5 This document is the primary reference for the Rust programming language. It
6 provides three kinds of material:
8 - Chapters that informally describe each language construct and their use.
9 - Chapters that informally describe the memory model, concurrency model,
10 runtime services, linkage model and debugging facilities.
11 - Appendix chapters providing rationale and references to languages that
12 influenced the design.
14 This document does not serve as an introduction to the language. Background
15 familiarity with the language is assumed. A separate [book] is available to
16 help acquire such background familiarity.
18 This document also does not serve as a reference to the [standard] library
19 included in the language distribution. Those libraries are documented
20 separately by extracting documentation attributes from their source code. Many
21 of the features that one might expect to be language features are library
22 features in Rust, so what you're looking for may be there, not here.
24 Finally, this document is not normative. It may include details that are
25 specific to `rustc` itself, and should not be taken as a specification for
26 the Rust language. We intend to produce such a document someday, but this
27 is what we have for now.
29 You may also be interested in the [grammar].
31 [book]: book/index.html
32 [standard]: std/index.html
33 [grammar]: grammar.html
37 ## Unicode productions
39 A few productions in Rust's grammar permit Unicode code points outside the
40 ASCII range. We define these productions in terms of character properties
41 specified in the Unicode standard, rather than in terms of ASCII-range code
42 points. The grammar has a [Special Unicode Productions][unicodeproductions]
43 section that lists these productions.
45 [unicodeproductions]: grammar.html#special-unicode-productions
47 ## String table productions
49 Some rules in the grammar — notably [unary
50 operators](#unary-operator-expressions), [binary
51 operators](#binary-operator-expressions), and [keywords][keywords] — are
52 given in a simplified form: as a listing of a table of unquoted, printable
53 whitespace-separated strings. These cases form a subset of the rules regarding
54 the [token](#tokens) rule, and are assumed to be the result of a
55 lexical-analysis phase feeding the parser, driven by a DFA, operating over the
56 disjunction of all such string table entries.
58 [keywords]: grammar.html#keywords
60 When such a string enclosed in double-quotes (`"`) occurs inside the grammar,
61 it is an implicit reference to a single member of such a string table
62 production. See [tokens](#tokens) for more information.
68 Rust input is interpreted as a sequence of Unicode code points encoded in UTF-8.
69 Most Rust grammar rules are defined in terms of printable ASCII-range
70 code points, but a small number are defined in terms of Unicode properties or
71 explicit code point lists. [^inputformat]
73 [^inputformat]: Substitute definitions for the special Unicode productions are
74 provided to the grammar verifier, restricted to ASCII range, when verifying the
75 grammar in this document.
79 An identifier is any nonempty Unicode[^non_ascii_idents] string of the following form:
81 [^non_ascii_idents]: Non-ASCII characters in identifiers are currently feature
82 gated. This is expected to improve soon.
86 * The first character has property `XID_start`
87 * The remaining characters have property `XID_continue`
91 * The first character is `_`
92 * The identifier is more than one character, `_` alone is not an identifier
93 * The remaining characters have property `XID_continue`
95 that does _not_ occur in the set of [keywords][keywords].
97 > **Note**: `XID_start` and `XID_continue` as character properties cover the
98 > character ranges used to form the more familiar C and Java language-family
103 Comments in Rust code follow the general C++ style of line (`//`) and
104 block (`/* ... */`) comment forms. Nested block comments are supported.
106 Line comments beginning with exactly _three_ slashes (`///`), and block
107 comments (`/** ... */`), are interpreted as a special syntax for `doc`
108 [attributes](#attributes). That is, they are equivalent to writing
109 `#[doc="..."]` around the body of the comment, i.e., `/// Foo` turns into
112 Line comments beginning with `//!` and block comments `/*! ... */` are
113 doc comments that apply to the parent of the comment, rather than the item
114 that follows. That is, they are equivalent to writing `#![doc="..."]` around
115 the body of the comment. `//!` comments are usually used to document
116 modules that occupy a source file.
118 Non-doc comments are interpreted as a form of whitespace.
122 Whitespace is any non-empty string containing only characters that have the
123 `Pattern_White_Space` Unicode property, namely:
125 - `U+0009` (horizontal tab, `'\t'`)
126 - `U+000A` (line feed, `'\n'`)
127 - `U+000B` (vertical tab)
128 - `U+000C` (form feed)
129 - `U+000D` (carriage return, `'\r'`)
130 - `U+0020` (space, `' '`)
131 - `U+0085` (next line)
132 - `U+200E` (left-to-right mark)
133 - `U+200F` (right-to-left mark)
134 - `U+2028` (line separator)
135 - `U+2029` (paragraph separator)
137 Rust is a "free-form" language, meaning that all forms of whitespace serve only
138 to separate _tokens_ in the grammar, and have no semantic significance.
140 A Rust program has identical meaning if each whitespace element is replaced
141 with any other legal whitespace element, such as a single space character.
145 Tokens are primitive productions in the grammar defined by regular
146 (non-recursive) languages. "Simple" tokens are given in [string table
147 production](#string-table-productions) form, and occur in the rest of the
148 grammar as double-quoted strings. Other tokens have exact rules given.
152 A literal is an expression consisting of a single token, rather than a sequence
153 of tokens, that immediately and directly denotes the value it evaluates to,
154 rather than referring to it by name or some other evaluation rule. A literal is
155 a form of constant expression, so is evaluated (primarily) at compile time.
159 ##### Characters and strings
161 | | Example | `#` sets | Characters | Escapes |
162 |----------------------------------------------|-----------------|------------|-------------|---------------------|
163 | [Character](#character-literals) | `'H'` | `N/A` | All Unicode | [Quote](#quote-escapes) & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
164 | [String](#string-literals) | `"hello"` | `N/A` | All Unicode | [Quote](#quote-escapes) & [Byte](#byte-escapes) & [Unicode](#unicode-escapes) |
165 | [Raw](#raw-string-literals) | `r#"hello"#` | `0...` | All Unicode | `N/A` |
166 | [Byte](#byte-literals) | `b'H'` | `N/A` | All ASCII | [Quote](#quote-escapes) & [Byte](#byte-escapes) |
167 | [Byte string](#byte-string-literals) | `b"hello"` | `N/A` | All ASCII | [Quote](#quote-escapes) & [Byte](#byte-escapes) |
168 | [Raw byte string](#raw-byte-string-literals) | `br#"hello"#` | `0...` | All ASCII | `N/A` |
174 | `\x7F` | 8-bit character code (exactly 2 digits) |
176 | `\r` | Carriage return |
181 ##### Unicode escapes
184 | `\u{7FFF}` | 24-bit Unicode character code (up to 6 digits) |
189 | `\'` | Single quote |
190 | `\"` | Double quote |
194 | [Number literals](#number-literals)`*` | Example | Exponentiation | Suffixes |
195 |----------------------------------------|---------|----------------|----------|
196 | Decimal integer | `98_222` | `N/A` | Integer suffixes |
197 | Hex integer | `0xff` | `N/A` | Integer suffixes |
198 | Octal integer | `0o77` | `N/A` | Integer suffixes |
199 | Binary integer | `0b1111_0000` | `N/A` | Integer suffixes |
200 | Floating-point | `123.0E+77` | `Optional` | Floating-point suffixes |
202 `*` All number literals allow `_` as a visual separator: `1_234.0E+18f64`
205 | Integer | Floating-point |
206 |---------|----------------|
207 | `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`, `isize`, `usize` | `f32`, `f64` |
209 #### Character and string literals
211 ##### Character literals
213 A _character literal_ is a single Unicode character enclosed within two
214 `U+0027` (single-quote) characters, with the exception of `U+0027` itself,
215 which must be _escaped_ by a preceding `U+005C` character (`\`).
217 ##### String literals
219 A _string literal_ is a sequence of any Unicode characters enclosed within two
220 `U+0022` (double-quote) characters, with the exception of `U+0022` itself,
221 which must be _escaped_ by a preceding `U+005C` character (`\`).
223 Line-break characters are allowed in string literals. Normally they represent
224 themselves (i.e. no translation), but as a special exception, when an unescaped
225 `U+005C` character (`\`) occurs immediately before the newline (`U+000A`), the
226 `U+005C` character, the newline, and all whitespace at the beginning of the
227 next line are ignored. Thus `a` and `b` are equal:
237 ##### Character escapes
239 Some additional _escapes_ are available in either character or non-raw string
240 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
243 * An _8-bit code point escape_ starts with `U+0078` (`x`) and is
244 followed by exactly two _hex digits_. It denotes the Unicode code point
245 equal to the provided hex value.
246 * A _24-bit code point escape_ starts with `U+0075` (`u`) and is followed
247 by up to six _hex digits_ surrounded by braces `U+007B` (`{`) and `U+007D`
248 (`}`). It denotes the Unicode code point equal to the provided hex value.
249 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
250 (`r`), or `U+0074` (`t`), denoting the Unicode values `U+000A` (LF),
251 `U+000D` (CR) or `U+0009` (HT) respectively.
252 * The _null escape_ is the character `U+0030` (`0`) and denotes the Unicode
253 value `U+0000` (NUL).
254 * The _backslash escape_ is the character `U+005C` (`\`) which must be
255 escaped in order to denote *itself*.
257 ##### Raw string literals
259 Raw string literals do not process any escapes. They start with the character
260 `U+0072` (`r`), followed by zero or more of the character `U+0023` (`#`) and a
261 `U+0022` (double-quote) character. The _raw string body_ can contain any sequence
262 of Unicode characters and is terminated only by another `U+0022` (double-quote)
263 character, followed by the same number of `U+0023` (`#`) characters that preceded
264 the opening `U+0022` (double-quote) character.
266 All Unicode characters contained in the raw string body represent themselves,
267 the characters `U+0022` (double-quote) (except when followed by at least as
268 many `U+0023` (`#`) characters as were used to start the raw string literal) or
269 `U+005C` (`\`) do not have any special meaning.
271 Examples for string literals:
274 "foo"; r"foo"; // foo
275 "\"foo\""; r#""foo""#; // "foo"
278 r##"foo #"# bar"##; // foo #"# bar
280 "\x52"; "R"; r"R"; // R
281 "\\x52"; r"\x52"; // \x52
284 #### Byte and byte string literals
288 A _byte literal_ is a single ASCII character (in the `U+0000` to `U+007F`
289 range) or a single _escape_ preceded by the characters `U+0062` (`b`) and
290 `U+0027` (single-quote), and followed by the character `U+0027`. If the character
291 `U+0027` is present within the literal, it must be _escaped_ by a preceding
292 `U+005C` (`\`) character. It is equivalent to a `u8` unsigned 8-bit integer
295 ##### Byte string literals
297 A non-raw _byte string literal_ is a sequence of ASCII characters and _escapes_,
298 preceded by the characters `U+0062` (`b`) and `U+0022` (double-quote), and
299 followed by the character `U+0022`. If the character `U+0022` is present within
300 the literal, it must be _escaped_ by a preceding `U+005C` (`\`) character.
301 Alternatively, a byte string literal can be a _raw byte string literal_, defined
302 below. A byte string literal of length `n` is equivalent to a `&'static [u8; n]` borrowed fixed-sized array
303 of unsigned 8-bit integers.
305 Some additional _escapes_ are available in either byte or non-raw byte string
306 literals. An escape starts with a `U+005C` (`\`) and continues with one of the
309 * A _byte escape_ escape starts with `U+0078` (`x`) and is
310 followed by exactly two _hex digits_. It denotes the byte
311 equal to the provided hex value.
312 * A _whitespace escape_ is one of the characters `U+006E` (`n`), `U+0072`
313 (`r`), or `U+0074` (`t`), denoting the bytes values `0x0A` (ASCII LF),
314 `0x0D` (ASCII CR) or `0x09` (ASCII HT) respectively.
315 * The _null escape_ is the character `U+0030` (`0`) and denotes the byte
316 value `0x00` (ASCII NUL).
317 * The _backslash escape_ is the character `U+005C` (`\`) which must be
318 escaped in order to denote its ASCII encoding `0x5C`.
320 ##### Raw byte string literals
322 Raw byte string literals do not process any escapes. They start with the
323 character `U+0062` (`b`), followed by `U+0072` (`r`), followed by zero or more
324 of the character `U+0023` (`#`), and a `U+0022` (double-quote) character. The
325 _raw string body_ can contain any sequence of ASCII characters and is terminated
326 only by another `U+0022` (double-quote) character, followed by the same number of
327 `U+0023` (`#`) characters that preceded the opening `U+0022` (double-quote)
328 character. A raw byte string literal can not contain any non-ASCII byte.
330 All characters contained in the raw string body represent their ASCII encoding,
331 the characters `U+0022` (double-quote) (except when followed by at least as
332 many `U+0023` (`#`) characters as were used to start the raw string literal) or
333 `U+005C` (`\`) do not have any special meaning.
335 Examples for byte string literals:
338 b"foo"; br"foo"; // foo
339 b"\"foo\""; br#""foo""#; // "foo"
342 br##"foo #"# bar"##; // foo #"# bar
344 b"\x52"; b"R"; br"R"; // R
345 b"\\x52"; br"\x52"; // \x52
350 A _number literal_ is either an _integer literal_ or a _floating-point
351 literal_. The grammar for recognizing the two kinds of literals is mixed.
353 ##### Integer literals
355 An _integer literal_ has one of four forms:
357 * A _decimal literal_ starts with a *decimal digit* and continues with any
358 mixture of *decimal digits* and _underscores_.
359 * A _hex literal_ starts with the character sequence `U+0030` `U+0078`
360 (`0x`) and continues as any mixture of hex digits and underscores.
361 * An _octal literal_ starts with the character sequence `U+0030` `U+006F`
362 (`0o`) and continues as any mixture of octal digits and underscores.
363 * A _binary literal_ starts with the character sequence `U+0030` `U+0062`
364 (`0b`) and continues as any mixture of binary digits and underscores.
366 Like any literal, an integer literal may be followed (immediately,
367 without any spaces) by an _integer suffix_, which forcibly sets the
368 type of the literal. The integer suffix must be the name of one of the
369 integral types: `u8`, `i8`, `u16`, `i16`, `u32`, `i32`, `u64`, `i64`,
372 The type of an _unsuffixed_ integer literal is determined by type inference:
374 * If an integer type can be _uniquely_ determined from the surrounding
375 program context, the unsuffixed integer literal has that type.
377 * If the program context under-constrains the type, it defaults to the
378 signed 32-bit integer `i32`.
380 * If the program context over-constrains the type, it is considered a
383 Examples of integer literals of various forms:
390 0o70_i16; // type i16
391 0b1111_1111_1001_0000_i32; // type i32
392 0usize; // type usize
395 Note that the Rust syntax considers `-1i8` as an application of the [unary minus
396 operator](#unary-operator-expressions) to an integer literal `1i8`, rather than
397 a single integer literal.
399 ##### Floating-point literals
401 A _floating-point literal_ has one of two forms:
403 * A _decimal literal_ followed by a period character `U+002E` (`.`). This is
404 optionally followed by another decimal literal, with an optional _exponent_.
405 * A single _decimal literal_ followed by an _exponent_.
407 Like integer literals, a floating-point literal may be followed by a
408 suffix, so long as the pre-suffix part does not end with `U+002E` (`.`).
409 The suffix forcibly sets the type of the literal. There are two valid
410 _floating-point suffixes_, `f32` and `f64` (the 32-bit and 64-bit floating point
411 types), which explicitly determine the type of the literal.
413 The type of an _unsuffixed_ floating-point literal is determined by
416 * If a floating-point type can be _uniquely_ determined from the
417 surrounding program context, the unsuffixed floating-point literal
420 * If the program context under-constrains the type, it defaults to `f64`.
422 * If the program context over-constrains the type, it is considered a
425 Examples of floating-point literals of various forms:
428 123.0f64; // type f64
431 12E+99_f64; // type f64
432 let x: f64 = 2.; // type f64
435 This last example is different because it is not possible to use the suffix
436 syntax with a floating point literal ending in a period. `2.f64` would attempt
437 to call a method named `f64` on `2`.
439 The representation semantics of floating-point numbers are described in
440 ["Machine Types"](#machine-types).
442 #### Boolean literals
444 The two values of the boolean type are written `true` and `false`.
448 Symbols are a general class of printable [tokens](#tokens) that play structural
449 roles in a variety of grammar productions. They are a
450 set of remaining miscellaneous printable tokens that do not
451 otherwise appear as [unary operators](#unary-operator-expressions), [binary
452 operators](#binary-operator-expressions), or [keywords][keywords].
453 They are catalogued in [the Symbols section][symbols] of the Grammar document.
455 [symbols]: grammar.html#symbols
460 A _path_ is a sequence of one or more path components _logically_ separated by
461 a namespace qualifier (`::`). If a path consists of only one component, it may
462 refer to either an [item](#items) or a [variable](#variables) in a local control
463 scope. If a path has multiple components, it refers to an item.
465 Every item has a _canonical path_ within its crate, but the path naming an item
466 is only meaningful within a given crate. There is no global namespace across
467 crates; an item's canonical path merely identifies it within the crate.
469 Two examples of simple paths consisting of only identifier components:
476 Path components are usually [identifiers](#identifiers), but they may
477 also include angle-bracket-enclosed lists of type arguments. In
478 [expression](#expressions) context, the type argument list is given
479 after a `::` namespace qualifier in order to disambiguate it from a
480 relational expression involving the less-than symbol (`<`). In type
481 expression context, the final namespace qualifier is omitted.
483 Two examples of paths with type arguments:
486 # struct HashMap<K, V>(K,V);
488 # fn id<T>(t: T) -> T { t }
489 type T = HashMap<i32,String>; // Type arguments used in a type expression
490 let x = id::<i32>(10); // Type arguments used in a call expression
494 Paths can be denoted with various leading qualifiers to change the meaning of
497 * Paths starting with `::` are considered to be global paths where the
498 components of the path start being resolved from the crate root. Each
499 identifier in the path must resolve to an item.
507 ::a::foo(); // call a's foo function
513 * Paths starting with the keyword `super` begin resolution relative to the
514 parent module. Each further identifier must resolve to an item.
522 super::a::foo(); // call a's foo function
528 * Paths starting with the keyword `self` begin resolution relative to the
529 current module. Each further identifier must resolve to an item.
539 Additionally keyword `super` may be repeated several times after the first
540 `super` or `self` to refer to ancestor modules.
549 super::super::foo(); // call a's foo function
550 self::super::super::foo(); // call a's foo function
560 A number of minor features of Rust are not central enough to have their own
561 syntax, and yet are not implementable as functions. Instead, they are given
562 names, and invoked through a consistent syntax: `some_extension!(...)`.
564 Users of `rustc` can define new syntax extensions in two ways:
566 * [Compiler plugins][plugin] can include arbitrary Rust code that
567 manipulates syntax trees at compile time. Note that the interface
568 for compiler plugins is considered highly unstable.
570 * [Macros](book/macros.html) define new syntax in a higher-level,
575 `macro_rules` allows users to define syntax extension in a declarative way. We
576 call such extensions "macros by example" or simply "macros" — to be distinguished
577 from the "procedural macros" defined in [compiler plugins][plugin].
579 Currently, macros can expand to expressions, statements, items, or patterns.
581 (A `sep_token` is any token other than `*` and `+`. A `non_special_token` is
582 any token other than a delimiter or `$`.)
584 The macro expander looks up macro invocations by name, and tries each macro
585 rule in turn. It transcribes the first successful match. Matching and
586 transcription are closely related to each other, and we will describe them
591 The macro expander matches and transcribes every token that does not begin with
592 a `$` literally, including delimiters. For parsing reasons, delimiters must be
593 balanced, but they are otherwise not special.
595 In the matcher, `$` _name_ `:` _designator_ matches the nonterminal in the Rust
596 syntax named by _designator_. Valid designators are:
598 * `item`: an [item](#items)
599 * `block`: a [block](#block-expressions)
600 * `stmt`: a [statement](#statements)
601 * `pat`: a [pattern](#match-expressions)
602 * `expr`: an [expression](#expressions)
603 * `ty`: a [type](#types)
604 * `ident`: an [identifier](#identifiers)
605 * `path`: a [path](#paths)
606 * `tt`: a token tree (a single [token](#tokens) or a sequence of token trees surrounded
607 by matching `()`, `[]`, or `{}`)
608 * `meta`: the contents of an [attribute](#attributes)
610 In the transcriber, the
611 designator is already known, and so only the name of a matched nonterminal comes
612 after the dollar sign.
614 In both the matcher and transcriber, the Kleene star-like operator indicates
615 repetition. The Kleene star operator consists of `$` and parentheses, optionally
616 followed by a separator token, followed by `*` or `+`. `*` means zero or more
617 repetitions, `+` means at least one repetition. The parentheses are not matched or
618 transcribed. On the matcher side, a name is bound to _all_ of the names it
619 matches, in a structure that mimics the structure of the repetition encountered
620 on a successful match. The job of the transcriber is to sort that structure
623 The rules for transcription of these repetitions are called "Macro By Example".
624 Essentially, one "layer" of repetition is discharged at a time, and all of them
625 must be discharged by the time a name is transcribed. Therefore, `( $( $i:ident
626 ),* ) => ( $i )` is an invalid macro, but `( $( $i:ident ),* ) => ( $( $i:ident
627 ),* )` is acceptable (if trivial).
629 When Macro By Example encounters a repetition, it examines all of the `$`
630 _name_ s that occur in its body. At the "current layer", they all must repeat
631 the same number of times, so ` ( $( $i:ident ),* ; $( $j:ident ),* ) => ( $(
632 ($i,$j) ),* )` is valid if given the argument `(a,b,c ; d,e,f)`, but not
633 `(a,b,c ; d,e)`. The repetition walks through the choices at that layer in
634 lockstep, so the former input transcribes to `(a,d), (b,e), (c,f)`.
636 Nested repetitions are allowed.
638 ### Parsing limitations
640 The parser used by the macro system is reasonably powerful, but the parsing of
641 Rust syntax is restricted in two ways:
643 1. Macro definitions are required to include suitable separators after parsing
644 expressions and other bits of the Rust grammar. This implies that
645 a macro definition like `$i:expr [ , ]` is not legal, because `[` could be part
646 of an expression. A macro definition like `$i:expr,` or `$i:expr;` would be legal,
647 however, because `,` and `;` are legal separators. See [RFC 550] for more information.
648 2. The parser must have eliminated all ambiguity by the time it reaches a `$`
649 _name_ `:` _designator_. This requirement most often affects name-designator
650 pairs when they occur at the beginning of, or immediately after, a `$(...)*`;
651 requiring a distinctive token in front can solve the problem.
653 [RFC 550]: https://github.com/rust-lang/rfcs/blob/master/text/0550-macro-future-proofing.md
655 # Crates and source files
657 Although Rust, like any other language, can be implemented by an interpreter as
658 well as a compiler, the only existing implementation is a compiler,
660 always been designed to be compiled. For these reasons, this section assumes a
663 Rust's semantics obey a *phase distinction* between compile-time and
664 run-time.[^phase-distinction] Semantic rules that have a *static
665 interpretation* govern the success or failure of compilation, while
667 that have a *dynamic interpretation* govern the behavior of the program at
670 [^phase-distinction]: This distinction would also exist in an interpreter.
671 Static checks like syntactic analysis, type checking, and lints should
672 happen before the program is executed regardless of when it is executed.
674 The compilation model centers on artifacts called _crates_. Each compilation
675 processes a single crate in source form, and if successful, produces a single
676 crate in binary form: either an executable or some sort of
677 library.[^cratesourcefile]
679 [^cratesourcefile]: A crate is somewhat analogous to an *assembly* in the
680 ECMA-335 CLI model, a *library* in the SML/NJ Compilation Manager, a *unit*
681 in the Owens and Flatt module system, or a *configuration* in Mesa.
683 A _crate_ is a unit of compilation and linking, as well as versioning,
684 distribution and runtime loading. A crate contains a _tree_ of nested
685 [module](#modules) scopes. The top level of this tree is a module that is
686 anonymous (from the point of view of paths within the module) and any item
687 within a crate has a canonical [module path](#paths) denoting its location
688 within the crate's module tree.
690 The Rust compiler is always invoked with a single source file as input, and
691 always produces a single output crate. The processing of that source file may
692 result in other source files being loaded as modules. Source files have the
695 A Rust source file describes a module, the name and location of which —
696 in the module tree of the current crate — are defined from outside the
697 source file: either by an explicit `mod_item` in a referencing source file, or
698 by the name of the crate itself. Every source file is a module, but not every
699 module needs its own source file: [module definitions](#modules) can be nested
702 Each source file contains a sequence of zero or more `item` definitions, and
703 may optionally begin with any number of [attributes](#items-and-attributes)
704 that apply to the containing module, most of which influence the behavior of
705 the compiler. The anonymous crate module can have additional attributes that
706 apply to the crate as a whole.
709 // Specify the crate name.
710 #![crate_name = "projx"]
712 // Specify the type of output artifact.
713 #![crate_type = "lib"]
715 // Turn on a warning.
716 // This can be done in any module, not just the anonymous crate module.
717 #![warn(non_camel_case_types)]
720 A crate that contains a `main` function can be compiled to an executable. If a
721 `main` function is present, its return type must be `()`
722 ("[unit](#tuple-types)") and it must take no arguments.
724 # Items and attributes
726 Crates contain [items](#items), each of which may have some number of
727 [attributes](#attributes) attached to it.
731 An _item_ is a component of a crate. Items are organized within a crate by a
732 nested set of [modules](#modules). Every crate has a single "outermost"
733 anonymous module; all further items within the crate have [paths](#paths)
734 within the module tree of the crate.
736 Items are entirely determined at compile-time, generally remain fixed during
737 execution, and may reside in read-only memory.
739 There are several kinds of item:
741 * [`extern crate` declarations](#extern-crate-declarations)
742 * [`use` declarations](#use-declarations)
743 * [modules](#modules)
744 * [function definitions](#functions)
745 * [`extern` blocks](#external-blocks)
746 * [type definitions](grammar.html#type-definitions)
747 * [struct definitions](#structs)
748 * [enumeration definitions](#enumerations)
749 * [constant items](#constant-items)
750 * [static items](#static-items)
751 * [trait definitions](#traits)
752 * [implementations](#implementations)
754 Some items form an implicit scope for the declaration of sub-items. In other
755 words, within a function or module, declarations of items can (in many cases)
756 be mixed with the statements, control blocks, and similar artifacts that
757 otherwise compose the item body. The meaning of these scoped items is the same
758 as if the item was declared outside the scope — it is still a static item
759 — except that the item's *path name* within the module namespace is
760 qualified by the name of the enclosing item, or is private to the enclosing
761 item (in the case of functions). The grammar specifies the exact locations in
762 which sub-item declarations may appear.
766 All items except modules, constants and statics may be *parameterized* by type.
767 Type parameters are given as a comma-separated list of identifiers enclosed in
768 angle brackets (`<...>`), after the name of the item and before its definition.
769 The type parameters of an item are considered "part of the name", not part of
770 the type of the item. A referencing [path](#paths) must (in principle) provide
771 type arguments as a list of comma-separated types enclosed within angle
772 brackets, in order to refer to the type-parameterized item. In practice, the
773 type-inference system can usually infer such argument types from context. There
774 are no general type-parametric types, only type-parametric items. That is, Rust
775 has no notion of type abstraction: there are no higher-ranked (or "forall") types
776 abstracted over other types, though higher-ranked types do exist for lifetimes.
780 A module is a container for zero or more [items](#items).
782 A _module item_ is a module, surrounded in braces, named, and prefixed with the
783 keyword `mod`. A module item introduces a new, named module into the tree of
784 modules making up a crate. Modules can nest arbitrarily.
786 An example of a module:
790 type Complex = (f64, f64);
791 fn sin(f: f64) -> f64 {
795 fn cos(f: f64) -> f64 {
799 fn tan(f: f64) -> f64 {
806 Modules and types share the same namespace. Declaring a named type with
807 the same name as a module in scope is forbidden: that is, a type definition,
808 trait, struct, enumeration, or type parameter can't shadow the name of a module
809 in scope, or vice versa.
811 A module without a body is loaded from an external file, by default with the
812 same name as the module, plus the `.rs` extension. When a nested submodule is
813 loaded from an external file, it is loaded from a subdirectory path that
814 mirrors the module hierarchy.
817 // Load the `vec` module from `vec.rs`
821 // Load the `local_data` module from `thread/local_data.rs`
822 // or `thread/local_data/mod.rs`.
827 The directories and files used for loading external file modules can be
828 influenced with the `path` attribute.
831 #[path = "thread_files"]
833 // Load the `local_data` module from `thread_files/tls.rs`
839 #### Extern crate declarations
841 An _`extern crate` declaration_ specifies a dependency on an external crate.
842 The external crate is then bound into the declaring scope as the `ident`
843 provided in the `extern_crate_decl`.
845 The external crate is resolved to a specific `soname` at compile time, and a
846 runtime linkage requirement to that `soname` is passed to the linker for
847 loading at runtime. The `soname` is resolved at compile time by scanning the
848 compiler's library path and matching the optional `crateid` provided against
849 the `crateid` attributes that were declared on the external crate when it was
850 compiled. If no `crateid` is provided, a default `name` attribute is assumed,
851 equal to the `ident` given in the `extern_crate_decl`.
853 Three examples of `extern crate` declarations:
858 extern crate std; // equivalent to: extern crate std as std;
860 extern crate std as ruststd; // linking to 'std' under another name
863 When naming Rust crates, hyphens are disallowed. However, Cargo packages may
864 make use of them. In such case, when `Cargo.toml` doesn't specify a crate name,
865 Cargo will transparently replace `-` with `_` (Refer to [RFC 940] for more
871 // Importing the Cargo package hello-world
872 extern crate hello_world; // hyphen replaced with an underscore
875 [RFC 940]: https://github.com/rust-lang/rfcs/blob/master/text/0940-hyphens-considered-harmful.md
877 #### Use declarations
879 A _use declaration_ creates one or more local name bindings synonymous with
880 some other [path](#paths). Usually a `use` declaration is used to shorten the
881 path required to refer to a module item. These declarations may appear in
882 [modules](#modules) and [blocks](grammar.html#block-expressions), usually at the top.
884 > **Note**: Unlike in many languages,
885 > `use` declarations in Rust do *not* declare linkage dependency with external crates.
886 > Rather, [`extern crate` declarations](#extern-crate-declarations) declare linkage dependencies.
888 Use declarations support a number of convenient shortcuts:
890 * Rebinding the target name as a new local name, using the syntax `use p::q::r as x;`
891 * Simultaneously binding a list of paths differing only in their final element,
892 using the glob-like brace syntax `use a::b::{c,d,e,f};`
893 * Binding all paths matching a given prefix, using the asterisk wildcard syntax
895 * Simultaneously binding a list of paths differing only in their final element
896 and their immediate parent module, using the `self` keyword, such as
897 `use a::b::{self, c, d};`
899 An example of `use` declarations:
902 use std::option::Option::{Some, None};
903 use std::collections::hash_map::{self, HashMap};
906 fn bar(map1: HashMap<String, usize>, map2: hash_map::HashMap<String, usize>){}
909 // Equivalent to 'foo(vec![std::option::Option::Some(1.0f64),
910 // std::option::Option::None]);'
911 foo(vec![Some(1.0f64), None]);
913 // Both `hash_map` and `HashMap` are in scope.
914 let map1 = HashMap::new();
915 let map2 = hash_map::HashMap::new();
920 Like items, `use` declarations are private to the containing module, by
921 default. Also like items, a `use` declaration can be public, if qualified by
922 the `pub` keyword. Such a `use` declaration serves to _re-export_ a name. A
923 public `use` declaration can therefore _redirect_ some public name to a
924 different target definition: even a definition with a private canonical path,
925 inside a different module. If a sequence of such redirections form a cycle or
926 cannot be resolved unambiguously, they represent a compile-time error.
928 An example of re-exporting:
933 pub use quux::foo::{bar, baz};
942 In this example, the module `quux` re-exports two public names defined in
945 Also note that the paths contained in `use` items are relative to the crate
946 root. So, in the previous example, the `use` refers to `quux::foo::{bar,
947 baz}`, and not simply to `foo::{bar, baz}`. This also means that top-level
948 module declarations should be at the crate root if direct usage of the declared
949 modules within `use` items is desired. It is also possible to use `self` and
950 `super` at the beginning of a `use` item to refer to the current and direct
951 parent modules respectively. All rules regarding accessing declared modules in
952 `use` declarations apply to both module declarations and `extern crate`
955 An example of what will and will not work for `use` items:
958 # #![allow(unused_imports)]
959 use foo::baz::foobaz; // good: foo is at the root of the crate
967 use foo::example::iter; // good: foo is at crate root
968 // use example::iter; // bad: example is not at the crate root
969 use self::baz::foobaz; // good: self refers to module 'foo'
970 use foo::bar::foobar; // good: foo is at crate root
977 use super::bar::foobar; // good: super refers to module 'foo'
987 A _function item_ defines a sequence of [statements](#statements) and a
988 final [expression](#expressions), along with a name and a set of
989 parameters. Other than a name, all these are optional.
990 Functions are declared with the keyword `fn`. Functions may declare a
991 set of *input* [*variables*](#variables) as parameters, through which the caller
992 passes arguments into the function, and the *output* [*type*](#types)
993 of the value the function will return to its caller on completion.
995 A function may also be copied into a first-class *value*, in which case the
996 value has the corresponding [*function type*](#function-types), and can be used
997 otherwise exactly as a function item (with a minor additional cost of calling
998 the function indirectly).
1000 Every control path in a function logically ends with a `return` expression or a
1001 diverging expression. If the outermost block of a function has a
1002 value-producing expression in its final-expression position, that expression is
1003 interpreted as an implicit `return` expression applied to the final-expression.
1005 An example of a function:
1008 fn add(x: i32, y: i32) -> i32 {
1013 As with `let` bindings, function arguments are irrefutable patterns, so any
1014 pattern that is valid in a let binding is also valid as an argument.
1017 fn first((value, _): (i32, i32)) -> i32 { value }
1021 #### Generic functions
1023 A _generic function_ allows one or more _parameterized types_ to appear in its
1024 signature. Each type parameter must be explicitly declared in an
1025 angle-bracket-enclosed and comma-separated list, following the function name.
1028 // foo is generic over A and B
1030 fn foo<A, B>(x: A, y: B) {
1033 Inside the function signature and body, the name of the type parameter can be
1034 used as a type name. [Trait](#traits) bounds can be specified for type parameters
1035 to allow methods with that trait to be called on values of that type. This is
1036 specified using the `where` syntax:
1039 fn foo<T>(x: T) where T: Debug {
1042 When a generic function is referenced, its type is instantiated based on the
1043 context of the reference. For example, calling the `foo` function here:
1046 use std::fmt::Debug;
1048 fn foo<T>(x: &[T]) where T: Debug {
1056 will instantiate type parameter `T` with `i32`.
1058 The type parameters can also be explicitly supplied in a trailing
1059 [path](#paths) component after the function name. This might be necessary if
1060 there is not sufficient context to determine the type parameters. For example,
1061 `mem::size_of::<u32>() == 4`.
1063 #### Diverging functions
1065 A special kind of function can be declared with a `!` character where the
1066 output type would normally be. For example:
1069 fn my_err(s: &str) -> ! {
1075 We call such functions "diverging" because they never return a value to the
1076 caller. Every control path in a diverging function must end with a `panic!()` or
1077 a call to another diverging function on every control path. The `!` annotation
1078 does *not* denote a type.
1080 It might be necessary to declare a diverging function because as mentioned
1081 previously, the typechecker checks that every control path in a function ends
1082 with a [`return`](#return-expressions) or diverging expression. So, if `my_err`
1083 were declared without the `!` annotation, the following code would not
1087 # fn my_err(s: &str) -> ! { panic!() }
1089 fn f(i: i32) -> i32 {
1094 my_err("Bad number!");
1099 This will not compile without the `!` annotation on `my_err`, since the `else`
1100 branch of the conditional in `f` does not return an `i32`, as required by the
1101 signature of `f`. Adding the `!` annotation to `my_err` informs the
1102 typechecker that, should control ever enter `my_err`, no further type judgments
1103 about `f` need to hold, since control will never resume in any context that
1104 relies on those judgments. Thus the return type on `f` only needs to reflect
1105 the `if` branch of the conditional.
1107 #### Extern functions
1109 Extern functions are part of Rust's foreign function interface, providing the
1110 opposite functionality to [external blocks](#external-blocks). Whereas
1111 external blocks allow Rust code to call foreign code, extern functions with
1112 bodies defined in Rust code _can be called by foreign code_. They are defined
1113 in the same way as any other Rust function, except that they have the `extern`
1117 // Declares an extern fn, the ABI defaults to "C"
1118 extern fn new_i32() -> i32 { 0 }
1120 // Declares an extern fn with "stdcall" ABI
1121 extern "stdcall" fn new_i32_stdcall() -> i32 { 0 }
1124 Unlike normal functions, extern fns have type `extern "ABI" fn()`. This is the
1125 same type as the functions declared in an extern block.
1128 # extern fn new_i32() -> i32 { 0 }
1129 let fptr: extern "C" fn() -> i32 = new_i32;
1132 Extern functions may be called directly from Rust code as Rust uses large,
1133 contiguous stack segments like C.
1137 A _type alias_ defines a new name for an existing [type](#types). Type
1138 aliases are declared with the keyword `type`. Every value has a single,
1139 specific type, but may implement several different traits, or be compatible with
1140 several different type constraints.
1142 For example, the following defines the type `Point` as a synonym for the type
1143 `(u8, u8)`, the type of pairs of unsigned 8 bit integers:
1146 type Point = (u8, u8);
1147 let p: Point = (41, 68);
1150 Currently a type alias to an enum type cannot be used to qualify the
1156 let _: F = E::A; // OK
1157 // let _: F = F::A; // Doesn't work
1162 A _struct_ is a nominal [struct type](#struct-types) defined with the
1165 An example of a `struct` item and its use:
1168 struct Point {x: i32, y: i32}
1169 let p = Point {x: 10, y: 11};
1173 A _tuple struct_ is a nominal [tuple type](#tuple-types), also defined with
1174 the keyword `struct`. For example:
1177 struct Point(i32, i32);
1178 let p = Point(10, 11);
1179 let px: i32 = match p { Point(x, _) => x };
1182 A _unit-like struct_ is a struct without any fields, defined by leaving off
1183 the list of fields entirely. Such a struct implicitly defines a constant of
1184 its type with the same name. For example:
1188 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1195 const Cookie: Cookie = Cookie {};
1196 let c = [Cookie, Cookie {}, Cookie, Cookie {}];
1199 The precise memory layout of a struct is not specified. One can specify a
1200 particular layout using the [`repr` attribute](#ffi-attributes).
1204 An _enumeration_ is a simultaneous definition of a nominal [enumerated
1205 type](#enumerated-types) as well as a set of *constructors*, that can be used
1206 to create or pattern-match values of the corresponding enumerated type.
1208 Enumerations are declared with the keyword `enum`.
1210 An example of an `enum` item and its use:
1218 let mut a: Animal = Animal::Dog;
1222 Enumeration constructors can have either named or unnamed fields:
1227 Cat { name: String, weight: f64 },
1230 let mut a: Animal = Animal::Dog("Cocoa".to_string(), 37.2);
1231 a = Animal::Cat { name: "Spotty".to_string(), weight: 2.7 };
1234 In this example, `Cat` is a _struct-like enum variant_,
1235 whereas `Dog` is simply called an enum variant.
1237 Each enum value has a _discriminant_ which is an integer associated to it. You
1238 can specify it explicitly:
1246 The right hand side of the specification is interpreted as an `isize` value,
1247 but the compiler is allowed to use a smaller type in the actual memory layout.
1248 The [`repr` attribute](#ffi-attributes) can be added in order to change
1249 the type of the right hand side and specify the memory layout.
1251 If a discriminant isn't specified, they start at zero, and add one for each
1254 You can cast an enum to get its discriminant:
1257 # enum Foo { Bar = 123 }
1258 let x = Foo::Bar as u32; // x is now 123u32
1261 This only works as long as none of the variants have data attached. If
1262 it were `Bar(i32)`, this is disallowed.
1266 A *constant item* is a named _constant value_ which is not associated with a
1267 specific memory location in the program. Constants are essentially inlined
1268 wherever they are used, meaning that they are copied directly into the relevant
1269 context when used. References to the same constant are not necessarily
1270 guaranteed to refer to the same memory address.
1272 Constant values must not have destructors, and otherwise permit most forms of
1273 data. Constants may refer to the address of other constants, in which case the
1274 address will have the `static` lifetime. (See below on [static lifetime
1275 elision](#static-lifetime-elision).) The compiler is, however, still at
1276 liberty to translate the constant many times, so the address referred to may not
1279 Constants must be explicitly typed. The type may be `bool`, `char`, a number, or
1280 a type derived from those primitive types. The derived types are references with
1281 the `static` lifetime, fixed-size arrays, tuples, enum variants, and structs.
1284 const BIT1: u32 = 1 << 0;
1285 const BIT2: u32 = 1 << 1;
1287 const BITS: [u32; 2] = [BIT1, BIT2];
1288 const STRING: &'static str = "bitstring";
1290 struct BitsNStrings<'a> {
1295 const BITS_N_STRINGS: BitsNStrings<'static> = BitsNStrings {
1303 A *static item* is similar to a *constant*, except that it represents a precise
1304 memory location in the program. A static is never "inlined" at the usage site,
1305 and all references to it refer to the same memory location. Static items have
1306 the `static` lifetime, which outlives all other lifetimes in a Rust program.
1307 Static items may be placed in read-only memory if they do not contain any
1308 interior mutability.
1310 Statics may contain interior mutability through the `UnsafeCell` language item.
1311 All access to a static is safe, but there are a number of restrictions on
1314 * Statics may not contain any destructors.
1315 * The types of static values must ascribe to `Sync` to allow thread-safe access.
1316 * Statics may not refer to other statics by value, only by reference.
1317 * Constants cannot refer to statics.
1319 Constants should in general be preferred over statics, unless large amounts of
1320 data are being stored, or single-address and mutability properties are required.
1322 #### Mutable statics
1324 If a static item is declared with the `mut` keyword, then it is allowed to
1325 be modified by the program. One of Rust's goals is to make concurrency bugs
1326 hard to run into, and this is obviously a very large source of race conditions
1327 or other bugs. For this reason, an `unsafe` block is required when either
1328 reading or writing a mutable static variable. Care should be taken to ensure
1329 that modifications to a mutable static are safe with respect to other threads
1330 running in the same process.
1332 Mutable statics are still very useful, however. They can be used with C
1333 libraries and can also be bound from C libraries (in an `extern` block).
1336 # fn atomic_add(_: &mut u32, _: u32) -> u32 { 2 }
1338 static mut LEVELS: u32 = 0;
1340 // This violates the idea of no shared state, and this doesn't internally
1341 // protect against races, so this function is `unsafe`
1342 unsafe fn bump_levels_unsafe1() -> u32 {
1348 // Assuming that we have an atomic_add function which returns the old value,
1349 // this function is "safe" but the meaning of the return value may not be what
1350 // callers expect, so it's still marked as `unsafe`
1351 unsafe fn bump_levels_unsafe2() -> u32 {
1352 return atomic_add(&mut LEVELS, 1);
1356 Mutable statics have the same restrictions as normal statics, except that the
1357 type of the value is not required to ascribe to `Sync`.
1359 #### `'static` lifetime elision
1361 [Unstable] Both constant and static declarations of reference types have
1362 *implicit* `'static` lifetimes unless an explicit lifetime is specified. As
1363 such, the constant declarations involving `'static` above may be written
1364 without the lifetimes. Returning to our previous example:
1367 #[feature(static_in_const)]
1368 const BIT1: u32 = 1 << 0;
1369 const BIT2: u32 = 1 << 1;
1371 const BITS: [u32; 2] = [BIT1, BIT2];
1372 const STRING: &str = "bitstring";
1374 struct BitsNStrings<'a> {
1379 const BITS_N_STRINGS: BitsNStrings = BitsNStrings {
1387 A _trait_ describes an abstract interface that types can
1388 implement. This interface consists of associated items, which come in
1395 Associated functions whose first parameter is named `self` are called
1396 methods and may be invoked using `.` notation (e.g., `x.foo()`).
1398 All traits define an implicit type parameter `Self` that refers to
1399 "the type that is implementing this interface". Traits may also
1400 contain additional type parameters. These type parameters (including
1401 `Self`) may be constrained by other traits and so forth as usual.
1403 Trait bounds on `Self` are considered "supertraits". These are
1404 required to be acyclic. Supertraits are somewhat different from other
1405 constraints in that they affect what methods are available in the
1406 vtable when the trait is used as a [trait object](#trait-objects).
1408 Traits are implemented for specific types through separate
1409 [implementations](#implementations).
1411 Consider the following trait:
1414 # type Surface = i32;
1415 # type BoundingBox = i32;
1417 fn draw(&self, Surface);
1418 fn bounding_box(&self) -> BoundingBox;
1422 This defines a trait with two methods. All values that have
1423 [implementations](#implementations) of this trait in scope can have their
1424 `draw` and `bounding_box` methods called, using `value.bounding_box()`
1425 [syntax](#method-call-expressions).
1427 Traits can include default implementations of methods, as in:
1432 fn baz(&self) { println!("We called baz."); }
1436 Here the `baz` method has a default implementation, so types that implement
1437 `Foo` need only implement `bar`. It is also possible for implementing types
1438 to override a method that has a default implementation.
1440 Type parameters can be specified for a trait to make it generic. These appear
1441 after the trait name, using the same syntax used in [generic
1442 functions](#generic-functions).
1446 fn len(&self) -> u32;
1447 fn elt_at(&self, n: u32) -> T;
1448 fn iter<F>(&self, F) where F: Fn(T);
1452 It is also possible to define associated types for a trait. Consider the
1453 following example of a `Container` trait. Notice how the type is available
1454 for use in the method signatures:
1460 fn insert(&mut self, Self::E);
1464 In order for a type to implement this trait, it must not only provide
1465 implementations for every method, but it must specify the type `E`. Here's
1466 an implementation of `Container` for the standard library type `Vec`:
1471 # fn empty() -> Self;
1472 # fn insert(&mut self, Self::E);
1474 impl<T> Container for Vec<T> {
1476 fn empty() -> Vec<T> { Vec::new() }
1477 fn insert(&mut self, x: T) { self.push(x); }
1481 Generic functions may use traits as _bounds_ on their type parameters. This
1482 will have two effects:
1484 - Only types that have the trait may instantiate the parameter.
1485 - Within the generic function, the methods of the trait can be
1486 called on values that have the parameter's type.
1491 # type Surface = i32;
1492 # trait Shape { fn draw(&self, Surface); }
1493 fn draw_twice<T: Shape>(surface: Surface, sh: T) {
1499 Traits also define a [trait object](#trait-objects) with the same
1500 name as the trait. Values of this type are created by coercing from a
1501 pointer of some specific type to a pointer of trait type. For example,
1502 `&T` could be coerced to `&Shape` if `T: Shape` holds (and similarly
1503 for `Box<T>`). This coercion can either be implicit or
1504 [explicit](#type-cast-expressions). Here is an example of an explicit
1509 impl Shape for i32 { }
1510 let mycircle = 0i32;
1511 let myshape: Box<Shape> = Box::new(mycircle) as Box<Shape>;
1514 The resulting value is a box containing the value that was cast, along with
1515 information that identifies the methods of the implementation that was used.
1516 Values with a trait type can have [methods called](#method-call-expressions) on
1517 them, for any method in the trait, and can be used to instantiate type
1518 parameters that are bounded by the trait.
1520 Trait methods may be static, which means that they lack a `self` argument.
1521 This means that they can only be called with function call syntax (`f(x)`) and
1522 not method call syntax (`obj.f()`). The way to refer to the name of a static
1523 method is to qualify it with the trait name, treating the trait name like a
1524 module. For example:
1528 fn from_i32(n: i32) -> Self;
1531 fn from_i32(n: i32) -> f64 { n as f64 }
1533 let x: f64 = Num::from_i32(42);
1536 Traits may inherit from other traits. Consider the following example:
1539 trait Shape { fn area(&self) -> f64; }
1540 trait Circle : Shape { fn radius(&self) -> f64; }
1543 The syntax `Circle : Shape` means that types that implement `Circle` must also
1544 have an implementation for `Shape`. Multiple supertraits are separated by `+`,
1545 `trait Circle : Shape + PartialEq { }`. In an implementation of `Circle` for a
1546 given type `T`, methods can refer to `Shape` methods, since the typechecker
1547 checks that any type with an implementation of `Circle` also has an
1548 implementation of `Shape`:
1553 trait Shape { fn area(&self) -> f64; }
1554 trait Circle : Shape { fn radius(&self) -> f64; }
1555 impl Shape for Foo {
1556 fn area(&self) -> f64 {
1560 impl Circle for Foo {
1561 fn radius(&self) -> f64 {
1562 println!("calling area: {}", self.area());
1572 In type-parameterized functions, methods of the supertrait may be called on
1573 values of subtrait-bound type parameters. Referring to the previous example of
1574 `trait Circle : Shape`:
1577 # trait Shape { fn area(&self) -> f64; }
1578 # trait Circle : Shape { fn radius(&self) -> f64; }
1579 fn radius_times_area<T: Circle>(c: T) -> f64 {
1580 // `c` is both a Circle and a Shape
1581 c.radius() * c.area()
1585 Likewise, supertrait methods may also be called on trait objects.
1588 # trait Shape { fn area(&self) -> f64; }
1589 # trait Circle : Shape { fn radius(&self) -> f64; }
1590 # impl Shape for i32 { fn area(&self) -> f64 { 0.0 } }
1591 # impl Circle for i32 { fn radius(&self) -> f64 { 0.0 } }
1592 # let mycircle = 0i32;
1593 let mycircle = Box::new(mycircle) as Box<Circle>;
1594 let nonsense = mycircle.radius() * mycircle.area();
1599 An _implementation_ is an item that implements a [trait](#traits) for a
1602 Implementations are defined with the keyword `impl`.
1605 # #[derive(Copy, Clone)]
1606 # struct Point {x: f64, y: f64};
1607 # type Surface = i32;
1608 # struct BoundingBox {x: f64, y: f64, width: f64, height: f64};
1609 # trait Shape { fn draw(&self, Surface); fn bounding_box(&self) -> BoundingBox; }
1610 # fn do_draw_circle(s: Surface, c: Circle) { }
1616 impl Copy for Circle {}
1618 impl Clone for Circle {
1619 fn clone(&self) -> Circle { *self }
1622 impl Shape for Circle {
1623 fn draw(&self, s: Surface) { do_draw_circle(s, *self); }
1624 fn bounding_box(&self) -> BoundingBox {
1625 let r = self.radius;
1627 x: self.center.x - r,
1628 y: self.center.y - r,
1636 It is possible to define an implementation without referring to a trait. The
1637 methods in such an implementation can only be used as direct calls on the values
1638 of the type that the implementation targets. In such an implementation, the
1639 trait type and `for` after `impl` are omitted. Such implementations are limited
1640 to nominal types (enums, structs, trait objects), and the implementation must
1641 appear in the same crate as the `self` type:
1644 struct Point {x: i32, y: i32}
1648 println!("Point is at ({}, {})", self.x, self.y);
1652 let my_point = Point {x: 10, y:11};
1656 When a trait _is_ specified in an `impl`, all methods declared as part of the
1657 trait must be implemented, with matching types and type parameter counts.
1659 An implementation can take type parameters, which can be different from the
1660 type parameters taken by the trait it implements. Implementation parameters
1661 are written after the `impl` keyword.
1664 # trait Seq<T> { fn dummy(&self, _: T) { } }
1665 impl<T> Seq<T> for Vec<T> {
1668 impl Seq<bool> for u32 {
1669 /* Treat the integer as a sequence of bits */
1675 External blocks form the basis for Rust's foreign function interface.
1676 Declarations in an external block describe symbols in external, non-Rust
1679 Functions within external blocks are declared in the same way as other Rust
1680 functions, with the exception that they may not have a body and are instead
1681 terminated by a semicolon.
1683 Functions within external blocks may be called by Rust code, just like
1684 functions defined in Rust. The Rust compiler automatically translates between
1685 the Rust ABI and the foreign ABI.
1687 Functions within external blocks may be variadic by specifying `...` after one
1688 or more named arguments in the argument list:
1692 fn foo(x: i32, ...);
1696 A number of [attributes](#ffi-attributes) control the behavior of external blocks.
1698 By default external blocks assume that the library they are calling uses the
1699 standard C ABI on the specific platform. Other ABIs may be specified using an
1700 `abi` string, as shown here:
1703 // Interface to the Windows API
1704 extern "stdcall" { }
1707 There are three ABI strings which are cross-platform, and which all compilers
1708 are guaranteed to support:
1710 * `extern "Rust"` -- The default ABI when you write a normal `fn foo()` in any
1712 * `extern "C"` -- This is the same as `extern fn foo()`; whatever the default
1713 your C compiler supports.
1714 * `extern "system"` -- Usually the same as `extern "C"`, except on Win32, in
1715 which case it's `"stdcall"`, or what you should use to link to the Windows API
1718 There are also some platform-specific ABI strings:
1720 * `extern "cdecl"` -- The default for x86\_32 C code.
1721 * `extern "stdcall"` -- The default for the Win32 API on x86\_32.
1722 * `extern "win64"` -- The default for C code on x86\_64 Windows.
1723 * `extern "sysv64"` -- The default for C code on non-Windows x86\_64.
1724 * `extern "aapcs"` -- The default for ARM.
1725 * `extern "fastcall"` -- The `fastcall` ABI -- corresponds to MSVC's
1726 `__fastcall` and GCC and clang's `__attribute__((fastcall))`
1727 * `extern "vectorcall"` -- The `vectorcall` ABI -- corresponds to MSVC's
1728 `__vectorcall` and clang's `__attribute__((vectorcall))`
1730 Finally, there are some rustc-specific ABI strings:
1732 * `extern "rust-intrinsic"` -- The ABI of rustc intrinsics.
1733 * `extern "rust-call"` -- The ABI of the Fn::call trait functions.
1734 * `extern "platform-intrinsic"` -- Specific platform intrinsics -- like, for
1735 example, `sqrt` -- have this ABI. You should never have to deal with it.
1737 The `link` attribute allows the name of the library to be specified. When
1738 specified the compiler will attempt to link against the native library of the
1742 #[link(name = "crypto")]
1746 The type of a function declared in an extern block is `extern "abi" fn(A1, ...,
1747 An) -> R`, where `A1...An` are the declared types of its arguments and `R` is
1748 the declared return type.
1750 It is valid to add the `link` attribute on an empty extern block. You can use
1751 this to satisfy the linking requirements of extern blocks elsewhere in your code
1752 (including upstream crates) instead of adding the attribute to each extern block.
1754 ## Visibility and Privacy
1756 These two terms are often used interchangeably, and what they are attempting to
1757 convey is the answer to the question "Can this item be used at this location?"
1759 Rust's name resolution operates on a global hierarchy of namespaces. Each level
1760 in the hierarchy can be thought of as some item. The items are one of those
1761 mentioned above, but also include external crates. Declaring or defining a new
1762 module can be thought of as inserting a new tree into the hierarchy at the
1763 location of the definition.
1765 To control whether interfaces can be used across modules, Rust checks each use
1766 of an item to see whether it should be allowed or not. This is where privacy
1767 warnings are generated, or otherwise "you used a private item of another module
1768 and weren't allowed to."
1770 By default, everything in Rust is *private*, with two exceptions: Associated
1771 items in a `pub` Trait are public by default; Enum variants
1772 in a `pub` enum are also public by default. When an item is declared as `pub`,
1773 it can be thought of as being accessible to the outside world. For example:
1777 // Declare a private struct
1780 // Declare a public struct with a private field
1785 // Declare a public enum with two public variants
1787 PubliclyAccessibleState,
1788 PubliclyAccessibleState2,
1792 With the notion of an item being either public or private, Rust allows item
1793 accesses in two cases:
1795 1. If an item is public, then it can be used externally through any of its
1797 2. If an item is private, it may be accessed by the current module and its
1800 These two cases are surprisingly powerful for creating module hierarchies
1801 exposing public APIs while hiding internal implementation details. To help
1802 explain, here's a few use cases and what they would entail:
1804 * A library developer needs to expose functionality to crates which link
1805 against their library. As a consequence of the first case, this means that
1806 anything which is usable externally must be `pub` from the root down to the
1807 destination item. Any private item in the chain will disallow external
1810 * A crate needs a global available "helper module" to itself, but it doesn't
1811 want to expose the helper module as a public API. To accomplish this, the
1812 root of the crate's hierarchy would have a private module which then
1813 internally has a "public API". Because the entire crate is a descendant of
1814 the root, then the entire local crate can access this private module through
1817 * When writing unit tests for a module, it's often a common idiom to have an
1818 immediate child of the module to-be-tested named `mod test`. This module
1819 could access any items of the parent module through the second case, meaning
1820 that internal implementation details could also be seamlessly tested from the
1823 In the second case, it mentions that a private item "can be accessed" by the
1824 current module and its descendants, but the exact meaning of accessing an item
1825 depends on what the item is. Accessing a module, for example, would mean
1826 looking inside of it (to import more items). On the other hand, accessing a
1827 function would mean that it is invoked. Additionally, path expressions and
1828 import statements are considered to access an item in the sense that the
1829 import/expression is only valid if the destination is in the current visibility
1832 Here's an example of a program which exemplifies the three cases outlined
1836 // This module is private, meaning that no external crate can access this
1837 // module. Because it is private at the root of this current crate, however, any
1838 // module in the crate may access any publicly visible item in this module.
1839 mod crate_helper_module {
1841 // This function can be used by anything in the current crate
1842 pub fn crate_helper() {}
1844 // This function *cannot* be used by anything else in the crate. It is not
1845 // publicly visible outside of the `crate_helper_module`, so only this
1846 // current module and its descendants may access it.
1847 fn implementation_detail() {}
1850 // This function is "public to the root" meaning that it's available to external
1851 // crates linking against this one.
1852 pub fn public_api() {}
1854 // Similarly to 'public_api', this module is public so external crates may look
1857 use crate_helper_module;
1859 pub fn my_method() {
1860 // Any item in the local crate may invoke the helper module's public
1861 // interface through a combination of the two rules above.
1862 crate_helper_module::crate_helper();
1865 // This function is hidden to any module which is not a descendant of
1867 fn my_implementation() {}
1873 fn test_my_implementation() {
1874 // Because this module is a descendant of `submodule`, it's allowed
1875 // to access private items inside of `submodule` without a privacy
1877 super::my_implementation();
1885 For a Rust program to pass the privacy checking pass, all paths must be valid
1886 accesses given the two rules above. This includes all use statements,
1887 expressions, types, etc.
1889 ### Re-exporting and Visibility
1891 Rust allows publicly re-exporting items through a `pub use` directive. Because
1892 this is a public directive, this allows the item to be used in the current
1893 module through the rules above. It essentially allows public access into the
1894 re-exported item. For example, this program is valid:
1897 pub use self::implementation::api;
1899 mod implementation {
1908 This means that any external crate referencing `implementation::api::f` would
1909 receive a privacy violation, while the path `api::f` would be allowed.
1911 When re-exporting a private item, it can be thought of as allowing the "privacy
1912 chain" being short-circuited through the reexport instead of passing through
1913 the namespace hierarchy as it normally would.
1917 Any item declaration may have an _attribute_ applied to it. Attributes in Rust
1918 are modeled on Attributes in ECMA-335, with the syntax coming from ECMA-334
1919 (C#). An attribute is a general, free-form metadatum that is interpreted
1920 according to name, convention, and language and compiler version. Attributes
1921 may appear as any of:
1923 * A single identifier, the attribute name
1924 * An identifier followed by the equals sign '=' and a literal, providing a
1926 * An identifier followed by a parenthesized list of sub-attribute arguments
1928 Attributes with a bang ("!") after the hash ("#") apply to the item that the
1929 attribute is declared within. Attributes that do not have a bang after the hash
1930 apply to the item that follows the attribute.
1932 An example of attributes:
1935 // General metadata applied to the enclosing module or crate.
1936 #![crate_type = "lib"]
1938 // A function marked as a unit test
1944 // A conditionally-compiled module
1945 #[cfg(target_os="linux")]
1950 // A lint attribute used to suppress a warning/error
1951 #[allow(non_camel_case_types)]
1955 > **Note:** At some point in the future, the compiler will distinguish between
1956 > language-reserved and user-available attributes. Until then, there is
1957 > effectively no difference between an attribute handled by a loadable syntax
1958 > extension and the compiler.
1960 ### Crate-only attributes
1962 - `crate_name` - specify the crate's crate name.
1963 - `crate_type` - see [linkage](#linkage).
1964 - `feature` - see [compiler features](#compiler-features).
1965 - `no_builtins` - disable optimizing certain code patterns to invocations of
1966 library functions that are assumed to exist
1967 - `no_main` - disable emitting the `main` symbol. Useful when some other
1968 object being linked to defines `main`.
1969 - `no_start` - disable linking to the `native` crate, which specifies the
1970 "start" language item.
1971 - `no_std` - disable linking to the `std` crate.
1972 - `plugin` - load a list of named crates as compiler plugins, e.g.
1973 `#![plugin(foo, bar)]`. Optional arguments for each plugin,
1974 i.e. `#![plugin(foo(... args ...))]`, are provided to the plugin's
1975 registrar function. The `plugin` feature gate is required to use
1977 - `recursion_limit` - Sets the maximum depth for potentially
1978 infinitely-recursive compile-time operations like
1979 auto-dereference or macro expansion. The default is
1980 `#![recursion_limit="64"]`.
1982 ### Module-only attributes
1984 - `no_implicit_prelude` - disable injecting `use std::prelude::*` in this
1986 - `path` - specifies the file to load the module from. `#[path="foo.rs"] mod
1987 bar;` is equivalent to `mod bar { /* contents of foo.rs */ }`. The path is
1988 taken relative to the directory that the current module is in.
1990 ### Function-only attributes
1992 - `main` - indicates that this function should be passed to the entry point,
1993 rather than the function in the crate root named `main`.
1994 - `plugin_registrar` - mark this function as the registration point for
1995 [compiler plugins][plugin], such as loadable syntax extensions.
1996 - `start` - indicates that this function should be used as the entry point,
1997 overriding the "start" language item. See the "start" [language
1998 item](#language-items) for more details.
1999 - `test` - indicates that this function is a test function, to only be compiled
2000 in case of `--test`.
2001 - `should_panic` - indicates that this test function should panic, inverting the success condition.
2002 - `cold` - The function is unlikely to be executed, so optimize it (and calls
2004 - `naked` - The function utilizes a custom ABI or custom inline ASM that requires
2005 epilogue and prologue to be skipped.
2007 ### Static-only attributes
2009 - `thread_local` - on a `static mut`, this signals that the value of this
2010 static may change depending on the current thread. The exact consequences of
2011 this are implementation-defined.
2015 On an `extern` block, the following attributes are interpreted:
2017 - `link_args` - specify arguments to the linker, rather than just the library
2018 name and type. This is feature gated and the exact behavior is
2019 implementation-defined (due to variety of linker invocation syntax).
2020 - `link` - indicate that a native library should be linked to for the
2021 declarations in this block to be linked correctly. `link` supports an optional
2022 `kind` key with three possible values: `dylib`, `static`, and `framework`. See
2023 [external blocks](#external-blocks) for more about external blocks. Two
2024 examples: `#[link(name = "readline")]` and
2025 `#[link(name = "CoreFoundation", kind = "framework")]`.
2026 - `linked_from` - indicates what native library this block of FFI items is
2027 coming from. This attribute is of the form `#[linked_from = "foo"]` where
2028 `foo` is the name of a library in either `#[link]` or a `-l` flag. This
2029 attribute is currently required to export symbols from a Rust dynamic library
2030 on Windows, and it is feature gated behind the `linked_from` feature.
2032 On declarations inside an `extern` block, the following attributes are
2035 - `link_name` - the name of the symbol that this function or static should be
2037 - `linkage` - on a static, this specifies the [linkage
2038 type](http://llvm.org/docs/LangRef.html#linkage-types).
2042 - `repr` - on C-like enums, this sets the underlying type used for
2043 representation. Takes one argument, which is the primitive
2044 type this enum should be represented for, or `C`, which specifies that it
2045 should be the default `enum` size of the C ABI for that platform. Note that
2046 enum representation in C is undefined, and this may be incorrect when the C
2047 code is compiled with certain flags.
2051 - `repr` - specifies the representation to use for this struct. Takes a list
2052 of options. The currently accepted ones are `C` and `packed`, which may be
2053 combined. `C` will use a C ABI compatible struct layout, and `packed` will
2054 remove any padding between fields (note that this is very fragile and may
2055 break platforms which require aligned access).
2057 ### Macro-related attributes
2059 - `macro_use` on a `mod` — macros defined in this module will be visible in the
2060 module's parent, after this module has been included.
2062 - `macro_use` on an `extern crate` — load macros from this crate. An optional
2063 list of names `#[macro_use(foo, bar)]` restricts the import to just those
2064 macros named. The `extern crate` must appear at the crate root, not inside
2065 `mod`, which ensures proper function of the [`$crate` macro
2066 variable](book/macros.html#the-variable-crate).
2068 - `macro_reexport` on an `extern crate` — re-export the named macros.
2070 - `macro_export` - export a macro for cross-crate usage.
2072 - `no_link` on an `extern crate` — even if we load this crate for macros, don't
2073 link it into the output.
2075 See the [macros section of the
2076 book](book/macros.html#scoping-and-macro-importexport) for more information on
2080 ### Miscellaneous attributes
2082 - `deprecated` - mark the item as deprecated; the full attribute is `#[deprecated(since = "crate version", note = "...")`, where both arguments are optional.
2083 - `export_name` - on statics and functions, this determines the name of the
2085 - `link_section` - on statics and functions, this specifies the section of the
2086 object file that this item's contents will be placed into.
2087 - `no_mangle` - on any item, do not apply the standard name mangling. Set the
2088 symbol for this item to its identifier.
2089 - `simd` - on certain tuple structs, derive the arithmetic operators, which
2090 lower to the target's SIMD instructions, if any; the `simd` feature gate
2091 is necessary to use this attribute.
2092 - `unsafe_destructor_blind_to_params` - on `Drop::drop` method, asserts that the
2093 destructor code (and all potential specializations of that code) will
2094 never attempt to read from nor write to any references with lifetimes
2095 that come in via generic parameters. This is a constraint we cannot
2096 currently express via the type system, and therefore we rely on the
2097 programmer to assert that it holds. Adding this to a Drop impl causes
2098 the associated destructor to be considered "uninteresting" by the
2099 Drop-Check rule, and thus it can help sidestep data ordering
2100 constraints that would otherwise be introduced by the Drop-Check
2101 rule. Such sidestepping of the constraints, if done incorrectly, can
2102 lead to undefined behavior (in the form of reading or writing to data
2103 outside of its dynamic extent), and thus this attribute has the word
2104 "unsafe" in its name. To use this, the
2105 `unsafe_destructor_blind_to_params` feature gate must be enabled.
2106 - `doc` - Doc comments such as `/// foo` are equivalent to `#[doc = "foo"]`.
2107 - `rustc_on_unimplemented` - Write a custom note to be shown along with the error
2108 when the trait is found to be unimplemented on a type.
2109 You may use format arguments like `{T}`, `{A}` to correspond to the
2110 types at the point of use corresponding to the type parameters of the
2111 trait of the same name. `{Self}` will be replaced with the type that is supposed
2112 to implement the trait but doesn't. To use this, the `on_unimplemented` feature gate
2114 - `must_use` - on structs and enums, will warn if a value of this type isn't used or
2115 assigned to a variable. You may also include an optional message by using
2116 `#[must_use = "message"]` which will be given alongside the warning.
2118 ### Conditional compilation
2120 Sometimes one wants to have different compiler outputs from the same code,
2121 depending on build target, such as targeted operating system, or to enable
2124 There are two kinds of configuration options, one that is either defined or not
2125 (`#[cfg(foo)]`), and the other that contains a string that can be checked
2126 against (`#[cfg(bar = "baz")]`). Currently, only compiler-defined configuration
2127 options can have the latter form.
2130 // The function is only included in the build when compiling for OSX
2131 #[cfg(target_os = "macos")]
2136 // This function is only included when either foo or bar is defined
2137 #[cfg(any(foo, bar))]
2138 fn needs_foo_or_bar() {
2142 // This function is only included when compiling for a unixish OS with a 32-bit
2144 #[cfg(all(unix, target_pointer_width = "32"))]
2145 fn on_32bit_unix() {
2149 // This function is only included when foo is not defined
2151 fn needs_not_foo() {
2156 This illustrates some conditional compilation can be achieved using the
2157 `#[cfg(...)]` attribute. `any`, `all` and `not` can be used to assemble
2158 arbitrarily complex configurations through nesting.
2160 The following configurations must be defined by the implementation:
2162 * `target_arch = "..."` - Target CPU architecture, such as `"x86"`,
2163 `"x86_64"` `"mips"`, `"powerpc"`, `"powerpc64"`, `"arm"`, or
2164 `"aarch64"`. This value is closely related to the first element of
2165 the platform target triple, though it is not identical.
2166 * `target_os = "..."` - Operating system of the target, examples
2167 include `"windows"`, `"macos"`, `"ios"`, `"linux"`, `"android"`,
2168 `"freebsd"`, `"dragonfly"`, `"bitrig"` , `"openbsd"` or
2169 `"netbsd"`. This value is closely related to the second and third
2170 element of the platform target triple, though it is not identical.
2171 * `target_family = "..."` - Operating system family of the target, e. g.
2172 `"unix"` or `"windows"`. The value of this configuration option is defined
2173 as a configuration itself, like `unix` or `windows`.
2174 * `unix` - See `target_family`.
2175 * `windows` - See `target_family`.
2176 * `target_env = ".."` - Further disambiguates the target platform with
2177 information about the ABI/libc. Presently this value is either
2178 `"gnu"`, `"msvc"`, `"musl"`, or the empty string. For historical
2179 reasons this value has only been defined as non-empty when needed
2180 for disambiguation. Thus on many GNU platforms this value will be
2181 empty. This value is closely related to the fourth element of the
2182 platform target triple, though it is not identical. For example,
2183 embedded ABIs such as `gnueabihf` will simply define `target_env` as
2185 * `target_endian = "..."` - Endianness of the target CPU, either `"little"` or
2187 * `target_pointer_width = "..."` - Target pointer width in bits. This is set
2188 to `"32"` for targets with 32-bit pointers, and likewise set to `"64"` for
2190 * `target_has_atomic = "..."` - Set of integer sizes on which the target can perform
2191 atomic operations. Values are `"8"`, `"16"`, `"32"`, `"64"` and `"ptr"`.
2192 * `target_vendor = "..."` - Vendor of the target, for example `apple`, `pc`, or
2194 * `test` - Enabled when compiling the test harness (using the `--test` flag).
2195 * `debug_assertions` - Enabled by default when compiling without optimizations.
2196 This can be used to enable extra debugging code in development but not in
2197 production. For example, it controls the behavior of the standard library's
2198 `debug_assert!` macro.
2200 You can also set another attribute based on a `cfg` variable with `cfg_attr`:
2206 Will be the same as `#[b]` if `a` is set by `cfg`, and nothing otherwise.
2208 ### Lint check attributes
2210 A lint check names a potentially undesirable coding pattern, such as
2211 unreachable code or omitted documentation, for the static entity to which the
2214 For any lint check `C`:
2216 * `allow(C)` overrides the check for `C` so that violations will go
2218 * `deny(C)` signals an error after encountering a violation of `C`,
2219 * `forbid(C)` is the same as `deny(C)`, but also forbids changing the lint
2221 * `warn(C)` warns about violations of `C` but continues compilation.
2223 The lint checks supported by the compiler can be found via `rustc -W help`,
2224 along with their default settings. [Compiler
2225 plugins](book/compiler-plugins.html#lint-plugins) can provide additional lint checks.
2229 // Missing documentation is ignored here
2230 #[allow(missing_docs)]
2231 pub fn undocumented_one() -> i32 { 1 }
2233 // Missing documentation signals a warning here
2234 #[warn(missing_docs)]
2235 pub fn undocumented_too() -> i32 { 2 }
2237 // Missing documentation signals an error here
2238 #[deny(missing_docs)]
2239 pub fn undocumented_end() -> i32 { 3 }
2243 This example shows how one can use `allow` and `warn` to toggle a particular
2247 #[warn(missing_docs)]
2249 #[allow(missing_docs)]
2251 // Missing documentation is ignored here
2252 pub fn undocumented_one() -> i32 { 1 }
2254 // Missing documentation signals a warning here,
2255 // despite the allow above.
2256 #[warn(missing_docs)]
2257 pub fn undocumented_two() -> i32 { 2 }
2260 // Missing documentation signals a warning here
2261 pub fn undocumented_too() -> i32 { 3 }
2265 This example shows how one can use `forbid` to disallow uses of `allow` for
2269 #[forbid(missing_docs)]
2271 // Attempting to toggle warning signals an error here
2272 #[allow(missing_docs)]
2274 pub fn undocumented_too() -> i32 { 2 }
2280 Some primitive Rust operations are defined in Rust code, rather than being
2281 implemented directly in C or assembly language. The definitions of these
2282 operations have to be easy for the compiler to find. The `lang` attribute
2283 makes it possible to declare these operations. For example, the `str` module
2284 in the Rust standard library defines the string equality function:
2288 pub fn eq_slice(a: &str, b: &str) -> bool {
2293 The name `str_eq` has a special meaning to the Rust compiler, and the presence
2294 of this definition means that it will use this definition when generating calls
2295 to the string equality function.
2297 The set of language items is currently considered unstable. A complete
2298 list of the built-in language items will be added in the future.
2300 ### Inline attributes
2302 The inline attribute suggests that the compiler should place a copy of
2303 the function or static in the caller, rather than generating code to
2304 call the function or access the static where it is defined.
2306 The compiler automatically inlines functions based on internal heuristics.
2307 Incorrectly inlining functions can actually make the program slower, so it
2308 should be used with care.
2310 `#[inline]` and `#[inline(always)]` always cause the function to be serialized
2311 into the crate metadata to allow cross-crate inlining.
2313 There are three different types of inline attributes:
2315 * `#[inline]` hints the compiler to perform an inline expansion.
2316 * `#[inline(always)]` asks the compiler to always perform an inline expansion.
2317 * `#[inline(never)]` asks the compiler to never perform an inline expansion.
2321 The `derive` attribute allows certain traits to be automatically implemented
2322 for data structures. For example, the following will create an `impl` for the
2323 `PartialEq` and `Clone` traits for `Foo`, the type parameter `T` will be given
2324 the `PartialEq` or `Clone` constraints for the appropriate `impl`:
2327 #[derive(PartialEq, Clone)]
2334 The generated `impl` for `PartialEq` is equivalent to
2337 # struct Foo<T> { a: i32, b: T }
2338 impl<T: PartialEq> PartialEq for Foo<T> {
2339 fn eq(&self, other: &Foo<T>) -> bool {
2340 self.a == other.a && self.b == other.b
2343 fn ne(&self, other: &Foo<T>) -> bool {
2344 self.a != other.a || self.b != other.b
2349 ### Compiler Features
2351 Certain aspects of Rust may be implemented in the compiler, but they're not
2352 necessarily ready for every-day use. These features are often of "prototype
2353 quality" or "almost production ready", but may not be stable enough to be
2354 considered a full-fledged language feature.
2356 For this reason, Rust recognizes a special crate-level attribute of the form:
2359 #![feature(feature1, feature2, feature3)]
2362 This directive informs the compiler that the feature list: `feature1`,
2363 `feature2`, and `feature3` should all be enabled. This is only recognized at a
2364 crate-level, not at a module-level. Without this directive, all features are
2365 considered off, and using the features will result in a compiler error.
2367 The currently implemented features of the reference compiler are:
2369 * `advanced_slice_patterns` - See the [match expressions](#match-expressions)
2370 section for discussion; the exact semantics of
2371 slice patterns are subject to change, so some types
2374 * `slice_patterns` - OK, actually, slice patterns are just scary and
2375 completely unstable.
2377 * `asm` - The `asm!` macro provides a means for inline assembly. This is often
2378 useful, but the exact syntax for this feature along with its
2379 semantics are likely to change, so this macro usage must be opted
2382 * `associated_consts` - Allows constants to be defined in `impl` and `trait`
2383 blocks, so that they can be associated with a type or
2384 trait in a similar manner to methods and associated
2387 * `box_patterns` - Allows `box` patterns, the exact semantics of which
2388 is subject to change.
2390 * `box_syntax` - Allows use of `box` expressions, the exact semantics of which
2391 is subject to change.
2393 * `cfg_target_vendor` - Allows conditional compilation using the `target_vendor`
2394 matcher which is subject to change.
2396 * `cfg_target_has_atomic` - Allows conditional compilation using the `target_has_atomic`
2397 matcher which is subject to change.
2399 * `concat_idents` - Allows use of the `concat_idents` macro, which is in many
2400 ways insufficient for concatenating identifiers, and may be
2401 removed entirely for something more wholesome.
2403 * `custom_attribute` - Allows the usage of attributes unknown to the compiler
2404 so that new attributes can be added in a backwards compatible
2407 * `custom_derive` - Allows the use of `#[derive(Foo,Bar)]` as sugar for
2408 `#[derive_Foo] #[derive_Bar]`, which can be user-defined syntax
2411 * `inclusive_range_syntax` - Allows use of the `a...b` and `...b` syntax for inclusive ranges.
2413 * `inclusive_range` - Allows use of the types that represent desugared inclusive ranges.
2415 * `intrinsics` - Allows use of the "rust-intrinsics" ABI. Compiler intrinsics
2416 are inherently unstable and no promise about them is made.
2418 * `lang_items` - Allows use of the `#[lang]` attribute. Like `intrinsics`,
2419 lang items are inherently unstable and no promise about them
2422 * `link_args` - This attribute is used to specify custom flags to the linker,
2423 but usage is strongly discouraged. The compiler's usage of the
2424 system linker is not guaranteed to continue in the future, and
2425 if the system linker is not used then specifying custom flags
2426 doesn't have much meaning.
2428 * `link_llvm_intrinsics` – Allows linking to LLVM intrinsics via
2429 `#[link_name="llvm.*"]`.
2431 * `linkage` - Allows use of the `linkage` attribute, which is not portable.
2433 * `log_syntax` - Allows use of the `log_syntax` macro attribute, which is a
2434 nasty hack that will certainly be removed.
2436 * `main` - Allows use of the `#[main]` attribute, which changes the entry point
2437 into a Rust program. This capability is subject to change.
2439 * `macro_reexport` - Allows macros to be re-exported from one crate after being imported
2440 from another. This feature was originally designed with the sole
2441 use case of the Rust standard library in mind, and is subject to
2444 * `non_ascii_idents` - The compiler supports the use of non-ascii identifiers,
2445 but the implementation is a little rough around the
2446 edges, so this can be seen as an experimental feature
2447 for now until the specification of identifiers is fully
2450 * `no_std` - Allows the `#![no_std]` crate attribute, which disables the implicit
2451 `extern crate std`. This typically requires use of the unstable APIs
2452 behind the libstd "facade", such as libcore and libcollections. It
2453 may also cause problems when using syntax extensions, including
2456 * `on_unimplemented` - Allows the `#[rustc_on_unimplemented]` attribute, which allows
2457 trait definitions to add specialized notes to error messages
2458 when an implementation was expected but not found.
2460 * `optin_builtin_traits` - Allows the definition of default and negative trait
2461 implementations. Experimental.
2463 * `plugin` - Usage of [compiler plugins][plugin] for custom lints or syntax extensions.
2464 These depend on compiler internals and are subject to change.
2466 * `plugin_registrar` - Indicates that a crate provides [compiler plugins][plugin].
2468 * `quote` - Allows use of the `quote_*!` family of macros, which are
2469 implemented very poorly and will likely change significantly
2470 with a proper implementation.
2472 * `rustc_attrs` - Gates internal `#[rustc_*]` attributes which may be
2473 for internal use only or have meaning added to them in the future.
2475 * `rustc_diagnostic_macros`- A mysterious feature, used in the implementation
2476 of rustc, not meant for mortals.
2478 * `simd` - Allows use of the `#[simd]` attribute, which is overly simple and
2479 not the SIMD interface we want to expose in the long term.
2481 * `simd_ffi` - Allows use of SIMD vectors in signatures for foreign functions.
2482 The SIMD interface is subject to change.
2484 * `start` - Allows use of the `#[start]` attribute, which changes the entry point
2485 into a Rust program. This capability, especially the signature for the
2486 annotated function, is subject to change.
2488 * `thread_local` - The usage of the `#[thread_local]` attribute is experimental
2489 and should be seen as unstable. This attribute is used to
2490 declare a `static` as being unique per-thread leveraging
2491 LLVM's implementation which works in concert with the kernel
2492 loader and dynamic linker. This is not necessarily available
2493 on all platforms, and usage of it is discouraged.
2495 * `trace_macros` - Allows use of the `trace_macros` macro, which is a nasty
2496 hack that will certainly be removed.
2498 * `unboxed_closures` - Rust's new closure design, which is currently a work in
2499 progress feature with many known bugs.
2501 * `allow_internal_unstable` - Allows `macro_rules!` macros to be tagged with the
2502 `#[allow_internal_unstable]` attribute, designed
2503 to allow `std` macros to call
2504 `#[unstable]`/feature-gated functionality
2505 internally without imposing on callers
2506 (i.e. making them behave like function calls in
2507 terms of encapsulation).
2509 * `default_type_parameter_fallback` - Allows type parameter defaults to
2510 influence type inference.
2512 * `stmt_expr_attributes` - Allows attributes on expressions.
2514 * `type_ascription` - Allows type ascription expressions `expr: Type`.
2516 * `abi_vectorcall` - Allows the usage of the vectorcall calling convention
2517 (e.g. `extern "vectorcall" func fn_();`)
2519 * `abi_sysv64` - Allows the usage of the system V AMD64 calling convention
2520 (e.g. `extern "sysv64" func fn_();`)
2522 If a feature is promoted to a language feature, then all existing programs will
2523 start to receive compilation warnings about `#![feature]` directives which enabled
2524 the new feature (because the directive is no longer necessary). However, if a
2525 feature is decided to be removed from the language, errors will be issued (if
2526 there isn't a parser error first). The directive in this case is no longer
2527 necessary, and it's likely that existing code will break if the feature isn't
2530 If an unknown feature is found in a directive, it results in a compiler error.
2531 An unknown feature is one which has never been recognized by the compiler.
2533 # Statements and expressions
2535 Rust is _primarily_ an expression language. This means that most forms of
2536 value-producing or effect-causing evaluation are directed by the uniform syntax
2537 category of _expressions_. Each kind of expression can typically _nest_ within
2538 each other kind of expression, and rules for evaluation of expressions involve
2539 specifying both the value produced by the expression and the order in which its
2540 sub-expressions are themselves evaluated.
2542 In contrast, statements in Rust serve _mostly_ to contain and explicitly
2543 sequence expression evaluation.
2547 A _statement_ is a component of a block, which is in turn a component of an
2548 outer [expression](#expressions) or [function](#functions).
2550 Rust has two kinds of statement: [declaration
2551 statements](#declaration-statements) and [expression
2552 statements](#expression-statements).
2554 ### Declaration statements
2556 A _declaration statement_ is one that introduces one or more *names* into the
2557 enclosing statement block. The declared names may denote new variables or new
2560 #### Item declarations
2562 An _item declaration statement_ has a syntactic form identical to an
2563 [item](#items) declaration within a module. Declaring an item — a
2564 function, enumeration, struct, type, static, trait, implementation or module
2565 — locally within a statement block is simply a way of restricting its
2566 scope to a narrow region containing all of its uses; it is otherwise identical
2567 in meaning to declaring the item outside the statement block.
2569 > **Note**: there is no implicit capture of the function's dynamic environment when
2570 > declaring a function-local item.
2572 #### `let` statements
2574 A _`let` statement_ introduces a new set of variables, given by a pattern. The
2575 pattern may be followed by a type annotation, and/or an initializer expression.
2576 When no type annotation is given, the compiler will infer the type, or signal
2577 an error if insufficient type information is available for definite inference.
2578 Any variables introduced by a variable declaration are visible from the point of
2579 declaration until the end of the enclosing block scope.
2581 ### Expression statements
2583 An _expression statement_ is one that evaluates an [expression](#expressions)
2584 and ignores its result. The type of an expression statement `e;` is always
2585 `()`, regardless of the type of `e`. As a rule, an expression statement's
2586 purpose is to trigger the effects of evaluating its expression.
2590 An expression may have two roles: it always produces a *value*, and it may have
2591 *effects* (otherwise known as "side effects"). An expression *evaluates to* a
2592 value, and has effects during *evaluation*. Many expressions contain
2593 sub-expressions (operands). The meaning of each kind of expression dictates
2596 * Whether or not to evaluate the sub-expressions when evaluating the expression
2597 * The order in which to evaluate the sub-expressions
2598 * How to combine the sub-expressions' values to obtain the value of the expression
2600 In this way, the structure of expressions dictates the structure of execution.
2601 Blocks are just another kind of expression, so blocks, statements, expressions,
2602 and blocks again can recursively nest inside each other to an arbitrary depth.
2604 #### Lvalues, rvalues and temporaries
2606 Expressions are divided into two main categories: _lvalues_ and _rvalues_.
2607 Likewise within each expression, sub-expressions may occur in _lvalue context_
2608 or _rvalue context_. The evaluation of an expression depends both on its own
2609 category and the context it occurs within.
2611 An lvalue is an expression that represents a memory location. These expressions
2612 are [paths](#path-expressions) (which refer to local variables, function and
2613 method arguments, or static variables), dereferences (`*expr`), [indexing
2614 expressions](#index-expressions) (`expr[expr]`), and [field
2615 references](#field-expressions) (`expr.f`). All other expressions are rvalues.
2617 The left operand of an [assignment](#assignment-expressions) or
2618 [compound-assignment](#compound-assignment-expressions) expression is
2619 an lvalue context, as is the single operand of a unary
2620 [borrow](#unary-operator-expressions). The discriminant or subject of
2621 a [match expression](#match-expressions) may be an lvalue context, if
2622 ref bindings are made, but is otherwise an rvalue context. All other
2623 expression contexts are rvalue contexts.
2625 When an lvalue is evaluated in an _lvalue context_, it denotes a memory
2626 location; when evaluated in an _rvalue context_, it denotes the value held _in_
2627 that memory location.
2629 ##### Temporary lifetimes
2631 When an rvalue is used in an lvalue context, a temporary un-named
2632 lvalue is created and used instead. The lifetime of temporary values
2633 is typically the innermost enclosing statement; the tail expression of
2634 a block is considered part of the statement that encloses the block.
2636 When a temporary rvalue is being created that is assigned into a `let`
2637 declaration, however, the temporary is created with the lifetime of
2638 the enclosing block instead, as using the enclosing statement (the
2639 `let` declaration) would be a guaranteed error (since a pointer to the
2640 temporary would be stored into a variable, but the temporary would be
2641 freed before the variable could be used). The compiler uses simple
2642 syntactic rules to decide which values are being assigned into a `let`
2643 binding, and therefore deserve a longer temporary lifetime.
2645 Here are some examples:
2647 - `let x = foo(&temp())`. The expression `temp()` is an rvalue. As it
2648 is being borrowed, a temporary is created which will be freed after
2649 the innermost enclosing statement (the `let` declaration, in this case).
2650 - `let x = temp().foo()`. This is the same as the previous example,
2651 except that the value of `temp()` is being borrowed via autoref on a
2652 method-call. Here we are assuming that `foo()` is an `&self` method
2653 defined in some trait, say `Foo`. In other words, the expression
2654 `temp().foo()` is equivalent to `Foo::foo(&temp())`.
2655 - `let x = &temp()`. Here, the same temporary is being assigned into
2656 `x`, rather than being passed as a parameter, and hence the
2657 temporary's lifetime is considered to be the enclosing block.
2658 - `let x = SomeStruct { foo: &temp() }`. As in the previous case, the
2659 temporary is assigned into a struct which is then assigned into a
2660 binding, and hence it is given the lifetime of the enclosing block.
2661 - `let x = [ &temp() ]`. As in the previous case, the
2662 temporary is assigned into an array which is then assigned into a
2663 binding, and hence it is given the lifetime of the enclosing block.
2664 - `let ref x = temp()`. In this case, the temporary is created using a ref binding,
2665 but the result is the same: the lifetime is extended to the enclosing block.
2667 #### Moved and copied types
2669 When a [local variable](#variables) is used as an
2670 [rvalue](#lvalues-rvalues-and-temporaries), the variable will be copied
2671 if its type implements `Copy`. All others are moved.
2673 ### Literal expressions
2675 A _literal expression_ consists of one of the [literal](#literals) forms
2676 described earlier. It directly describes a number, character, string, boolean
2677 value, or the unit value.
2681 "hello"; // string type
2682 '5'; // character type
2686 ### Path expressions
2688 A [path](#paths) used as an expression context denotes either a local variable
2689 or an item. Path expressions are [lvalues](#lvalues-rvalues-and-temporaries).
2691 ### Tuple expressions
2693 Tuples are written by enclosing zero or more comma-separated expressions in
2694 parentheses. They are used to create [tuple-typed](#tuple-types) values.
2698 ("a", 4usize, true);
2701 You can disambiguate a single-element tuple from a value in parentheses with a
2705 (0,); // single-element tuple
2706 (0); // zero in parentheses
2709 ### Struct expressions
2711 There are several forms of struct expressions. A _struct expression_
2712 consists of the [path](#paths) of a [struct item](#structs), followed by
2713 a brace-enclosed list of zero or more comma-separated name-value pairs,
2714 providing the field values of a new instance of the struct. A field name
2715 can be any identifier, and is separated from its value expression by a colon.
2716 The location denoted by a struct field is mutable if and only if the
2717 enclosing struct is mutable.
2719 A _tuple struct expression_ consists of the [path](#paths) of a [struct
2720 item](#structs), followed by a parenthesized list of one or more
2721 comma-separated expressions (in other words, the path of a struct item
2722 followed by a tuple expression). The struct item must be a tuple struct
2725 A _unit-like struct expression_ consists only of the [path](#paths) of a
2726 [struct item](#structs).
2728 The following are examples of struct expressions:
2731 # struct Point { x: f64, y: f64 }
2732 # struct NothingInMe { }
2733 # struct TuplePoint(f64, f64);
2734 # mod game { pub struct User<'a> { pub name: &'a str, pub age: u32, pub score: usize } }
2735 # struct Cookie; fn some_fn<T>(t: T) {}
2736 Point {x: 10.0, y: 20.0};
2738 TuplePoint(10.0, 20.0);
2739 let u = game::User {name: "Joe", age: 35, score: 100_000};
2740 some_fn::<Cookie>(Cookie);
2743 A struct expression forms a new value of the named struct type. Note
2744 that for a given *unit-like* struct type, this will always be the same
2747 A struct expression can terminate with the syntax `..` followed by an
2748 expression to denote a functional update. The expression following `..` (the
2749 base) must have the same struct type as the new struct type being formed.
2750 The entire expression denotes the result of constructing a new struct (with
2751 the same type as the base expression) with the given values for the fields that
2752 were explicitly specified and the values in the base expression for all other
2756 # struct Point3d { x: i32, y: i32, z: i32 }
2757 let base = Point3d {x: 1, y: 2, z: 3};
2758 Point3d {y: 0, z: 10, .. base};
2761 ### Block expressions
2763 A _block expression_ is similar to a module in terms of the declarations that
2764 are possible. Each block conceptually introduces a new namespace scope. Use
2765 items can bring new names into scopes and declared items are in scope for only
2768 A block will execute each statement sequentially, and then execute the
2769 expression (if given). If the block ends in a statement, its value is `()`:
2772 let x: () = { println!("Hello."); };
2775 If it ends in an expression, its value and type are that of the expression:
2778 let x: i32 = { println!("Hello."); 5 };
2783 ### Method-call expressions
2785 A _method call_ consists of an expression followed by a single dot, an
2786 identifier, and a parenthesized expression-list. Method calls are resolved to
2787 methods on specific traits, either statically dispatching to a method if the
2788 exact `self`-type of the left-hand-side is known, or dynamically dispatching if
2789 the left-hand-side expression is an indirect [trait object](#trait-objects).
2791 ### Field expressions
2793 A _field expression_ consists of an expression followed by a single dot and an
2794 identifier, when not immediately followed by a parenthesized expression-list
2795 (the latter is a [method call expression](#method-call-expressions)). A field
2796 expression denotes a field of a [struct](#struct-types).
2801 (Struct {a: 10, b: 20}).a;
2804 A field access is an [lvalue](#lvalues-rvalues-and-temporaries) referring to
2805 the value of that field. When the type providing the field inherits mutability,
2806 it can be [assigned](#assignment-expressions) to.
2808 Also, if the type of the expression to the left of the dot is a
2809 pointer, it is automatically dereferenced as many times as necessary
2810 to make the field access possible. In cases of ambiguity, we prefer
2811 fewer autoderefs to more.
2813 ### Array expressions
2815 An [array](#array-and-slice-types) _expression_ is written by enclosing zero
2816 or more comma-separated expressions of uniform type in square brackets.
2818 In the `[expr ';' expr]` form, the expression after the `';'` must be a
2819 constant expression that can be evaluated at compile time, such as a
2820 [literal](#literals) or a [static item](#static-items).
2824 ["a", "b", "c", "d"];
2825 [0; 128]; // array with 128 zeros
2826 [0u8, 0u8, 0u8, 0u8];
2829 ### Index expressions
2831 [Array](#array-and-slice-types)-typed expressions can be indexed by
2832 writing a square-bracket-enclosed expression (the index) after them. When the
2833 array is mutable, the resulting [lvalue](#lvalues-rvalues-and-temporaries) can
2836 Indices are zero-based, and may be of any integral type. Vector access is
2837 bounds-checked at compile-time for constant arrays being accessed with a constant index value.
2838 Otherwise a check will be performed at run-time that will put the thread in a _panicked state_ if it fails.
2843 let x = (["a", "b"])[10]; // compiler error: const index-expr is out of bounds
2846 let y = (["a", "b"])[n]; // panics
2848 let arr = ["a", "b"];
2852 Also, if the type of the expression to the left of the brackets is a
2853 pointer, it is automatically dereferenced as many times as necessary
2854 to make the indexing possible. In cases of ambiguity, we prefer fewer
2857 ### Range expressions
2859 The `..` operator will construct an object of one of the `std::ops::Range` variants.
2862 1..2; // std::ops::Range
2863 3..; // std::ops::RangeFrom
2864 ..4; // std::ops::RangeTo
2865 ..; // std::ops::RangeFull
2868 The following expressions are equivalent.
2871 let x = std::ops::Range {start: 0, end: 10};
2877 Similarly, the `...` operator will construct an object of one of the
2878 `std::ops::RangeInclusive` variants.
2881 # #![feature(inclusive_range_syntax)]
2882 1...2; // std::ops::RangeInclusive
2883 ...4; // std::ops::RangeToInclusive
2886 The following expressions are equivalent.
2889 # #![feature(inclusive_range_syntax, inclusive_range)]
2890 let x = std::ops::RangeInclusive::NonEmpty {start: 0, end: 10};
2896 ### Unary operator expressions
2898 Rust defines the following unary operators. With the exception of `?`, they are
2899 all written as prefix operators, before the expression they apply to.
2902 : Negation. Signed integer types and floating-point types support negation. It
2903 is an error to apply negation to unsigned types; for example, the compiler
2906 : Dereference. When applied to a [pointer](#pointer-types) it denotes the
2907 pointed-to location. For pointers to mutable locations, the resulting
2908 [lvalue](#lvalues-rvalues-and-temporaries) can be assigned to.
2909 On non-pointer types, it calls the `deref` method of the `std::ops::Deref`
2910 trait, or the `deref_mut` method of the `std::ops::DerefMut` trait (if
2911 implemented by the type and required for an outer expression that will or
2912 could mutate the dereference), and produces the result of dereferencing the
2913 `&` or `&mut` borrowed pointer returned from the overload method.
2915 : Logical negation. On the boolean type, this flips between `true` and
2916 `false`. On integer types, this inverts the individual bits in the
2917 two's complement representation of the value.
2919 : Borrowing. When applied to an lvalue, these operators produce a
2920 reference (pointer) to the lvalue. The lvalue is also placed into
2921 a borrowed state for the duration of the reference. For a shared
2922 borrow (`&`), this implies that the lvalue may not be mutated, but
2923 it may be read or shared again. For a mutable borrow (`&mut`), the
2924 lvalue may not be accessed in any way until the borrow expires.
2925 If the `&` or `&mut` operators are applied to an rvalue, a
2926 temporary value is created; the lifetime of this temporary value
2927 is defined by [syntactic rules](#temporary-lifetimes).
2929 : Propagating errors if applied to `Err(_)` and unwrapping if
2930 applied to `Ok(_)`. Only works on the `Result<T, E>` type,
2931 and written in postfix notation.
2933 ### Binary operator expressions
2935 Binary operators expressions are given in terms of [operator
2936 precedence](#operator-precedence).
2938 #### Arithmetic operators
2940 Binary arithmetic expressions are syntactic sugar for calls to built-in traits,
2941 defined in the `std::ops` module of the `std` library. This means that
2942 arithmetic operators can be overridden for user-defined types. The default
2943 meaning of the operators on standard types is given here.
2946 : Addition and array/string concatenation.
2947 Calls the `add` method on the `std::ops::Add` trait.
2950 Calls the `sub` method on the `std::ops::Sub` trait.
2953 Calls the `mul` method on the `std::ops::Mul` trait.
2956 Calls the `div` method on the `std::ops::Div` trait.
2959 Calls the `rem` method on the `std::ops::Rem` trait.
2961 #### Bitwise operators
2963 Like the [arithmetic operators](#arithmetic-operators), bitwise operators are
2964 syntactic sugar for calls to methods of built-in traits. This means that
2965 bitwise operators can be overridden for user-defined types. The default
2966 meaning of the operators on standard types is given here. Bitwise `&`, `|` and
2967 `^` applied to boolean arguments are equivalent to logical `&&`, `||` and `!=`
2968 evaluated in non-lazy fashion.
2972 Calls the `bitand` method of the `std::ops::BitAnd` trait.
2974 : Bitwise inclusive OR.
2975 Calls the `bitor` method of the `std::ops::BitOr` trait.
2977 : Bitwise exclusive OR.
2978 Calls the `bitxor` method of the `std::ops::BitXor` trait.
2981 Calls the `shl` method of the `std::ops::Shl` trait.
2983 : Right shift (arithmetic).
2984 Calls the `shr` method of the `std::ops::Shr` trait.
2986 #### Lazy boolean operators
2988 The operators `||` and `&&` may be applied to operands of boolean type. The
2989 `||` operator denotes logical 'or', and the `&&` operator denotes logical
2990 'and'. They differ from `|` and `&` in that the right-hand operand is only
2991 evaluated when the left-hand operand does not already determine the result of
2992 the expression. That is, `||` only evaluates its right-hand operand when the
2993 left-hand operand evaluates to `false`, and `&&` only when it evaluates to
2996 #### Comparison operators
2998 Comparison operators are, like the [arithmetic
2999 operators](#arithmetic-operators), and [bitwise operators](#bitwise-operators),
3000 syntactic sugar for calls to built-in traits. This means that comparison
3001 operators can be overridden for user-defined types. The default meaning of the
3002 operators on standard types is given here.
3006 Calls the `eq` method on the `std::cmp::PartialEq` trait.
3009 Calls the `ne` method on the `std::cmp::PartialEq` trait.
3012 Calls the `lt` method on the `std::cmp::PartialOrd` trait.
3015 Calls the `gt` method on the `std::cmp::PartialOrd` trait.
3017 : Less than or equal.
3018 Calls the `le` method on the `std::cmp::PartialOrd` trait.
3020 : Greater than or equal.
3021 Calls the `ge` method on the `std::cmp::PartialOrd` trait.
3023 #### Type cast expressions
3025 A type cast expression is denoted with the binary operator `as`.
3027 Executing an `as` expression casts the value on the left-hand side to the type
3028 on the right-hand side.
3030 An example of an `as` expression:
3033 # fn sum(values: &[f64]) -> f64 { 0.0 }
3034 # fn len(values: &[f64]) -> i32 { 0 }
3036 fn average(values: &[f64]) -> f64 {
3037 let sum: f64 = sum(values);
3038 let size: f64 = len(values) as f64;
3043 Some of the conversions which can be done through the `as` operator
3044 can also be done implicitly at various points in the program, such as
3045 argument passing and assignment to a `let` binding with an explicit
3046 type. Implicit conversions are limited to "harmless" conversions that
3047 do not lose information and which have minimal or no risk of
3048 surprising side-effects on the dynamic execution semantics.
3050 #### Assignment expressions
3052 An _assignment expression_ consists of an
3053 [lvalue](#lvalues-rvalues-and-temporaries) expression followed by an equals
3054 sign (`=`) and an [rvalue](#lvalues-rvalues-and-temporaries) expression.
3056 Evaluating an assignment expression [either copies or
3057 moves](#moved-and-copied-types) its right-hand operand to its left-hand
3066 #### Compound assignment expressions
3068 The `+`, `-`, `*`, `/`, `%`, `&`, `|`, `^`, `<<`, and `>>` operators may be
3069 composed with the `=` operator. The expression `lval OP= val` is equivalent to
3070 `lval = lval OP val`. For example, `x = x + 1` may be written as `x += 1`.
3072 Any such expression always has the [`unit`](#tuple-types) type.
3074 #### Operator precedence
3076 The precedence of Rust binary operators is ordered as follows, going from
3079 ```{.text .precedence}
3095 Operators at the same precedence level are evaluated left-to-right. [Unary
3096 operators](#unary-operator-expressions) have the same precedence level and are
3097 stronger than any of the binary operators.
3099 ### Grouped expressions
3101 An expression enclosed in parentheses evaluates to the result of the enclosed
3102 expression. Parentheses can be used to explicitly specify evaluation order
3103 within an expression.
3105 An example of a parenthesized expression:
3108 let x: i32 = (2 + 3) * 4;
3112 ### Call expressions
3114 A _call expression_ invokes a function, providing zero or more input variables
3115 and an optional location to move the function's output into. If the function
3116 eventually returns, then the expression completes.
3118 Some examples of call expressions:
3121 # fn add(x: i32, y: i32) -> i32 { 0 }
3123 let x: i32 = add(1i32, 2i32);
3124 let pi: Result<f32, _> = "3.14".parse();
3127 ### Lambda expressions
3129 A _lambda expression_ (sometimes called an "anonymous function expression")
3130 defines a function and denotes it as a value, in a single expression. A lambda
3131 expression is a pipe-symbol-delimited (`|`) list of identifiers followed by an
3134 A lambda expression denotes a function that maps a list of parameters
3135 (`ident_list`) onto the expression that follows the `ident_list`. The
3136 identifiers in the `ident_list` are the parameters to the function. These
3137 parameters' types need not be specified, as the compiler infers them from
3140 Lambda expressions are most useful when passing functions as arguments to other
3141 functions, as an abbreviation for defining and capturing a separate function.
3143 Significantly, lambda expressions _capture their environment_, which regular
3144 [function definitions](#functions) do not. The exact type of capture depends
3145 on the [function type](#function-types) inferred for the lambda expression. In
3146 the simplest and least-expensive form (analogous to a ```|| { }``` expression),
3147 the lambda expression captures its environment by reference, effectively
3148 borrowing pointers to all outer variables mentioned inside the function.
3149 Alternately, the compiler may infer that a lambda expression should copy or
3150 move values (depending on their type) from the environment into the lambda
3151 expression's captured environment. A lambda can be forced to capture its
3152 environment by moving values by prefixing it with the `move` keyword.
3154 In this example, we define a function `ten_times` that takes a higher-order
3155 function argument, and we then call it with a lambda expression as an argument,
3156 followed by a lambda expression that moves values from its environment.
3159 fn ten_times<F>(f: F) where F: Fn(i32) {
3160 for index in 0..10 {
3165 ten_times(|j| println!("hello, {}", j));
3167 let word = "konnichiwa".to_owned();
3168 ten_times(move |j| println!("{}, {}", word, j));
3173 A `loop` expression denotes an infinite loop.
3175 A `loop` expression may optionally have a _label_. The label is written as
3176 a lifetime preceding the loop expression, as in `'foo: loop{ }`. If a
3177 label is present, then labeled `break` and `continue` expressions nested
3178 within this loop may exit out of this loop or return control to its head.
3179 See [break expressions](#break-expressions) and [continue
3180 expressions](#continue-expressions).
3182 ### `break` expressions
3184 A `break` expression has an optional _label_. If the label is absent, then
3185 executing a `break` expression immediately terminates the innermost loop
3186 enclosing it. It is only permitted in the body of a loop. If the label is
3187 present, then `break 'foo` terminates the loop with label `'foo`, which need not
3188 be the innermost label enclosing the `break` expression, but must enclose it.
3190 ### `continue` expressions
3192 A `continue` expression has an optional _label_. If the label is absent, then
3193 executing a `continue` expression immediately terminates the current iteration
3194 of the innermost loop enclosing it, returning control to the loop *head*. In
3195 the case of a `while` loop, the head is the conditional expression controlling
3196 the loop. In the case of a `for` loop, the head is the call-expression
3197 controlling the loop. If the label is present, then `continue 'foo` returns
3198 control to the head of the loop with label `'foo`, which need not be the
3199 innermost label enclosing the `continue` expression, but must enclose it.
3201 A `continue` expression is only permitted in the body of a loop.
3205 A `while` loop begins by evaluating the boolean loop conditional expression.
3206 If the loop conditional expression evaluates to `true`, the loop body block
3207 executes and control returns to the loop conditional expression. If the loop
3208 conditional expression evaluates to `false`, the `while` expression completes.
3221 Like `loop` expressions, `while` loops can be controlled with `break` or
3222 `continue`, and may optionally have a _label_. See [infinite
3223 loops](#infinite-loops), [break expressions](#break-expressions), and
3224 [continue expressions](#continue-expressions) for more information.
3226 ### `for` expressions
3228 A `for` expression is a syntactic construct for looping over elements provided
3229 by an implementation of `std::iter::IntoIterator`.
3231 An example of a `for` loop over the contents of an array:
3235 # fn bar(f: &Foo) { }
3240 let v: &[Foo] = &[a, b, c];
3247 An example of a for loop over a series of integers:
3250 # fn bar(b:usize) { }
3256 Like `loop` expressions, `for` loops can be controlled with `break` or
3257 `continue`, and may optionally have a _label_. See [infinite
3258 loops](#infinite-loops), [break expressions](#break-expressions), and
3259 [continue expressions](#continue-expressions) for more information.
3261 ### `if` expressions
3263 An `if` expression is a conditional branch in program control. The form of an
3264 `if` expression is a condition expression, followed by a consequent block, any
3265 number of `else if` conditions and blocks, and an optional trailing `else`
3266 block. The condition expressions must have type `bool`. If a condition
3267 expression evaluates to `true`, the consequent block is executed and any
3268 subsequent `else if` or `else` block is skipped. If a condition expression
3269 evaluates to `false`, the consequent block is skipped and any subsequent `else
3270 if` condition is evaluated. If all `if` and `else if` conditions evaluate to
3271 `false` then any `else` block is executed.
3273 ### `match` expressions
3275 A `match` expression branches on a *pattern*. The exact form of matching that
3276 occurs depends on the pattern. Patterns consist of some combination of
3277 literals, destructured arrays or enum constructors, structs and tuples,
3278 variable binding specifications, wildcards (`..`), and placeholders (`_`). A
3279 `match` expression has a *head expression*, which is the value to compare to
3280 the patterns. The type of the patterns must equal the type of the head
3283 In a pattern whose head expression has an `enum` type, a placeholder (`_`)
3284 stands for a *single* data field, whereas a wildcard `..` stands for *all* the
3285 fields of a particular variant.
3287 A `match` behaves differently depending on whether or not the head expression
3288 is an [lvalue or an rvalue](#lvalues-rvalues-and-temporaries). If the head
3289 expression is an rvalue, it is first evaluated into a temporary location, and
3290 the resulting value is sequentially compared to the patterns in the arms until
3291 a match is found. The first arm with a matching pattern is chosen as the branch
3292 target of the `match`, any variables bound by the pattern are assigned to local
3293 variables in the arm's block, and control enters the block.
3295 When the head expression is an lvalue, the match does not allocate a temporary
3296 location (however, a by-value binding may copy or move from the lvalue). When
3297 possible, it is preferable to match on lvalues, as the lifetime of these
3298 matches inherits the lifetime of the lvalue, rather than being restricted to
3299 the inside of the match.
3301 An example of a `match` expression:
3307 1 => println!("one"),
3308 2 => println!("two"),
3309 3 => println!("three"),
3310 4 => println!("four"),
3311 5 => println!("five"),
3312 _ => println!("something else"),
3316 Patterns that bind variables default to binding to a copy or move of the
3317 matched value (depending on the matched value's type). This can be changed to
3318 bind to a reference by using the `ref` keyword, or to a mutable reference using
3321 Subpatterns can also be bound to variables by the use of the syntax `variable @
3322 subpattern`. For example:
3328 e @ 1 ... 5 => println!("got a range element {}", e),
3329 _ => println!("anything"),
3333 Patterns can also dereference pointers by using the `&`, `&mut` and `box`
3334 symbols, as appropriate. For example, these two matches on `x: &i32` are
3339 let y = match *x { 0 => "zero", _ => "some" };
3340 let z = match x { &0 => "zero", _ => "some" };
3345 Multiple match patterns may be joined with the `|` operator. A range of values
3346 may be specified with `...`. For example:
3351 let message = match x {
3352 0 | 1 => "not many",
3358 Range patterns only work on scalar types (like integers and characters; not
3359 like arrays and structs, which have sub-components). A range pattern may not
3360 be a sub-range of another range pattern inside the same `match`.
3362 Finally, match patterns can accept *pattern guards* to further refine the
3363 criteria for matching a case. Pattern guards appear after the pattern and
3364 consist of a bool-typed expression following the `if` keyword. A pattern guard
3365 may refer to the variables bound within the pattern they follow.
3368 # let maybe_digit = Some(0);
3369 # fn process_digit(i: i32) { }
3370 # fn process_other(i: i32) { }
3372 let message = match maybe_digit {
3373 Some(x) if x < 10 => process_digit(x),
3374 Some(x) => process_other(x),
3379 ### `if let` expressions
3381 An `if let` expression is semantically identical to an `if` expression but in
3382 place of a condition expression it expects a `let` statement with a refutable
3383 pattern. If the value of the expression on the right hand side of the `let`
3384 statement matches the pattern, the corresponding block will execute, otherwise
3385 flow proceeds to the first `else` block that follows.
3388 let dish = ("Ham", "Eggs");
3390 // this body will be skipped because the pattern is refuted
3391 if let ("Bacon", b) = dish {
3392 println!("Bacon is served with {}", b);
3395 // this body will execute
3396 if let ("Ham", b) = dish {
3397 println!("Ham is served with {}", b);
3401 ### `while let` loops
3403 A `while let` loop is semantically identical to a `while` loop but in place of
3404 a condition expression it expects `let` statement with a refutable pattern. If
3405 the value of the expression on the right hand side of the `let` statement
3406 matches the pattern, the loop body block executes and control returns to the
3407 pattern matching statement. Otherwise, the while expression completes.
3409 ### `return` expressions
3411 Return expressions are denoted with the keyword `return`. Evaluating a `return`
3412 expression moves its argument into the designated output location for the
3413 current function call, destroys the current function activation frame, and
3414 transfers control to the caller frame.
3416 An example of a `return` expression:
3419 fn max(a: i32, b: i32) -> i32 {
3431 Every variable, item and value in a Rust program has a type. The _type_ of a
3432 *value* defines the interpretation of the memory holding it.
3434 Built-in types and type-constructors are tightly integrated into the language,
3435 in nontrivial ways that are not possible to emulate in user-defined types.
3436 User-defined types have limited capabilities.
3440 The primitive types are the following:
3442 * The boolean type `bool` with values `true` and `false`.
3443 * The machine types (integer and floating-point).
3444 * The machine-dependent integer types.
3452 The machine types are the following:
3454 * The unsigned word types `u8`, `u16`, `u32` and `u64`, with values drawn from
3455 the integer intervals [0, 2^8 - 1], [0, 2^16 - 1], [0, 2^32 - 1] and
3456 [0, 2^64 - 1] respectively.
3458 * The signed two's complement word types `i8`, `i16`, `i32` and `i64`, with
3459 values drawn from the integer intervals [-(2^(7)), 2^7 - 1],
3460 [-(2^(15)), 2^15 - 1], [-(2^(31)), 2^31 - 1], [-(2^(63)), 2^63 - 1]
3463 * The IEEE 754-2008 `binary32` and `binary64` floating-point types: `f32` and
3464 `f64`, respectively.
3466 #### Machine-dependent integer types
3468 The `usize` type is an unsigned integer type with the same number of bits as the
3469 platform's pointer type. It can represent every memory address in the process.
3471 The `isize` type is a signed integer type with the same number of bits as the
3472 platform's pointer type. The theoretical upper bound on object and array size
3473 is the maximum `isize` value. This ensures that `isize` can be used to calculate
3474 differences between pointers into an object or array and can address every byte
3475 within an object along with one byte past the end.
3479 The types `char` and `str` hold textual data.
3481 A value of type `char` is a [Unicode scalar value](
3482 http://www.unicode.org/glossary/#unicode_scalar_value) (i.e. a code point that
3483 is not a surrogate), represented as a 32-bit unsigned word in the 0x0000 to
3484 0xD7FF or 0xE000 to 0x10FFFF range. A `[char]` array is effectively an UCS-4 /
3487 A value of type `str` is a Unicode string, represented as an array of 8-bit
3488 unsigned bytes holding a sequence of UTF-8 code points. Since `str` is of
3489 unknown size, it is not a _first-class_ type, but can only be instantiated
3490 through a pointer type, such as `&str`.
3494 A tuple *type* is a heterogeneous product of other types, called the *elements*
3495 of the tuple. It has no nominal name and is instead structurally typed.
3497 Tuple types and values are denoted by listing the types or values of their
3498 elements, respectively, in a parenthesized, comma-separated list.
3500 Because tuple elements don't have a name, they can only be accessed by
3501 pattern-matching or by using `N` directly as a field to access the
3504 An example of a tuple type and its use:
3507 type Pair<'a> = (i32, &'a str);
3508 let p: Pair<'static> = (10, "ten");
3512 assert_eq!(b, "ten");
3513 assert_eq!(p.0, 10);
3514 assert_eq!(p.1, "ten");
3517 For historical reasons and convenience, the tuple type with no elements (`()`)
3518 is often called ‘unit’ or ‘the unit type’.
3520 ### Array, and Slice types
3522 Rust has two different types for a list of items:
3524 * `[T; N]`, an 'array'
3527 An array has a fixed size, and can be allocated on either the stack or the
3530 A slice is a 'view' into an array. It doesn't own the data it points
3536 // A stack-allocated array
3537 let array: [i32; 3] = [1, 2, 3];
3539 // A heap-allocated array
3540 let vector: Vec<i32> = vec![1, 2, 3];
3542 // A slice into an array
3543 let slice: &[i32] = &vector[..];
3546 As you can see, the `vec!` macro allows you to create a `Vec<T>` easily. The
3547 `vec!` macro is also part of the standard library, rather than the language.
3549 All in-bounds elements of arrays and slices are always initialized, and access
3550 to an array or slice is always bounds-checked.
3554 A `struct` *type* is a heterogeneous product of other types, called the
3555 *fields* of the type.[^structtype]
3557 [^structtype]: `struct` types are analogous to `struct` types in C,
3558 the *record* types of the ML family,
3559 or the *struct* types of the Lisp family.
3561 New instances of a `struct` can be constructed with a [struct
3562 expression](#struct-expressions).
3564 The memory layout of a `struct` is undefined by default to allow for compiler
3565 optimizations like field reordering, but it can be fixed with the
3566 `#[repr(...)]` attribute. In either case, fields may be given in any order in
3567 a corresponding struct *expression*; the resulting `struct` value will always
3568 have the same memory layout.
3570 The fields of a `struct` may be qualified by [visibility
3571 modifiers](#visibility-and-privacy), to allow access to data in a
3572 struct outside a module.
3574 A _tuple struct_ type is just like a struct type, except that the fields are
3577 A _unit-like struct_ type is like a struct type, except that it has no
3578 fields. The one value constructed by the associated [struct
3579 expression](#struct-expressions) is the only value that inhabits such a
3582 ### Enumerated types
3584 An *enumerated type* is a nominal, heterogeneous disjoint union type, denoted
3585 by the name of an [`enum` item](#enumerations). [^enumtype]
3587 [^enumtype]: The `enum` type is analogous to a `data` constructor declaration in
3588 ML, or a *pick ADT* in Limbo.
3590 An [`enum` item](#enumerations) declares both the type and a number of *variant
3591 constructors*, each of which is independently named and takes an optional tuple
3594 New instances of an `enum` can be constructed by calling one of the variant
3595 constructors, in a [call expression](#call-expressions).
3597 Any `enum` value consumes as much memory as the largest variant constructor for
3598 its corresponding `enum` type.
3600 Enum types cannot be denoted *structurally* as types, but must be denoted by
3601 named reference to an [`enum` item](#enumerations).
3605 Nominal types — [enumerations](#enumerated-types) and
3606 [structs](#struct-types) — may be recursive. That is, each `enum`
3607 constructor or `struct` field may refer, directly or indirectly, to the
3608 enclosing `enum` or `struct` type itself. Such recursion has restrictions:
3610 * Recursive types must include a nominal type in the recursion
3611 (not mere [type definitions](grammar.html#type-definitions),
3612 or other structural types such as [arrays](#array-and-slice-types) or [tuples](#tuple-types)).
3613 * A recursive `enum` item must have at least one non-recursive constructor
3614 (in order to give the recursion a basis case).
3615 * The size of a recursive type must be finite;
3616 in other words the recursive fields of the type must be [pointer types](#pointer-types).
3617 * Recursive type definitions can cross module boundaries, but not module *visibility* boundaries,
3618 or crate boundaries (in order to simplify the module system and type checker).
3620 An example of a *recursive* type and its use:
3625 Cons(T, Box<List<T>>)
3628 let a: List<i32> = List::Cons(7, Box::new(List::Cons(13, Box::new(List::Nil))));
3633 All pointers in Rust are explicit first-class values. They can be copied,
3634 stored into data structs, and returned from functions. There are two
3635 varieties of pointer in Rust:
3638 : These point to memory _owned by some other value_.
3639 A reference type is written `&type`,
3640 or `&'a type` when you need to specify an explicit lifetime.
3641 Copying a reference is a "shallow" operation:
3642 it involves only copying the pointer itself.
3643 Releasing a reference has no effect on the value it points to,
3644 but a reference of a temporary value will keep it alive during the scope
3645 of the reference itself.
3647 * Raw pointers (`*`)
3648 : Raw pointers are pointers without safety or liveness guarantees.
3649 Raw pointers are written as `*const T` or `*mut T`,
3650 for example `*const i32` means a raw pointer to a 32-bit integer.
3651 Copying or dropping a raw pointer has no effect on the lifecycle of any
3652 other value. Dereferencing a raw pointer or converting it to any other
3653 pointer type is an [`unsafe` operation](#unsafe-functions).
3654 Raw pointers are generally discouraged in Rust code;
3655 they exist to support interoperability with foreign code,
3656 and writing performance-critical or low-level functions.
3658 The standard library contains additional 'smart pointer' types beyond references
3663 The function type constructor `fn` forms new function types. A function type
3664 consists of a possibly-empty set of function-type modifiers (such as `unsafe`
3665 or `extern`), a sequence of input types and an output type.
3667 An example of a `fn` type:
3670 fn add(x: i32, y: i32) -> i32 {
3674 let mut x = add(5,7);
3676 type Binop = fn(i32, i32) -> i32;
3677 let bo: Binop = add;
3681 #### Function types for specific items
3683 Internal to the compiler, there are also function types that are specific to a particular
3684 function item. In the following snippet, for example, the internal types of the functions
3685 `foo` and `bar` are different, despite the fact that they have the same signature:
3692 The types of `foo` and `bar` can both be implicitly coerced to the fn
3693 pointer type `fn()`. There is currently no syntax for unique fn types,
3694 though the compiler will emit a type like `fn() {foo}` in error
3695 messages to indicate "the unique fn type for the function `foo`".
3699 A [lambda expression](#lambda-expressions) produces a closure value with
3700 a unique, anonymous type that cannot be written out.
3702 Depending on the requirements of the closure, its type implements one or
3703 more of the closure traits:
3706 : The closure can be called once. A closure called as `FnOnce`
3707 can move out values from its environment.
3710 : The closure can be called multiple times as mutable. A closure called as
3711 `FnMut` can mutate values from its environment. `FnMut` inherits from
3712 `FnOnce` (i.e. anything implementing `FnMut` also implements `FnOnce`).
3715 : The closure can be called multiple times through a shared reference.
3716 A closure called as `Fn` can neither move out from nor mutate values
3717 from its environment. `Fn` inherits from `FnMut`, which itself
3718 inherits from `FnOnce`.
3723 In Rust, a type like `&SomeTrait` or `Box<SomeTrait>` is called a _trait object_.
3724 Each instance of a trait object includes:
3726 - a pointer to an instance of a type `T` that implements `SomeTrait`
3727 - a _virtual method table_, often just called a _vtable_, which contains, for
3728 each method of `SomeTrait` that `T` implements, a pointer to `T`'s
3729 implementation (i.e. a function pointer).
3731 The purpose of trait objects is to permit "late binding" of methods. Calling a
3732 method on a trait object results in virtual dispatch at runtime: that is, a
3733 function pointer is loaded from the trait object vtable and invoked indirectly.
3734 The actual implementation for each vtable entry can vary on an object-by-object
3737 Note that for a trait object to be instantiated, the trait must be
3738 _object-safe_. Object safety rules are defined in [RFC 255].
3740 [RFC 255]: https://github.com/rust-lang/rfcs/blob/master/text/0255-object-safety.md
3742 Given a pointer-typed expression `E` of type `&T` or `Box<T>`, where `T`
3743 implements trait `R`, casting `E` to the corresponding pointer type `&R` or
3744 `Box<R>` results in a value of the _trait object_ `R`. This result is
3745 represented as a pair of pointers: the vtable pointer for the `T`
3746 implementation of `R`, and the pointer value of `E`.
3748 An example of a trait object:
3752 fn stringify(&self) -> String;
3755 impl Printable for i32 {
3756 fn stringify(&self) -> String { self.to_string() }
3759 fn print(a: Box<Printable>) {
3760 println!("{}", a.stringify());
3764 print(Box::new(10) as Box<Printable>);
3768 In this example, the trait `Printable` occurs as a trait object in both the
3769 type signature of `print`, and the cast expression in `main`.
3773 Within the body of an item that has type parameter declarations, the names of
3774 its type parameters are types:
3777 fn to_vec<A: Clone>(xs: &[A]) -> Vec<A> {
3781 let first: A = xs[0].clone();
3782 let mut rest: Vec<A> = to_vec(&xs[1..]);
3783 rest.insert(0, first);
3788 Here, `first` has type `A`, referring to `to_vec`'s `A` type parameter; and `rest`
3789 has type `Vec<A>`, a vector with element type `A`.
3793 The special type `Self` has a meaning within traits and impls. In a trait definition, it refers
3794 to an implicit type parameter representing the "implementing" type. In an impl,
3795 it is an alias for the implementing type. For example, in:
3802 impl From<i32> for String {
3803 fn from(x: i32) -> Self {
3809 The notation `Self` in the impl refers to the implementing type: `String`. In another
3814 fn make_string(&self) -> String;
3817 impl Printable for String {
3818 fn make_string(&self) -> String {
3824 The notation `&self` is a shorthand for `self: &Self`. In this case,
3825 in the impl, `Self` refers to the value of type `String` that is the
3826 receiver for a call to the method `make_string`.
3830 Subtyping is implicit and can occur at any stage in type checking or
3831 inference. Subtyping in Rust is very restricted and occurs only due to
3832 variance with respect to lifetimes and between types with higher ranked
3833 lifetimes. If we were to erase lifetimes from types, then the only subtyping
3834 would be due to type equality.
3836 Consider the following example: string literals always have `'static`
3837 lifetime. Nevertheless, we can assign `s` to `t`:
3841 let s: &'static str = "hi";
3845 Since `'static` "lives longer" than `'a`, `&'static str` is a subtype of
3850 Coercions are defined in [RFC 401]. A coercion is implicit and has no syntax.
3852 [RFC 401]: https://github.com/rust-lang/rfcs/blob/master/text/0401-coercions.md
3856 A coercion can only occur at certain coercion sites in a program; these are
3857 typically places where the desired type is explicit or can be derived by
3858 propagation from explicit types (without type inference). Possible coercion
3861 * `let` statements where an explicit type is given.
3863 For example, `42` is coerced to have type `i8` in the following:
3869 * `static` and `const` statements (similar to `let` statements).
3871 * Arguments for function calls
3873 The value being coerced is the actual parameter, and it is coerced to
3874 the type of the formal parameter.
3876 For example, `42` is coerced to have type `i8` in the following:
3886 * Instantiations of struct or variant fields
3888 For example, `42` is coerced to have type `i8` in the following:
3891 struct Foo { x: i8 }
3898 * Function results, either the final line of a block if it is not
3899 semicolon-terminated or any expression in a `return` statement
3901 For example, `42` is coerced to have type `i8` in the following:
3909 If the expression in one of these coercion sites is a coercion-propagating
3910 expression, then the relevant sub-expressions in that expression are also
3911 coercion sites. Propagation recurses from these new coercion sites.
3912 Propagating expressions and their relevant sub-expressions are:
3914 * Array literals, where the array has type `[U; n]`. Each sub-expression in
3915 the array literal is a coercion site for coercion to type `U`.
3917 * Array literals with repeating syntax, where the array has type `[U; n]`. The
3918 repeated sub-expression is a coercion site for coercion to type `U`.
3920 * Tuples, where a tuple is a coercion site to type `(U_0, U_1, ..., U_n)`.
3921 Each sub-expression is a coercion site to the respective type, e.g. the
3922 zeroth sub-expression is a coercion site to type `U_0`.
3924 * Parenthesized sub-expressions (`(e)`): if the expression has type `U`, then
3925 the sub-expression is a coercion site to `U`.
3927 * Blocks: if a block has type `U`, then the last expression in the block (if
3928 it is not semicolon-terminated) is a coercion site to `U`. This includes
3929 blocks which are part of control flow statements, such as `if`/`else`, if
3930 the block has a known type.
3934 Coercion is allowed between the following types:
3936 * `T` to `U` if `T` is a subtype of `U` (*reflexive case*)
3938 * `T_1` to `T_3` where `T_1` coerces to `T_2` and `T_2` coerces to `T_3`
3941 Note that this is not fully supported yet
3945 * `*mut T` to `*const T`
3947 * `&T` to `*const T`
3949 * `&mut T` to `*mut T`
3951 * `&T` to `&U` if `T` implements `Deref<Target = U>`. For example:
3954 use std::ops::Deref;
3956 struct CharContainer {
3960 impl Deref for CharContainer {
3963 fn deref<'a>(&'a self) -> &'a char {
3968 fn foo(arg: &char) {}
3971 let x = &mut CharContainer { value: 'y' };
3972 foo(x); //&mut CharContainer is coerced to &char.
3976 * `&mut T` to `&mut U` if `T` implements `DerefMut<Target = U>`.
3978 * TyCtor(`T`) to TyCtor(coerce_inner(`T`)), where TyCtor(`T`) is one of
3986 - coerce_inner(`[T, ..n]`) = `[T]`
3987 - coerce_inner(`T`) = `U` where `T` is a concrete type which implements the
3990 In the future, coerce_inner will be recursively extended to tuples and
3991 structs. In addition, coercions from sub-traits to super-traits will be
3992 added. See [RFC 401] for more details.
3996 Several traits define special evaluation behavior.
4000 The `Copy` trait changes the semantics of a type implementing it. Values whose
4001 type implements `Copy` are copied rather than moved upon assignment.
4003 ## The `Sized` trait
4005 The `Sized` trait indicates that the size of this type is known at compile-time.
4009 The `Drop` trait provides a destructor, to be run whenever a value of this type
4012 ## The `Deref` trait
4014 The `Deref<Target = U>` trait allows a type to implicitly implement all the methods
4015 of the type `U`. When attempting to resolve a method call, the compiler will search
4016 the top-level type for the implementation of the called method. If no such method is
4017 found, `.deref()` is called and the compiler continues to search for the method
4018 implementation in the returned type `U`.
4022 The `Send` trait indicates that a value of this type is safe to send from one
4027 The `Sync` trait indicates that a value of this type is safe to share between
4032 A Rust program's memory consists of a static set of *items* and a *heap*.
4033 Immutable portions of the heap may be safely shared between threads, mutable
4034 portions may not be safely shared, but several mechanisms for effectively-safe
4035 sharing of mutable values, built on unsafe code but enforcing a safe locking
4036 discipline, exist in the standard library.
4038 Allocations in the stack consist of *variables*, and allocations in the heap
4041 ### Memory allocation and lifetime
4043 The _items_ of a program are those functions, modules and types that have their
4044 value calculated at compile-time and stored uniquely in the memory image of the
4045 rust process. Items are neither dynamically allocated nor freed.
4047 The _heap_ is a general term that describes boxes. The lifetime of an
4048 allocation in the heap depends on the lifetime of the box values pointing to
4049 it. Since box values may themselves be passed in and out of frames, or stored
4050 in the heap, heap allocations may outlive the frame they are allocated within.
4051 An allocation in the heap is guaranteed to reside at a single location in the
4052 heap for the whole lifetime of the allocation - it will never be relocated as
4053 a result of moving a box value.
4055 ### Memory ownership
4057 When a stack frame is exited, its local allocations are all released, and its
4058 references to boxes are dropped.
4062 A _variable_ is a component of a stack frame, either a named function parameter,
4063 an anonymous [temporary](#lvalues-rvalues-and-temporaries), or a named local
4066 A _local variable_ (or *stack-local* allocation) holds a value directly,
4067 allocated within the stack's memory. The value is a part of the stack frame.
4069 Local variables are immutable unless declared otherwise like: `let mut x = ...`.
4071 Function parameters are immutable unless declared with `mut`. The `mut` keyword
4072 applies only to the following parameter (so `|mut x, y|` and `fn f(mut x:
4073 Box<i32>, y: Box<i32>)` declare one mutable variable `x` and one immutable
4076 Methods that take either `self` or `Box<Self>` can optionally place them in a
4077 mutable variable by prefixing them with `mut` (similar to regular arguments):
4080 trait Changer: Sized {
4081 fn change(mut self) {}
4082 fn modify(mut self: Box<Self>) {}
4086 Local variables are not initialized when allocated; the entire frame worth of
4087 local variables are allocated at once, on frame-entry, in an uninitialized
4088 state. Subsequent statements within a function may or may not initialize the
4089 local variables. Local variables can be used only after they have been
4090 initialized; this is enforced by the compiler.
4094 The Rust compiler supports various methods to link crates together both
4095 statically and dynamically. This section will explore the various methods to
4096 link Rust crates together, and more information about native libraries can be
4097 found in the [FFI section of the book][ffi].
4099 In one session of compilation, the compiler can generate multiple artifacts
4100 through the usage of either command line flags or the `crate_type` attribute.
4101 If one or more command line flags are specified, all `crate_type` attributes will
4102 be ignored in favor of only building the artifacts specified by command line.
4104 * `--crate-type=bin`, `#[crate_type = "bin"]` - A runnable executable will be
4105 produced. This requires that there is a `main` function in the crate which
4106 will be run when the program begins executing. This will link in all Rust and
4107 native dependencies, producing a distributable binary.
4109 * `--crate-type=lib`, `#[crate_type = "lib"]` - A Rust library will be produced.
4110 This is an ambiguous concept as to what exactly is produced because a library
4111 can manifest itself in several forms. The purpose of this generic `lib` option
4112 is to generate the "compiler recommended" style of library. The output library
4113 will always be usable by rustc, but the actual type of library may change from
4114 time-to-time. The remaining output types are all different flavors of
4115 libraries, and the `lib` type can be seen as an alias for one of them (but the
4116 actual one is compiler-defined).
4118 * `--crate-type=dylib`, `#[crate_type = "dylib"]` - A dynamic Rust library will
4119 be produced. This is different from the `lib` output type in that this forces
4120 dynamic library generation. The resulting dynamic library can be used as a
4121 dependency for other libraries and/or executables. This output type will
4122 create `*.so` files on linux, `*.dylib` files on osx, and `*.dll` files on
4125 * `--crate-type=staticlib`, `#[crate_type = "staticlib"]` - A static system
4126 library will be produced. This is different from other library outputs in that
4127 the Rust compiler will never attempt to link to `staticlib` outputs. The
4128 purpose of this output type is to create a static library containing all of
4129 the local crate's code along with all upstream dependencies. The static
4130 library is actually a `*.a` archive on linux and osx and a `*.lib` file on
4131 windows. This format is recommended for use in situations such as linking
4132 Rust code into an existing non-Rust application because it will not have
4133 dynamic dependencies on other Rust code.
4135 * `--crate-type=cdylib`, `#[crate_type = "cdylib"]` - A dynamic system
4136 library will be produced. This is used when compiling Rust code as
4137 a dynamic library to be loaded from another language. This output type will
4138 create `*.so` files on Linux, `*.dylib` files on OSX, and `*.dll` files on
4141 * `--crate-type=rlib`, `#[crate_type = "rlib"]` - A "Rust library" file will be
4142 produced. This is used as an intermediate artifact and can be thought of as a
4143 "static Rust library". These `rlib` files, unlike `staticlib` files, are
4144 interpreted by the Rust compiler in future linkage. This essentially means
4145 that `rustc` will look for metadata in `rlib` files like it looks for metadata
4146 in dynamic libraries. This form of output is used to produce statically linked
4147 executables as well as `staticlib` outputs.
4149 Note that these outputs are stackable in the sense that if multiple are
4150 specified, then the compiler will produce each form of output at once without
4151 having to recompile. However, this only applies for outputs specified by the
4152 same method. If only `crate_type` attributes are specified, then they will all
4153 be built, but if one or more `--crate-type` command line flags are specified,
4154 then only those outputs will be built.
4156 With all these different kinds of outputs, if crate A depends on crate B, then
4157 the compiler could find B in various different forms throughout the system. The
4158 only forms looked for by the compiler, however, are the `rlib` format and the
4159 dynamic library format. With these two options for a dependent library, the
4160 compiler must at some point make a choice between these two formats. With this
4161 in mind, the compiler follows these rules when determining what format of
4162 dependencies will be used:
4164 1. If a static library is being produced, all upstream dependencies are
4165 required to be available in `rlib` formats. This requirement stems from the
4166 reason that a dynamic library cannot be converted into a static format.
4168 Note that it is impossible to link in native dynamic dependencies to a static
4169 library, and in this case warnings will be printed about all unlinked native
4170 dynamic dependencies.
4172 2. If an `rlib` file is being produced, then there are no restrictions on what
4173 format the upstream dependencies are available in. It is simply required that
4174 all upstream dependencies be available for reading metadata from.
4176 The reason for this is that `rlib` files do not contain any of their upstream
4177 dependencies. It wouldn't be very efficient for all `rlib` files to contain a
4178 copy of `libstd.rlib`!
4180 3. If an executable is being produced and the `-C prefer-dynamic` flag is not
4181 specified, then dependencies are first attempted to be found in the `rlib`
4182 format. If some dependencies are not available in an rlib format, then
4183 dynamic linking is attempted (see below).
4185 4. If a dynamic library or an executable that is being dynamically linked is
4186 being produced, then the compiler will attempt to reconcile the available
4187 dependencies in either the rlib or dylib format to create a final product.
4189 A major goal of the compiler is to ensure that a library never appears more
4190 than once in any artifact. For example, if dynamic libraries B and C were
4191 each statically linked to library A, then a crate could not link to B and C
4192 together because there would be two copies of A. The compiler allows mixing
4193 the rlib and dylib formats, but this restriction must be satisfied.
4195 The compiler currently implements no method of hinting what format a library
4196 should be linked with. When dynamically linking, the compiler will attempt to
4197 maximize dynamic dependencies while still allowing some dependencies to be
4198 linked in via an rlib.
4200 For most situations, having all libraries available as a dylib is recommended
4201 if dynamically linking. For other situations, the compiler will emit a
4202 warning if it is unable to determine which formats to link each library with.
4204 In general, `--crate-type=bin` or `--crate-type=lib` should be sufficient for
4205 all compilation needs, and the other options are just available if more
4206 fine-grained control is desired over the output format of a Rust crate.
4210 Unsafe operations are those that potentially violate the memory-safety
4211 guarantees of Rust's static semantics.
4213 The following language level features cannot be used in the safe subset of
4216 - Dereferencing a [raw pointer](#pointer-types).
4217 - Reading or writing a [mutable static variable](#mutable-statics).
4218 - Calling an unsafe function (including an intrinsic or foreign function).
4222 Unsafe functions are functions that are not safe in all contexts and/or for all
4223 possible inputs. Such a function must be prefixed with the keyword `unsafe` and
4224 can only be called from an `unsafe` block or another `unsafe` function.
4228 A block of code can be prefixed with the `unsafe` keyword, to permit calling
4229 `unsafe` functions or dereferencing raw pointers within a safe function.
4231 When a programmer has sufficient conviction that a sequence of potentially
4232 unsafe operations is actually safe, they can encapsulate that sequence (taken
4233 as a whole) within an `unsafe` block. The compiler will consider uses of such
4234 code safe, in the surrounding context.
4236 Unsafe blocks are used to wrap foreign libraries, make direct use of hardware
4237 or implement features not directly present in the language. For example, Rust
4238 provides the language features necessary to implement memory-safe concurrency
4239 in the language but the implementation of threads and message passing is in the
4242 Rust's type system is a conservative approximation of the dynamic safety
4243 requirements, so in some cases there is a performance cost to using safe code.
4244 For example, a doubly-linked list is not a tree structure and can only be
4245 represented with reference-counted pointers in safe code. By using `unsafe`
4246 blocks to represent the reverse links as raw pointers, it can be implemented
4249 ## Behavior considered undefined
4251 The following is a list of behavior which is forbidden in all Rust code,
4252 including within `unsafe` blocks and `unsafe` functions. Type checking provides
4253 the guarantee that these issues are never caused by safe code.
4256 * Dereferencing a null/dangling raw pointer
4257 * Reads of [undef](http://llvm.org/docs/LangRef.html#undefined-values)
4258 (uninitialized) memory
4259 * Breaking the [pointer aliasing
4260 rules](http://llvm.org/docs/LangRef.html#pointer-aliasing-rules)
4261 with raw pointers (a subset of the rules used by C)
4262 * `&mut T` and `&T` follow LLVM’s scoped [noalias] model, except if the `&T`
4263 contains an `UnsafeCell<U>`. Unsafe code must not violate these aliasing
4265 * Mutating non-mutable data (that is, data reached through a shared reference or
4266 data owned by a `let` binding), unless that data is contained within an `UnsafeCell<U>`.
4267 * Invoking undefined behavior via compiler intrinsics:
4268 * Indexing outside of the bounds of an object with `std::ptr::offset`
4269 (`offset` intrinsic), with
4270 the exception of one byte past the end which is permitted.
4271 * Using `std::ptr::copy_nonoverlapping_memory` (`memcpy32`/`memcpy64`
4272 intrinsics) on overlapping buffers
4273 * Invalid values in primitive types, even in private fields/locals:
4274 * Dangling/null references or boxes
4275 * A value other than `false` (0) or `true` (1) in a `bool`
4276 * A discriminant in an `enum` not included in the type definition
4277 * A value in a `char` which is a surrogate or above `char::MAX`
4278 * Non-UTF-8 byte sequences in a `str`
4279 * Unwinding into Rust from foreign code or unwinding from Rust into foreign
4280 code. Rust's failure system is not compatible with exception handling in
4281 other languages. Unwinding must be caught and handled at FFI boundaries.
4283 [noalias]: http://llvm.org/docs/LangRef.html#noalias
4285 ## Behavior not considered unsafe
4287 This is a list of behavior not considered *unsafe* in Rust terms, but that may
4291 * Leaks of memory and other resources
4292 * Exiting without calling destructors
4294 - Overflow is considered "unexpected" behavior and is always user-error,
4295 unless the `wrapping` primitives are used. In non-optimized builds, the compiler
4296 will insert debug checks that panic on overflow, but in optimized builds overflow
4297 instead results in wrapped values. See [RFC 560] for the rationale and more details.
4299 [RFC 560]: https://github.com/rust-lang/rfcs/blob/master/text/0560-integer-overflow.md
4301 # Appendix: Influences
4303 Rust is not a particularly original language, with design elements coming from
4304 a wide range of sources. Some of these are listed below (including elements
4305 that have since been removed):
4307 * SML, OCaml: algebraic data types, pattern matching, type inference,
4308 semicolon statement separation
4309 * C++: references, RAII, smart pointers, move semantics, monomorphization,
4311 * ML Kit, Cyclone: region based memory management
4312 * Haskell (GHC): typeclasses, type families
4313 * Newsqueak, Alef, Limbo: channels, concurrency
4314 * Erlang: message passing, thread failure, ~~linked thread failure~~,
4315 ~~lightweight concurrency~~
4316 * Swift: optional bindings
4317 * Scheme: hygienic macros
4319 * Ruby: ~~block syntax~~
4320 * NIL, Hermes: ~~typestate~~
4321 * [Unicode Annex #31](http://www.unicode.org/reports/tr31/): identifier and
4324 [ffi]: book/ffi.html
4325 [plugin]: book/compiler-plugins.html