From: Ralf Jung Date: Tue, 16 Jun 2015 08:53:34 +0000 (+0200) Subject: write a lot on Copy and Clone. Now part 05 is too long... X-Git-Url: https://git.ralfj.de/rust-101.git/commitdiff_plain/63b086dfbd9a5e90700f595c2e6ef7ee10522559 write a lot on Copy and Clone. Now part 05 is too long... --- diff --git a/src/main.rs b/src/main.rs index 15b798c..bf1c08c 100644 --- a/src/main.rs +++ b/src/main.rs @@ -68,7 +68,8 @@ // * [Part 01](part01.html) // * [Part 02](part02.html) // * [Part 03](part03.html) -// * [Part 04](part04.html) (WIP) +// * [Part 04](part04.html) +// * [Part 05](part05.html) (WIP) // * (to be continued) #![allow(dead_code, unused_imports, unused_variables, unused_mut)] mod part00; diff --git a/src/part01.rs b/src/part01.rs index 1603c14..8b14050 100644 --- a/src/part01.rs +++ b/src/part01.rs @@ -87,7 +87,7 @@ pub fn main() { // You will have to replace `part00` by `part01` in the `main` function in // `main.rs` to run this code. -// **Exercise**: Write a funtion `vec_avg` that computes the average value of a `Vec`. +// **Exercise 01.1**: Write a funtion `vec_avg` that computes the average value of a `Vec`. // // *Hint*: `vec.len()` returns the length of a vector `vec`. diff --git a/src/part02.rs b/src/part02.rs index 233fa0f..d0cbbf7 100644 --- a/src/part02.rs +++ b/src/part02.rs @@ -30,7 +30,7 @@ type NumberOrNothing = SomethingOrNothing; // The types are so similar, that we can provide a generic function to construct a `SomethingOrNothing` // from an `Option`, and vice versa. -// **Exercise**: Implement such functions! I provided a skeleton of the solution. Here, +// **Exercise 02.1**: Implement such functions! I provided a skeleton of the solution. Here, // `panic!` is another macro. This one terminates execution with the given message. // // Notice the syntax for giving generic implementations to generic types: Think of the first `` diff --git a/src/part03.rs b/src/part03.rs index e34ef58..ddaa47e 100644 --- a/src/part03.rs +++ b/src/part03.rs @@ -88,7 +88,7 @@ pub fn main() { min.print(); } -// **Exercise**: Define a trait `Print` to write a generic version of `SomethingOrNothing::print`. +// **Exercise 03.1**: Define a trait `Print` to write a generic version of `SomethingOrNothing::print`. // Implement that trait for `i32`, and change the code above to use it. // I will again provide a skeleton for this solution. It also shows how to attach bounds to generic // implementations (just compare it to the `impl` block from the previous exercise). diff --git a/src/part05.rs b/src/part05.rs index 25c98e2..f031f09 100644 --- a/src/part05.rs +++ b/src/part05.rs @@ -1,4 +1,4 @@ -// Rust-101, Part 05: Copy, Clone +// Rust-101, Part 05: Clone, Copy // ============================== use std::cmp; @@ -86,16 +86,28 @@ impl Clone for BigInt { // Making a type clonable is such a common exercise that Rust can even help you doing it: // If you add `#[derive(Clone)]' right in front of the definition of `BigInt`, Rust will // generate an implementation of `clone` that simply clones all the fields. Try it! -// -// To put this in perspective, `clone` in Rust corresponds to what people usually manually do in -// the copy constructor of a C++ class: It creates new, independent instance containing the -// same values. Contrary to that, if you pass something to a function normally (like the -// second call to `from_vec` in `clone_demo`), only a *shallow* copy is created: The fields -// are copied, but pointers are simply duplicated. This corresponds to the default copy -// constructor in C++. Rust assumes that after such a copy, the old value is useless -// (as the new one uses the same pointers), and hence considers the data semantically -// moved to the copy. That's another explanation of why Rust does not let you access -// a vector anymore after you passed ownership to some function. + +// We can also make the type `SomethingOrNothing` implement `Clone`. However, that +// can only work if `T` is `Clone`! So we have to add this bound to `T` when we introduce +// the type variable. +use part02::{SomethingOrNothing,Something,Nothing}; +impl Clone for SomethingOrNothing { + fn clone(&self) -> Self { + match *self { + Nothing => Nothing, + // In the second arm of the match, we need to talk about the value `v` + // that's stored in `self`. However, if we would write the pattern as + // `Something(v)`, that would indicate that we *own* `v` in the code + // after the arrow. That can't work though, we have to leave `v` owned by + // whoever called us - after all, we don't even own `self`, we just borrowed it. + // By writing `Something(ref v)`, we just borrow `v` for the duration of the match + // arm. That's good enough for cloning it. + Something(ref v) => Something(v.clone()), + } + } +} +// Again, Rust will generate this implementation automatically if you add +// `#[derive(Clone)]` right before the definition of `SomethingOrNothing`. // With `BigInt` being about numbers, we should be able to write a version of `vec_min` // that computes the minimum of a list of `BigInt`. We start by writing `min` for @@ -118,16 +130,16 @@ impl BigInt { } else if self.data.len() > other.data.len() { other } else { - // **Exercise**: Fill in this code. + // **Exercise 05.1**: Fill in this code. panic!("Not yet implemented."); } } } +// Now we can write `vec_min`. In order to make it type-check, we have to write it as follows. fn vec_min(v: &Vec) -> Option { let mut min: Option = None; for e in v { - // In the loop, `e` now has type `&i32`, so we have to dereference it. min = Some(match min { None => e.clone(), Some(n) => e.clone().min(n) @@ -135,3 +147,68 @@ fn vec_min(v: &Vec) -> Option { } min } +// Now, what's happening here? Why do we have to write `clone()`, and why did we not +// have to write that in our previous version? +// +// The answer is already hidden in the type of `vec_min`: `v` is just borrowed, but +// the Option that it returns is *owned*. We can't just return one +// of the elements of `v`, as that would mean that it is no longer in the vector! +// In our code, this comes up when we update the intermediate variable `min`, which +// also has type `Option`. If you replace `e.clone()` in the `None` arm +// with `*e`, Rust will complain "Cannot move out of borrowed content". That's because +// `e` is a `&BigInt`. Assigning `min` to `*e` works just like a function call: +// Ownership of the underlying data (in this case, the digits) is transferred from +// the vector to `min`. But that's not allowed, since we must retain the vector +// in its existing state. After cloning `e`, we own the copy that was created, +// and hence we can store it in `min`.
+// Of course, making such a full copy is expensive, so we'd like to avoid it. +// That's going to happen in the next part. +// +// But before we go there, I should answer the second question I brought up above: +// Why did our old `vec_min` work? We stored the minimal `i32` locally without +// cloning, and Rust did not complain. That's because there isn't really much +// of an "ownership" when it comes to types like `i32` or `bool`: If you move +// the value from one place to another, then both instance are independent +// and complete instances of their type. This is in stark contrast to types +// like `Vec`, where merely moving the value results in both the old +// and the new vector to point to the same underlying buffer. +// +// Rust calls types like `i32` that can be freely duplicated `Copy` types. +// `Copy` is another trait, and it is implemented for the basic types of +// the language. Remember how we defined the trait `Minimum` by writing +// `trait Minimum : Copy { ...`? This tells Rust that every type that +// implements `Minimum` must also implement `Copy`, and that's why Rust +// accepted our generic `vec_min` in part 02. +// +// Curiously, `Copy` is a trait that does not require any method to +// be implemented. Implementing `Copy` is merely a semantic statement, +// saying that the idea of ownership does not really apply to this type. +// Traits without methods are called *marker traits*. We will see some +// more examples of such traits later. +// +// If you try to implement `Copy` for `BigInt`, you will notice that Rust +// does not let you do that. A type can only be `Copy` if all its elements +// are `Copy`, and that's not the case for `BigInt`. However, we can make +// `SomethingOrNothing` copy if `T` is `Copy`. +impl Copy for SomethingOrNothing{} +// Again, Rust can generate implementations of `Copy` automatically. If +// you add `#[derive(Copy,Clone)]` right before the definition of `SomethingOrNothing`, +// both `Copy` and `Clone` will automatically be implemented. + +// In closing this part, I'd like to give you another perspective on the +// move semantics (i.e., ownership passing) that Rust applies, and how +// `Copy` and `Clone` fit.
+// When Rust code is executed, passing a value (like `i32` or `Vec`) +// to a function will always result in a shallow copy being performed: Rust +// just copies the bytes representing that value, and considers itself done. +// That's just like the default copy constructor in C++. Rust, however, will +// consider this a destructive operation: After copying the bytes elsewhere, +// the original value must no longer be used. After all, the two could not +// share a pointer! If, however, you mark a type `Copy`, then Rust will *not* +// consider a move destructive, and just like in C++, the old and new value +// can happily coexist. Now, Rust does not allow to to overload the copy +// constructor. This means that passing a value around will always be a fast +// operation, no allocation or copying of large data of the heap will happen. +// In the situations where you would write a copy constructor in C++ (and hence +// incur a hidden cost on every copy of this type), you'd have the type *not* +// implement `Copy`, but only `Clone`. This makes the cost explicit.