1 % The Rust Ownership Guide
3 This guide presents Rust's ownership system. This is one of Rust's most unique
4 and compelling features, with which Rust developers should become quite
5 acquainted. Ownership is how Rust achieves its largest goal, memory safety.
6 The ownership system has a few distinct concepts: **ownership**, **borrowing**,
7 and **lifetimes**. We'll talk about each one in turn.
11 Before we get to the details, two important notes about the ownership system.
13 Rust has a focus on safety and speed. It accomplishes these goals through many
14 "zero cost abstractions," which means that in Rust, abstractions cost as little
15 as possible in order to make them work. The ownership system is a prime example
16 of a zero cost abstraction. All of the analysis we'll talk about in this guide
17 is _done at compile time_. You do not pay any run-time cost for any of these
20 However, this system does have a certain cost: learning curve. Many new users
21 to Rust experience something we like to call "fighting with the borrow
22 checker," where the Rust compiler refuses to compile a program that the author
23 thinks is valid. This often happens because the programmer's mental model of
24 how ownership should work doesn't match the actual rules that Rust implements.
25 You probably will experience similar things at first. There is good news,
26 however: more experienced Rust developers report that once they work with the
27 rules of the ownership system for a period of time, they fight the borrow
28 checker less and less.
30 With that in mind, let's learn about ownership.
34 At its core, ownership is about 'resources.' For the purposes of the vast
35 majority of this guide, we will talk about a specific resource: memory. The
36 concept generalizes to any kind of resource, like a file handle, but to make it
37 more concrete, we'll focus on memory.
39 When your program allocates some memory, it needs some way to deallocate that
40 memory. Imagine a function `foo` that allocates four bytes of memory, and then
41 never deallocates that memory. We call this problem 'leaking' memory, because
42 each time we call `foo`, we're allocating another four bytes. Eventually, with
43 enough calls to `foo`, we will run our system out of memory. That's no good. So
44 we need some way for `foo` to deallocate those four bytes. It's also important
45 that we don't deallocate too many times, either. Without getting into the
46 details, attempting to deallocate memory multiple times can lead to problems.
47 In other words, any time some memory is allocated, we need to make sure that we
48 deallocate that memory once and only once. Too many times is bad, not enough
49 times is bad. The counts must match.
51 There's one other important detail with regards to allocating memory. Whenever
52 we request some amount of memory, what we are given is a handle to that memory.
53 This handle (often called a 'pointer', when we're referring to memory) is how
54 we interact with the allocated memory. As long as we have that handle, we can
55 do something with the memory. Once we're done with the handle, we're also done
56 with the memory, as we can't do anything useful without a handle to it.
58 Historically, systems programming languages require you to track these
59 allocations, deallocations, and handles yourself. For example, if we want some
60 memory from the heap in a language like C, we do this:
64 int *x = malloc(sizeof(int));
66 // we can now do stuff with our handle x
73 The call to `malloc` allocates some memory. The call to `free` deallocates the
74 memory. There's also bookkeeping about allocating the correct amount of memory.
76 Rust combines these two aspects of allocating memory (and other resources) into
77 a concept called 'ownership.' Whenever we request some memory, that handle we
78 receive is called the 'owning handle.' Whenever that handle goes out of scope,
79 Rust knows that you cannot do anything with the memory anymore, and so
80 therefore deallocates the memory for you. Here's the equivalent example in
84 # use std::boxed::Box;
90 The `Box::new` function creates a `Box<T>` (specifically `Box<int>` in this
91 case) by allocating a small segment of memory on the heap with enough space to
92 fit an `int`. But where in the code is the box deallocated? We said before that
93 we must have a deallocation for each allocation. Rust handles this for you. It
94 knows that our handle, `x`, is the owning reference to our box. Rust knows that
95 `x` will go out of scope at the end of the block, and so it inserts a call to
96 deallocate the memory at the end of the scope. Because the compiler does this
97 for us, it's impossible to forget. We always have exactly one deallocation
98 paired with each of our allocations.
100 This is pretty straightforward, but what happens when we want to pass our box
101 to a function? Let's look at some code:
104 # use std::boxed::Box;
106 let x = Box::new(5i);
111 fn add_one(mut num: Box<int>) {
116 This code works, but it's not ideal. For example, let's add one more line of
117 code, where we print out the value of `x`:
120 # use std::boxed::Box;
122 let x = Box::new(5i);
129 fn add_one(mut num: Box<int>) {
134 This does not compile, and gives us an error:
137 error: use of moved value: `x`
142 Remember, we need one deallocation for every allocation. When we try to pass
143 our box to `add_one`, we would have two handles to the memory: `x` in `main`,
144 and `num` in `add_one`. If we deallocated the memory when each handle went out
145 of scope, we would have two deallocations and one allocation, and that's wrong.
146 So when we call `add_one`, Rust defines `num` as the owner of the handle. And
147 so, now that we've given ownership to `num`, `x` is invalid. `x`'s value has
148 "moved" from `x` to `num`. Hence the error: use of moved value `x`.
150 To fix this, we can have `add_one` give ownership back when it's done with the
154 # use std::boxed::Box;
156 let x = Box::new(5i);
163 fn add_one(mut num: Box<int>) -> Box<int> {
170 This code will compile and run just fine. Now, we return a `box`, and so the
171 ownership is transferred back to `y` in `main`. We only have ownership for the
172 duration of our function before giving it back. This pattern is very common,
173 and so Rust introduces a concept to describe a handle which temporarily refers
174 to something another handle owns. It's called "borrowing," and it's done with
175 "references", designated by the `&` symbol.
179 Here's the current state of our `add_one` function:
182 fn add_one(mut num: Box<int>) -> Box<int> {
189 This function takes ownership, because it takes a `Box`, which owns its
190 contents. But then we give ownership right back.
192 In the physical world, you can give one of your possessions to someone for a
193 short period of time. You still own your possession, you're just letting someone
194 else use it for a while. We call that 'lending' something to someone, and that
195 person is said to be 'borrowing' that something from you.
197 Rust's ownership system also allows an owner to lend out a handle for a limited
198 period. This is also called 'borrowing.' Here's a version of `add_one` which
199 borrows its argument rather than taking ownership:
202 fn add_one(num: &mut int) {
207 This function borrows an `int` from its caller, and then increments it. When
208 the function is over, and `num` goes out of scope, the borrow is over.
212 Lending out a reference to a resource that someone else owns can be
213 complicated, however. For example, imagine this set of operations:
215 1. I acquire a handle to some kind of resource.
216 2. I lend you a reference to the resource.
217 3. I decide I'm done with the resource, and deallocate it, while you still have
219 4. You decide to use the resource.
221 Uh oh! Your reference is pointing to an invalid resource. This is called a
222 "dangling pointer" or "use after free," when the resource is memory.
224 To fix this, we have to make sure that step four never happens after step
225 three. The ownership system in Rust does this through a concept called
226 "lifetimes," which describe the scope that a reference is valid for.
228 Let's look at that function which borrows an `int` again:
231 fn add_one(num: &int) -> int {
236 Rust has a feature called 'lifetime elision,' which allows you to not write
237 lifetime annotations in certain circumstances. This is one of them. We will
238 cover the others later. Without eliding the lifetimes, `add_one` looks like
242 fn add_one<'a>(num: &'a int) -> int {
247 The `'a` is called a **lifetime**. Most lifetimes are used in places where
248 short names like `'a`, `'b` and `'c` are clearest, but it's often useful to
249 have more descriptive names. Let's dig into the syntax in a bit more detail:
255 This part _declares_ our lifetimes. This says that `add_one` has one lifetime,
256 `'a`. If we had two, it would look like this:
259 fn add_two<'a, 'b>(...)
262 Then in our parameter list, we use the lifetimes we've named:
265 ...(num: &'a int) -> ...
268 If you compare `&int` to `&'a int`, they're the same, it's just that the
269 lifetime `'a` has snuck in between the `&` and the `int`. We read `&int` as "a
270 reference to an int" and `&'a int` as "a reference to an int with the lifetime 'a.'"
272 Why do lifetimes matter? Well, for example, here's some code:
280 let y = &5i; // this is the same as `let _y = 5; let y = &_y;
281 let f = Foo { x: y };
287 As you can see, `struct`s can also have lifetimes. In a similar way to functions,
295 declares a lifetime, and
303 uses it. So why do we need a lifetime here? We need to ensure that any reference
304 to a `Foo` cannot outlive the reference to an `int` it contains.
306 ## Thinking in scopes
308 A way to think about lifetimes is to visualize the scope that a reference is
309 valid for. For example:
313 let y = &5i; // -+ y goes into scope
317 } // -+ y goes out of scope
328 let y = &5i; // -+ y goes into scope
329 let f = Foo { x: y }; // -+ f goes into scope
332 } // -+ f and y go out of scope
335 Our `f` lives within the scope of `y`, so everything works. What if it didn't?
336 This code won't work:
344 let x; // -+ x goes into scope
347 let y = &5i; // ---+ y goes into scope
348 let f = Foo { x: y }; // ---+ f goes into scope
349 x = &f.x; // | | error here
350 } // ---+ f and y go out of scope
352 println!("{}", x); // |
353 } // -+ x goes out of scope
356 Whew! As you can see here, the scopes of `f` and `y` are smaller than the scope
357 of `x`. But when we do `x = &f.x`, we make `x` a reference to something that's
358 about to go out of scope.
360 Named lifetimes are a way of giving these scopes a name. Giving something a
361 name is the first step towards being able to talk about it.
365 The lifetime named 'static' is a special lifetime. It signals that something
366 has the lifetime of the entire program. Most Rust programmers first come across
367 `'static` when dealing with strings:
370 let x: &'static str = "Hello, world.";
373 String literals have the type `&'static str` because the reference is always
374 alive: they are baked into the data segment of the final binary. Another
378 static FOO: int = 5i;
379 let x: &'static int = &FOO;
382 This adds an `int` to the data segment of the binary, and FOO is a reference to
387 In all the examples we've considered so far, we've assumed that each handle has
388 a singular owner. But sometimes, this doesn't work. Consider a car. Cars have
389 four wheels. We would want a wheel to know which car it was attached to. But
403 let car = Car { name: "DeLorean".to_string() };
405 for _ in range(0u, 4) {
406 Wheel { size: 360, owner: car };
411 We try to make four `Wheel`s, each with a `Car` that it's attached to. But the
412 compiler knows that on the second iteration of the loop, there's a problem:
415 error: use of moved value: `car`
416 Wheel { size: 360, owner: car };
418 note: `car` moved here because it has type `Car`, which is non-copyable
419 Wheel { size: 360, owner: car };
423 We need our `Car` to be pointed to by multiple `Wheel`s. We can't do that with
424 `Box<T>`, because it has a single owner. We can do it with `Rc<T>` instead:
439 let car = Car { name: "DeLorean".to_string() };
441 let car_owner = Rc::new(car);
443 for _ in range(0u, 4) {
444 Wheel { size: 360, owner: car_owner.clone() };
449 We wrap our `Car` in an `Rc<T>`, getting an `Rc<Car>`, and then use the
450 `clone()` method to make new references. We've also changed our `Wheel` to have
451 an `Rc<Car>` rather than just a `Car`.
453 This is the simplest kind of multiple ownership possible. For example, there's
454 also `Arc<T>`, which uses more expensive atomic instructions to be the
455 thread-safe counterpart of `Rc<T>`.
459 Earlier, we mentioned 'lifetime elision,' a feature of Rust which allows you to
460 not write lifetime annotations in certain circumstances. All references have a
461 lifetime, and so if you elide a lifetime (like `&T` instead of `&'a T`), Rust
462 will do three things to determine what those lifetimes should be.
464 When talking about lifetime elision, we use the term 'input lifetime' and
465 'output lifetime'. An 'input lifetime' is a lifetime associated with a parameter
466 of a function, and an 'output lifetime' is a lifetime associated with the return
467 value of a function. For example, this function has an input lifetime:
470 fn foo<'a>(bar: &'a str)
473 This one has an output lifetime:
476 fn foo<'a>() -> &'a str
479 This one has a lifetime in both positions:
482 fn foo<'a>(bar: &'a str) -> &'a str
485 Here are the three rules:
487 * Each elided lifetime in a function's arguments becomes a distinct lifetime
490 * If there is exactly one input lifetime, elided or not, that lifetime is
491 assigned to all elided lifetimes in the return values of that function..
493 * If there are multiple input lifetimes, but one of them is `&self` or `&mut
494 self`, the lifetime of `self` is assigned to all elided output lifetimes.
496 Otherwise, it is an error to elide an output lifetime.
500 Here are some examples of functions with elided lifetimes, and the version of
501 what the elided lifetimes are expand to:
504 fn print(s: &str); // elided
505 fn print<'a>(s: &'a str); // expanded
507 fn debug(lvl: uint, s: &str); // elided
508 fn debug<'a>(lvl: uint, s: &'a str); // expanded
510 // In the preceeding example, `lvl` doesn't need a lifetime because it's not a
511 // reference (`&`). Only things relating to references (such as a `struct`
512 // which contains a reference) need lifetimes.
514 fn substr(s: &str, until: uint) -> &str; // elided
515 fn substr<'a>(s: &'a str, until: uint) -> &'a str; // expanded
517 fn get_str() -> &str; // ILLEGAL, no inputs
519 fn frob(s: &str, t: &str) -> &str; // ILLEGAL, two inputs
521 fn get_mut(&mut self) -> &mut T; // elided
522 fn get_mut<'a>(&'a mut self) -> &'a mut T; // expanded
524 fn args<T:ToCStr>(&mut self, args: &[T]) -> &mut Command // elided
525 fn args<'a, 'b, T:ToCStr>(&'a mut self, args: &'b [T]) -> &'a mut Command // expanded
527 fn new(buf: &mut [u8]) -> BufWriter; // elided
528 fn new<'a>(buf: &'a mut [u8]) -> BufWriter<'a> // expanded