1 // Rust-101, Part 05: Clone
2 // ========================
6 //@ In the course of the next few parts, we are going to build a data-structure for computations
7 //@ with *big* numbers. We would like to not have an upper bound to how large these numbers can
8 //@ get, with the memory of the machine being the only limit.
10 //@ We start by deciding how to represent such big numbers. One possibility here is to use a vector
11 //@ "digits" of the number. This is like "1337" being a vector of four digits (1, 3, 3, 7), except
12 //@ that we will use `u64` as type of our digits, meaning we have 2^64 individual digits. Now we
13 //@ just have to decide the order in which we store numbers. I decided that we will store the least
14 //@ significant digit first. This means that "1337" would actually become (7, 3, 3, 1). <br/>
15 //@ Finally, we declare that there must not be any trailing zeros (corresponding to
16 //@ useless leading zeros in our usual way of writing numbers). This is to ensure that
17 //@ the same number can only be stored in one way.
19 //@ To write this down in Rust, we use a `struct`, which is a lot like structs in C:
20 //@ Just a bunch of named fields. Every field can be private to the current module (which is the
21 //@ default), or public (which is indicated by a `pub` in front of the name). For the sake of the
22 //@ tutorial, we make `data` public - otherwise, the next parts of this course could not work on
23 //@ `BigInt`s. Of course, in a real program, one would make the field private to ensure that the
24 //@ invariant (no trailing zeros) is maintained.
26 pub data: Vec<u64>, // least significant digit first, no trailing zeros
29 // Now that we fixed the data representation, we can start implementing methods on it.
31 //@ Let's start with a constructor, creating a `BigInt` from an ordinary integer.
32 //@ To create an instance of a struct, we write its name followed by a list of
33 //@ fields and initial values assigned to them.
34 pub fn new(x: u64) -> Self {
36 BigInt { data: vec![] } /*@*/
38 BigInt { data: vec![x] } /*@*/
42 //@ It can often be useful to encode the invariant of a data-structure in code, so here
43 //@ is a check that detects useless trailing zeros.
44 pub fn test_invariant(&self) -> bool {
45 if self.data.len() == 0 {
48 self.data[self.data.len() - 1] != 0 /*@*/
52 // We can convert any little-endian vector of digits (i.e., least-significant digit first) into
53 // a number, by removing trailing zeros. The `mut` declaration for `v` here is just like the
54 // one in `let mut ...`: We completely own `v`, but Rust still asks us to make our intention of
55 // modifying it explicit. This `mut` is *not* part of the type of `from_vec` - the caller has
56 // to give up ownership of `v` anyway, so they don't care anymore what you do to it.
58 // **Exercise 05.1**: Implement this function.
60 // *Hint*: You can use `pop` to remove the last element of a vector.
61 pub fn from_vec(mut v: Vec<u64>) -> Self {
67 //@ If you take a close look at the type of `BigInt::from_vec`, you will notice that it consumes
68 //@ the vector `v`. The caller hence loses access to its vector. However, there is something we can
69 //@ do if we don't want that to happen: We can explicitly `clone` the vector, which means that a
70 //@ full (or *deep*) copy will be performed. Technically, `clone` takes a borrowed vector in the
71 //@ form of a shared reference, and returns a fully owned one.
73 let v = vec![0,1 << 16];
74 let b1 = BigInt::from_vec((&v).clone());
75 let b2 = BigInt::from_vec(v);
77 //@ Rust has special treatment for methods that borrow their `self` argument (like `clone`, or
78 //@ like `test_invariant` above): It is not necessary to explicitly borrow the receiver of the
79 //@ method. Hence you could replace `(&v).clone()` by `v.clone()` above. Just try it!
81 //@ To be clonable is a property of a type, and as such, naturally expressed with a trait.
82 //@ In fact, Rust already comes with a trait `Clone` for exactly this purpose. We can hence
83 //@ make our `BigInt` clonable as well.
84 impl Clone for BigInt {
85 fn clone(&self) -> Self {
86 BigInt { data: self.data.clone() } /*@*/
89 //@ Making a type clonable is such a common exercise that Rust can even help you doing it:
90 //@ If you add `#[derive(Clone)]` right in front of the definition of `BigInt`, Rust will
91 //@ generate an implementation of `Clone` that simply clones all the fields. Try it!
92 //@ These `#[...]` annotations at types (and functions, modules, crates) are called *attributes*.
93 //@ We will see some more examples of attributes later.
95 // We can also make the type `SomethingOrNothing<T>` implement `Clone`.
96 //@ However, that can only work if `T` is `Clone`! So we have to add this bound to `T` when we
97 //@ introduce the type variable.
98 use part02::{SomethingOrNothing,Something,Nothing};
99 impl<T: Clone> Clone for SomethingOrNothing<T> {
100 fn clone(&self) -> Self {
102 Nothing => Nothing, /*@*/
103 //@ In the second arm of the match, we need to talk about the value `v`
104 //@ that's stored in `self`. However, if we were to write the pattern as
105 //@ `Something(v)`, that would indicate that we *own* `v` in the code
106 //@ after the arrow. That can't work though, we have to leave `v` owned by
107 //@ whoever called us - after all, we don't even own `self`, we just borrowed it.
108 //@ By writing `Something(ref v)`, we borrow `v` for the duration of the match
109 //@ arm. That's good enough for cloning it.
110 Something(ref v) => Something(v.clone()), /*@*/
114 //@ Again, Rust will generate this implementation automatically if you add
115 //@ `#[derive(Clone)]` right before the definition of `SomethingOrNothing`.
117 // **Exercise 05.2**: Write some more functions on `BigInt`. What about a function that returns the
118 // number of digits? The number of non-zero digits? The smallest/largest digit? Of course, these
119 // should all take `self` as a shared reference (i.e., in borrowed form).
121 // ## Mutation + aliasing considered harmful (part 2)
122 //@ Now that we know how to create references to contents of an `enum` (like `v` above), there's
123 //@ another example we can look at for why we have to rule out mutation in the presence of
124 //@ aliasing. First, we define an `enum` that can hold either a number, or a string.
129 //@ Now consider the following piece of code. Like above, `n` will be a reference to a part of
130 //@ `var`, and since we wrote `ref mut`, the reference will be unique and mutable. In other words,
131 //@ right after the match, `ptr` points to the number that's stored in `var`, where `var` is a
132 //@ `Number`. Remember that `_` means "we don't care".
133 fn work_on_variant(mut var: Variant, text: String) {
134 let mut ptr: &mut i32;
136 Variant::Number(ref mut n) => ptr = n,
137 Variant::Text(_) => return,
139 /* var = Variant::Text(text); */ /* BAD! */
142 //@ Now, imagine what would happen if we were permitted to also mutate `var`. We could, for
143 //@ example, make it a `Text`. However, `ptr` still points to the old location! Hence `ptr` now
144 //@ points somewhere into the representation of a `String`. By changing `ptr`, we manipulate the
145 //@ string in completely unpredictable ways, and anything could happen if we were to use it again!
146 //@ (Technically, the first field of a `String` is a pointer to its character data, so by
147 //@ overwriting that pointer with an integer, we make it a completely invalid address. When the
148 //@ destructor of `var` runs, it would try to deallocate that address, and Rust would eat your
149 //@ laundry - or whatever.)
151 //@ I hope this example clarifies why Rust has to rule out mutation in the presence of aliasing
152 //@ *in general*, not just for the specific case of a buffer being reallocated, and old pointers
153 //@ becoming hence invalid.
155 //@ [index](main.html) | [previous](part04.html) | [raw source](workspace/src/part05.rs) |
156 //@ [next](part06.html)