2021-01-01 11:28:10 +00:00
<!DOCTYPE HTML>
< html lang = "en" class = "sidebar-visible no-js rust" >
< head >
<!-- Book generated using mdBook -->
< meta charset = "UTF-8" >
< title > Rust Design Patterns< / title >
< meta name = "robots" content = "noindex" / >
<!-- Custom HTML head -->
< meta content = "text/html; charset=utf-8" http-equiv = "Content-Type" >
< meta name = "description" content = "A catalogue of Rust design patterns, anti-patterns and idioms" >
< meta name = "viewport" content = "width=device-width, initial-scale=1" >
< meta name = "theme-color" content = "#ffffff" / >
< link rel = "icon" href = "favicon.svg" >
< link rel = "shortcut icon" href = "favicon.png" >
< link rel = "stylesheet" href = "css/variables.css" >
< link rel = "stylesheet" href = "css/general.css" >
< link rel = "stylesheet" href = "css/chrome.css" >
< link rel = "stylesheet" href = "css/print.css" media = "print" >
<!-- Fonts -->
< link rel = "stylesheet" href = "FontAwesome/css/font-awesome.css" >
< link rel = "stylesheet" href = "fonts/fonts.css" >
<!-- Highlight.js Stylesheets -->
< link rel = "stylesheet" href = "highlight.css" >
< link rel = "stylesheet" href = "tomorrow-night.css" >
< link rel = "stylesheet" href = "ayu-highlight.css" >
<!-- Custom theme stylesheets -->
< / head >
< body >
<!-- Provide site root to javascript -->
< script type = "text/javascript" >
var path_to_root = "";
var default_theme = window.matchMedia("(prefers-color-scheme: dark)").matches ? "navy" : "rust";
< / script >
<!-- Work around some values being stored in localStorage wrapped in quotes -->
< script type = "text/javascript" >
try {
var theme = localStorage.getItem('mdbook-theme');
var sidebar = localStorage.getItem('mdbook-sidebar');
if (theme.startsWith('"') & & theme.endsWith('"')) {
localStorage.setItem('mdbook-theme', theme.slice(1, theme.length - 1));
}
if (sidebar.startsWith('"') & & sidebar.endsWith('"')) {
localStorage.setItem('mdbook-sidebar', sidebar.slice(1, sidebar.length - 1));
}
} catch (e) { }
< / script >
<!-- Set the theme before any content is loaded, prevents flash -->
< script type = "text/javascript" >
var theme;
try { theme = localStorage.getItem('mdbook-theme'); } catch(e) { }
if (theme === null || theme === undefined) { theme = default_theme; }
var html = document.querySelector('html');
html.classList.remove('no-js')
html.classList.remove('rust')
html.classList.add(theme);
html.classList.add('js');
< / script >
<!-- Hide / unhide sidebar before it is displayed -->
< script type = "text/javascript" >
var html = document.querySelector('html');
var sidebar = 'hidden';
if (document.body.clientWidth >= 1080) {
try { sidebar = localStorage.getItem('mdbook-sidebar'); } catch(e) { }
sidebar = sidebar || 'visible';
}
html.classList.remove('sidebar-visible');
html.classList.add("sidebar-" + sidebar);
< / script >
< nav id = "sidebar" class = "sidebar" aria-label = "Table of contents" >
< div class = "sidebar-scrollbox" >
2021-01-07 08:04:55 +00:00
< ol class = "chapter" > < li class = "chapter-item expanded " > < a href = "intro.html" > < strong aria-hidden = "true" > 1.< / strong > Introduction< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/index.html" > < strong aria-hidden = "true" > 2.< / strong > Idioms< / a > < / li > < li > < ol class = "section" > < li class = "chapter-item expanded " > < a href = "idioms/coercion-arguments.html" > < strong aria-hidden = "true" > 2.1.< / strong > Use borrowed types for arguments< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/concat-format.html" > < strong aria-hidden = "true" > 2.2.< / strong > Concatenating Strings with format!< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/ctor.html" > < strong aria-hidden = "true" > 2.3.< / strong > Constructor< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/default.html" > < strong aria-hidden = "true" > 2.4.< / strong > The Default Trait< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/deref.html" > < strong aria-hidden = "true" > 2.5.< / strong > Collections Are Smart Pointers< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/dtor-finally.html" > < strong aria-hidden = "true" > 2.6.< / strong > Finalisation in Destructors< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/mem-replace.html" > < strong aria-hidden = "true" > 2.7.< / strong > mem::replace(_)< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/on-stack-dyn-dispatch.html" > < strong aria-hidden = "true" > 2.8.< / strong > On-Stack Dynamic Dispatch< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/ffi-intro.html" > < strong aria-hidden = "true" > 2.9.< / strong > Foreign function interface usage< / a > < / li > < li > < ol class = "section" > < li class = "chapter-item expanded " > < a href = "idioms/ffi-errors.html" > < strong aria-hidden = "true" > 2.9.1.< / strong > Idiomatic Errors< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/ffi-accepting-strings.html" > < strong aria-hidden = "true" > 2.9.2.< / strong > Accepting Strings< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/ffi-passing-strings.html" > < strong aria-hidden = "true" > 2.9.3.< / strong > Passing Strings< / a > < / li > < / ol > < / li > < li class = "chapter-item expanded " > < a href = "idioms/option-iter.html" > < strong aria-hidden = "true" > 2.10.< / strong > Iterating over an Option< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/pass-var-to-closure.html" > < strong aria-hidden = "true" > 2.11.< / strong > Pass Variables to Closure< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/priv-extend.html" > < strong aria-hidden = "true" > 2.12.< / strong > Privacy For Extensibility< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/rustdoc-init.html" > < strong aria-hidden = "true" > 2.13.< / strong > Easy doc initialization< / a > < / li > < li class = "chapter-item expanded " > < a href = "idioms/temporary-mutability.html" > < strong aria-hidden = "true" > 2.14.< / strong > Temporary mutability< / a > < / li > < / ol > < / li > < li class = "chapter-item expanded " > < a href = "patterns/index.html" > < strong aria-hidden = "true" > 3.< / strong > Design Patterns< / a > < / li > < li > < ol class = "section" > < li class = "chapter-item expanded " > < a href = "patterns/builder.html" > < strong aria-hidden = "true" > 3.1.< / strong > Builder< / a > < / li > < li class = "chapter-item expanded " > < a href = "patterns/compose-structs.html" > < strong aria-hidden = "true" > 3.2.< / strong > Compose Structs< / a > < / li > < li class = "chapter-item expanded " > < a href = "patterns/entry.html" > < strong aria-hidden = "true" > 3.3.< / strong > Entry API< / a > < / li > < li class = "chapter-item expanded " > < a href = "patterns/ffi-intro.html" > < strong aria-hidden = "true" > 3.4.< / strong > Foreign function interface usage< / a > < / li > < li > < ol class = "section" > < li class = "chapter-item expanded " > < a href = "patterns/ffi-export.html" > < strong aria-hidden = "true" > 3.4.1.< / strong > Object-Based APIs< / a > < / li > < li class = "chapter-item expanded " > < a href = "patterns/ffi-wrappers.html" > < strong aria-hidden = "true" > 3.4.2.< / strong > Type Consolidation into Wrappers< / a > < / li > < / ol > < / li > < li class = "chapter-item expanded " > < a href = "patterns/fold.html" > < strong aria-hidden = "true" > 3.5.< / strong > Fold< / a > < / li > < li class = "chapter-item expanded " > < a href = "patterns/newtype.html" > < strong aria-hidden = "true" > 3.6.< / strong > Newtype< / a > < / li > < li class = "chapter-item expanded " > < a
2021-01-01 11:28:10 +00:00
< / div >
< div id = "sidebar-resize-handle" class = "sidebar-resize-handle" > < / div >
< / nav >
< div id = "page-wrapper" class = "page-wrapper" >
< div class = "page" >
< div id = "menu-bar-hover-placeholder" > < / div >
< div id = "menu-bar" class = "menu-bar sticky bordered" >
< div class = "left-buttons" >
< button id = "sidebar-toggle" class = "icon-button" type = "button" title = "Toggle Table of Contents" aria-label = "Toggle Table of Contents" aria-controls = "sidebar" >
< i class = "fa fa-bars" > < / i >
< / button >
< button id = "theme-toggle" class = "icon-button" type = "button" title = "Change theme" aria-label = "Change theme" aria-haspopup = "true" aria-expanded = "false" aria-controls = "theme-list" >
< i class = "fa fa-paint-brush" > < / i >
< / button >
< ul id = "theme-list" class = "theme-popup" aria-label = "Themes" role = "menu" >
< li role = "none" > < button role = "menuitem" class = "theme" id = "light" > Light< / button > < / li >
< li role = "none" > < button role = "menuitem" class = "theme" id = "rust" > Rust (default)< / button > < / li >
< li role = "none" > < button role = "menuitem" class = "theme" id = "coal" > Coal< / button > < / li >
< li role = "none" > < button role = "menuitem" class = "theme" id = "navy" > Navy< / button > < / li >
< li role = "none" > < button role = "menuitem" class = "theme" id = "ayu" > Ayu< / button > < / li >
< / ul >
< button id = "search-toggle" class = "icon-button" type = "button" title = "Search. (Shortkey: s)" aria-label = "Toggle Searchbar" aria-expanded = "false" aria-keyshortcuts = "S" aria-controls = "searchbar" >
< i class = "fa fa-search" > < / i >
< / button >
< / div >
< h1 class = "menu-title" > Rust Design Patterns< / h1 >
< div class = "right-buttons" >
< a href = "print.html" title = "Print this book" aria-label = "Print this book" >
< i id = "print-button" class = "fa fa-print" > < / i >
< / a >
< a href = "https://github.com/rust-unofficial/patterns" title = "Git repository" aria-label = "Git repository" >
< i id = "git-repository-button" class = "fa fa-github" > < / i >
< / a >
< / div >
< / div >
< div id = "search-wrapper" class = "hidden" >
< form id = "searchbar-outer" class = "searchbar-outer" >
< input type = "search" name = "search" id = "searchbar" name = "searchbar" placeholder = "Search this book ..." aria-controls = "searchresults-outer" aria-describedby = "searchresults-header" >
< / form >
< div id = "searchresults-outer" class = "searchresults-outer hidden" >
< div id = "searchresults-header" class = "searchresults-header" > < / div >
< ul id = "searchresults" >
< / ul >
< / div >
< / div >
<!-- Apply ARIA attributes after the sidebar and the sidebar toggle button are added to the DOM -->
< script type = "text/javascript" >
document.getElementById('sidebar-toggle').setAttribute('aria-expanded', sidebar === 'visible');
document.getElementById('sidebar').setAttribute('aria-hidden', sidebar !== 'visible');
Array.from(document.querySelectorAll('#sidebar a')).forEach(function(link) {
link.setAttribute('tabIndex', sidebar === 'visible' ? 0 : -1);
});
< / script >
< div id = "content" class = "content" >
< main >
< h1 > < a class = "header" href = "#introduction" id = "introduction" > Introduction< / a > < / h1 >
2021-01-05 16:14:14 +00:00
< h2 > < a class = "header" href = "#participation" id = "participation" > Participation< / a > < / h2 >
2021-01-07 08:04:55 +00:00
< p > If you are interested in contributing to the Patterns catalogue, check out the
< a href = "https://github.com/rust-unofficial/patterns/blob/master/CONTRIBUTING.md" > contribution guidelines< / a > .< / p >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#design-patterns" id = "design-patterns" > Design patterns< / a > < / h2 >
2021-01-07 08:22:07 +00:00
< p > When developing programs, we have to solve many problems.
A program can be viewed as a solution to a problem.
It can also be viewed as a collection of solutions to many different problems.
All of these solutions work together to solve a bigger problem.< / p >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#design-patterns-in-rust" id = "design-patterns-in-rust" > Design patterns in Rust< / a > < / h2 >
2021-01-07 08:22:07 +00:00
< p > There are many problems that share the same form.
Due to the fact that Rust is not object-oriented design patterns vary with respect to other object-oriented programming languages.
While the details are different, since they have the same form they can be solved using the same fundamental methods.< / p >
2021-01-03 09:44:12 +00:00
< p > < a href = "./patterns/index.html" > Design patterns< / a > are methods to solve common problems when writing software.< / p >
2021-01-05 14:36:44 +00:00
< p > < a href = "./anti_patterns/index.html" > Anti-patterns< / a > are methods to solve these same common problems.< / p >
2021-01-02 12:32:53 +00:00
< p > However, while design patterns give us benefits, anti-patterns create more problems.< / p >
2021-01-07 08:22:07 +00:00
< p > < a href = "./idioms/index.html" > Idioms< / a > are guidelines to follow when coding.
They are social norms of the community.
2021-01-05 14:36:44 +00:00
You can break them, but if you do you should have a good reason for it.< / p >
2021-01-02 11:15:56 +00:00
< p > TODO: Mention why Rust is a bit special - functional elements, type system, borrow checker< / p >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#idioms" id = "idioms" > Idioms< / a > < / h1 >
2021-01-08 08:00:20 +00:00
< p > < a href = "https://en.wikipedia.org/wiki/Programming_idiom" > Idioms< / a > are commonly used styles and patterns largely agreed upon by a community.
They are guidelines.
Writing idiomatic code allows other developers to understand what is happening because they are familiar with the form that it has.< / p >
< p > The computer understands the machine code that is generated by the compiler.
The language is therefore mostly beneficial to the developer.
So, since we have this abstraction layer, why not put it to good use and make it simple?< / p >
< p > Remember the < a href = "https://en.wikipedia.org/wiki/KISS_principle" > KISS principle< / a > : " Keep It Simple, Stupid" .
It claims that " most systems work best if they are kept simple rather than made complicated;
therefore, simplicity should be a key goal in design, and unnecessary complexity should be avoided" .< / p >
2021-01-02 11:15:56 +00:00
< blockquote >
< p > Code is there for humans, not computers, to understand.< / p >
< / blockquote >
2021-01-04 08:32:25 +00:00
< h1 > < a class = "header" href = "#use-borrowed-types-for-arguments" id = "use-borrowed-types-for-arguments" > Use borrowed types for arguments< / a > < / h1 >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#description" id = "description" > Description< / a > < / h2 >
2021-01-04 08:32:25 +00:00
< p > Using a target of a deref coercion can increase the flexibility of your code when you are deciding which argument type to use for a function argument.
In this way, the function will accept more input types.< / p >
2021-01-08 08:00:20 +00:00
< p > This is not limited to slice-able or fat pointer types.
In fact you should always prefer using the < strong > borrowed type< / strong > over < strong > borrowing the owned type< / strong > .
Such as < code > & str< / code > over < code > & String< / code > , < code > & [T]< / code > over < code > & Vec< T> < / code > , or < code > & T< / code > over < code > & Box< T> < / code > .< / p >
< p > Using borrowed types you can avoid layers of indirection for those instances where the owned type already provides a layer of indirection.
For instance, a < code > String< / code > has a layer of indirection, so a < code > & String< / code > will have two layers of indrection.
2021-01-05 14:36:44 +00:00
We can avoid this by using < code > & str< / code > instead, and letting < code > & String< / code > coerce to a < code > & str< / code > whenever the function is invoked.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example" id = "example" > Example< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > For this example, we will illustrate some differences for using < code > & String< / code > as a function argument versus using a < code > & str< / code > ,
but the ideas apply as well to using < code > & Vec< T> < / code > versus using a < code > & [T]< / code > or using a < code > & T< / code > versus a < code > & Box< T> < / code > .< / p >
2021-01-04 08:32:25 +00:00
< p > Consider an example where we wish to determine if a word contains three consecutive vowels.
We don't need to own the string to determine this, so we will take a reference.< / p >
< p > The code might look something like this:< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" > fn three_vowels(word: & String) -> bool {
let mut vowel_count = 0;
for c in word.chars() {
match c {
'a' | 'e' | 'i' | 'o' | 'u' => {
vowel_count += 1;
if vowel_count > = 3 {
return true
}
}
_ => vowel_count = 0
}
}
false
}
fn main() {
let ferris = " Ferris" .to_string();
let curious = " Curious" .to_string();
println!(" {}: {}" , ferris, three_vowels(& ferris));
println!(" {}: {}" , curious, three_vowels(& curious));
// This works fine, but the following two lines would fail:
// println!(" Ferris: {}" , three_vowels(" Ferris" ));
// println!(" Curious: {}" , three_vowels(" Curious" ));
}
< / code > < / pre > < / pre >
< p > This works fine because we are passing a < code > & String< / code > type as a parameter.
2021-01-05 14:36:44 +00:00
If we comment in the last two lines this example fails because a < code > & str< / code > type will not coerce to a < code > & String< / code > type.
We can fix this by simply modifying the type for our argument.< / p >
2021-01-04 08:32:25 +00:00
< p > For instance, if we change our function declaration to:< / p >
< pre > < code class = "language-rust ignore" > fn three_vowels(word: & str) -> bool {
< / code > < / pre >
< p > then both versions will compile and print the same output.< / p >
< pre > < code class = "language-bash" > Ferris: false
Curious: true
< / code > < / pre >
2021-01-05 14:36:44 +00:00
< p > But wait, that's not all! There is more to this story.
2021-01-04 08:32:25 +00:00
It's likely that you may say to yourself: that doesn't matter, I will never be using a < code > & 'static str< / code > as an input anways (as we did when we used < code > " Ferris" < / code > ).
Even ignoring this special example, you may still find that using < code > & str< / code > will give you more flexibility than using a < code > & String< / code > .< / p >
2021-01-08 08:00:20 +00:00
< p > Let's now take an example where someone gives us a sentence, and we want to determine if any of the words
in the sentence has a word that contains three consecutive vowels.
2021-01-04 08:32:25 +00:00
We probably should make use of the function we have already defined and simply feed in each word from the sentence.< / p >
< p > An example of this could look like this:< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" > fn three_vowels(word: & str) -> bool {
let mut vowel_count = 0;
for c in word.chars() {
match c {
'a' | 'e' | 'i' | 'o' | 'u' => {
vowel_count += 1;
if vowel_count > = 3 {
return true
}
}
_ => vowel_count = 0
}
}
false
}
fn main() {
2021-01-05 14:57:36 +00:00
let sentence_string =
2021-01-04 08:32:25 +00:00
" Once upon a time, there was a friendly curious crab named Ferris" .to_string();
for word in sentence_string.split(' ') {
if three_vowels(word) {
println!(" {} has three consecutive vowels!" , word);
}
}
}
< / code > < / pre > < / pre >
< p > Running this example using our function declared with an argument type < code > & str< / code > will yield< / p >
< pre > < code class = "language-bash" > curious has three consecutive vowels!
< / code > < / pre >
2021-01-05 14:36:44 +00:00
< p > However, this example will not run when our function is declared with an argument type < code > & String< / code > .
2021-01-08 08:00:20 +00:00
This is because string slices are a < code > & str< / code > and not a < code > & String< / code > which would require an allocation to be
converted to < code > & String< / code > which is not implicit, whereas converting from < code > String< / code > to < code > & str< / code > is cheap and implicit.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also" id = "see-also" > See also< / a > < / h2 >
< ul >
< li > < a href = "https://doc.rust-lang.org/reference/type-coercions.html" > Rust Language Reference on Type Coercions< / a > < / li >
2021-01-08 08:00:20 +00:00
< li > For more discussion on how to handle < code > String< / code > and < code > & str< / code > see
< a href = "https://web.archive.org/web/20201112023149/https://hermanradtke.com/2015/05/03/string-vs-str-in-rust-functions.html" > this blog series (2015)< / a >
by Herman J. Radtke III< / li >
2021-01-04 08:32:25 +00:00
< / ul >
< h1 > < a class = "header" href = "#concatenating-strings-with-format" id = "concatenating-strings-with-format" > Concatenating strings with < code > format!< / code > < / a > < / h1 >
< h2 > < a class = "header" href = "#description-1" id = "description-1" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > It is possible to build up strings using the < code > push< / code > and < code > push_str< / code > methods on a
mutable < code > String< / code > , or using its < code > +< / code > operator. However, it is often more
convenient to use < code > format!< / code > , especially where there is a mix of literal and
non-literal strings.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-1" id = "example-1" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > fn say_hello(name: & str) -> String {
// We could construct the result string manually.
// let mut result = " Hello " .to_owned();
// result.push_str(name);
// result.push('!');
// result
// But using format! is better.
format!(" Hello {}!" , name)
}
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
< h2 > < a class = "header" href = "#advantages" id = "advantages" > Advantages< / a > < / h2 >
< p > Using < code > format!< / code > is usually the most succinct and readable way to combine strings.< / p >
< h2 > < a class = "header" href = "#disadvantages" id = "disadvantages" > Disadvantages< / a > < / h2 >
< p > It is usually not the most efficient way to combine strings - a series of < code > push< / code >
operations on a mutable string is usually the most efficient (especially if the
string has been pre-allocated to the expected size).< / p >
< h1 > < a class = "header" href = "#constructors" id = "constructors" > Constructors< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-2" id = "description-2" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Rust does not have constructors as a language construct. Instead, the convention
is to use a static < code > new< / code > method to create an object.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-2" id = "example-2" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > // A Rust vector, see liballoc/vec.rs
2021-01-01 11:28:10 +00:00
pub struct Vec< T> {
buf: RawVec< T> ,
len: usize,
}
impl< T> Vec< T> {
// Constructs a new, empty `Vec< T> `.
// Note this is a static method - no self.
// This constructor doesn't take any arguments, but some might in order to
// properly initialise an object
pub fn new() -> Vec< T> {
// Create a new Vec with fields properly initialised.
Vec {
// Note that here we are calling RawVec's constructor.
buf: RawVec::new(),
len: 0,
}
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-1" id = "see-also-1" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The < a href = "idioms/../patterns/builder.html" > builder pattern< / a > for constructing objects where there are multiple
2021-01-05 14:36:44 +00:00
configurations.< / p >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#the-default-trait" id = "the-default-trait" > The < code > Default< / code > Trait< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-3" id = "description-3" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Many types in Rust have a < a href = "idioms/ctor.html" > constructor< / a > . However, this is < em > specific< / em > to the
type; Rust cannot abstract over " everything that has a < code > new()< / code > method" . To
allow this, the < a href = "https://doc.rust-lang.org/stable/std/default/trait.Default.html" > < code > Default< / code > < / a > trait was conceived, which can be used with
containers and other generic types (e.g. see < a href = "https://doc.rust-lang.org/stable/std/option/enum.Option.html#method.unwrap_or_default" > < code > Option::unwrap_or_default()< / code > < / a > ).
Notably, some containers already implement it where applicable.< / p >
< p > Not only do one-element containers like < code > Cow< / code > , < code > Box< / code > or < code > Arc< / code > implement
< code > Default< / code > for contained < code > Default< / code > types, one can automatically
< code > #[derive(Default)]< / code > for structs whose fields all implement it, so the more
types implement < code > Default< / code > , the more useful it becomes.< / p >
< p > On the other hand, constructors can take multiple arguments, while the
< code > default()< / code > method does not. There can even be multiple constructors with
different names, but there can only be one < code > Default< / code > implementation per type.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-3" id = "example-3" > Example< / a > < / h2 >
2021-01-03 07:44:53 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" > use std::{path::PathBuf, time::Duration};
// note that we can simply auto-derive Default here.
#[derive(Default, Debug)]
2021-01-01 11:28:10 +00:00
struct MyConfiguration {
// Option defaults to None
2021-01-03 07:44:53 +00:00
output: Option< PathBuf> ,
2021-01-01 11:28:10 +00:00
// Vecs default to empty vector
2021-01-03 07:44:53 +00:00
search_path: Vec< PathBuf> ,
2021-01-01 11:28:10 +00:00
// Duration defaults to zero time
timeout: Duration,
// bool defaults to false
check: bool,
}
impl MyConfiguration {
// add setters here
}
fn main() {
// construct a new instance with default values
let mut conf = MyConfiguration::default();
2021-01-03 07:44:53 +00:00
// do something with conf here
conf.check = true;
println!(" conf = {:#?}" , conf);
2021-01-01 11:28:10 +00:00
}
2021-01-03 07:44:53 +00:00
< / code > < / pre > < / pre >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-2" id = "see-also-2" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > The < a href = "idioms/ctor.html" > constructor< / a > idiom is another way to generate instances that may or may
not be " default" < / li >
< li > The < a href = "https://doc.rust-lang.org/stable/std/default/trait.Default.html" > < code > Default< / code > < / a > documentation (scroll down for the list of implementors)< / li >
< li > < a href = "https://doc.rust-lang.org/stable/std/option/enum.Option.html#method.unwrap_or_default" > < code > Option::unwrap_or_default()< / code > < / a > < / li >
< li > < a href = "https://crates.io/crates/derive-new/" > < code > derive(new)< / code > < / a > < / li >
< / ul >
< h1 > < a class = "header" href = "#collections-are-smart-pointers" id = "collections-are-smart-pointers" > Collections are smart pointers< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-4" id = "description-4" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Use the < code > Deref< / code > trait to treat collections like smart pointers, offering owning
and borrowed views of data.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-4" id = "example-4" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > use std::ops::Deref;
struct Vec< T> {
data: T,
//..
2021-01-01 11:28:10 +00:00
}
impl< T> Deref for Vec< T> {
type Target = [T];
fn deref(& self) -> & [T] {
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > A < code > Vec< T> < / code > is an owning collection of < code > T< / code > s, a slice (< code > & [T]< / code > ) is a borrowed
collection of < code > T< / code > s. Implementing < code > Deref< / code > for < code > Vec< / code > allows implicit dereferencing
from < code > & Vec< T> < / code > to < code > & [T]< / code > and includes the relationship in auto-derefencing
searches. Most methods you might expect to be implemented for < code > Vec< / code > s are instead
implemented for slices.< / p >
< p > See also < code > String< / code > and < code > & str< / code > .< / p >
< h2 > < a class = "header" href = "#motivation" id = "motivation" > Motivation< / a > < / h2 >
< p > Ownership and borrowing are key aspects of the Rust language. Data structures
must account for these semantics properly in order to give a good user
experience. When implementing a data structure which owns its data, offering a
borrowed view of that data allows for more flexible APIs.< / p >
< h2 > < a class = "header" href = "#advantages-1" id = "advantages-1" > Advantages< / a > < / h2 >
< p > Most methods can be implemented only for the borrowed view, they are then
implicitly available for the owning view.< / p >
< p > Gives clients a choice between borrowing or taking ownership of data.< / p >
< h2 > < a class = "header" href = "#disadvantages-1" id = "disadvantages-1" > Disadvantages< / a > < / h2 >
< p > Methods and traits only available via dereferencing are not taken into account
when bounds checking, so generic programming with data structures using this
pattern can get complex (see the < code > Borrow< / code > and < code > AsRef< / code > traits, etc.).< / p >
< h2 > < a class = "header" href = "#discussion" id = "discussion" > Discussion< / a > < / h2 >
< p > Smart pointers and collections are analogous: a smart pointer points to a single
object, whereas a collection points to many objects. From the point of view of
the type system there is little difference between the two. A collection owns
its data if the only way to access each datum is via the collection and the
collection is responsible for deleting the data (even in cases of shared
ownership, some kind of borrowed view may be appropriate). If a collection owns
its data, it is usually useful to provide a view of the data as borrowed so that
it can be multiply referenced.< / p >
< p > Most smart pointers (e.g., < code > Foo< T> < / code > ) implement < code > Deref< Target=T> < / code > . However,
collections will usually dereference to a custom type. < code > [T]< / code > and < code > str< / code > have some
language support, but in the general case, this is not necessary. < code > Foo< T> < / code > can
implement < code > Deref< Target=Bar< T> > < / code > where < code > Bar< / code > is a dynamically sized type and
< code > & Bar< T> < / code > is a borrowed view of the data in < code > Foo< T> < / code > .< / p >
< p > Commonly, ordered collections will implement < code > Index< / code > for < code > Range< / code > s to provide
slicing syntax. The target will be the borrowed view.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-3" id = "see-also-3" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > < a href = "idioms/../anti_patterns/deref.html" > Deref polymorphism anti-pattern< / a > .< / p >
< p > < a href = "https://doc.rust-lang.org/std/ops/trait.Deref.html" > Documentation for < code > Deref< / code > trait< / a > .< / p >
< h1 > < a class = "header" href = "#finalisation-in-destructors" id = "finalisation-in-destructors" > Finalisation in destructors< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-5" id = "description-5" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Rust does not provide the equivalent to < code > finally< / code > blocks - code that will be
executed no matter how a function is exited. Instead an object's destructor can
be used to run code that must be run before exit.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-5" id = "example-5" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > fn bar() -> Result< (), ()> {
2021-01-01 11:28:10 +00:00
// These don't need to be defined inside the function.
struct Foo;
// Implement a destructor for Foo.
impl Drop for Foo {
fn drop(& mut self) {
println!(" exit" );
}
}
// The dtor of _exit will run however the function `bar` is exited.
let _exit = Foo;
// Implicit return with `?` operator.
baz()?;
// Normal return.
Ok(())
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#motivation-1" id = "motivation-1" > Motivation< / a > < / h2 >
< p > If a function has multiple return points, then executing code on exit becomes
difficult and repetitive (and thus bug-prone). This is especially the case where
return is implicit due to a macro. A common case is the < code > ?< / code > operator which
returns if the result is an < code > Err< / code > , but continues if it is < code > Ok< / code > . < code > ?< / code > is used as
an exception handling mechanism, but unlike Java (which has < code > finally< / code > ), there is
no way to schedule code to run in both the normal and exceptional cases.
Panicking will also exit a function early.< / p >
< h2 > < a class = "header" href = "#advantages-2" id = "advantages-2" > Advantages< / a > < / h2 >
< p > Code in destructors will (nearly) always be run - copes with panics, early
returns, etc.< / p >
< h2 > < a class = "header" href = "#disadvantages-2" id = "disadvantages-2" > Disadvantages< / a > < / h2 >
< p > It is not guaranteed that destructors will run. For example, if there is an
infinite loop in a function or if running a function crashes before exit.
Destructors are also not run in the case of a panic in an already panicking
thread. Therefore destructors cannot be relied on as finalisers where it is
absolutely essential that finalisation happens.< / p >
< p > This pattern introduces some hard to notice, implicit code. Reading a function
gives no clear indication of destructors to be run on exit. This can make
debugging tricky.< / p >
< p > Requiring an object and < code > Drop< / code > impl just for finalisation is heavy on boilerplate.< / p >
< h2 > < a class = "header" href = "#discussion-1" id = "discussion-1" > Discussion< / a > < / h2 >
< p > There is some subtlety about how exactly to store the object used as a
finaliser. It must be kept alive until the end of the function and must then be
destroyed. The object must always be a value or uniquely owned pointer (e.g.,
< code > Box< Foo> < / code > ). If a shared pointer (such as < code > Rc< / code > ) is used, then the finaliser can
be kept alive beyond the lifetime of the function. For similar reasons, the
finaliser should not be moved or returned.< / p >
< p > The finaliser must be assigned into a variable, otherwise it will be destroyed
immediately, rather than when it goes out of scope. The variable name must start
with < code > _< / code > if the variable is only used as a finaliser, otherwise the compiler
will warn that the finaliser is never used. However, do not call the variable
2021-01-04 04:54:57 +00:00
< code > _< / code > with no suffix - in that case it will be destroyed immediately.< / p >
2021-01-01 11:28:10 +00:00
< p > In Rust, destructors are run when an object goes out of scope. This happens
whether we reach the end of block, there is an early return, or the program
panics. When panicking, Rust unwinds the stack running destructors for each
object in each stack frame. So, destructors get called even if the panic happens
in a function being called.< / p >
< p > If a destructor panics while unwinding, there is no good action to take, so Rust
aborts the thread immediately, without running further destructors. This means
that desctructors are not absolutely guaranteed to run. It also means that you
must take extra care in your destructors not to panic, since it could leave
resources in an unexpected state.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-4" id = "see-also-4" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > < a href = "idioms/../patterns/RAII.html" > RAII< / a > .< / p >
< h1 > < a class = "header" href = "#memreplace-to-keep-owned-values-in-changed-enums" id = "memreplace-to-keep-owned-values-in-changed-enums" > < code > mem::replace< / code > to keep owned values in changed enums< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-6" id = "description-6" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Say we have a < code > & mut MyEnum< / code > which has (at least) two variants,
< code > A { name: String, x: u8 }< / code > and < code > B { name: String }< / code > . Now we want to change
< code > MyEnum::A< / code > to a < code > B< / code > if < code > x< / code > is zero, while keeping < code > MyEnum::B< / code > intact.< / p >
< p > We can do this without cloning the < code > name< / code > .< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-6" id = "example-6" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > use std::mem;
enum MyEnum {
A { name: String, x: u8 },
B { name: String }
}
fn a_to_b(e: & mut MyEnum) {
// we mutably borrow `e` here. This precludes us from changing it directly
// as in `*e = ...`, because the borrow checker won't allow it. Therefore
2021-01-05 14:57:36 +00:00
// the assignment to `e` must be outside the `if let` clause.
2021-01-01 11:28:10 +00:00
*e = if let MyEnum::A { ref mut name, x: 0 } = *e {
2021-01-05 14:57:36 +00:00
2021-01-01 11:28:10 +00:00
// this takes out our `name` and put in an empty String instead
// (note that empty strings don't allocate).
2021-01-05 14:57:36 +00:00
// Then, construct the new enum variant (which will
2021-01-01 11:28:10 +00:00
// be assigned to `*e`, because it is the result of the `if let` expression).
MyEnum::B { name: mem::replace(name, String::new()) }
2021-01-05 14:57:36 +00:00
2021-01-01 11:28:10 +00:00
// In all other cases, we return immediately, thus skipping the assignment
} else { return }
}
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
< p > This also works with more variants:< / p >
< pre > < code class = "language-Rust" > use std::mem;
enum MultiVariateEnum {
A { name: String },
B { name: String },
C,
D
}
fn swizzle(e: & mut MultiVariateEnum) {
use self::MultiVariateEnum::*;
*e = match *e {
// Ownership rules do not allow taking `name` by value, but we cannot
// take the value out of a mutable reference, unless we replace it:
A { ref mut name } => B { name: mem::replace(name, String::new()) },
B { ref mut name } => A { name: mem::replace(name, String::new()) },
C => D,
D => C
}
}
< / code > < / pre >
< h2 > < a class = "header" href = "#motivation-2" id = "motivation-2" > Motivation< / a > < / h2 >
< p > When working with enums, we may want to change an enum value in place, perhaps
to another variant. This is usually done in two phases to keep the borrow
checker happy. In the first phase, we observe the existing value and look at
its parts to decide what to do next. In the second phase we may conditionally
change the value (as in the example above).< / p >
< p > The borrow checker won't allow us to take out < code > name< / code > of the enum (because
< em > something< / em > must be there. We could of course < code > .clone()< / code > name and put the clone
into our < code > MyEnum::B< / code > , but that would be an instance of the [Clone to satisfy
the borrow checker] antipattern. Anyway, we can avoid the extra allocation by
changing < code > e< / code > with only a mutable borrow.< / p >
< p > < code > mem::replace< / code > lets us swap out the value, replacing it with something else. In
this case, we put in an empty < code > String< / code > , which does not need to allocate. As a
result, we get the original < code > name< / code > < em > as an owned value< / em > . We can then wrap this in
another enum.< / p >
< p > Note, however, that if we are using an < code > Option< / code > and want to replace its
value with a < code > None< / code > , < code > Option< / code > ’ < code > take()< / code > method provides a shorter and
more idiomatic alternative.< / p >
< h2 > < a class = "header" href = "#advantages-3" id = "advantages-3" > Advantages< / a > < / h2 >
< p > Look ma, no allocation! Also you may feel like Indiana Jones while doing it.< / p >
< h2 > < a class = "header" href = "#disadvantages-3" id = "disadvantages-3" > Disadvantages< / a > < / h2 >
< p > This gets a bit wordy. Getting it wrong repeatedly will make you hate the
borrow checker. The compiler may fail to optimize away the double store,
resulting in reduced performance as opposed to what you'd do in unsafe
languages.< / p >
< h2 > < a class = "header" href = "#discussion-2" id = "discussion-2" > Discussion< / a > < / h2 >
< p > This pattern is only of interest in Rust. In GC'd languages, you'd take the
reference to the value by default (and the GC would keep track of refs), and in
other low-level languages like C you'd simply alias the pointer and fix things
later.< / p >
< p > However, in Rust, we have to do a little more work to do this. An owned value
may only have one owner, so to take it out, we need to put something back in –
like Indiana Jones, replacing the artifact with a bag of sand.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-5" id = "see-also-5" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > This gets rid of the [Clone to satisfy the borrow checker] antipattern in a
specific case.< / p >
< p > [Clone to satisfy the borrow checker](TODO: Hinges on PR #23)< / p >
< h1 > < a class = "header" href = "#on-stack-dynamic-dispatch" id = "on-stack-dynamic-dispatch" > On-Stack Dynamic Dispatch< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-7" id = "description-7" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > We can dynamically dispatch over multiple values, however, to do so, we need
to declare multiple variables to bind differently-typed objects. To extend the
lifetime as necessary, we can use deferred conditional initialization, as seen
below:< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-7" id = "example-7" > Example< / a > < / h2 >
2021-01-03 07:40:16 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" > use std::io;
use std::fs;
2021-01-02 11:41:18 +00:00
2021-01-03 07:40:16 +00:00
< span class = "boring" > fn main() -> Result< (), Box< dyn std::error::Error> > {
< / span > < span class = "boring" > let arg = " -" ;
< / span >
2021-01-02 11:41:18 +00:00
// These must live longer than `readable`, and thus are declared first:
2021-01-01 11:28:10 +00:00
let (mut stdin_read, mut file_read);
// We need to ascribe the type to get dynamic dispatch.
2021-01-03 07:40:16 +00:00
let readable: & mut dyn io::Read = if arg == " -" {
2021-01-01 11:28:10 +00:00
stdin_read = io::stdin();
& mut stdin_read
} else {
file_read = fs::File::open(arg)?;
& mut file_read
};
// Read from `readable` here.
2021-01-03 07:40:16 +00:00
< span class = "boring" > Ok(())
< / span > < span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#motivation-3" id = "motivation-3" > Motivation< / a > < / h2 >
< p > Rust monomorphises code by default. This means a copy of the code will be
generated for each type it is used with and optimized independently. While this
allows for very fast code on the hot path, it also bloats the code in places
where performance is not of the essence, thus costing compile time and cache
usage.< / p >
< p > Luckily, Rust allows us to use dynamic dispatch, but we have to explicitly ask
for it.< / p >
< h2 > < a class = "header" href = "#advantages-4" id = "advantages-4" > Advantages< / a > < / h2 >
< p > We do not need to allocate anything on the heap. Neither do we need to
initialize something we won't use later, nor do we need to monomorphize the
2021-01-03 02:26:25 +00:00
whole code that follows to work with both < code > File< / code > or < code > Stdin< / code > .< / p >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#disadvantages-4" id = "disadvantages-4" > Disadvantages< / a > < / h2 >
< p > The code needs more moving parts than the < code > Box< / code > -based version:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > // We still need to ascribe the type for dynamic dispatch.
2021-01-01 11:28:10 +00:00
let readable: Box< dyn io::Read> = if arg == " -" {
Box::new(io::stdin())
} else {
Box::new(fs::File::open(arg)?)
};
// Read from `readable` here.
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#discussion-3" id = "discussion-3" > Discussion< / a > < / h2 >
< p > Rust newcomers will usually learn that Rust requires all variables to be
initialized < em > before use< / em > , so it's easy to overlook the fact that < em > unused< / em >
variables may well be uninitialized. Rust works quite hard to ensure that this
works out fine and only the initialized values are dropped at the end of their
scope.< / p >
< p > The example meets all the constraints Rust places on us:< / p >
< ul >
< li > All variables are initialized before using (in this case borrowing) them< / li >
< li > Each variable only holds values of a single type. In our example, < code > stdin< / code > is
of type < code > Stdin< / code > , < code > file< / code > is of type < code > File< / code > and < code > readable< / code > is of type < code > & mut dyn Read< / code > < / li >
< li > Each borrowed value outlives all the references borrowed from it< / li >
< / ul >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-6" id = "see-also-6" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
2021-01-05 14:36:44 +00:00
< li > < a href = "idioms/dtor-finally.html" > Finalisation in destructors< / a > and
2021-01-01 11:28:10 +00:00
< a href = "idioms/../patterns/RAII.html" > RAII guards< / a > can benefit from tight control over lifetimes.< / li >
< li > For conditionally filled < code > Option< & T> < / code > s of (mutable) references, one can
initialize an < code > Option< T> < / code > directly and use its < a href = "https://doc.rust-lang.org/std/option/enum.Option.html#method.as_ref" > < code > .as_ref()< / code > < / a > method to get an
optional reference.< / li >
< / ul >
2021-01-05 15:44:58 +00:00
< h1 > < a class = "header" href = "#ffi-idioms" id = "ffi-idioms" > FFI Idioms< / a > < / h1 >
< p > Writing FFI code is an entire course in itself.
However, there are several idioms here that can act as pointers, and avoid traps for inexperienced users of < code > unsafe< / code > Rust.< / p >
< p > This section contains idioms that may be useful when doing FFI.< / p >
< ol >
< li >
< p > < a href = "idioms/./ffi-errors.html" > Idiomatic Errors< / a > - Error handling with integer codes and sentinel return values (such as < code > NULL< / code > pointers)< / p >
< / li >
< li >
< p > < a href = "idioms/./ffi-accepting-strings.html" > Accepting Strings< / a > with minimal unsafe code< / p >
< / li >
< li >
< p > < a href = "idioms/./ffi-passing-strings.html" > Passing Strings< / a > to FFI functions< / p >
< / li >
< / ol >
< h1 > < a class = "header" href = "#error-handling-in-ffi" id = "error-handling-in-ffi" > Error Handling in FFI< / a > < / h1 >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#description-8" id = "description-8" > Description< / a > < / h2 >
2021-01-05 15:44:58 +00:00
< p > In foreign languages like C, errors are represented by return codes.
However, Rust's type system allows much more rich error information to be captured a propogated through a full type.< / p >
< p > This best practice shows different kinds of error codes, and how to expose them in a usable way:< / p >
< ol >
< li > Flat Enums should be converted to integers and returned as codes.< / li >
< li > Structured Enums should be converted to an integer code with a string error message for detail.< / li >
< li > Custom Error Types should become " transparent" , with a C representation.< / li >
< / ol >
< h2 > < a class = "header" href = "#code-example" id = "code-example" > Code Example< / a > < / h2 >
< h3 > < a class = "header" href = "#flat-enums" id = "flat-enums" > Flat Enums< / a > < / h3 >
< pre > < code class = "language-rust ignore" > enum DatabaseError {
IsReadOnly = 1, // user attempted a write operation
IOError = 2, // user should read the C errno() for what it was
FileCorrupted = 3, // user should run a repair tool to recover it
}
impl From< DatabaseError> for libc::c_int {
fn from(e: DatabaseError) -> libc::c_int {
(e as i8).into()
}
}
< / code > < / pre >
< h3 > < a class = "header" href = "#structured-enums" id = "structured-enums" > Structured Enums< / a > < / h3 >
< pre > < code class = "language-rust ignore" > pub mod errors {
enum DatabaseError {
IsReadOnly,
IOError(std::io::Error),
FileCorrupted(String), // message describing the issue
}
impl From< DatabaseError> for libc::c_int {
fn from(e: DatabaseError) -> libc::c_int {
match e {
DatabaseError::IsReadOnly => 1,
DatabaseError::IOError(_) => 2,
DatabaseError::FileCorrupted(_) => 3,
}
}
}
}
pub mod c_api {
use super::errors::DatabaseError;
#[no_mangle]
pub extern " C" fn db_error_description(
e: *const DatabaseError
) -> *mut libc::c_char {
let error: & DatabaseError = unsafe {
/* SAFETY: pointer lifetime is greater than the current stack frame */
& *e
};
let error_str: String = match error {
DatabaseError::IsReadOnly => {
format!(" cannot write to read-only database" );
}
DatabaseError::IOError(e) => {
format!(" I/O Error: {}" , e);
}
DatabaseError::FileCorrupted(s) => {
format!(" File corrupted, run repair: {}" , & s);
}
};
let c_error = unsafe {
// SAFETY: copying error_str to an allocated buffer with a NUL
// character at the end
let mut malloc: *mut u8 = libc::malloc(error_str.len() + 1) as *mut _;
if malloc.is_null() {
return std::ptr::null_mut();
}
let src = error_str.as_bytes().as_ptr();
std::ptr::copy_nonoverlapping(src, malloc, error_str.len());
std::ptr::write(malloc.add(error_str.len()), 0);
malloc as *mut libc::c_char
};
c_error
}
}
< / code > < / pre >
< h3 > < a class = "header" href = "#custom-error-types" id = "custom-error-types" > Custom Error Types< / a > < / h3 >
< pre > < code class = "language-rust ignore" > struct ParseError {
expected: char,
line: u32,
ch: u16
}
impl ParseError { /* ... */ }
/* Create a second version which is exposed as a C structure */
#[repr(C)]
pub struct parse_error {
pub expected: libc::c_char,
pub line: u32,
pub ch: u16
}
impl From< ParseError> for parse_error {
fn from(e: ParseError) -> parse_error {
let ParseError { expected, line, ch } = e;
parse_error { expected, line, ch }
}
}
< / code > < / pre >
< h2 > < a class = "header" href = "#advantages-5" id = "advantages-5" > Advantages< / a > < / h2 >
< p > This ensures that the foreign language has clear access to error information while not compromising the Rust code's API at all.< / p >
< h2 > < a class = "header" href = "#disadvantages-5" id = "disadvantages-5" > Disadvantages< / a > < / h2 >
< p > It's a lot of typing, and some types may not be able to be converted easily to C.< / p >
< h1 > < a class = "header" href = "#accepting-strings" id = "accepting-strings" > Accepting Strings< / a > < / h1 >
< h2 > < a class = "header" href = "#description-9" id = "description-9" > Description< / a > < / h2 >
< p > When accepting strings via FFI through pointers, there are two principles that should be followed:< / p >
< ol >
< li > Keep foreign strings " borrowed" , rather than copying them directly.< / li >
< li > Minimize < code > unsafe< / code > code during the conversion.< / li >
< / ol >
< h2 > < a class = "header" href = "#motivation-4" id = "motivation-4" > Motivation< / a > < / h2 >
< p > Rust has built-in support for C-style strings with its < code > CString< / code > and < code > CStr< / code > types.
However, there are different approaches one can take with strings that are being accepted from a foreign caller of a Rust function.< / p >
< p > The best practice is simple: use < code > CStr< / code > in such a way as to minimize unsafe code, and create a borrowed slice.
If an owned String is needed, call < code > to_string()< / code > on the string slice.< / p >
< h2 > < a class = "header" href = "#code-example-1" id = "code-example-1" > Code Example< / a > < / h2 >
< pre > < code class = "language-rust ignore" > pub mod unsafe_module {
// other module content
#[no_mangle]
pub extern " C" fn mylib_log(msg: *const libc::c_char, level: libc::c_int) {
let level: crate::LogLevel = match level { /* ... */ };
let msg_str: & str = unsafe {
// SAFETY: accessing raw pointers expected to live for the call,
// and creating a shared reference that does not outlive the current
// stack frame.
match std::ffi::CStr::from_ptr(msg).to_str() {
Ok(s) => s,
Err(e) => {
crate::log_error(" FFI string conversion failed" );
return;
}
}
};
crate::log(msg_str, level);
}
}
< / code > < / pre >
< h2 > < a class = "header" href = "#advantages-6" id = "advantages-6" > Advantages< / a > < / h2 >
< p > The example is is written to ensure that:< / p >
< ol >
< li > The < code > unsafe< / code > block is as small as possible.< / li >
< li > The pointer with an " untracked" lifetime becomes a " tracked" shared reference< / li >
< / ol >
< p > Consider an alternative, where the string is actually copied:< / p >
< pre > < code class = "language-rust ignore" > pub mod unsafe_module {
// other module content
pub extern " C" fn mylib_log(msg: *const libc::c_char, level: libc::c_int) {
/* DO NOT USE THIS CODE. IT IS UGLY, VERBOSE, AND CONTAINS A SUBTLE BUG. */
let level: crate::LogLevel = match level { /* ... */ };
let msg_len = unsafe { /* SAFETY: strlen is what it is, I guess? */
libc::strlen(msg)
};
let mut msg_data = Vec::with_capacity(msg_len + 1);
let msg_cstr: std::ffi::CString = unsafe {
// SAFETY: copying from a foreign pointer expected to live
// for the entire stack frame into owned memory
std::ptr::copy_nonoverlapping(msg, msg_data.as_mut(), msg_len);
msg_data.set_len(msg_len + 1);
std::ffi::CString::from_vec_with_nul(msg_data).unwrap()
}
let msg_str: String = unsafe {
match msg_cstr.into_string() {
Ok(s) => s,
Err(e) => {
crate::log_error(" FFI string conversion failed" );
return;
}
}
};
crate::log(& msg_str, level);
}
}
< / code > < / pre >
< p > This code in inferior to the original in two respects:< / p >
< ol >
< li > There is much more < code > unsafe< / code > code, and more importantly, more invariants it must uphold.< / li >
< li > Due to the extensive arithmetic required, there is a bug in this version that cases Rust < code > undefined behaviour< / code > .< / li >
< / ol >
< p > The bug here is a simple mistake in pointer arithmetic: the string was copied, all < code > msg_len< / code > bytes of it.
However, the < code > NUL< / code > terminator at the end was not.< / p >
2021-01-08 08:00:20 +00:00
< p > The Vector then had its size < em > set< / em > to the length of the < em > zero padded string< / em > --
rather than < em > resized< / em > to it, which could have added a zero at the end.
As a result, the last byte in the Vector is uninitialized memory.
2021-01-05 15:44:58 +00:00
When the < code > CString< / code > is created at the bottom of the block, its read of the Vector will cause < code > undefined behaviour< / code > !< / p >
< p > Like many such issues, this would be difficult issue to track down.
2021-01-08 08:00:20 +00:00
Sometimes it would panic because the string was not < code > UTF-8< / code > , sometimes it would put a weird character at the end of the string,
sometimes it would just completely crash.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-6" id = "disadvantages-6" > Disadvantages< / a > < / h2 >
< p > None?< / p >
< h1 > < a class = "header" href = "#passing-strings" id = "passing-strings" > Passing Strings< / a > < / h1 >
< h2 > < a class = "header" href = "#description-10" id = "description-10" > Description< / a > < / h2 >
< p > When passing strings to FFI functions, there are four principles that should be followed:< / p >
< ol >
< li > Make the lifetime of owned strings as long as possible.< / li >
< li > Minimize < code > unsafe< / code > code during the conversion.< / li >
< li > If the C code can modify the string data, use < code > Vec< / code > instead of < code > CString< / code > .< / li >
< li > Unless the Foreign Function API requires it, the ownership of the string should not transfer to the callee.< / li >
< / ol >
< h2 > < a class = "header" href = "#motivation-5" id = "motivation-5" > Motivation< / a > < / h2 >
< p > Rust has built-in support for C-style strings with its < code > CString< / code > and < code > CStr< / code > types.
However, there are different approaches one can take with strings that are being sent to a foreign function call from a Rust function.< / p >
< p > The best practice is simple: use < code > CString< / code > in such a way as to minimize < code > unsafe< / code > code.
However, a secondary caveat is that < em > the object must live long enough< / em > , meaning the lifetime should be maximized.
In addition, the documentation explains that " round-tripping" a < code > CString< / code > after modification is UB, so additional work is necessary in that case.< / p >
< h2 > < a class = "header" href = "#code-example-2" id = "code-example-2" > Code Example< / a > < / h2 >
< pre > < code class = "language-rust ignore" > pub mod unsafe_module {
// other module content
extern " C" {
fn seterr(message: *const libc::c_char);
fn geterr(buffer: *mut libc::c_char, size: libc::c_int) -> libc::c_int;
}
fn report_error_to_ffi< S: Into< String> > (
err: S
) -> Result< (), std::ffi::NulError> {
let c_err = std::ffi::CString::new(err.into())?;
unsafe {
// SAFETY: calling an FFI whose documentation says the pointer is
// const, so no modificationshould occur
seterr(c_err.as_ptr());
}
Ok(())
// The lifetime of c_err continues until here
}
fn get_error_from_ffi() -> Result< String, std::ffi::IntoStringError> {
let mut buffer = vec![0u8; 1024];
unsafe {
// SAFETY: calling an FFI whose documentation implies
// that the input need only live as long as the call
let written: usize = geterr(buffer.as_mut_ptr(), 1023).into();
buffer.truncate(written + 1);
}
std::ffi::CString::new(buffer).unwrap().into_string()
}
}
< / code > < / pre >
< h2 > < a class = "header" href = "#advantages-7" id = "advantages-7" > Advantages< / a > < / h2 >
< p > The example is written in a way to ensure that:< / p >
< ol >
< li > The < code > unsafe< / code > block is as small as possible.< / li >
< li > The < code > CString< / code > lives long enough.< / li >
< li > Errors with typecasts are always propagated when possible.< / li >
< / ol >
< p > A common mistake (so common it's in the documentation) is to not use the variable in the first block:< / p >
< pre > < code class = "language-rust ignore" > pub mod unsafe_module {
// other module content
fn report_error< S: Into< String> > (err: S) -> Result< (), std::ffi::NulError> {
unsafe {
// SAFETY: whoops, this contains a dangling pointer!
seterr(std::ffi::CString::new(err.into())?.as_ptr());
}
Ok(())
}
}
< / code > < / pre >
2021-01-08 08:00:20 +00:00
< p > This code will result in a dangling pointer, because the lifetime of the < code > CString< / code > is not extended
by the pointer creation, unlike if a reference were created.< / p >
2021-01-05 15:44:58 +00:00
< p > Another issue frequently raised is that the initialization of a 1k vector of zeroes is " slow" .
2021-01-08 08:00:20 +00:00
However, recent versions of Rust actually optimize that particular macro to a call to < code > zmalloc< / code > ,
meaning it is as fast as the operating system's ability to return zeroed memory (which is quite fast).< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-7" id = "disadvantages-7" > Disadvantages< / a > < / h2 >
< p > None?< / p >
< h1 > < a class = "header" href = "#iterating-over-an-option" id = "iterating-over-an-option" > Iterating over an < code > Option< / code > < / a > < / h1 >
< h2 > < a class = "header" href = "#description-11" id = "description-11" > Description< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > < code > Option< / code > can be viewed as a container that contains either zero or one elements.
In particular, it implements the < code > IntoIterator< / code > trait, and as such can be used with generic code that needs such a type.< / p >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#examples" id = "examples" > Examples< / a > < / h2 >
< p > Since < code > Option< / code > implements < code > IntoIterator< / code > , it can be used as an argument to < a href = "https://doc.rust-lang.org/std/iter/trait.Extend.html#tymethod.extend" > < code > .extend()< / code > < / a > :< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > let turing = Some(" Turing" );
let mut logicians = vec![" Curry" , " Kleene" , " Markov" ];
logicians.extend(turing);
// equivalent to
if let Some(turing_inner) = turing {
logicians.push(turing_inner);
}
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
< p > If you need to tack an < code > Option< / code > to the end of an existing iterator, you can pass it to < a href = "https://doc.rust-lang.org/std/iter/trait.Iterator.html#method.chain" > < code > .chain()< / code > < / a > :< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > let turing = Some(" Turing" );
let logicians = vec![" Curry" , " Kleene" , " Markov" ];
for logician in logicians.iter().chain(turing.iter()) {
println!(" {} is a logician" , logician);
}
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-08 08:00:20 +00:00
< p > Note that if the < code > Option< / code > is always < code > Some< / code > , then it is more idiomatic to use < a href = "https://doc.rust-lang.org/std/iter/fn.once.html" > < code > std::iter::once< / code > < / a >
on the element instead.< / p >
< p > Also, since < code > Option< / code > implements < code > IntoIterator< / code > , it's possible to iterate over it using a < code > for< / code > loop.
This is equivalent to matching it with < code > if let Some(..)< / code > , and in most cases you should prefer the latter.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-7" id = "see-also-7" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li >
2021-01-08 08:00:20 +00:00
< p > < a href = "https://doc.rust-lang.org/std/iter/fn.once.html" > < code > std::iter::once< / code > < / a > is an iterator which yields exactly one element.
It's a more readable alternative to < code > Some(foo).into_iter()< / code > .< / p >
2021-01-01 11:28:10 +00:00
< / li >
< li >
2021-01-08 08:00:20 +00:00
< p > < a href = "https://doc.rust-lang.org/std/iter/trait.Iterator.html#method.filter_map" > < code > Iterator::filter_map< / code > < / a > is a version of
< a href = "https://doc.rust-lang.org/std/iter/trait.Iterator.html#method.flat_map" > < code > Iterator::flat_map< / code > < / a > , specialized to mapping functions which
return < code > Option< / code > .< / p >
2021-01-01 11:28:10 +00:00
< / li >
< li >
< p > The < a href = "https://crates.io/crates/ref_slice" > < code > ref_slice< / code > < / a > crate provides functions for converting an < code > Option< / code > to a zero- or one-element slice.< / p >
< / li >
< li >
< p > < a href = "https://doc.rust-lang.org/std/option/enum.Option.html" > Documentation for < code > Option< T> < / code > < / a > < / p >
< / li >
< / ul >
< h1 > < a class = "header" href = "#pass-variables-to-closure" id = "pass-variables-to-closure" > Pass variables to closure< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-12" id = "description-12" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > By default, closures capture their environment by borrowing. Or you can use < code > move< / code > -closure
to move whole environment. However, often you want to move just some variables to closure,
give it copy of some data, pass it by reference, or perform some other transformation.< / p >
< p > Use variable rebinding in separate scope for that.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-8" id = "example-8" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Use< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
2021-01-02 11:41:18 +00:00
< / span > use std::rc::Rc;
let num1 = Rc::new(1);
2021-01-01 11:28:10 +00:00
let num2 = Rc::new(2);
let num3 = Rc::new(3);
let closure = {
// `num1` is moved
let num2 = num2.clone(); // `num2` is cloned
let num3 = num3.as_ref(); // `num3` is borrowed
move || {
*num1 + *num2 + *num3;
}
};
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
< p > instead of< / p >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
2021-01-02 11:41:18 +00:00
< / span > use std::rc::Rc;
let num1 = Rc::new(1);
2021-01-01 11:28:10 +00:00
let num2 = Rc::new(2);
let num3 = Rc::new(3);
let num2_cloned = num2.clone();
let num3_borrowed = num3.as_ref();
let closure = move || {
2021-01-06 13:56:19 +00:00
*num1 + *num2_cloned + *num3_borrowed;
2021-01-01 11:28:10 +00:00
};
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-8" id = "advantages-8" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Copied data are grouped together with closure definition, so their purpose is more clear
and they will be dropped immediately even if they are not consumed by closure.< / p >
< p > Closure uses same variable names as surrounding code whether data are copied or moved.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-8" id = "disadvantages-8" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Additional indentation of closure body.< / p >
< h1 > < a class = "header" href = "#privacy-for-extensibility" id = "privacy-for-extensibility" > Privacy for extensibility< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-13" id = "description-13" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Use a private field to ensure that a struct is extensible without breaking
stability guarantees.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-9" id = "example-9" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > mod a {
2021-01-01 11:28:10 +00:00
// Public struct.
pub struct S {
pub foo: i32,
// Private field.
bar: i32,
}
}
fn main(s: a::S) {
// Because S::bar is private, it cannot be named here and we must use `..`
// in the pattern.
let a::S { foo: _, ..} = s;
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#discussion-4" id = "discussion-4" > Discussion< / a > < / h2 >
2021-01-06 13:58:52 +00:00
< p > Adding a field to a struct is a mostly backwards compatible change.
2021-01-08 08:00:20 +00:00
However, if a client uses a pattern to deconstruct a struct instance, they might name all the fields in the struct
and adding a new one would break that pattern.
2021-01-06 13:58:52 +00:00
The client could name some of the fields and use < code > ..< / code > in the pattern, in which case adding another field is backwards compatible.
Making at least one of the struct's fields private forces clients to use the latter form of patterns, ensuring that the struct is future-proof.< / p >
< p > The downside of this approach is that you might need to add an otherwise unneeded field to the struct.
You can use the < code > ()< / code > type so that there is no runtime overhead and prepend < code > _< / code > to the field name to avoid the unused field warning.< / p >
< p > If Rust allowed private variants of enums, we could use the same trick to make adding a variant to an enum backwards compatible.
The problem there is exhaustive match expressions.
2021-01-06 14:19:00 +00:00
A private variant would force clients to have a < code > _< / code > wildcard pattern.
A common way to implement this instead is using the < a href = "https://doc.rust-lang.org/reference/attributes/type_system.html" > #[non_exhaustive]< / a > attribute.< / p >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#easy-doc-initialization" id = "easy-doc-initialization" > Easy doc initialization< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-14" id = "description-14" > Description< / a > < / h2 >
2021-01-05 14:36:44 +00:00
< p > If a struct takes significant effort to initialize, when writing docs, it can be quicker to wrap your example with a
2021-01-01 11:28:10 +00:00
function which takes the struct as an argument.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-6" id = "motivation-6" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Sometimes there is a struct with multiple or complicated parameters and several methods.
2021-01-05 14:36:44 +00:00
Each of these methods should have examples.< / p >
2021-01-01 11:28:10 +00:00
< p > For example:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > struct Connection {
2021-01-01 11:28:10 +00:00
name: String,
stream: TcpStream,
}
impl Connection {
/// Sends a request over the connection.
///
/// # Example
/// ```no_run
/// # // Boilerplate are required to get an example working.
/// # let stream = TcpStream::connect(" 127.0.0.1:34254" );
/// # let connection = Connection { name: " foo" .to_owned(), stream };
/// # let request = Request::new(" RequestId" , RequestType::Get, " payload" );
/// let response = connection.send_request(request);
/// assert!(response.is_ok());
/// ```
fn send_request(& self, request: Request) -> Result< Status, SendErr> {
// ...
}
2021-01-05 14:57:36 +00:00
2021-01-01 11:28:10 +00:00
/// Oh no, all that boilerplate needs to be repeated here!
fn check_status(& self) -> Status {
// ...
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-10" id = "example-10" > Example< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > Instead of typing all of this boiler plate to create an < code > Connection< / code > and < code > Request< / code > it is easier to just
create a wrapping dummy function which takes them as arguments:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > struct Connection {
2021-01-01 11:28:10 +00:00
name: String,
stream: TcpStream,
}
impl Connection {
/// Sends a request over the connection.
///
/// # Example
/// ```
/// # fn call_send(connection: Connection, request: Request) {
2021-01-05 15:02:50 +00:00
/// let response = connection.send_request(request);
2021-01-05 14:57:36 +00:00
/// assert!(response.is_ok());
2021-01-01 11:28:10 +00:00
/// # }
/// ```
fn send_request(& self, request: Request) {
// ...
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-08 08:00:20 +00:00
< p > < strong > Note< / strong > in the above example the line < code > assert!(response.is_ok());< / code > will not actually run while testing
because it is inside of a function which is never invoked.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-9" id = "advantages-9" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > This is much more concise and avoids repetitive code in examples.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-9" id = "disadvantages-9" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > As example is in a function, the code will not be tested. (Though it still will checked to make sure it compiles when running a < code > cargo test< / code > )
So this pattern is most useful when need < code > no_run< / code > . With this, you do not need to add < code > no_run< / code > .< / p >
< h2 > < a class = "header" href = "#discussion-5" id = "discussion-5" > Discussion< / a > < / h2 >
2021-01-05 14:36:44 +00:00
< p > If assertions are not required this pattern works well.< / p >
2021-01-08 08:00:20 +00:00
< p > If they are, an alternative can be to create a public method to create a dummy instance which is annotated
with < code > #[doc(hidden)]< / code > (so that users won't see it).
2021-01-01 11:28:10 +00:00
Then this method can be called inside of rustdoc because it is part of the crate's public API.< / p >
< h1 > < a class = "header" href = "#temporary-mutability" id = "temporary-mutability" > Temporary mutability< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-15" id = "description-15" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Often it is necessary to prepare and process some data, but after that data are only inspected
and never modified. The intention can be made explicit by redefining the mutable variable as immutable.< / p >
< p > It can be done either by processing data within nested block or by redefining variable.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-11" id = "example-11" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Say, vector must be sorted before usage.< / p >
< p > Using nested block:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > let data = {
2021-01-06 13:56:19 +00:00
let mut data = get_vec();
data.sort();
data
2021-01-01 11:28:10 +00:00
};
// Here `data` is immutable.
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > Using variable rebinding:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > let mut data = get_vec();
2021-01-01 11:28:10 +00:00
data.sort();
let data = data;
// Here `data` is immutable.
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-10" id = "advantages-10" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Compiler ensures that you don't accidentally mutate data after some point.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-10" id = "disadvantages-10" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Nested block requires additional indentation of block body.
One more line to return data from block or redefine variable.< / p >
< h1 > < a class = "header" href = "#design-patterns-1" id = "design-patterns-1" > Design Patterns< / a > < / h1 >
2021-01-08 08:00:20 +00:00
< p > < a href = "https://en.wikipedia.org/wiki/Software_design_pattern" > Design patterns< / a > are " general reusable
solutions to a commonly occurring problem within a given context in software design" .
Design patterns are a great way to describe some of the culture and 'tribal knowledge'
of programming in a language.
Design patterns are very language-specific - what is a pattern in one language may be
unnecessary in another due to a language feature, or impossible to express due to a missing feature.< / p >
< p > If overused, design patterns can add unnecessary complexity to programs.
However, they are a great way to share intermediate and advanced level knowledge about a programming language.< / p >
2021-01-02 11:15:56 +00:00
< h2 > < a class = "header" href = "#design-patterns-in-rust-1" id = "design-patterns-in-rust-1" > Design patterns in Rust< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > Rust has many very unique features. These features give us great benefit by removing whole classes of problems.
Some of them are also patterns that are < em > unique< / em > to Rust.< / p >
2021-01-02 11:15:56 +00:00
< h2 > < a class = "header" href = "#yagni" id = "yagni" > YAGNI< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > If you're not familiar with it, YAGNI is an acronym that stands for < code > You Aren't Going to Need It< / code > .
It's an important software design principle to apply as you write code.< / p >
2021-01-02 11:15:56 +00:00
< blockquote >
< p > The best code I ever wrote was code I never wrote.< / p >
< / blockquote >
2021-01-08 08:00:20 +00:00
< p > If we apply YAGNI to design patterns, we see that the features of Rust allow us to throw out many patterns.
For instance, there is no need for the < a href = "https://en.wikipedia.org/wiki/Strategy_pattern" > strategy pattern< / a > in Rust
because we can just use < a href = "https://doc.rust-lang.org/book/traits.html" > traits< / a > .< / p >
2021-01-02 11:15:56 +00:00
< p > TODO: Maybe include some code to illustrate the traits.< / p >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#builder" id = "builder" > Builder< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-16" id = "description-16" > Description< / a > < / h2 >
2021-01-03 10:35:52 +00:00
< p > Construct an object with calls to a builder helper.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-12" id = "example-12" > Example< / a > < / h2 >
2021-01-03 10:35:52 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > #[derive(Debug, PartialEq)]
pub struct Foo {
2021-01-01 11:28:10 +00:00
// Lots of complicated fields.
2021-01-03 10:35:52 +00:00
bar: String,
2021-01-01 11:28:10 +00:00
}
2021-01-03 10:35:52 +00:00
pub struct FooBuilder {
2021-01-01 11:28:10 +00:00
// Probably lots of optional fields.
2021-01-03 10:35:52 +00:00
bar: String,
2021-01-01 11:28:10 +00:00
}
impl FooBuilder {
2021-01-03 10:35:52 +00:00
pub fn new(/* ... */) -> FooBuilder {
2021-01-01 11:28:10 +00:00
// Set the minimally required fields of Foo.
2021-01-03 10:35:52 +00:00
FooBuilder {
bar: String::from(" X" ),
}
2021-01-01 11:28:10 +00:00
}
2021-01-03 10:35:52 +00:00
pub fn name(mut self, bar: String) -> FooBuilder {
2021-01-01 11:28:10 +00:00
// Set the name on the builder itself, and return the builder by value.
2021-01-03 10:35:52 +00:00
self.bar = bar;
self
2021-01-01 11:28:10 +00:00
}
// If we can get away with not consuming the Builder here, that is an
2021-01-03 10:35:52 +00:00
// advantage. It means we can use the FooBuilder as a template for constructing many Foos.
pub fn build(self) -> Foo {
2021-01-01 11:28:10 +00:00
// Create a Foo from the FooBuilder, applying all settings in FooBuilder to Foo.
2021-01-03 10:35:52 +00:00
Foo { bar: self.bar }
2021-01-01 11:28:10 +00:00
}
}
2021-01-03 10:35:52 +00:00
#[test]
fn builder_test() {
let foo = Foo {
bar: String::from(" Y" ),
};
let foo_from_builder: Foo = FooBuilder::new().name(String::from(" Y" )).build();
assert_eq!(foo, foo_from_builder);
2021-01-01 11:28:10 +00:00
}
2021-01-03 10:35:52 +00:00
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-7" id = "motivation-7" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Useful when you would otherwise require many different constructors or where
construction has side effects.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-11" id = "advantages-11" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Separates methods for building from other methods.< / p >
< p > Prevents proliferation of constructors< / p >
< p > Can be used for one-liner initialisation as well as more complex construction.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-11" id = "disadvantages-11" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > More complex than creating a struct object directly, or a simple constructor
function.< / p >
< h2 > < a class = "header" href = "#discussion-6" id = "discussion-6" > Discussion< / a > < / h2 >
< p > This pattern is seen more frequently in Rust (and for simpler objects) than in
many other languages because Rust lacks overloading. Since you can only have a
single method with a given name, having multiple constructors is less nice in
Rust than in C++, Java, or others.< / p >
< p > This pattern is often used where the builder object is useful in its own right,
rather than being just a builder. For example, see
< a href = "https://doc.rust-lang.org/std/process/struct.Command.html" > < code > std::process::Command< / code > < / a >
is a builder for < a href = "https://doc.rust-lang.org/std/process/struct.Child.html" > < code > Child< / code > < / a >
(a process). In these cases, the < code > T< / code > and < code > TBuilder< / code > pattern
of naming is not used.< / p >
< p > The example takes and returns the builder by value. It is often more ergonomic
(and more efficient) to take and return the builder as a mutable reference. The
borrow checker makes this work naturally. This approach has the advantage that
one can write code like< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > let mut fb = FooBuilder::new();
2021-01-01 11:28:10 +00:00
fb.a();
fb.b();
2021-01-03 10:35:52 +00:00
let f = fb.build();
2021-01-01 11:28:10 +00:00
< / code > < / pre >
2021-01-03 10:35:52 +00:00
< p > as well as the < code > FooBuilder::new().a().b().build()< / code > style.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#see-also-8" id = "see-also-8" > See also< / a > < / h2 >
2021-01-05 16:24:27 +00:00
< ul >
< li > < a href = "https://web.archive.org/web/20210104103100/https://doc.rust-lang.org/1.12.0/style/ownership/builders.html" > Description in the style guide< / a > < / li >
< li > < a href = "https://crates.io/crates/derive_builder" > derive_builder< / a > , a crate for automatically implementing this pattern while avoiding the boilerplate.< / li >
< li > < a href = "patterns/../idioms/ctor.html" > Constructor pattern< / a > for when construction is simpler.< / li >
< li > < a href = "https://en.wikipedia.org/wiki/Builder_pattern" > Builder pattern (wikipedia)< / a > < / li >
2021-01-08 08:00:20 +00:00
< li > < a href = "https://web.archive.org/web/20210104103000/https://rust-lang.github.io/api-guidelines/type-safety.html#c-builder" > Construction of complex values< / a > < / li >
2021-01-05 16:24:27 +00:00
< / ul >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#compose-structs-together-for-better-borrowing" id = "compose-structs-together-for-better-borrowing" > Compose structs together for better borrowing< / a > < / h1 >
< p > TODO - this is not a very snappy name< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-17" id = "description-17" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Sometimes a large struct will cause issues with the borrow checker - although
fields can be borrowed independently, sometimes the whole struct ends up being
used at once, preventing other uses. A solution might be to decompose the struct
into several smaller structs. Then compose these together into the original
struct. Then each struct can be borrowed separately and have more flexible
behaviour.< / p >
< p > This will often lead to a better design in other ways: applying this design
pattern often reveals smaller units of functionality.< / p >
2021-01-04 08:32:25 +00:00
< h2 > < a class = "header" href = "#example-13" id = "example-13" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Here is a contrived example of where the borrow checker foils us in our plan to
use a struct:< / p >
2021-01-03 07:34:40 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > struct A {
2021-01-01 11:28:10 +00:00
f1: u32,
f2: u32,
f3: u32,
}
fn foo(a: & mut A) -> & u32 { & a.f2 }
fn bar(a: & mut A) -> u32 { a.f1 + a.f3 }
2021-01-02 11:41:18 +00:00
fn baz(a: & mut A) {
2021-01-03 07:34:40 +00:00
// The later usage of x causes a to be borrowed for the rest of the function.
2021-01-01 11:28:10 +00:00
let x = foo(a);
2021-01-03 07:34:40 +00:00
// Borrow checker error:
// let y = bar(a); // ~ ERROR: cannot borrow `*a` as mutable more than once at a time
println!(" {}" , x);
2021-01-01 11:28:10 +00:00
}
2021-01-03 07:34:40 +00:00
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-01 11:28:10 +00:00
< p > We can apply this design pattern and refactor < code > A< / code > into two smaller structs, thus
solving the borrow checking issue:< / p >
2021-01-02 11:41:18 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > // A is now composed of two structs - B and C.
2021-01-01 11:28:10 +00:00
struct A {
b: B,
c: C,
}
struct B {
f2: u32,
}
struct C {
f1: u32,
f3: u32,
}
// These functions take a B or C, rather than A.
fn foo(b: & mut B) -> & u32 { & b.f2 }
fn bar(c: & mut C) -> u32 { c.f1 + c.f3 }
2021-01-02 11:41:18 +00:00
fn baz(a: & mut A) {
2021-01-01 11:28:10 +00:00
let x = foo(& mut a.b);
// Now it's OK!
let y = bar(& mut a.c);
2021-01-03 07:34:40 +00:00
println!(" {}" , x);
2021-01-01 11:28:10 +00:00
}
2021-01-02 11:41:18 +00:00
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-8" id = "motivation-8" > Motivation< / a > < / h2 >
2021-01-03 02:43:25 +00:00
< p > TODO Why and where you should use the pattern< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-12" id = "advantages-12" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Lets you work around limitations in the borrow checker.< / p >
< p > Often produces a better design.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-12" id = "disadvantages-12" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Leads to more verbose code.< / p >
< p > Sometimes, the smaller structs are not good abstractions, and so we end up with
a worse design. That is probably a 'code smell', indicating that the program
should be refactored in some way.< / p >
< h2 > < a class = "header" href = "#discussion-7" id = "discussion-7" > Discussion< / a > < / h2 >
< p > This pattern is not required in languages that don't have a borrow checker, so
in that sense is unique to Rust. However, making smaller units of functionality
often leads to cleaner code: a widely acknowledged principle of software
engineering, independent of the language.< / p >
< p > This pattern relies on Rust's borrow checker to be able to borrow fields
independently of each other. In the example, the borrow checker knows that < code > a.b< / code >
and < code > a.c< / code > are distinct and can be borrowed independently, it does not try to
borrow all of < code > a< / code > , which would make this pattern useless.< / p >
2021-01-05 15:44:58 +00:00
< h1 > < a class = "header" href = "#entry-api" id = "entry-api" > Entry API< / a > < / h1 >
< h2 > < a class = "header" href = "#description-18" id = "description-18" > Description< / a > < / h2 >
< p > A short, prose description of the pattern.< / p >
< h2 > < a class = "header" href = "#example-14" id = "example-14" > Example< / a > < / h2 >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > // An example of the pattern in action, should be mostly code, commented
// liberally.
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
< h2 > < a class = "header" href = "#motivation-9" id = "motivation-9" > Motivation< / a > < / h2 >
< p > Why and where you should use the pattern< / p >
< h2 > < a class = "header" href = "#advantages-13" id = "advantages-13" > Advantages< / a > < / h2 >
< p > Good things about this pattern.< / p >
< h2 > < a class = "header" href = "#disadvantages-13" id = "disadvantages-13" > Disadvantages< / a > < / h2 >
< p > Bad things about this pattern. Possible contraindications.< / p >
< h2 > < a class = "header" href = "#discussion-8" id = "discussion-8" > Discussion< / a > < / h2 >
< p > TODO vs insert_or_update etc.< / p >
< h2 > < a class = "header" href = "#see-also-9" id = "see-also-9" > See also< / a > < / h2 >
< p > < a href = "https://github.com/rust-lang/rfcs/blob/master/text/0216-collection-views.md" > RFC< / a >
< a href = "https://github.com/rust-lang/rfcs/blob/8e2d3a3341da533f846f61f10335b72c9a9f4740/text/0921-entry_v3.md" > RFC< / a > < / p >
< p > < a href = "https://doc.rust-lang.org/std/collections/struct.HashMap.html#method.entry" > Hashmap::entry docs< / a > < / p >
< h1 > < a class = "header" href = "#ffi-patterns" id = "ffi-patterns" > FFI Patterns< / a > < / h1 >
< p > Writing FFI code is an entire course in itself.
However, there are several idioms here that can act as pointers, and avoid traps for inexperienced users of unsafe Rust.< / p >
< p > This section contains design patterns that may be useful when doing FFI.< / p >
< ol >
< li >
< p > < a href = "patterns/./ffi-export.html" > Object-Based API< / a > design that has good memory safety characteristics, and a clean boundary of what is safe and what is unsafe< / p >
< / li >
< li >
< p > < a href = "patterns/./ffi-wrappers.html" > Type Consolidation into Wrappers< / a > - group multiple Rust types together into an opaque " object" < / p >
< / li >
< / ol >
< h1 > < a class = "header" href = "#object-based-apis" id = "object-based-apis" > Object-Based APIs< / a > < / h1 >
< h2 > < a class = "header" href = "#description-19" id = "description-19" > Description< / a > < / h2 >
< p > When designing APIs in Rust which are exposed to other languages, there are some important design principles which are contrary to normal Rust API design:< / p >
< ol >
< li > All Encapsulated types should be < em > owned< / em > by Rust, < em > managed< / em > by the user, and < em > opaque< / em > .< / li >
< li > All Transactional data types should be < em > owned< / em > by the user, and < em > transparent< / em > .< / li >
< li > All library behavior should be functions acting upon Encapsulated types.< / li >
< li > All library behavior should be encapsulated into types not based on structure, but < em > provenance/lifetime< / em > .< / li >
< / ol >
< h2 > < a class = "header" href = "#motivation-10" id = "motivation-10" > Motivation< / a > < / h2 >
< p > Rust has built-in FFI support to other languages.
It does this by providing a way for crate authors to provide C-compatible APIs through different ABIs (though that is unimportant to this practice).< / p >
2021-01-08 08:00:20 +00:00
< p > Well-designed Rust FFI follows C API design principles, while compromising the design in Rust as little as possible.
There are three goals with any foreign API:< / p >
2021-01-05 15:44:58 +00:00
< ol >
< li > Make it easy to use in the target language.< / li >
< li > Avoid the API dictating internal unsafety on the Rust side as much as possible.< / li >
< li > Keep the potential for memory unsafety and Rust < code > undefined behaviour< / code > as small as possible.< / li >
< / ol >
< p > Rust code must trust the memory safety of the foreign language beyond a certain point.
However, every bit of < code > unsafe< / code > code on the Rust side is an opportunity for bugs, or to exacerbate < code > undefined behaviour< / code > .< / p >
< p > For example, if a pointer provenance is wrong, that may be a segfault due to invalid memory access.
But if it is manipulated by unsafe code, it could become full-blown heap corruption.< / p >
< p > The Object-Based API design allows for writing shims that have good memory safety characteristics, and a clean boundary of what is safe and what is < code > unsafe< / code > .< / p >
< h2 > < a class = "header" href = "#code-example-3" id = "code-example-3" > Code Example< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > The POSIX standard defines the API to access an on-file database, known as < a href = "https://web.archive.org/web/20210105035602/https://www.mankier.com/0p/ndbm.h" > DBM< / a > .
It is an excellent example of an " object-based" API.< / p >
2021-01-05 15:44:58 +00:00
< p > Here is the definition in C, which hopefully should be easy to read for those involved in FFI.
The commentary below should help explaining it for those who miss the subtleties.< / p >
< pre > < code class = "language-C" > struct DBM;
typedef struct { void *dptr, size_t dsize } datum;
int dbm_clearerr(DBM *);
void dbm_close(DBM *);
int dbm_delete(DBM *, datum);
int dbm_error(DBM *);
datum dbm_fetch(DBM *, datum);
datum dbm_firstkey(DBM *);
datum dbm_nextkey(DBM *);
DBM *dbm_open(const char *, int, mode_t);
int dbm_store(DBM *, datum, datum, int);
< / code > < / pre >
< p > This API defines two types: < code > DBM< / code > and < code > datum< / code > .< / p >
< p > The < code > DBM< / code > type was called an " encapsulated" type above.
It is designed to contain internal state, and acts as an entry point for the library's behavior.< / p >
< p > It is completely opaque to the user, who cannot create a < code > DBM< / code > themselves since they don't know its size or layout.
Instead, they must call < code > dbm_open< / code > , and that only gives them < em > a pointer to one< / em > .< / p >
2021-01-08 08:00:20 +00:00
< p > This means all < code > DBM< / code > s are " owned" by the library in a Rust sense.
The internal state of unknown size is kept in memory controlled by the library, not the user.
2021-01-05 15:44:58 +00:00
The user can only manage its life cycle with < code > open< / code > and < code > close< / code > , and perform operations on it with the other functions.< / p >
2021-01-08 08:00:20 +00:00
< p > The < code > datum< / code > type was called a " transactional" type above.
It is designed to facilitate the exchange of information between the library and its user.< / p >
2021-01-05 15:44:58 +00:00
< p > The database is designed to store " unstructured data" , with no pre-defined length or meaning.
As a result, the < code > datum< / code > is the C equivalent of a Rust slice: a bunch of bytes, and a count of how many there are.
The main difference is that there is no type information, which is what < code > void< / code > indicates.< / p >
< p > Keep in mind that this header is written from the library's point of view.
The user likely has some type they are using, which has a known size.
But the library does not care, and by the rules of C casting, any type behind a pointer can be cast to < code > void< / code > .< / p >
< p > As noted earlier, this type is < em > transparent< / em > to the user. But also, this type is < em > owned< / em > by the user.
This has subtle ramifications, due to that pointer inside it.
The question is, who owns the memory that pointer points to?< / p >
< p > The answer for best memory safety is, " the user" .
But in cases such as retrieving a value, the user does not know how to allocate it correctly (since they don't know how long the value is).
2021-01-08 08:00:20 +00:00
In this case, the library code is expected to use the heap that the user has access to --
such as the C library < code > malloc< / code > and < code > free< / code > -- and then < em > transfer ownership< / em > in the Rust sense.< / p >
2021-01-05 15:44:58 +00:00
< p > This may all seem speculative, but this is what a pointer means in C.
It means the same thing as Rust: " user defined lifetime."
The user of the library needs to read the documentation in order to use it correctly.
That said, there are some decisions that have fewer or greater consequences if users do it wrong.
Minimizing those is what this best practice is about, and the key is to < em > transfer ownership of everything that is transparent< / em > .< / p >
< h2 > < a class = "header" href = "#advantages-14" id = "advantages-14" > Advantages< / a > < / h2 >
< p > This minimizes the number of memory safety guarantees the user must uphold to a relatively small number:< / p >
< ol >
< li > Do not call any function with a pointer not returned by < code > dbm_open< / code > (invalid access or corruption).< / li >
< li > Do not call any function on a pointer after close (use after free).< / li >
< li > The < code > dptr< / code > on any < code > datum< / code > must be < code > NULL< / code > , or point to a valid slice of memory at the advertised length.< / li >
< / ol >
< p > In addition, it avoids a lot of pointer provenance issues.
To understand why, let us consider an alternative in some depth: key iteration.< / p >
< p > Rust is well known for its iterators.
When implementing one, the programmer makes a separate type with a bounded lifetime to its owner, and implements the < code > Iterator< / code > trait.< / p >
< p > Here is how iteration would be done in Rust for < code > DBM< / code > :< / p >
< pre > < code class = "language-rust ignore" > struct Dbm { ... }
impl Dbm {
/* ... */
pub fn keys< 'it> (& 'it self) -> DbmKeysIter< 'it> { ... }
/* ... */
}
struct DbmKeysIter< 'it> {
owner: & 'it Dbm,
}
impl< 'it> Iterator for DbmKeysIter< 'it> { ... }
< / code > < / pre >
< p > This is clean, idiomatic, and safe. thanks to Rust's guarantees.
However, consider what a straightforward API translation would look like:< / p >
< pre > < code class = "language-rust ignore" > #[no_mangle]
pub extern " C" fn dbm_iter_new(owner: *const Dbm) -> *mut DbmKeysIter {
/* THIS API IS A BAD IDEA! For real applications, use object-based design instead. */
}
#[no_mangle]
pub extern " C" fn dbm_iter_next(iter: *mut DbmKeysIter, key_out: *const datum) -> libc::c_int {
/* THIS API IS A BAD IDEA! For real applications, use object-based design instead. */
}
#[no_mangle]
pub extern " C" fn dbm_iter_del(*mut DbmKeysIter) {
/* THIS API IS A BAD IDEA! For real applications, use object-based design instead. */
}
< / code > < / pre >
< p > This API loses a key piece of information: the lifetime of the iterator must not exceed the lifetime of the < code > Dbm< / code > object that owns it.
A user of the library could use it in a way which causes the iterator to outlive the data it is iterating on, resulting in reading uninitialized memory.< / p >
< p > This example written in C contains a bug that will be explained afterwards:< / p >
< pre > < code class = "language-C" > int count_key_sizes(DBM *db) {
/* DO NOT USE THIS FUNCTION. IT HAS A SUBTLE BUT SERIOUS BUG! */
datum key;
int len = 0;
if (!dbm_iter_new(db)) {
dbm_close(db);
return -1;
}
int l;
while ((l = dbm_iter_next(owner, & key)) > = 0) { // an error is indicated by -1
free(key.dptr);
len += key.dsize;
if (l == 0) { // end of the iterator
dbm_close(owner);
}
}
if l > = 0 {
return -1;
} else {
return len;
}
}
< / code > < / pre >
< p > This bug is a classic. Here's what happens when the iterator returns the end-of-iteration marker:< / p >
< ol >
< li > The loop condition sets < code > l< / code > to zero, and enters the loop because < code > 0 > = 0< / code > .< / li >
< li > The length is incremented, in this case by zero.< / li >
< li > The if statement is true, so the database is closed. There should be a break statement here.< / li >
< li > The loop condition executes again, causing a < code > next< / code > call on the closed object.< / li >
< / ol >
< p > The worst part about this bug?
If the Rust implementation was careful, this code will work most of the time!
2021-01-08 08:00:20 +00:00
If the memory for the < code > Dbm< / code > object is not immediately reused, an internal check will almost certainly fail,
resulting in the iterator returning a < code > -1< / code > indicating an error.
2021-01-05 15:44:58 +00:00
But occasionally, it will cause a segmentation fault, or even worse, nonsensical memory corruption!< / p >
< p > None of this can be avoided by Rust.
2021-01-08 08:00:20 +00:00
From its perspective, it put those objects on its heap, returned pointers to them, and gave up control of their lifetimes.
The C code simply must " play nice" .< / p >
2021-01-05 15:44:58 +00:00
< p > The programmer must read and understand the API documentation.
While some consider that par for the course in C, a good API design can mitigate this risk.
The POSIX API for < code > DBM< / code > did this by < em > consolidating the ownership< / em > of the iterator with its parent:< / p >
< pre > < code class = "language-C" > datum dbm_firstkey(DBM *);
datum dbm_nextkey(DBM *);
< / code > < / pre >
< p > Thus, all of the lifetimes were bound together, and such unsafety was prevented.< / p >
< h2 > < a class = "header" href = "#disadvantages-14" id = "disadvantages-14" > Disadvantages< / a > < / h2 >
< p > However, this design choice also has a number of drawbacks, which should be considered as well.< / p >
< p > First, the API itself becomes less expressive.
With POSIX DBM, there is only one iterator per object, and every call changes its state.
This is much more restrictive than iterators in almost any language, even though it is safe.
Perhaps with other related objects, whose lifetimes are less hierarchical, this limitation is more of a cost than the safety.< / p >
< p > Second, depending on the relationships of the API's parts, significant design effort may be involved.
Many of the easier design points have other patterns associated with them:< / p >
< ul >
< li >
< p > < a href = "patterns/./ffi-wrappers.html" > Wrapper Type Consolidation< / a > groups multiple Rust types together into an opaque " object" < / p >
< / li >
< li >
< p > < a href = "patterns/../idioms/ffi-errors.html" > FFI Error Passing< / a > explains error handling with integer codes and sentinel return values (such as < code > NULL< / code > pointers)< / p >
< / li >
< li >
2021-01-08 08:00:20 +00:00
< p > < a href = "patterns/../idioms/ffi-accepting-strings.html" > Accepting Foreign Strings< / a > allows accepting strings with minimal unsafe code,
and is easier to get right than < a href = "patterns/../idioms/ffi-passing-strings.html" > Passing Strings to FFI< / a > < / p >
2021-01-05 15:44:58 +00:00
< / li >
< / ul >
< p > However, not every API can be done this way.
It is up to the best judgement of the programmer as to who their audience is.< / p >
< h1 > < a class = "header" href = "#type-consolidation-into-wrappers" id = "type-consolidation-into-wrappers" > Type Consolidation into Wrappers< / a > < / h1 >
< h2 > < a class = "header" href = "#description-20" id = "description-20" > Description< / a > < / h2 >
< p > This pattern is designed to allow gracefully handling multiple related types, while minimizing the surface area for memory unsafety.< / p >
< p > One of the cornerstones of Rust's aliasing rules is lifetimes.
This ensures that many patterns of access between types can be memory safe, data race safety included.< / p >
< p > However, when Rust types are exported to other languages, they are usually transformed into pointers.
In Rust, a pointer means " the user manages the lifetime of the pointee." It is their responsibility to avoid memory unsafety.< / p >
< p > Some level of trust in the user code is thus required, notably around use-after-free which Rust can do nothing about.
However, some API designs place higher burdens than others on the code written in the other language.< / p >
2021-01-08 08:00:20 +00:00
< p > The lowest risk API is the " consolidated wrapper" , where all possible interactions with an object
are folded into a " wrapper type" , while keeping the Rust API clean.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#code-example-4" id = "code-example-4" > Code Example< / a > < / h2 >
< p > To understand this, let us look at a classic example of an API to export: iteration through a collection.< / p >
< p > That API looks like this:< / p >
< ol >
< li > The iterator is initialized with < code > first_key< / code > .< / li >
< li > Each call to < code > next_key< / code > will advance the iterator.< / li >
< li > Calls to < code > next_key< / code > if the iterator is at the end will do nothing.< / li >
< li > As noted above, the iterator is " wrapped into" the collection (unlike the native Rust API).< / li >
< / ol >
< p > If the iterator implements < code > nth()< / code > efficiently, then it is possible to make it ephemeral to each function call:< / p >
< pre > < code class = "language-rust ignore" > struct MySetWrapper {
myset: MySet,
iter_next: usize,
}
impl MySetWrapper {
pub fn first_key(& mut self) -> Option< & Key> {
self.iter_next = 0;
self.next_key()
}
pub fn next_key(& mut self) -> Option< & Key> {
if let Some(next) = self.myset.keys().nth(self.iter_next) {
self.iter_next += 1;
Some(next)
} else {
None
}
}
}
< / code > < / pre >
< p > As a result, the wrapper is simple and contains no < code > unsafe< / code > code.< / p >
< h2 > < a class = "header" href = "#advantages-15" id = "advantages-15" > Advantages< / a > < / h2 >
< p > This makes APIs safer to use, avoiding issues with lifetimes between types.
See < a href = "patterns/./ffi-export.html" > Object-Based APIs< / a > for more on the advantages and pitfalls this avoids.< / p >
< h2 > < a class = "header" href = "#disadvantages-15" id = "disadvantages-15" > Disadvantages< / a > < / h2 >
< p > Often, wrapping types is quite difficult, and sometimes a Rust API compromise would make things easier.< / p >
< p > As an example, consider an iterator which does not efficiently implement < code > nth()< / code > .
2021-01-08 08:00:20 +00:00
It would definitely be worth putting in special logic to make the object handle iteration internally,
or to support a different access pattern efficiently that only the Foreign Function API will use.< / p >
2021-01-05 15:44:58 +00:00
< h3 > < a class = "header" href = "#trying-to-wrap-iterators-and-failing" id = "trying-to-wrap-iterators-and-failing" > Trying to Wrap Iterators (and Failing)< / a > < / h3 >
2021-01-08 08:00:20 +00:00
< p > To wrap any type of iterator into the API correctly, the wrapper would need to do what a C version of
the code would do: erase the lifetime of the iterator, and manage it manually.< / p >
2021-01-05 15:44:58 +00:00
< p > Suffice it to say, this is < em > incredibly< / em > difficult.< / p >
< p > Here is an illustration of just < em > one< / em > pitfall.< / p >
< p > A first version of < code > MySetWrapper< / code > would look like this:< / p >
< pre > < code class = "language-rust ignore" > struct MySetWrapper {
myset: MySet,
iter_next: usize,
// created from a transmuted Box< KeysIter + 'self>
iterator: Option< NonNull< KeysIter< 'static> > > ,
}
< / code > < / pre >
< p > With < code > transmute< / code > being used to extend a lifetime, and a pointer to hide it, it's ugly already.
But it gets even worse: < em > any other operation can cause Rust < code > undefined behaviour< / code > < / em > .< / p >
2021-01-08 08:00:20 +00:00
< p > Consider that the < code > MySet< / code > in the wrapper could be manipulated by other functions during iteration,
such as storing a new value to the key it was iterating over.
2021-01-05 15:44:58 +00:00
The API doesn't discourage this, and in fact some similar C libraries expect it.< / p >
< p > A simple implementation of < code > myset_store< / code > would be:< / p >
< pre > < code class = "language-rust ignore" > pub mod unsafe_module {
// other module content
pub fn myset_store(
myset: *mut MySetWrapper,
key: datum,
value: datum) -> libc::c_int {
/* DO NOT USE THIS CODE. IT IS UNSAFE TO DEMONSTRATE A PROLBEM. */
let myset: & mut MySet = unsafe { // SAFETY: whoops, UB occurs in here!
& mut (*myset).myset
};
/* ...check and cast key and value data... */
match myset.store(casted_key, casted_value) {
Ok(_) => 0,
Err(e) => e.into()
}
}
}
< / code > < / pre >
< p > If the iterator exists when this function is called, we have violated one of Rust's aliasing rules.
According to Rust, the mutable reference in this block must have < em > exclusive< / em > access to the object.
If the iterator simply exists, it's not exclusive, so we have < code > undefined behaviour< / code > ! < sup class = "footnote-reference" > < a href = "#1" > 1< / a > < / sup > < / p >
< p > To avoid this, we must have a way of ensuring that mutable reference really is exclusive.
That basically means clearing out the iterator's shared reference while it exists, and then reconstructing it.
In most cases, that will still be less efficient than the C version.< / p >
< p > Some may ask: how can C do this more efficiently?
The answer is, it cheats. Rust's aliasing rules are the problem, and C simply ignores them for its pointers.
In exchange, it is common to see code that is declared in the manual as " not thread safe" under some or all circumstances.
In fact, < a href = "https://manpages.debian.org/buster/manpages/attributes.7.en.html" > The GNU C library has an entire lexicon dedicated to concurrent behavior!< / a > < / p >
< p > Rust would rather make everything memory safe all the time, for both safety and optimizations that C code cannot attain.
Being denied access to certain shortcuts is the price Rust programmers need to pay.< / p >
< div class = "footnote-definition" id = "1" > < sup class = "footnote-definition-label" > 1< / sup >
< p > For the C programmers out there scratching their heads, the iterator need not be read < em > during< / em > this code cause the UB.
2021-01-08 08:00:20 +00:00
The exclusivity rule also enables compiler optimizations which may cause inconsistent observations by the iterator's
shared reference (e.g. stack spills or reordering instructions for efficiency).
2021-01-05 15:44:58 +00:00
These observations may happen < em > any time after< / em > the mutable reference is created.< / p >
< / div >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#fold" id = "fold" > Fold< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-21" id = "description-21" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Run an algorithm over each item in a collection of data to create a new item,
thus creating a whole new collection.< / p >
< p > The etymology here is unclear to me. The terms 'fold' and 'folder' are used
in the Rust compiler, although it appears to me to be more like a map than a
fold in the usual sense. See the discussion below for more details.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-15" id = "example-15" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > // The data we will fold, a simple AST.
2021-01-01 11:28:10 +00:00
mod ast {
pub enum Stmt {
Expr(Box< Expr> ),
Let(Box< Name> , Box< Expr> ),
}
pub struct Name {
value: String,
}
pub enum Expr {
IntLit(i64),
Add(Box< Expr> , Box< Expr> ),
Sub(Box< Expr> , Box< Expr> ),
}
}
// The abstract folder
mod fold {
use ast::*;
pub trait Folder {
// A leaf node just returns the node itself. In some cases, we can do this
// to inner nodes too.
fn fold_name(& mut self, n: Box< Name> ) -> Box< Name> { n }
// Create a new inner node by folding its children.
fn fold_stmt(& mut self, s: Box< Stmt> ) -> Box< Stmt> {
match *s {
Stmt::Expr(e) => Box::new(Stmt::Expr(self.fold_expr(e))),
Stmt::Let(n, e) => Box::new(Stmt::Let(self.fold_name(n), self.fold_expr(e))),
}
}
fn fold_expr(& mut self, e: Box< Expr> ) -> Box< Expr> { ... }
}
}
use fold::*;
use ast::*;
// An example concrete implementation - renames every name to 'foo'.
struct Renamer;
impl Folder for Renamer {
fn fold_name(& mut self, n: Box< Name> ) -> Box< Name> {
Box::new(Name { value: " foo" .to_owned() })
}
// Use the default methods for the other nodes.
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > The result of running the < code > Renamer< / code > on an AST is a new AST identical to the old
one, but with every name changed to < code > foo< / code > . A real life folder might have some
state preserved between nodes in the struct itself.< / p >
< p > A folder can also be defined to map one data structure to a different (but
usually similar) data structure. For example, we could fold an AST into a HIR
tree (HIR stands for high-level intermediate representation).< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-11" id = "motivation-11" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > It is common to want to map a data structure by performing some operation on
each node in the structure. For simple operations on simple data structures,
this can be done using < code > Iterator::map< / code > . For more complex operations, perhaps
where earlier nodes can affect the operation on later nodes, or where iteration
over the data structure is non-trivial, using the fold pattern is more
appropriate.< / p >
< p > Like the visitor pattern, the fold pattern allows us to separate traversal of a
data structure from the operations performed to each node.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#discussion-9" id = "discussion-9" > Discussion< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Mapping data structures in this fashion is common in functional languages. In OO
languages, it would be more common to mutate the data structure in place. The
'functional' approach is common in Rust, mostly due to the preference for
immutability. Using fresh data structures, rather than mutating old ones, makes
reasoning about the code easier in most circumstances.< / p >
< p > The trade-off between efficiency and reusability can be tweaked by changing how
nodes are accepted by the < code > fold_*< / code > methods.< / p >
< p > In the above example we operate on < code > Box< / code > pointers. Since these own their data
exclusively, the original copy of the data structure cannot be re-used. On the
other hand if a node is not changed, reusing it is very efficient.< / p >
< p > If we were to operate on borrowed references, the original data structure can be
2021-01-01 13:23:56 +00:00
reused; however, a node must be cloned even if unchanged, which can be
2021-01-01 11:28:10 +00:00
expensive.< / p >
< p > Using a reference counted pointer gives the best of both worlds - we can reuse
the original data structure and we don't need to clone unchanged nodes. However,
they are less ergonomic to use and mean that the data structures cannot be
mutable.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-10" id = "see-also-10" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Iterators have a < code > fold< / code > method, however this folds a data structure into a
value, rather than into a new data structure. An iterator's < code > map< / code > is more like
this fold pattern.< / p >
< p > In other languages, fold is usually used in the sense of Rust's iterators,
rather than this pattern. Some functional languages have powerful constructs for
performing flexible maps over data structures.< / p >
< p > The < a href = "patterns/visitor.html" > visitor< / a > pattern is closely related to fold. They share the
concept of walking a data structure performing an operation on each node.
However, the visitor does not create a new data structure nor consume the old
one.< / p >
< h1 > < a class = "header" href = "#newtype" id = "newtype" > Newtype< / a > < / h1 >
2021-01-05 14:36:44 +00:00
< p > What if in some cases we want a type to behave similar to another type or
enforce some behaviour at compile time where using only type aliases would
not be enough?< / p >
< p > For example, if we want to create a custom < code > Display< / code > implementation for < code > String< / code >
2021-01-05 03:33:00 +00:00
due to security considerations (e.g. passwords).< / p >
< p > For such cases we could use the < code > Newtype< / code > pattern to provide < strong > type safety< / strong > and < strong > encapsulation< / strong > .< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-22" id = "description-22" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Use a tuple struct with a single field to make an opaque wrapper for a type.
This creates a new type, rather than an alias to a type (< code > type< / code > items).< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-16" id = "example-16" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > // Some type, not necessarily in the same module or even crate.
2021-01-01 11:28:10 +00:00
struct Foo {
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
impl Foo {
// These functions are not present on Bar.
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
// The newtype.
pub struct Bar(Foo);
impl Bar {
// Constructor.
2021-01-02 11:41:18 +00:00
pub fn new(
//..
) -> Bar {
2021-01-05 14:57:36 +00:00
2021-01-02 11:41:18 +00:00
//..
2021-01-05 14:57:36 +00:00
2021-01-01 11:28:10 +00:00
}
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
fn main() {
let b = Bar::new(...);
// Foo and Bar are type incompatible, the following do not type check.
// let f: Foo = b;
// let b: Bar = Foo { ... };
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-12" id = "motivation-12" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The primary motivation for newtypes is abstraction. It allows you to share
implementation details between types while precisely controlling the interface.
By using a newtype rather than exposing the implementation type as part of an
API, it allows you to change implementation backwards compatibly.< / p >
< p > Newtypes can be used for distinguishing units, e.g., wrapping < code > f64< / code > to give
distinguishable < code > Miles< / code > and < code > Kms< / code > .< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-16" id = "advantages-16" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The wrapped and wrapper types are not type compatible (as opposed to using
< code > type< / code > ), so users of the newtype will never 'confuse' the wrapped and wrapper
types.< / p >
< p > Newtypes are a zero-cost abstraction - there is no runtime overhead.< / p >
< p > The privacy system ensures that users cannot access the wrapped type (if the
field is private, which it is by default).< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-16" id = "disadvantages-16" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The downside of newtypes (especially compared with type aliases), is that there
is no special language support. This means there can be < em > a lot< / em > of boilerplate.
You need a 'pass through' method for every method you want to expose on the
wrapped type, and an impl for every trait you want to also be implemented for
the wrapper type.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#discussion-10" id = "discussion-10" > Discussion< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Newtypes are very common in Rust code. Abstraction or representing units are the
most common uses, but they can be used for other reasons:< / p >
< ul >
< li > restricting functionality (reduce the functions exposed or traits implemented),< / li >
< li > making a type with copy semantics have move semantics,< / li >
< li > abstraction by providing a more concrete type and thus hiding internal types, e.g.,< / li >
< / ul >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > pub struct Foo(Bar< T1, T2> );
< / code > < / pre >
2021-01-08 08:00:20 +00:00
< p > Here, < code > Bar< / code > might be some public, generic type and < code > T1< / code > and < code > T2< / code > are some internal types.
Users of our module shouldn't know that we implement < code > Foo< / code > by using a < code > Bar< / code > , but what we're
really hiding here is the types < code > T1< / code > and < code > T2< / code > , and how they are used with < code > Bar< / code > .< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-11" id = "see-also-11" > See also< / a > < / h2 >
2021-01-03 13:16:54 +00:00
< ul >
2021-01-05 03:33:00 +00:00
< li > < a href = "https://doc.rust-lang.org/book/ch19-04-advanced-types.html?highlight=newtype#using-the-newtype-pattern-for-type-safety-and-abstraction" > Advanced Types in the book< / a > < / li >
2021-01-03 13:16:54 +00:00
< li > < a href = "https://wiki.haskell.org/Newtype" > Newtypes in Haskell< / a > < / li >
< li > < a href = "https://doc.rust-lang.org/stable/book/ch19-04-advanced-types.html#creating-type-synonyms-with-type-aliases" > Type aliases< / a > < / li >
< li > < a href = "https://crates.io/crates/derive_more" > derive_more< / a > , a crate for deriving many builtin traits on newtypes.< / li >
2021-01-03 14:03:46 +00:00
< li > < a href = "https://www.worthe-it.co.za/blog/2020-10-31-newtype-pattern-in-rust.html" > The Newtype Pattern In Rust< / a > < / li >
2021-01-03 13:16:54 +00:00
< / ul >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#raii-with-guards" id = "raii-with-guards" > RAII with guards< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-23" id = "description-23" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > < a href = "https://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization" > RAII< / a > stands for " Resource Acquisition is Initialisation" which is a terrible
name. The essence of the pattern is that resource initialisation is done in the
constructor of an object and finalisation in the destructor. This pattern is
extended in Rust by using an RAII object as a guard of some resource and relying
on the type system to ensure that access is always mediated by the guard object.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-17" id = "example-17" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Mutex guards are the classic example of this pattern from the std library (this
is a simplified version of the real implementation):< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > use std::ops::Deref;
struct Foo {}
struct Mutex< T> {
2021-01-01 11:28:10 +00:00
// We keep a reference to our data: T here.
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
struct MutexGuard< 'a, T: 'a> {
data: & 'a T,
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
// Locking the mutex is explicit.
impl< T> Mutex< T> {
fn lock(& self) -> MutexGuard< T> {
// Lock the underlying OS mutex.
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
// MutexGuard keeps a reference to self
2021-01-05 14:57:36 +00:00
MutexGuard {
data: self,
//..
2021-01-02 11:41:18 +00:00
}
2021-01-01 11:28:10 +00:00
}
}
// Destructor for unlocking the mutex.
impl< 'a, T> Drop for MutexGuard< 'a, T> {
fn drop(& mut self) {
// Unlock the underlying OS mutex.
2021-01-02 11:41:18 +00:00
//..
2021-01-01 11:28:10 +00:00
}
}
// Implementing Deref means we can treat MutexGuard like a pointer to T.
impl< 'a, T> Deref for MutexGuard< 'a, T> {
type Target = T;
fn deref(& self) -> & T {
self.data
}
}
2021-01-02 11:41:18 +00:00
fn baz(x: Mutex< Foo> ) {
2021-01-01 11:28:10 +00:00
let xx = x.lock();
xx.foo(); // foo is a method on Foo.
// The borrow checker ensures we can't store a reference to the underlying
// Foo which will outlive the guard xx.
// x is unlocked when we exit this function and xx's destructor is executed.
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-13" id = "motivation-13" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Where a resource must be finalised after use, RAII can be used to do this
finalisation. If it is an error to access that resource after finalisation, then
this pattern can be used to prevent such errors.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-17" id = "advantages-17" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Prevents errors where a resource is not finalised and where a resource is used
after finalisation.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#discussion-11" id = "discussion-11" > Discussion< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > RAII is a useful pattern for ensuring resources are properly deallocated or
finalised. We can make use of the borrow checker in Rust to statically prevent
errors stemming from using resources after finalisation takes place.< / p >
< p > The core aim of the borrow checker is to ensure that references to data do not
outlive that data. The RAII guard pattern works because the guard object
contains a reference to the underlying resource and only exposes such
references. Rust ensures that the guard cannot outlive the underlying resource
and that references to the resource mediated by the guard cannot outlive the
guard. To see how this works it is helpful to examine the signature of < code > deref< / code >
without lifetime elision:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > fn deref< 'a> (& 'a self) -> & 'a T {
//..
}
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > The returned reference to the resource has the same lifetime as < code > self< / code > (< code > 'a< / code > ).
The borrow checker therefore ensures that the lifetime of the reference to < code > T< / code >
is shorter than the lifetime of < code > self< / code > .< / p >
< p > Note that implementing < code > Deref< / code > is not a core part of this pattern, it only makes
using the guard object more ergonomic. Implementing a < code > get< / code > method on the guard
works just as well.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-12" id = "see-also-12" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > < a href = "patterns/../idioms/dtor-finally.html" > Finalisation in destructors idiom< / a > < / p >
< p > RAII is a common pattern in C++: < a href = "http://en.cppreference.com/w/cpp/language/raii" > cppreference.com< / a > ,
< a href = "https://en.wikipedia.org/wiki/Resource_Acquisition_Is_Initialization" > wikipedia< / a > .< / p >
< p > < a href = "https://doc.rust-lang.org/1.0.0/style/ownership/raii.html" > Style guide entry< / a >
(currently just a placeholder).< / p >
< h1 > < a class = "header" href = "#prefer-small-crates" id = "prefer-small-crates" > Prefer small crates< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-24" id = "description-24" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Prefer small crates that do one thing well.< / p >
2021-01-06 13:58:52 +00:00
< p > Cargo and crates.io make it easy to add third-party libraries, much more so than in say C or C++.
Moreover, since packages on crates.io cannot be edited or removed after publication, any build that works now should continue to work in the future.
We should take advantage of this tooling, and use smaller, more fine-grained dependencies.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-18" id = "advantages-18" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > Small crates are easier to understand, and encourage more modular code.< / li >
2021-01-06 13:58:52 +00:00
< li > Crates allow for re-using code between projects.
For example, the < code > url< / code > crate was developed as part of the Servo browser engine, but has since found wide use outside the project.< / li >
2021-01-01 11:28:10 +00:00
< li > Since the compilation unit of Rust is the crate, splitting a project into multiple crates can allow more of the code to be built in parallel.< / li >
< / ul >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-17" id = "disadvantages-17" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
2021-01-06 13:58:52 +00:00
< li > This can lead to " dependency hell" , when a project depends on multiple conflicting versions of a crate at the same time.
For example, the < code > url< / code > crate has both versions 1.0 and 0.5.
2021-01-08 08:00:20 +00:00
Since the < code > Url< / code > from < code > url:1.0< / code > and the < code > Url< / code > from < code > url:0.5< / code > are different types,
an HTTP client that uses < code > url:0.5< / code > would not accept < code > Url< / code > values from a web scraper that uses < code > url:1.0< / code > .< / li >
2021-01-01 11:28:10 +00:00
< li > Packages on crates.io are not curated. A crate may be poorly written, have unhelpful documentation, or be outright malicious.< / li >
< li > Two small crates may be less optimized than one large one, since the compiler does not perform link-time optimization (LTO) by default.< / li >
< / ul >
< h2 > < a class = "header" href = "#examples-1" id = "examples-1" > Examples< / a > < / h2 >
< p > The < a href = "https://crates.io/crates/ref_slice" > < code > ref_slice< / code > < / a > crate provides functions for converting < code > & T< / code > to < code > & [T]< / code > .< / p >
< p > The < a href = "https://crates.io/crates/url" > < code > url< / code > < / a > crate provides tools for working with URLs.< / p >
< p > The < a href = "https://crates.io/crates/num_cpus" > < code > num_cpus< / code > < / a > crate provides a function to query the number of CPUs on a machine.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-13" id = "see-also-13" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > < a href = "https://crates.io/" > crates.io: The Rust community crate host< / a > < / li >
< / ul >
< h1 > < a class = "header" href = "#contain-unsafety-in-small-modules" id = "contain-unsafety-in-small-modules" > Contain unsafety in small modules< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-25" id = "description-25" > Description< / a > < / h2 >
2021-01-06 13:58:52 +00:00
< p > If you have < code > unsafe< / code > code, create the smallest possible module that can uphold the needed invariants to build a minimal safe interface upon the unsafety.
Embed this into a larger module that contains only safe code and presents an ergonomic interface.
Note that the outer module can contain unsafe functions and methods that call directly into the unsafe code.
Users may use this to gain speed benefits.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-19" id = "advantages-19" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > This restricts the unsafe code that must be audited< / li >
< li > Writing the outer module is much easier, since you can count on the guarantees of the inner module< / li >
< / ul >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-18" id = "disadvantages-18" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > Sometimes, it may be hard to find a suitable interface.< / li >
< li > The abstraction may introduce inefficiencies.< / li >
< / ul >
< h2 > < a class = "header" href = "#examples-2" id = "examples-2" > Examples< / a > < / h2 >
< ul >
< li > The < a href = "https://docs.rs/toolshed" > < code > toolshed< / code > < / a > crate contains its unsafe operations in submodules, presenting a safe interface to users.< / li >
2021-01-06 13:58:52 +00:00
< li > < code > std< / code > s < code > String< / code > class is a wrapper over < code > Vec< u8> < / code > with the added invariant that the contents must be valid UTF-8.
The operations on < code > String< / code > ensure this behavior.
However, users have the option of using an < code > unsafe< / code > method to create a < code > String< / code > , in which case the onus is on them to guarantee the validity of the contents.< / li >
2021-01-01 11:28:10 +00:00
< / ul >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-14" id = "see-also-14" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > < a href = "https://www.ralfj.de/blog/2018/08/22/two-kinds-of-invariants.html" > Ralf Jung's Blog about invariants in unsafe code< / a > < / li >
< / ul >
< h1 > < a class = "header" href = "#visitor" id = "visitor" > Visitor< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-26" id = "description-26" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > A visitor encapsulates an algorithm that operates over a heterogeneous
collection of objects. It allows multiple different algorithms to be written
over the same data without having to modify the data (or their primary
behaviour).< / p >
< p > Furthermore, the visitor pattern allows separating the traversal of
a collection of objects from the operations performed on each object.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-18" id = "example-18" > Example< / a > < / h2 >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > // The data we will visit
2021-01-01 11:28:10 +00:00
mod ast {
pub enum Stmt {
Expr(Expr),
Let(Name, Expr),
}
pub struct Name {
value: String,
}
pub enum Expr {
IntLit(i64),
Add(Box< Expr> , Box< Expr> ),
Sub(Box< Expr> , Box< Expr> ),
}
}
// The abstract visitor
mod visit {
use ast::*;
pub trait Visitor< T> {
fn visit_name(& mut self, n: & Name) -> T;
fn visit_stmt(& mut self, s: & Stmt) -> T;
fn visit_expr(& mut self, e: & Expr) -> T;
}
}
use visit::*;
use ast::*;
// An example concrete implementation - walks the AST interpreting it as code.
struct Interpreter;
impl Visitor< i64> for Interpreter {
fn visit_name(& mut self, n: & Name) -> i64 { panic!() }
fn visit_stmt(& mut self, s: & Stmt) -> i64 {
match *s {
Stmt::Expr(ref e) => self.visit_expr(e),
Stmt::Let(..) => unimplemented!(),
}
}
fn visit_expr(& mut self, e: & Expr) -> i64 {
match *e {
Expr::IntLit(n) => n,
Expr::Add(ref lhs, ref rhs) => self.visit_expr(lhs) + self.visit_expr(rhs),
Expr::Sub(ref lhs, ref rhs) => self.visit_expr(lhs) - self.visit_expr(rhs),
}
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > One could implement further visitors, for example a type checker, without having
to modify the AST data.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#motivation-14" id = "motivation-14" > Motivation< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The visitor pattern is useful anywhere that you want to apply an algorithm to
heterogeneous data. If data is homogeneous, you can use an iterator-like pattern.
Using a visitor object (rather than a functional approach) allows the visitor to
be stateful and thus communicate information between nodes.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#discussion-12" id = "discussion-12" > Discussion< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > It is common for the < code > visit_*< / code > methods to return void (as opposed to in the
example). In that case it is possible to factor out the traversal code and share
it between algorithms (and also to provide noop default methods). In Rust, the
common way to do this is to provide < code > walk_*< / code > functions for each datum. For
example,< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > pub fn walk_expr(visitor: & mut Visitor, e: & Expr) {
2021-01-01 11:28:10 +00:00
match *e {
Expr::IntLit(_) => {},
Expr::Add(ref lhs, ref rhs) => {
visitor.visit_expr(lhs);
visitor.visit_expr(rhs);
}
Expr::Sub(ref lhs, ref rhs) => {
visitor.visit_expr(lhs);
visitor.visit_expr(rhs);
}
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > In other languages (e.g., Java) it is common for data to have an < code > accept< / code > method
which performs the same duty.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-15" id = "see-also-15" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > The visitor pattern is a common pattern in most OO languages.< / p >
< p > < a href = "https://en.wikipedia.org/wiki/Visitor_pattern" > Wikipedia article< / a > < / p >
< p > The < a href = "patterns/fold.html" > fold< / a > pattern is similar to visitor but produces a new version of
the visited data structure.< / p >
< h1 > < a class = "header" href = "#anti-patterns" id = "anti-patterns" > Anti-patterns< / a > < / h1 >
2021-01-08 08:00:20 +00:00
< p > An < a href = "https://en.wikipedia.org/wiki/Anti-pattern" > anti-pattern< / a > is a solution to a
" recurring problem that is usually ineffective and risks being highly counterproductive" .
2021-01-02 11:15:56 +00:00
Just as valuable as knowing how to solve a problem, is knowing how < em > not< / em > to solve it.
Anti-patterns give us great counter-examples to consider relative to design patterns.
Anti-patterns are not confined to code. For example, a process can be an anti-pattern, too.< / p >
2021-01-01 11:28:10 +00:00
< h1 > < a class = "header" href = "#denywarnings" id = "denywarnings" > < code > #![deny(warnings)]< / code > < / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-27" id = "description-27" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > A well-intentioned crate author wants to ensure their code builds without
warnings. So they annotate their crate root with the following:< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-19" id = "example-19" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > #![deny(warnings)]
< span class = "boring" > fn main() {
< / span > // All is well.
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-20" id = "advantages-20" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > It is short and will stop the build if anything is amiss.< / p >
< h2 > < a class = "header" href = "#drawbacks" id = "drawbacks" > Drawbacks< / a > < / h2 >
< p > By disallowing the compiler to build with warnings, a crate author opts out of
Rust's famed stability. Sometimes new features or old misfeatures need a change
in how things are done, thus lints are written that < code > warn< / code > for a certain grace
period before being turned to < code > deny< / code > .< / p >
< p > For example, it was discovered that a type could have two < code > impl< / code > s with the same
method. This was deemed a bad idea, but in order to make the transition smooth,
the < code > overlapping-inherent-impls< / code > lint was introduced to give a warning to those
stumbling on this fact, before it becomes a hard error in a future release.< / p >
< p > Also sometimes APIs get deprecated, so their use will emit a warning where
before there was none.< / p >
< p > All this conspires to potentially break the build whenever something changes.< / p >
< p > Furthermore, crates that supply additional lints (e.g. < a href = "https://github.com/Manishearth/rust-clippy" > rust-clippy< / a > ) can no
longer be used unless the annotation is removed. This is mitigated with
2021-01-02 12:03:57 +00:00
< a href = "https://doc.rust-lang.org/rustc/lints/levels.html#capping-lints" > --cap-lints< / a > . The < code > --cap-lints=warn< / code > command line argument, turns all < code > deny< / code >
lint errors into warnings. But be aware that < code > forbid< / code > lints are stronger than
< code > deny< / code > hence the 'forbid' level cannot be overridden to be anything lower than
an error. As a result < code > forbid< / code > lints will still stop compilation.< / p >
2021-01-01 11:28:10 +00:00
< h2 > < a class = "header" href = "#alternatives" id = "alternatives" > Alternatives< / a > < / h2 >
< p > There are two ways of tackling this problem: First, we can decouple the build
setting from the code, and second, we can name the lints we want to deny
explicitly.< / p >
< p > The following command line will build with all warnings set to < code > deny< / code > :< / p >
< p > < code > RUSTFLAGS=" -D warnings" cargo build< / code > < / p >
< p > This can be done by any individual developer (or be set in a CI tool like
Travis, but remember that this may break the build when something changes)
without requiring a change to the code.< / p >
< p > Alternatively, we can specify the lints that we want to < code > deny< / code > in the code.
Here is a list of warning lints that is (hopefully) safe to deny:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > #[deny(bad-style,
2021-01-01 11:28:10 +00:00
const-err,
dead-code,
extra-requirement-in-impl,
improper-ctypes,
legacy-directory-ownership,
non-shorthand-field-patterns,
no-mangle-generic-items,
overflowing-literals,
path-statements ,
patterns-in-fns-without-body,
plugin-as-library,
private-in-public,
private-no-mangle-fns,
private-no-mangle-statics,
raw-pointer-derive,
safe-extern-statics,
unconditional-recursion,
unions-with-drop-fields,
unused,
unused-allocation,
unused-comparisons,
unused-parens,
while-true)]
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > In addition, the following < code > allow< / code > ed lints may be a good idea to < code > deny< / code > :< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > #[deny(missing-debug-implementations,
2021-01-01 11:28:10 +00:00
missing-docs,
trivial-casts,
trivial-numeric-casts,
unused-extern-crates,
unused-import-braces,
unused-qualifications,
unused-results)]
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > Some may also want to add < code > missing-copy-implementations< / code > to their list.< / p >
< p > Note that we explicitly did not add the < code > deprecated< / code > lint, as it is fairly
certain that there will be more deprecated APIs in the future.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-16" id = "see-also-16" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< ul >
< li > < a href = "https://doc.rust-lang.org/reference/attributes.html#deprecation" > deprecate attribute< / a > documentation< / li >
< li > Type < code > rustc -W help< / code > for a list of lints on your system. Also type
< code > rustc --help< / code > for a general list of options< / li >
< li > < a href = "https://github.com/Manishearth/rust-clippy" > rust-clippy< / a > is a collection of lints for better Rust code< / li >
< / ul >
< h1 > < a class = "header" href = "#deref-polymorphism" id = "deref-polymorphism" > < code > Deref< / code > polymorphism< / a > < / h1 >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#description-28" id = "description-28" > Description< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Abuse the < code > Deref< / code > trait to emulate inheritance between structs, and thus reuse
methods.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#example-20" id = "example-20" > Example< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Sometimes we want to emulate the following common pattern from OO languages such
as Java:< / p >
< pre > < code class = "language-java" > class Foo {
void m() { ... }
}
class Bar extends Foo {}
public static void main(String[] args) {
Bar b = new Bar();
b.m();
}
< / code > < / pre >
< p > We can use the deref polymorphism anti-pattern to do so:< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > use std::ops::Deref;
struct Foo {}
2021-01-01 11:28:10 +00:00
impl Foo {
2021-01-05 14:57:36 +00:00
fn m(& self) {
//..
2021-01-02 11:41:18 +00:00
}
2021-01-01 11:28:10 +00:00
}
struct Bar {
f: Foo
}
impl Deref for Bar {
type Target = Foo;
fn deref(& self) -> & Foo {
& self.f
}
}
fn main() {
let b = Bar { Foo {} };
b.m();
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-01 11:28:10 +00:00
< p > There is no struct inheritance in Rust. Instead we use composition and include
an instance of < code > Foo< / code > in < code > Bar< / code > (since the field is a value, it is stored inline,
so if there were fields, they would have the same layout in memory as the Java
version (probably, you should use < code > #[repr(C)]< / code > if you want to be sure)).< / p >
< p > In order to make the method call work we implement < code > Deref< / code > for < code > Bar< / code > with < code > Foo< / code >
as the target (returning the embedded < code > Foo< / code > field). That means that when we
dereference a < code > Bar< / code > (for example, using < code > *< / code > ) then we will get a < code > Foo< / code > . That is
pretty weird. Dereferencing usually gives a < code > T< / code > from a reference to < code > T< / code > , here we
have two unrelated types. However, since the dot operator does implicit
dereferencing, it means that the method call will search for methods on < code > Foo< / code > as
well as < code > Bar< / code > .< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#advantages-21" id = "advantages-21" > Advantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > You save a little boilerplate, e.g.,< / p >
2021-01-02 11:41:18 +00:00
< pre > < code class = "language-rust ignore" > impl Bar {
2021-01-05 14:57:36 +00:00
fn m(& self) {
2021-01-01 11:28:10 +00:00
self.f.m()
}
}
2021-01-02 11:41:18 +00:00
< / code > < / pre >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#disadvantages-19" id = "disadvantages-19" > Disadvantages< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > Most importantly this is a surprising idiom - future programmers reading this in
code will not expect this to happen. That's because we are abusing the < code > Deref< / code >
trait rather than using it as intended (and documented, etc.). It's also because
the mechanism here is completely implicit.< / p >
< p > This pattern does not introduce subtyping between < code > Foo< / code > and < code > Bar< / code > like
inheritance in Java or C++ does. Furthermore, traits implemented by < code > Foo< / code > are
not automatically implemented for < code > Bar< / code > , so this pattern interacts badly with
bounds checking and thus generic programming.< / p >
< p > Using this pattern gives subtly different semantics from most OO languages with
regards to < code > self< / code > . Usually it remains a reference to the sub-class, with this
pattern it will be the 'class' where the method is defined.< / p >
< p > Finally, this pattern only supports single inheritance, and has no notion of
interfaces, class-based privacy, or other inheritance-related features. So, it
gives an experience that will be subtly surprising to programmers used to Java
inheritance, etc.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#discussion-13" id = "discussion-13" > Discussion< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > There is no one good alternative. Depending on the exact circumstances it might
be better to re-implement using traits or to write out the facade methods to
dispatch to < code > Foo< / code > manually. We do intend to add a mechanism for inheritance
similar to this to Rust, but it is likely to be some time before it reaches
stable Rust. See these < a href = "http://aturon.github.io/blog/2015/09/18/reuse/" > blog< / a >
< a href = "http://smallcultfollowing.com/babysteps/blog/2015/10/08/virtual-structs-part-4-extended-enums-and-thin-traits/" > posts< / a >
and this < a href = "https://github.com/rust-lang/rfcs/issues/349" > RFC issue< / a > for more details.< / p >
< p > The < code > Deref< / code > trait is designed for the implementation of custom pointer types.
The intention is that it will take a pointer-to-< code > T< / code > to a < code > T< / code > , not convert
between different types. It is a shame that this isn't (probably cannot be)
enforced by the trait definition.< / p >
< p > Rust tries to strike a careful balance between explicit and implicit mechanisms,
favouring explicit conversions between types. Automatic dereferencing in the dot
operator is a case where the ergonomics strongly favour an implicit mechanism,
but the intention is that this is limited to degrees of indirection, not
conversion between arbitrary types.< / p >
2021-01-05 15:44:58 +00:00
< h2 > < a class = "header" href = "#see-also-17" id = "see-also-17" > See also< / a > < / h2 >
2021-01-01 11:28:10 +00:00
< p > < a href = "anti_patterns/../idioms/deref.html" > Collections are smart pointers idiom< / a > .< / p >
< p > < a href = "https://doc.rust-lang.org/std/ops/trait.Deref.html" > Documentation for < code > Deref< / code > trait< / a > .< / p >
2021-01-02 11:15:56 +00:00
< h1 > < a class = "header" href = "#functional-usage-of-rust" id = "functional-usage-of-rust" > Functional Usage of Rust< / a > < / h1 >
2021-01-06 13:58:52 +00:00
< p > Rust is an imperative language, but it follows many functional programming paradigms.
One of the biggest hurdles to understanding functional programs when coming from an imperative background is the shift in thinking.
Imperative programs describe < strong > how< / strong > to do something, whereas declarative programs describe < strong > what< / strong > to do.
Let's sum the numbers from 1 to 10 to show this.< / p >
2021-01-02 11:15:56 +00:00
< h2 > < a class = "header" href = "#imperative" id = "imperative" > Imperative< / a > < / h2 >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > let mut sum = 0;
for i in 1..11 {
2021-01-06 13:56:19 +00:00
sum += i;
2021-01-02 11:15:56 +00:00
}
println!(" {}" , sum);
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-06 13:58:52 +00:00
< p > With imperative programs, we have to play compiler to see what is happening.
Here, we start with a < code > sum< / code > of < code > 0< / code > .
Next, we iterate through the range from 1 to 10.
Each time through the loop, we add the corresponding value in the range.
Then we print it out.< / p >
2021-01-03 15:20:26 +00:00
< table > < thead > < tr > < th align = "center" > < code > i< / code > < / th > < th align = "center" > < code > sum< / code > < / th > < / tr > < / thead > < tbody >
< tr > < td align = "center" > 1< / td > < td align = "center" > 1< / td > < / tr >
< tr > < td align = "center" > 2< / td > < td align = "center" > 3< / td > < / tr >
< tr > < td align = "center" > 3< / td > < td align = "center" > 6< / td > < / tr >
< tr > < td align = "center" > 4< / td > < td align = "center" > 10< / td > < / tr >
< tr > < td align = "center" > 5< / td > < td align = "center" > 15< / td > < / tr >
< tr > < td align = "center" > 6< / td > < td align = "center" > 21< / td > < / tr >
< tr > < td align = "center" > 7< / td > < td align = "center" > 28< / td > < / tr >
< tr > < td align = "center" > 8< / td > < td align = "center" > 36< / td > < / tr >
< tr > < td align = "center" > 9< / td > < td align = "center" > 45< / td > < / tr >
< tr > < td align = "center" > 10< / td > < td align = "center" > 55< / td > < / tr >
2021-01-02 11:15:56 +00:00
< / tbody > < / table >
< p > This is how most of us start out programming. We learn that a program is a set of steps.< / p >
< h2 > < a class = "header" href = "#declarative" id = "declarative" > Declarative< / a > < / h2 >
< pre > < pre class = "playground" > < code class = "language-rust edition2018" >
< span class = "boring" > #![allow(unused)]
< / span > < span class = "boring" > fn main() {
< / span > println!(" {}" , (1..11).fold(0, |a, b| a + b));
< span class = "boring" > }
< / span > < / code > < / pre > < / pre >
2021-01-08 08:00:20 +00:00
< p > Whoa! This is really different! What's going on here? Remember that with declarative programs
we are describing < strong > what< / strong > to do, rather than < strong > how< / strong > to do it.
2021-01-06 13:58:52 +00:00
< code > fold< / code > is a function that < a href = "https://en.wikipedia.org/wiki/Function_composition" > composes< / a > functions. The name is a convention from Haskell.< / p >
< p > Here, we are composing functions of addition (this closure: < code > |a, b| a + b)< / code > ) with a range from 1 to 10.
The < code > 0< / code > is the starting point, so < code > a< / code > is < code > 0< / code > at first.
< code > b< / code > is the first element of the range, < code > 1< / code > . < code > 0 + 1 = 1< / code > is the result.
So now we < code > fold< / code > again, with < code > a = 1< / code > , < code > b = 2< / code > and so < code > 1 + 2 = 3< / code > is the next result.
This process continues until we get to the last element in the range, < code > 10< / code > .< / p >
2021-01-03 15:20:26 +00:00
< table > < thead > < tr > < th align = "center" > < code > a< / code > < / th > < th align = "center" > < code > b< / code > < / th > < th align = "center" > result< / th > < / tr > < / thead > < tbody >
< tr > < td align = "center" > 0< / td > < td align = "center" > 1< / td > < td align = "center" > 1< / td > < / tr >
< tr > < td align = "center" > 1< / td > < td align = "center" > 2< / td > < td align = "center" > 3< / td > < / tr >
< tr > < td align = "center" > 3< / td > < td align = "center" > 3< / td > < td align = "center" > 6< / td > < / tr >
< tr > < td align = "center" > 6< / td > < td align = "center" > 4< / td > < td align = "center" > 10< / td > < / tr >
< tr > < td align = "center" > 10< / td > < td align = "center" > 5< / td > < td align = "center" > 15< / td > < / tr >
< tr > < td align = "center" > 15< / td > < td align = "center" > 6< / td > < td align = "center" > 21< / td > < / tr >
< tr > < td align = "center" > 21< / td > < td align = "center" > 7< / td > < td align = "center" > 28< / td > < / tr >
< tr > < td align = "center" > 28< / td > < td align = "center" > 8< / td > < td align = "center" > 36< / td > < / tr >
< tr > < td align = "center" > 36< / td > < td align = "center" > 9< / td > < td align = "center" > 45< / td > < / tr >
< tr > < td align = "center" > 45< / td > < td align = "center" > 10< / td > < td align = "center" > 55< / td > < / tr >
2021-01-02 11:15:56 +00:00
< / tbody > < / table >
2021-01-02 11:57:11 +00:00
< h1 > < a class = "header" href = "#additional-resources" id = "additional-resources" > Additional resources< / a > < / h1 >
< p > A collection of complementary helpful content< / p >
< h2 > < a class = "header" href = "#talks" id = "talks" > Talks< / a > < / h2 >
< ul >
2021-01-05 16:31:19 +00:00
< li > < a href = "https://www.youtube.com/watch?v=Pm_oO0N5B9k" > Design Patterns in Rust< / a > by Nicholas Cameron at the PDRust (2016)< / li >
< li > < a href = "https://www.youtube.com/watch?v=0zOg8_B71gE" > Writing Idiomatic Libraries in Rust< / a > by Pascal Hertleif at RustFest (2017)< / li >
< li > < a href = "https://www.youtube.com/watch?v=vqavdUGKeb4" > Rust Programming Techniques< / a > by Nicholas Cameron at LinuxConfAu (2018)< / li >
< / ul >
< h2 > < a class = "header" href = "#books-online" id = "books-online" > Books (Online)< / a > < / h2 >
< ul >
< li > < a href = "https://rust-lang.github.io/api-guidelines" > The Rust API Guidelines< / a > < / li >
2021-01-02 11:57:11 +00:00
< / ul >
2021-01-06 14:03:34 +00:00
< h1 > < a class = "header" href = "#design-principles" id = "design-principles" > Design principles< / a > < / h1 >
< h2 > < a class = "header" href = "#a-brief-overview-over-common-design-principles" id = "a-brief-overview-over-common-design-principles" > A brief overview over common design principles< / a > < / h2 >
< hr / >
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikisolidsolida" id = "a-hrefhttpsenwikipediaorgwikisolidsolida" > < a href = "https://en.wikipedia.org/wiki/SOLID" > SOLID< / a > < / a > < / h2 >
< ul >
< li > < a href = "https://en.wikipedia.org/wiki/Single-responsibility_principle" > Single Responsibility Principle (SRP)< / a > :
A class should only have a single responsibility, that is, only changes to one part of the software's
specification should be able to affect the specification of the class.< / li >
< li > < a href = "https://en.wikipedia.org/wiki/Open%E2%80%93closed_principle" > Open/Closed Principle (OCP)< / a > :
" Software entities ... should be open for extension, but closed for modification." < / li >
< li > < a href = "https://en.wikipedia.org/wiki/Liskov_substitution_principle" > Liscov Substitution Principle (LSP)< / a > :
" Objects in a program should be replaceable with instances of their subtypes without altering the correctness
of that program." < / li >
< li > < a href = "https://en.wikipedia.org/wiki/Interface_segregation_principle" > Interface Segregation Principle (ISP)< / a > :
" Many client-specific interfaces are better than one general-purpose interface." < / li >
< li > < a href = "https://en.wikipedia.org/wiki/Dependency_inversion_principle" > Dependency Inversion Principle (DIP)< / a > :
One should " depend upon abstractions, [not] concretions." < / li >
< / ul >
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikidon27t_repeat_yourselfdry-dont-repeat-yourselfa" id = "a-hrefhttpsenwikipediaorgwikidon27t_repeat_yourselfdry-dont-repeat-yourselfa" > < a href = "https://en.wikipedia.org/wiki/Don%27t_repeat_yourself" > DRY (Don’ < / a > < / a > < / h2 >
< p > " Every piece of knowledge must have a single, unambiguous, authoritative representation within a system" < / p >
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikikiss_principlekiss-principlea" id = "a-hrefhttpsenwikipediaorgwikikiss_principlekiss-principlea" > < a href = "https://en.wikipedia.org/wiki/KISS_principle" > KISS principle< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > most systems work best if they are kept simple rather than made complicated; therefore,
simplicity should be a key goal in design, and unnecessary complexity should be avoided< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikilaw_of_demeterlaw-of-demeter-loda" id = "a-hrefhttpsenwikipediaorgwikilaw_of_demeterlaw-of-demeter-loda" > < a href = "https://en.wikipedia.org/wiki/Law_of_Demeter" > Law of Demeter (LoD)< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > a given object should assume as little as possible about the structure or properties of
anything else (including its subcomponents), in accordance with the principle of " information hiding" < / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikidesign_by_contractdesign-by-contract-dbca" id = "a-hrefhttpsenwikipediaorgwikidesign_by_contractdesign-by-contract-dbca" > < a href = "https://en.wikipedia.org/wiki/Design_by_contract" > Design by contract (DbC)< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > software designers should define formal, precise and verifiable interface specifications
for software components, which extend the ordinary definition of abstract data types with
preconditions, postconditions and invariants< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikiencapsulation_computer_programmingencapsulationa" id = "a-hrefhttpsenwikipediaorgwikiencapsulation_computer_programmingencapsulationa" > < a href = "https://en.wikipedia.org/wiki/Encapsulation_(computer_programming)" > Encapsulation< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > bundling of data with the methods that operate on that data, or the restricting of
direct access to some of an object's components. Encapsulation is used to hide the
values or state of a structured data object inside a class, preventing unauthorized
parties' direct access to them.< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikicommande28093query_separationcommand-query-separationcqsa" id = "a-hrefhttpsenwikipediaorgwikicommande28093query_separationcommand-query-separationcqsa" > < a href = "https://en.wikipedia.org/wiki/Command%E2%80%93query_separation" > Command-Query-Separation(CQS)< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “Functions should not produce abstract side effects...only commands (procedures) will
be permitted to produce side effects.” - Bertrand Meyer: Object Oriented Software Construction< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#a-hrefhttpsenwikipediaorgwikiprinciple_of_least_astonishmentprinciple-of-least-astonishment-polaa" id = "a-hrefhttpsenwikipediaorgwikiprinciple_of_least_astonishmentprinciple-of-least-astonishment-polaa" > < a href = "https://en.wikipedia.org/wiki/Principle_of_least_astonishment" > Principle of least astonishment (POLA)< / a > < / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > a component of a system should behave in a way that most users will expect it to behave.
The behavior should not astonish or surprise users< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#linguistic-modular-units" id = "linguistic-modular-units" > Linguistic-Modular-Units< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “Modules must correspond to syntactic units in the language used.” - Bertrand Meyer:
Object Oriented Software Construction< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#self-documentation" id = "self-documentation" > Self-Documentation< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “The designer of a module should strive to make all information about the module part of
the module itself.” - Bertrand Meyer: Object Oriented Software Construction< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#uniform-access" id = "uniform-access" > Uniform-Access< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “All services offered by a module should be available through a uniform notation,
which does not betray whether they are implemented through storage or through computation.” -
Bertrand Meyer: Object Oriented Software Construction< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#single-choice" id = "single-choice" > Single-Choice< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “Whenever a software system must support a set of alternatives, one and only one module
in the system should know their exhaustive list.” - Bertrand Meyer: Object Oriented
Software Construction< / p >
2021-01-06 14:03:34 +00:00
< h2 > < a class = "header" href = "#persistence-closure" id = "persistence-closure" > Persistence-Closure< / a > < / h2 >
2021-01-08 08:00:20 +00:00
< p > “Whenever a storage mechanism stores an object, it must store with it the dependents
of that object. Whenever a retrieval mechanism retrieves a previously stored object,
it must also retrieve any dependent of that object that has not yet been retrieved.” -
Bertrand Meyer: Object Oriented Software Construction< / p >
2021-01-01 11:28:10 +00:00
< / main >
< nav class = "nav-wrapper" aria-label = "Page navigation" >
<!-- Mobile navigation buttons -->
< div style = "clear: both" > < / div >
< / nav >
< / div >
< / div >
< nav class = "nav-wide-wrapper" aria-label = "Page navigation" >
< / nav >
< / div >
< script type = "text/javascript" >
window.playground_copyable = true;
< / script >
< script src = "elasticlunr.min.js" type = "text/javascript" charset = "utf-8" > < / script >
< script src = "mark.min.js" type = "text/javascript" charset = "utf-8" > < / script >
< script src = "searcher.js" type = "text/javascript" charset = "utf-8" > < / script >
< script src = "clipboard.min.js" type = "text/javascript" charset = "utf-8" > < / script >
< script src = "highlight.js" type = "text/javascript" charset = "utf-8" > < / script >
< script src = "book.js" type = "text/javascript" charset = "utf-8" > < / script >
<!-- Custom JS scripts -->
< script type = "text/javascript" >
window.addEventListener('load', function() {
window.setTimeout(window.print, 100);
});
< / script >
< / body >
< / html >
