list.

Here is a graph of what it would look like Here we create a list a that contains 5 then 10.

We then create two more lists

• b that starts with 3
• c that starts with 4 Both b and c will continue on to the first a list containing 5 and 10.

Both lists will share the first list containing 5 and 10.

Trying to implement this with our definition of List with Box<T> will not work nor compile

enum List {
    Cons(i32, Box<List>),
    Nil,
}

use crate::List::{Cons, Nil};

fn main() {
    let a = Cons(5, Box::new(Cons(10, Box::new(Nil))));
    let b = Cons(3, Box::new(a));
    let c = Cons(4, Box::new(a));
}

When we try to compile this code we get this error

$ cargo run
   Compiling cons-list v0.1.0 (file:///projects/cons-list)
error[E0382]: use of moved value: `a`
  --> src/main.rs:11:30
   |
9  |     let a = Cons(5, Box::new(Cons(10, Box::new(Nil))));
   |         - move occurs because `a` has type `List`, which does not implement the `Copy` trait
10 |     let b = Cons(3, Box::new(a));
   |                              - value moved here
11 |     let c = Cons(4, Box::new(a));
   |                              ^ value used here after move

For more information about this error, try `rustc --explain E0382`.
error: could not compile `cons-list` (bin "cons-list") due to 1 previous error

The Cons variants own the data they hold so when we create the b list a is moved into b and b owns a.

We could change the definition of Cons to hold a reference instead. We would then have to specify lifetime parameters.

Then by specifying lifetime parameters we would be specifying that every element in the list will live at least as long as the entire list.

This would be the case for the example above, but not in every scenario.

Instead we will change our definition of List to use Rc<T> in place of Box<T>.

Each Cons variant will now hold a value and an Rc<T> pointing to a List.

Now when we create b, instead of taking ownership of a.

We will clone the Rc<List> that a is holding, this increases the number of references from 1 -> 2 and letting a and b share ownership of the data in that Rc<List>.

Next we will clone a when creating c, increasing again the number of references form 2 -> 3.

Every time we call Rc::clone, the references count to the data within the Rc<List> will increase, and the data will not be cleaned up unless there are zero references to it.

enum List {
    Cons(i32, Rc<List>),
    Nil,
}

use crate::List::{Cons, Nil};
use std::rc::Rc;

fn main() {
    let a = Rc::new(Cons(5, Rc::new(Cons(10, Rc::new(Nil)))));
    let b = Cons(3, Rc::clone(&a));
    let c = Cons(4, Rc::clone(&a));
}

We also need to add a use statement to bring Rc<T> into scope because it is not in the prelude.

Now in main we create the list holding 5 and 10 and store it in a new Rc<List> in a.

Then we create a b and c, we call the Rc::clone function and pass a reference to the Rc<List> in a as an arg.

We could have used a.clone() rather than Rc::clone(&a).

It is convention in Rust to use Rc::clone in this case.

The implementation of Rc::clone doesn't make a deep copy of all data, this is unlike most tpyes' implementations of clone.

The call to Rc::clone only increments the reference count, this doesn't take much time.

Deep copies take much more time than a shallow copy.

By using Rc::clone for reference counting, we can visually distinguish between the deep-copy kinds of clines and the kinds of clones that increase the reference count.

When looking for performance problems in the code, we only need to consider the deep copy clones and can disregard class to Rc::clone

Cloning an Rc<T> Increases the Reference Count

We will now change the example above to see the reference counts changing as we create and drop references to the Rc<List> in a

Here we changed it so that there is an inner scope around list c.

This is used so that we can see a difference when dropping c

enum List {
    Cons(i32, Rc<List>),
    Nil,
}

use crate::List::{Cons, Nil};
use std::rc::Rc;

fn main() {
    let a = Rc::new(Cons(5, Rc::new(Cons(10, Rc::new(Nil)))));
    println!("count after creating a = {}", Rc::strong_count(&a));
    let b = Cons(3, Rc::clone(&a));
    println!("count after creating b = {}", Rc::strong_count(&a));
    {
        let c = Cons(4, Rc::clone(&a));
        println!("count after creating c = {}", Rc::strong_count(&a));
    }
    println!("count after c goes out of scope = {}", Rc::strong_count(&a));
}

At each point in the program where the ref count changes we print out the count so that we can see it.

We do this by calling the Rc::strong_count function.

This is named string_count rather than count because the Rc<T> type also has a weak_count.

We will see weak_count later.

Here is the output

$ cargo run
   Compiling cons-list v0.1.0 (file:///projects/cons-list)
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.45s
     Running `target/debug/cons-list`
count after creating a = 1
count after creating b = 2
count after creating c = 3
count after c goes out of scope = 2

We cam see that the a of type Rc<List> has an initial count of 1.

Then each time we call clone, it goes up by 1.

When c goes out of scope, the count goes down by 1

We don't have to call a function to decrease the reference count like how we have to call Rc::clone to increase the ref count.

The implementation of the Drop trait decreases the ref count automatically when an Rc<T> value goes out of scope.

In this example we can't see what happens when b and a goes out of scope at the end of main, when the count is 0.

The Rc<List> is cleaned up completely.

Using Rc<T> allows a single value to have multiple owners, and the count ensures that the value remains valid as long as any of the owners still exist.

Using immutable references Rc<T> allows you to share data between multiple parts of your program for read only.

If Rc<T> allowed for multiple mutable references, you have the possibility of violating borrowing rules.

Multiple mutable borrows to the same place can cause data races and inconsistencies. Section Reference Here

Being able to mutate data is very useful.

The nest section will the interior mutability pattern and the RefCell<T> type that you can use in conjunction with an Rc<T> to work with this immutability restriction.

It can be found here.